Skip to content

[AIROCMLIR-891] LLVM Upstream Merge May#2395

Open
djramic wants to merge 8 commits into
developfrom
djramic-891
Open

[AIROCMLIR-891] LLVM Upstream Merge May#2395
djramic wants to merge 8 commits into
developfrom
djramic-891

Conversation

@djramic

@djramic djramic commented Jun 1, 2026

Copy link
Copy Markdown
Contributor

Motivation

Resolve: https://amd-hub.atlassian.net/browse/AIROCMLIR-891

Technical Details

Conflicts resolved:

  • AMDGPUAsmPrinter.cpp: Kept rocMLIR's commented-out occupancy-warning block and adopted upstream's removal of the isGFX11Plus InstPrefSize block (now handled via STM.hasInstPrefSize()).
  • AMDGPUToROCDL.cpp: Adopted the upstream conversion-pattern list, adding GlobalTransposeLoadOpLowering.
  • EmulateNarrowType.cpp: Adopted the upstream version that adds the ConvertMemRefCast pattern and passes assumeAligned to the ConvertMemRefSubview/ConvertMemRefReinterpretCast constructors.

Diff files:
llvm-diff.txt
mlir-diff.txt

Test Plan

External Tests

  • check-llvm
  • check-mlir
    CI
  • PR
  • Nightly
  • Weekly

Submission Checklist

djramic added 2 commits June 1, 2026 08:12
…2a0a

cb4240522a0a merge main into amd-staging (#2745)
df14c4a224a6 [Infra] Export python libs for fixing aomp failure (#2746)
8634976d50a6 [Infra][Nightly] Temp cherry-pick 2719 for hipcc and hiprtc test (#2727)
42c6a49126b8 merge main into amd-staging
019dab5844a1 merge main into amd-staging (#2744)
090308199b30 [BOLT][DebugInfo] Make parallel DWARF debug names generation deterministic(1/2) (#197670)
256d09201cf9 [AMDGPU] Fix missing SIMachineFunctionInfo::MinNumAGPRs from YAML mirror (#200647)
28be1eaf755a [MC][COFF] Mark .debug_*.dwo sections IMAGE_SCN_LNK_REMOVE (#199616)
f68ef5b4303d [clang-tidy] Fix FP/FN in cppcoreguidelines-missing-std-forward (#178651)
50f452f4b561 [InstCombine] Fix miscompilation when folding cttz with non-powers-of-2 (#200583)
1bbd613e6267 [InstCombine] Fold lshr 1, X into zext (X == 0) (#200669)
863502498715 [AArch64][SME] Add multi-vector load opcodes to getMemOpInfo (#200238)
6897c5e24ce5 [M68k][MC] Add MC support for PCI w/ base displacement addressing mode (#200696)
fae9a1c3e887 [clang-format] Recognize Verilog class item qualifiers (#199085)
0631ac764a56 [clang-format] Remove the blank line in the function try block (#199086)
4298580c5516 [ARM] Fix some fp16 Shuffle lowering without +fullfp16 (#200688)
62a23823a979 [AMDGPU] Use v_rsq_f32 for f16 rsqrt on targets without 16-bit insts (#200646)
cf3ae95c0155 merge main into amd-staging
5c4fba8645c0 merge main into amd-staging (#2741)
e3574d46e0de [NFC][LLVM] Fix Intrinsics.td to adhere to 80 col limit (#199346)
500071679810 [mlir][SPIR-V] Convert math.cttz to spirv.GLFindILsb (#200455)
77860f206812 [NFC][LLVM] Remove redundant verifier type checks for some intrinsics (#200658)
c42f1884730f [lldb] Fix up comment placement (NFC) (#200671)
a89a6596129c [TailCallElim] Drop poison-generating flags on reassociated accumulators (#200624)
07f9a7d2ab89 [CIR] Implement Atomic Complex to Complex cast (#199260)
cebf520118ae merge main into amd-staging
ac1620ec38b9 merge main into amd-staging (#2740)
025b9a27cf74 [TableGen] Add !switch operator (#199659)
5d20fd32807e [mlir] Fix typo in toy tutorial (#200640)
fe87c971bf07 [llubi] Fix the number of words to reserve in `Context::toBytes` (#200651)
b48743be091b [clang-tidy] `use-ranges`: preserve iterator results with `.begin()` (#196036)
8dc098589f31 [clang-tidy] Avoid unsafe `use-default-member-init` fixes (#191607)
37845da44408 [clang-tidy] `use-ranges`: preserve used `unique` results (#196035)
7966cbbdd02b [AArch64] Use ADDP tree for v16i8 to i16 bitmask extraction (#199812)
a67efda258fa [VPlan] Recognize shl by constant in getSCEVExprForVPValue. (#199359)
c78c9004a7f8 [AArch64] Add tests for bf16 convert with fast-math flags. NFC (#200643)
da3f152eebca [IR][NFC] Fast path for Constant::isNullValue (#200434)
80f6b7641ef9 [Object] Remove unused argument of DataExtractor constructor (NFC) (#197124)
b49a468051e1 [Profile] Remove unused argument of DataExtractor constructor (NFC) (#197119)
bc125b315414 merge main into amd-staging
e920ece6ae3b [RISCV] Prevent RISCVOptWInstrs from shrinking volatile LD instructions. (#200472)
1f2223d231a7 [ExpandIRInsts] Fix profile metadata after #200291 (#200629)
183fecb435bc merge main into amd-staging (#2738)
b17edc69b07d [ExpandIRInsts] Drop ProfcheckDisableMetadataFixes in some cases (#200627)
dbf01d40ca99 [bazel] Add CLANG_USE_EXPERIMENTAL_CONST_INTERP (fixup for #199396) (#200619)
16a127b531eb [IR] Fix !invariant.group in combineMetadataForCSE when K moves (#200551)
baa079610087 [asan] NFC: clang-format allocator-related files (#200478)
b51054818b78 [asan] NFC: tag ambiguous literal args at Allocate call sites (#200615)
7a07381d5511 [libc][wctype] Rename src/wctype/wctype.h to avoid name collision with system header. (#200613)
866945c20298 [libc++][ranges] P2542R8: Implement `views::concat` (#120920)
223ef1f31274 [IRBuilder] ConstFold unary intrinsics (#200496)
8d5d550242cf [VPlan] Factor vputils::getIntrinsicID (NFC) (#200174)
42d712ea62f5 [X86][TTI] Use dyn_cast_or_null in getGSVectorCost. (#200606)
6d2a90bd8bf3 [X86] matchBinaryPermuteShuffle - match to X86ISD::SHLD funnel shift patterns (REAPPLIED) (#200604)
1f10f1ca8af3 [DenseMap] Replace tombstone deletion with TAOCP 6.4 Algorithm R (#200595)
90779840d51a [LowerInvoke] Use createCallMatchingInvoke instead of rolling it ourselves (#200536)
56a9bba59817 [IR] SubclassOptionalData is only optional for Instruction (#200419)
a3815779ce34 [revPat] update reverts
9117da78a2f1 merge main into amd-staging
5cd6f6bcb2b6 Revert "[offload][OpenMP] Add strict flag for blocks and threads in kernel arguments (#199483)"   needs downstream integration : see rlieberm/SalaOffProblem
013ee58df53b Revert "Reapply "[PDB][llvm-pdbutil] Add DXContainer support for `pdb2yaml` and `yaml2pdb`"" (#200588)
2d9367ee9c39 merge main into amd-staging
1c4e91428545 merge main into amd-staging
b771021970fc [libc][ci] Clean up libc-fullbuild-tests precommit CI. (#200520)
aa1cf3205e63 [IR] Fix PoisoningVH relocation of a poisoned handle (#200540)
51256ed96304 [IR][NFC] Inline Type::isIntegerTy(n) (#200471)
2396aa8ac0e8 [CIR][NFC] Mark Complex visitExpr as unsupported (#197782)
0d1ed9383d1d [libc] Tag cpp::byte with gnu::may_alias (#200462)
470411cf18e3 [clang][bytecode] Loosen an assertion about operator delete (#200575)
f7d576010fc1 [mlir][nfc] Fix assertion text in IndexingUtils (#181826)
ef6a217703a8 [clang][bytecode] Fix a crash with an empty InitListExpr (#200366)
30ec1fa2a9d9 [clang][bytecode] Reject invalid UnaryOperators (#200394)
3c95dd29d834 [VPlan] Thread scalar type through VPBlend, VPExpression recipes. (NFC) (#200255)
2501cdd2cad9 [X86] Add VBMI2 shuffle test to track miscompile reported in #200136 (#200569)
d46e7dcee756 [clang][bytecode] Fix an assertion failure in AddSubNonNumber (#200393)
343f005ff3a8 [offload] *KernelInfo becomes KernelInfo
8d86dc22be4f [SLP] Fix extract-cost scale using NCD of all external-user sites
d9cbed5d47d9 [MLIR][GPU][NFC] Reformat GPU target attachment tests (#199339)
3467e7f9eb16 [SLP] Fix extract-cost scale for LCSSA-phi external users in nested loops
a59f9951f3e1 [SLP] Recompute copyable operand deps for duplicate copyable nodes
80ad13af63ef [bazel] Fix compiler-rt:interception (#200561)
f15904ec71a8 [offload][OpenMP] Add strict flag for blocks and threads in kernel arguments (#199483)
e3f80fe493c1 [mlir][SPIR-V] Convert math.clampf to spirv.GLFClamp and math.ctpop to spirv.BitCount (#200454)
fdfb99f63364 [libc++] std::byteswap support for _BitInt(N) (#196512)
02af47b6053d [Clang][test] Fix leading slash (#200549)
3c4bc6c9f7b5 [InstCombine] Handle "trunc nuw to i1" as "icmp ne,0" in foldSelectValueEquivalence (#198131)
c7614cd55070 [MergeICmps] Don't merge comparisons whose width isn't a byte multiple (#200346)
c22f5665a9d6 [LICM] Drop poison-generating flags when reassociating an icmp (#200344)
093c76a9b101 [SelectionDAG] Preserve IR alignment on atomicrmw/cmpxchg MMOs (#200332)
9e3c18494dfd Revert "[X86] matchBinaryPermuteShuffle - match to X86ISD::SHLD funnel shift patterns" (#200546)
9d0ce81b649f [Flang][OpenMP] Reject array sections and subobjects in LINEAR clause (#197430)
666a8cf6ebb7 [Flang][Parser] Handle compiler directives inside INTERFACE blocks (#198516)
799b8669e57a [SelectionDAG] Remove redundant asserts in WidenVecRes_ATOMIC_LOAD (#200159)
9b03d3f31968 [CIR][CodeGen] Replace errorNYI with assert for address space in emitAutoVarAlloca (#197506)
3c193d4d88c8 [lldb][Windows] Cache thread context in NativeRegisterContextWindows_arm64 (#197385)
467216904045 [Bazel] Add inl files to nanobind textual headers (#200516)
b5228cb04473 merge main into amd-staging
d38b5b3b990d [lit] Add __slots__ to hot shell-execution objects (#199668)
83f4fd3f3527 [compiler-rt] Restore GPU profile runtime build and fix multi-device crash (#2709)
80e04ff2745a merge main into amd-staging (#2733)
08d281fe7a90 [clang] Add CLANG_USE_EXPERIMENTAL_CONST_INTERP cmake option (#199396)
524734d88b3c [lldb/test] Trim @expectedFailureAll(remote=True) decorators (#200529)
0c94404f2135 [clang][bytecode] Reject bitcasts of objc block pointers (#200397)
522ac7582dda [dsymutil] Bump .dSYM bundle directory mtime after a rewrite (#199257)
639b44a5f03f [psdb] use linux-gfx942-1gpu-core42-ossci-rocm instead of linux-gfx942-1gpu-ossci-rocm
1dd08114eb3b [RISCV] Do not append duplicate Zilsd GPRPair CSR (#200463)
f561d59856c1 [LLDB] Simplify the hint when po-ing an object with no object descrip… (#200499)
720df36d34e7 [os_log] Fix a CodeGen crash for non-trivial C++ arguments (#200320)
cc8dd6ccd307 [X86] Use kmovw, not kmovq, for VK16 copies without BWI (#200337)
20f117f386e5 [lldb] Fix copy-paste typo in Symbol::operator= (#200528)
d46b9851c0e2 [ProfCheck] Exclude some tests (#200527)
796f1b332059 [CIR][HIP][NFC] Add device variable registration coverage (#200204)
95994d1e6307 [Support] Add path-based setLastAccessAndModificationTime overload (#199256)
a3241c53528a workflows/build-ci-container-windows: Remove template expansion (#200097)
4e47b560196a [ExpandIRInsts] Fix sitofp/uitofp to float producing garbage instead of inf (#200291)
9b77b22a0d14 [lldb][Darwin] Correct jGetLoadedDynamicLibrariesInfos typeo & fallback (#200515)
d70975c36995 Update Google Benchmark to v1.9.5 (#198964)
7565961d1906 workflows/issue-write: Do not read pr number from file for pull_request events (#200474)
f976a1d67108 [VPlan] Squelch unused warn in collectMemOps (NFC) (#200507)
3d24f9acc92b [lldb-server] Add accelerator plugin infrastructure for debugging hardware accelerators like gpus (#198907)
f958f1716fe3 Revert "[MLIR][Arith][WIE] Fix Conversions for Shift Left and Right Ops" (#200512)
9b5d0af42f8d [lldb] Halve the size of indexes in DemangledNameInfo (NFC) (#200506)
34da785839f7 [mlir][vector] Fix incorrect byte-alignment assumption in ConvertVectorStore (#189235)
62acf9378a4d Manual update of LLVM_MAIN_REVISION to 582231
50ed21d39c0d [CIR] Add RunCleanupsScope RAII around loop bodies (#200461)
02cf363d1748 [clang-doc] Add specialization info to record references (#191252)
807f92ed7eba merge main into amd-staging (#2729)
03d8fa70398d [MLIR][Arith][WIE] Fix Conversions for Shift Left and Right Ops (#198457)
0f8a64fb33ef [AtomicExpand] Preserve flags expanding loads/stores to cmpxchg/atomicrmw (#200324)
60290059386b [clang][scan-deps] Add option to disable caching stat failures (#200484)
37732ab65726 [OpenMP] Fix RPC register segfaulting without PM initialized (#200494)
61d17f0e69cd [NVPTX][FIX] Ensure memmove are kept if not lowered to loops  (#200469)
3a2b697fc101 [VPlan] Check type in collectGroupedReplicateMemOps (NFC) (#192781)
4ac26f45fa8f [DirectX] Update checks for more precise error messages (#200491)
0ed795a70b9d [docs] Rewrite the ghlink Sphinx plugin to use doctree (#199076)
ee872666e288 [VPlan] Expand VPExpandSCEVRecipes to VPInstructions before CSE. (#197643)
5d6ed5365c0f [flang][docs] Documented IS_CONTIGUOUS() extension for constant arrays (#200451)
59eb19d0c061 [InstSimplify] Expose simplifyUnaryIntrinsic (NFC) (#200444)
9b6611986828 Add yamlobj roundtrip tests for SM 6.8 features (#198403)
4e8c14df16a7 [ExpandIRInsts] Fix e.g. fptoui.sat.f32.i256's handling of inf (#200261)
f8545bed7415 [CIR] Implement cleanups of base classes for aggregates. (#200473)
3a7dd6e8ce76 [Comgr] Fix SPIR-V triple typo in setIsaName (#2712)
9b4e2dc29177 [Flang][OpenMP]Handling restrictions of using Memory-Order-Clause with Atomic-Clause (#199636)
c9cdbc4a30b8 [scudo] Improve performance of pushBlocks sort. (#200297)
d337e680ba32 [DirectX] Remove obsolete cbuffer layout test (#200307)
02997d75928b [IR] Introduce Instruction::getFastMathFlagsOrNone (NFC) (#200457)
e04dbbf19477 [libc][bazel] Add arm and riscv FEnvImpl.h to textual_hdrs. (#200479)
0fc5d0ab0cea [IR] Introduce `mem.cache_hint` metadata for composable memory cache control hints (#181612)
6a9694827181 [libc] Fix SSE2 check for x86_64/sqrt.h. (#200468)
a20c21b8bedc [clang] Enable GNU __attribute__((init_priority(...))) on z/OS. (#199573)
352fd0bd9b8e [lldb][test] Link test binaries with -random_uuid (#199385)
43511b564751 [CIR] Add Extend (signext/zeroext) handling to CallConvLowering (#195745)
550f61f191e5 [SimplifyCFG] Preserve atomicity when merging atomic conditional stores (#200327)
a02a959c88ee merge main into amd-staging
97d9ec9847d4 merge main into amd-staging (#2725)
ce78ca1bccb5 Revert "[Flang] Move builtin .mod generation into runtimes (Reapply #137828) (#171515)"
6da24e4022f0 [clang-sycl-linker] Improve --spirv-dump-device-code and harden CLI handling (#200096)
0a0ceaa698cf merge main into amd-staging
d5abd9c851e8 [ExpandIRInsts] Avoid redundant dyn_cast after #175864 (#200475)
3aed319af981 [SPIR-V] Insert service function when module has only declarations (#199964)
5945eacb9345 [libclc][test] Update math/cos.cl check lines after #199981 revert (#200353)
3e3871d6fecd [HLSL] Adding matrix support to splitdouble (#200257)
91f3d8d7030e [Docs] Fix build (#200467)
36264ec61518 [lit] Remove redundant f.close() in TestingConfig (#200459)
a10ca8ed6454 [offload] use argument pointer array in olLaunchKernel (#194333)
bd5c7247525f [CIR] Implement __sync_synchronize builtin (#200423)
ddef3a893ae3 LangRef: Clarify that unnamed_addr constants can be duplicated.
8918dd8b4dc6 [Clang][test] Fix space in ld path (#200012)
2c5dca550ae7 llvm-objdump: Support --disassemble= option.
3330752fa10a Fix MemProf imported alias handling after guid metadata changes (#200430)
70b6c3419e99 [Flang] Fix runtime module file dependency on tests (#200417)
2d411c58671d [alpha.webkit.UncountedLocalVarsChecker] Check uninitialized raw pointer. (#200309)
82840c51a843 [analyzer][webkit][NFC] Add more tests for nodelete checker (#200357)
8a3ed55e56dc [lldb] Tighten TestObjcMethods test case for arm64e (#200262)
098b05f32989 [RISC-V][MC] Introduce initial support for RVY (CHERI)
e5dc01128ac5 [AMDGPU][True16] Add legalization/selection handling for G_MERGE_VALUES of 2 s16 -> s32 (#200082)
df697a65e673 [CIR] Set the linkage properly for the _GLOBAL__SUB_I_ function (#200443)
2520a9bcd091 [APInt] Add `APIntOps::compressBits` and `APIntOps::expandBits` (#200114)
2f2122a14381 [libc][NFC] Add few options to allow users to skip building and running some tests. (#199474)
4472e935678a [LLDB] Skip test with older versions of clang (#200456)
77800224eba1 Reapply "[PDB][llvm-pdbutil] Add DXContainer support for `pdb2yaml` and `yaml2pdb`" (#200413)
c241374079f1 [clang] Fix @tparam warnings for concept template parameters (#199800)
3c34c34901d6 [aarch64] Add more MSVC intrinsics (#194632)
912f0d1055b8 [HLSL] `constexpr` vector element conversions (#195173)
5f71480b522e [flang][OpenMP] Fix USE-associated declare reduction symbol resolution (#200328)
0ad051a0a338 [AMDGPU][MIRFormatter] S_WAIT_LOADCNT_DSCNT human-readable mask (#200059)
398573eecb37 [lit] Make MetricValue a proper abstract base class (#200187)
bf8e8c221f26 [llubi] Map tags into provenances (#200236)
7dcce4c16da1 [lldb][windows] fix scope_exit warning (#200439)
a8a988bd6a5b [CIR]Fix CIR Test failures after unnamed_addr removed from vtables (#200435)
d043ec846b79 [lldb] Avoid trie prefix recompuation (NFC) (#200026)
15398bac0def Revert "[InstCombine] Fold zext into de-interleaving (factor=2) instructions (#195330)"
7ab48632ce24 merge main into amd-staging
c4346f2a09aa Ignore `!guid` metadata when merging globals or constants (#200323)
6c9542c3f99e Add NVPTX DW_AT_LLVM_language_dialect emission. (#200003)
99b5ebd51f49 [InstCombine] Replace undef with poison in a test (#199757)
dc40fccbc1f8 [WebAssembly] Avoid crash in LateEHPrepare with empty cleanup pads (#200322)
6ce211f02fcf Revert "[DenseMap] Replace tombstone deletion with TAOCP 6.4 Algorithm R" (#200421)
cea75bdb43b7 merge main into amd-staging
a57049a7f825 [AMDGPU][True16] Upstream GlobalISel True16 tests/run lines (#200288)
fa6f7426a399 [clang-sycl-linker] Decouple link step from SYCL; adopt -L / --bc-library options (#199777)
00062ed98225 Revert "[NFC][AMDGPU] Introduce AMDGPU::FlatVariant enum to replace uint64_t FLAT discriminant" (#200420)
8d4e7d9f4e2e Add new coro test to profcheck-xfail.txt (coro aren't yet profcheck ready (#200422)
22a95c1b204d merge main into amd-staging
c4a6fe766e2f [lld][LoongArch] Handle DTPREL relocations in debug sections (#199327)
170bdf55faef [NFC][AMDGPU] Introduce AMDGPU::FlatVariant enum to replace uint64_t FLAT discriminant (#200396)
bdcce4a12c82 [X86] combineMaskBitOp - fold vXi1 logicop(truncate(N0),truncate(N1)) -> truncate(logicop(X,Y)) (#200385)
4db74c054e2b [NFC][AMDGPU] Replace uint64_t SDWA inst-type discriminant with typed enum (#200388)
a8a91c01b734 [SYCL] Fix -nolibsycl and spurious spirv-link flags for SYCL offloading (#200252)
afe61654c0e8 [Clang][Sema] Fix crash when calling EvaluateForOverflow for UnaryOpe… (#200317)
3791c8b87737 [flang][FIRToMemRef] Non-box complex slices use shapeVec strides, not `fir.box_dims` (#200285)
733d188a9548 [AArch64][llvm] Restrict STSHH decoding to valid policy encodings (#200389)
f9207d7c4b48 [libc++][NFC] Move functions in shared_ptr.h into the class body (#200343)
7fba4dba57ac [NFC][AMDGPU] Replace getClampMask with hasSameClamp in SIInstrInfo (#200382)
0fd11e4a83e8 Honor two's complement signed overflow with -fms-compatibility (#198538)
a64bc13adcb5 [ValueMap] Simplify insert methods (NFC) (#200264)
4b703fb5fbb9 [lldb][windows] enumerate volumes via FindFirstVolumeW (#200231)
318852be473b [Comgr][Cmake] Fix incbin detection (#2702)
f545a3b3f558 [lldb] Make PTY initialization idempotent (#197717)
c7d072c0d046 Fix assert in SlotIndexes::getInstructionIndex in debug builds (#189370)
4a955e932e3a [AMDGPU][GIsel] Legalize rules for cvt_scalef32_sr_pk_fp4_{f16,f32} intrinsic (#193176)
13ccc7c5bfc6 [CIR] Handle vptr zero-init- (#200305)
a182b4bed2c4 [lldb][windows] fix command source hitting EOF (#194950)
3c0e4ece789d workflows/ci-post-commit-analyzer: Pin container image (#200293)
235516489f0d Merge branch 'amd-staging' into fix/gpu-profile-runtime
7c34ef68f75b merge main into amd-staging (#2716)
8466750c230d [flang][coarray] Fix error fir.store mismatch memory reference type in genCoBounds #193131 (#194870)
2359e773d7e4 [Mips] Add 'j' alias for r6 (#196706)
5e89f5217299 Revert "[PDB][llvm-pdbutil] Add DXContainer support for `pdb2yaml` and `yaml2pdb`" (#200380)
f9cab1481053 [lldb][windows] allow longer paths than MAX_PATH in GetFileNameByLoadAddress (#200225)
78eca55b542b [SystemZ] Fix off-by-one error in backend (#200141)
8785f7cea2ce [LLVM] Precise error message for intrinsic signature verification (2/n) (#199217)
040ee4af4d25 [VPlan] Add m_Shl matcher. (NFC) (#200376)
5154b695ec1c [clang][bytecode] Use in `VarDecl::evaluateDestruction()` (#199646)
8c4a5318ae93 [PDB][llvm-pdbutil] Add DXContainer support for `pdb2yaml` and `yaml2pdb` (#198351)
caa7b801d217 [X86] avx512-mask-op.ll - add #198162 test coverage based off "Eval4Inputs" example (#200373)
394377e08e4c [Infra] Fix to get the submodule name instead of path for teams notification (#2690)
421cfead959f [clang][bytecode] Reject dynamic_cast on constexpr-unknown pointers (#200363)
9f34092cbdc0 [C23] Combine two proposals on C status tracking page (#200239)
e0282749ff6b [LV][NFC] Factor out vectorization report functions to a separate namespace (#194951)
472e7ed971bd [lldb] Treat `__this` from CodeView/PDB as captured `this` (#200271)
1cd28beb7755 [AMDGPU] Have VCC as a first-class member of the SGPR pool. (#173870)
edb7a0487060 [DAG] isAllOnesOrAllOnesSplat - handle implicitly truncated constants (#200209)
fad51d3f4148 [libc++] Only push/pop diagnostics for ABI annotations pragmas if not in a system header (#200338)
458e23591c4d [LLVM][InstSimplify] Ensure scalar sincos constant fold only triggers for scalars. (#200229)
354cdd40be87 [compiler-rt] [ARM] Restore Windows specific names for fp->int functions (#200360)
1cf8469545ab [NFC][ValueType] Add is*VectorOf(VT) helpers. (#200189)
e7a1a60f63b0 [SLP][REVEC] Skip external-use extract cost for insertelement buildvector roots
cfeeab33e66c [SystemZ] Remove unused variable warning (#200359)
3843c160526f [clang][bytecode] `Pointer::computeOffsetForComparison()` can fail (#199987)
7a4d7733ae55 [ELF] Fix location counter used for output section address with MEMOR… (#197293)
afddfe074afb [lldb][NFCI] Cleanup AppleObjCClassDescriptorV2::class_ro_t API (#200213)
7527985fd76c [AMDGPU] Use direct lookup table for VOPD eligibility (#200241)
73080a6c3b5c [AMDGPU] Remove _dpp variants from VOPD tables (#200184)
2c49ef885bae [VPlan] Support URem in getSCEVExprForVPValue. (#199794)
0850f6e62ea0 [CodeGen][NFC] Refactor EmitAsmStmt to reduce header churn (#199377)
386dac00fba8 Revert "[Clang] Mark new as inaccessiblememonly if sane" (#200349)
0d8316a53da5 [Compiler-rt][CMake][BoundsSafety] Add `COMPILER_RT_ENABLE_TEST_SUITES` to control which test suites are enabled (#196385)
bafde6fbb860 [Flang] Move builtin .mod generation into runtimes (Reapply #137828) (#171515)
2e10a9f2e95a [clang][bytecode] Check Mulc op for numbers (#199657)
c2e39974c8e2 [CoroSplit] Never cross suspending points when `coro.is_in_ramp` is true (#198226)
f7fa98ea57dc [llvm][docs] Correct link to libc maintainers file (#200348)
2d046bd62066 [LSR] Improve the recognition of vscale immediates (#196998)
a0f7a62bb6a2 [clang][doc][SYCL] Add a very first SYCL release note (#200200)
d7703c6aab9f [libc] Add byteswap.h to Linux public header target lists (#200345)
e8772d99d89a [Maintainers] Fix Markdown heading level (#200340)
eaf5ba7478e7 [SPIR-V] Recognize reqd_sub_group_size metadata for SubgroupSize (#199521)
9d47197593c6 [libc++] Use variable templates instead of class templates in shared_ptr.h (#199481)
3ca81aa6b6d0 [VectorCombine] Don't scalarize atomic loads (#200263)
971ad593d6a2 [AArch64][GlobalISel] Add handling for scalar sqneg intrinsic (#200211)
3648f148c70b [LLVM] Add myself as maintainer for LoopInterchange and related passes (#200335)
f7a2267de9e3 [IR] Skip remove_if scan for empty maps in removeNotPreservedAnalysis (#200292)
64bc1fae11d6 [SystemZ] Global Stackprotector and associated location section (#169317)
8f45cf9822a5 [SystemZ] Handle pointer size correctly in getVectorBitmaskConversionCost(). (#200171)
b159ed6fa212 [LoongArch] Emit DTPREL relocations for TLS debug info (#199326)
01c93f511b4b [Clang] Mark new as inaccessiblememonly if sane (#197199)
b835d5a98f48 [IR] Only require nofree in canBeFreed() (#200191)
bbf8a33ad5f5 [mlir][spirv][tosa] Extend TOSA to SPIR-V TOSA op conversion (#200009)
4c1bc5936536 [clang][bytecode] Fix variadic operator calls with explicit this (#199978)
05bcb056f056 [RISCV] Combine isGPRPairCopyCandidateEven/isGPRPairCopyCandidateOdd in MoveMerger. NFC (#200254)
6bee6b09e21c [BPF] cli option to allow libcalls (#199542)
891b8f6fa8e7 [BPF] crash fix for memcpy with unsupported alignment (#199541)
7d7f98b11b9c [CodeGen] Stop setting vtable as unnamed_addr (#200108)
bf8dabd88282 [HIP][CI] Add HIP Kokkos bot recipe (#200201)
021b93e7e763 [lldb] Make TestSBModule a no-SHARED_BUILD_TESTCASE (#188265)
767f01e0910d [libclc] Use float sw_fma for spirv*-mesa-mesa3d targets (#199626)
51efc6333dc9 merge main into amd-staging
479818a38ccf [Profcheck] Update xfail list after b378c7162613afdc982e8a2e6ec478066ea04492 (#200326)
023e7decf625 [RISCV] Skip unrelated COPY during Zdinx/P instruction generation in RISCVMoveMerge (#200219)
b378c7162613 [profcheck] Fix profile metadata propagation for Large FP Operations (#175864)
01a92098491a [CIR] Add emitBuiltinWithOneOverloadedType helper (#199518)
561c528b5e19 [RISCV] Extend vp.reverse vp.load/vp.store combine to non-VP splice + reverse (#180907)
3cb8e7215f2d [AMDGPU] Address follow-up concerns from 189121 (#192322)
4c33844b4b56 [PGO][HIP] Skip ROCm interceptor in profile-only compiler-rt builds (#200111)
7672a992182d [Bazel] Fixes 7964b66 (#200316)
5047ae21b426 [Docs] Note change to Maintainers.md in DeveloperPolicy (#200315)
7964b66bf55d [Instrumentor] Move instrumentor stub library into its own header file (#199329)
4910eda2a04e [FlowSensitive] [StatusOr] Support ABSL_RETURN_IF_ERROR (#200287)
843d4a6a27a0 [Instrumentor] Add numeric operation instruction instrumentation support (#199749)
46d7b19bd77f [cov][prof] tweak amdgpu to not be in ubsan
3f140498c2d7 [Instrumentor] Improve stub printer (for C/C++ and value packs) (#198366)
a853917bbf7b [Instrumentor] Fix nondeterminism in indirect argument alloca usage  (#200289)
8d8ebb1cb1b5 [SLP][REVEC][NFC]Add a test with the missed fma vectorization, NFC
75fb8e47f617 [lld][nfc] Fix spelling (#200279)
f85729489de0 [Docs] Improve Documentation on Writing IR Test Cases (#200259)
7e200ebdd08c [AggressiveInstCombine] Support Len==8 in tryToRecognizePopCount2n3. (#199551)
ee40c5e35152 [flang] Fix WHERE self-updates on allocatable array sections (#200281)
58dff9282806 [llvm-offload-binary] Fix test failures on z/OS by splitting malformed tests (#200258)
b1b20686afeb [compiler-rt] Fix multi-device crash in GPU profile runtime and add HIP tests
437d68d4f855 [compiler-rt] Restore GPU profile runtime build flags dropped during merge
927c51d2434e [lldb] Improve frame variable handling of recognizer arguments (#200084)
075e16f74d7b [CIR] Correctly emit the size expr of a VLA with a 'bool' size. (#200066)
f7f3d9d0aea8 [lldb][Windows] Move `processthreadsapi.h` include (#200278)
01cfbaabef6c workflows/upload-release-artifact: Use require-release-manager action (#195374)
642bbbaf57ed [2/3][RegAlloc][LiveRegMatrix] Fix inconsistency when fold creates a CopyMI (#197776)
0d1e9a6177d9 [X86][TTI] Retrieve address space from intrinsic info. (#200265)
2f98f171c298 [PowerPC] Fix MSan failure in LowerBUILD_VECTOR (#200260)
23776bd325ec [llvm] Fix LLVMOrcTargetProcess symbol export with MinGW/Cygwin shared libs (#174266)
5c219c86b6ff [Instrumentor] Add cast instruction instrumentation support (#198224)
b41aef8b809c [CIR] Omit nsw/nuw on integer vector binops (#199123)
cdb606c37c8f [VPlan] Predicate SCEVs in getSCEVExprForVPValue (NFC) (#199994)
b2edcf9fb336 [Instrumentor] Introduce BasePointerIO to communicate base pointer information (#197607)
810e67f50d8f [flang][OpenMP] Fix crash on standalone ordered with depend(source|sink:) (#200193)
42010078e750 [MemCpyOpt] Keep volatile memset before memcpy (#200100)
63da2695b1a3 [VPlan] Strip IR-PatternMatch include in VPlanAnalysis (NFC) (#200186)
46c8c1e1320e [AggressiveInstCombine] Factor common code out of tryToRecognizePopCount and tryToRecognizePopCount2n3. (#199440)
e7ea47d98e7e worflows/premerge: Remove template expansion (#200068)
3df7e3707c2a [SandboxVec][DAG] Implement missing API for successors (#195301)
3769bd61d0b3 [RISCV] Use std::optional::operator* instead of value(). NFC (#200235)
f089890d5e20 [clang-tidy] Fixes false positive with a non-const method on a pointer (#188844)
4df74706ab47 [MLIR][NVVM] Split nvvm.barrier into nvvm.barrier and nvvm.barrier.reduction (#199404)
060ef6a90c39 [lit] Enforce read-only behavior on lit_config.maxIndividualTestTime (#198193)
d69c732e6e71 CI: ignore errors on HLSL Publish step (#200240)
5f49cff125e2 [AArch64] Add tests for and+xor+shift. NFC (#200250)
184a8f94757c [libc++] Applied `[[nodiscard]]` to `array::iterator` (#198492)
09ad10052903 [lldb][NativePDB] Use decl context from a method's class (#199221)
9c640504466a [AArch64] Fixup f16->bf16 in intrinsic defs. NFC (#200244)
a0ac75213425 [scudo] Create a non-static getErrorInfo function. (#199770)
742b750f5e2b [lldb][Darwin] Save "most recent SDK root" in local var (#200095)
e1d2b5b52d80 [z/OS][tests] using aliases on z/OS are not supported (#200233)
6dec1e526f03 [AArch64] Add more tests for generating mull from disjoint or. NFC (#200237)
192601e8b3ad [RISCV] Update the ProcResource used by vsetvli in SiFiveP400 model (#199802)
87a663b94aa6 merge main into amd-staging (#2706)
45304733791f [LifetimeSafety] Improve dangling field/global diagnostics (#200147)
a0150ce371d1 [flang] Canonicalize fir.array_coor for contiguous arrays. (#200106)
f30438b2ab6c [AMDGPU] Add an early exit in `tryMatchVOPDPairVariant` (#200148)
743cd53e67d2 [AMDGPU][GlobalIsel] Add regbank support for amdgcn_cvt_scalef32_sr_f8_f16/32 (#197031)
ab0178ce058b [ExpandVariadics] Fixup handeling of mismatched return types (#198843)
72871f6fa1f1 [libc++] Fix multi{map,set}::extract not returning the first matching element (#199703)
9092493752a2 [Clang][Sema] Use correct DeclContext when checking 'this' (#163243)
0aa9ec6a3610 [InstCombine] Fold zext into de-interleaving (factor=2) instructions (#195330)
3613e27b4c6f [LifetimeSafety] Add fix-it for misplaced lifetimebound attributes (#199149)
f1c26bce3f30 [libc] Implement towupper and towlower entrypoints (#198659)
218a21cee57f [gn build] Format all files (#200226)
58e35f930699 [gn build] Port commits (#200224)
934e5610ebfa [libc] Don't touch str_end in strto* and wcsto* functions when base is incorrect (#200073)
7e539a4497fb [VPlan] Thread scalar types through VPWiden(Cast)Recipe. (NFC) (#199572)
b76d08a8fc46 [RISCV] Factor out common scheduling model part between SiFive P400 and P500. NFCI (#199796)
f505ceabdd23 [clang] fix member specializations of class and variable partial specializations (#200092)
839091f0f298 [lldb-server][Windows] add support for thread name (#199983)
020b4dde0f8c [CI] Add 'lld' to LLVM projects in build script for lldb windows bot (#200207)
35a10827d093 [analyzer][NFC] Merge the class `LocationContext` into `StackFrame` (#198211)
ac2c011dd3b1 [offload] add SKIP_KNOWN_FAILURE unittest macro (#196275)
aff129e825b5 [AArch64][GlobalISel] Add support for pmul intrinsic (#198809)
10497d20fc2f [IR] Remove options for vector of null representation (#200172)
10f84995b430 [lit] Modernize ParserKind implementation using Python3 IntEnum (#199965)
2c8ca9679e85 [OMPT] Add callback for `omp_target_memset` calls (#194168)
ea4cd9228f87 Revert "[AMDGPU] Stop coercing structs with FP and int fields to inte… (#199981)
c4d820ce311b [DenseMap] Replace tombstone deletion with TAOCP 6.4 Algorithm R (#199615)
67ac1b07a38f [RISCV][P-ext] Support v4i16/v8i8 build_vector on RV32. (#199883)
c34220a5709d [RISCV] Use append TableGen feature in RISCVInstrInfoP.td. NFC (#200110)
8fea28089a91 [DirectX] Check DXContainer's header magic during parsing (#198586)
90aece05fdd4 [AMDGPU] Do not combine V_ASHRREV_I16* to make sdwa (#198491)
ba7d42b096cb [PowerPC] Optimize vec_splats of small FP values on Power8/9 (#199538)
7948d336a9eb [lit] Refactor super() calls in ResultCode (#199891)
96b6bb78aa04 [DAGCombiner] Loop unfreeze step in visitFREEZE freezeOtherUses (#200120)
3beee85e5244 [flang][OpenMP] Fix declare reduction accessibility in module scope (#197078)
7ac0a3cec9bf [lit] Replace zip(range(len(x)), x) with enumerate(x) in ProgressBar (#199884)
082ac6e3eaaa [lit] Optimize ShLexer string construction (#199641)
c20c666e7df2 [lldb][NFCI] Cleanup AppleObjCClassDescriptorV2::objc_class_t API (#200180)
3e92883e83f7 [clangd] Remove redundant symbol name from hierarchy item details (#170112)
f24ee22550de [mlir][EmitC] Include DeallocOp in AllocOp memref conversion tests (#198275)
8ab00f2c21e5 [z/OS][tests] XFAIL using aliases on z/OS (#200176)
a16511c2d969 [clang][modules-driver] Precompile std modules independently of -o and final phase (#199289)
5536348d0600 [lit] Handle config loading safely (#200168)
3c21a0cbad43 [SPIR-V] Add s128 to allPtrsScalarsAndVectors in legalizer (#199998)
b43dcbeffe02 Implement SPV_KHR_poison_freeze extension (#198037)
34e7ecaf872f [IR] Avoid caching a DenseMap reference across erase in handleOperandChangeImpl. NFC (#200179)
2766733764f4 [compiler-rt][profile] Add COMPILER_RT_BUILD_PROFILE_ROCM option (#200127)
992ee5934c11 Add support for Alpha to libsanitizer (#194161)
d5edbe411d8f workflows/upload-release-artifact: Validate input and remove template expansion (#199972)
5d29057a5876 actions/push-container: Remove template expansion (#200062)
c8ef5dcfae34 [VPlan] Directly pass SrcEltTy when constructing WideGEP (NFC) (#200043)
ebe8595c0012 [lldb][NFCI] Cleanup APIs in AppleObjCClassDescriptorV2 (#200122)
da4e5cc928c8 [SCEV] Canonicalise round-up idiom when some bits known (#197126)
ee3709b5db6f [AArch64] Lower SUMLA via two udot products on plain dotprod targets. (#199761)
7fb99eefbc6e Emit newline after IR-dump banner in PrintCallGraphPass (#199410)
34f294ccbca8 [flang][OpenMP] Event handles are not predetermined shared (#200055)
9f82a4e41f9a [AArch64] Do not generate indexed addressing mode for volatile accesses (#196305)
98c7814ca93b [llubi] Add basic support for provenance modeling (#185977)
4471b6236315 [GitHub] Add InstCombine Contributor Guide to new contributor greeting comment (#199730)
a829e4e3eb4f merge main into amd-staging
3e8fc3561fb5 [CIR] Implement 'coroutine' exception handling lowering (#200045)
61b5b08623e8 [libc] Move fixed buffer GPU test to an integration test (#200042)
9107dacd6e41 llvm: Fix most LLVM_ABI annotations in Analysis (#199019)
62ac21e810af merge main into amd-staging (#2698)
5b32c6e99e8d [InstCombine] Drop the correct assume when working on assume bundles (#198404)
8ecfd9b96df4 [BUILD]: Fix for af772866 (#200158)
5cac2751fb9c [LoongArch] Add `-fstack-clash-protection` support (#195595)
f6758a5664a6 [flang] Fix -E -dM macro dumping for stdin and .f90 inputs (#200144)
4683f6ba6152 [lldb-dap] Fix data race on disconnecting (#200017)
ff7e6bbb7f11 llvm: Fix most LLVM_ABI annotations in CodeGen (#199921)
dbdbb8d6afd8 [libc++] Simplify the implementation of conditional a bit (#199916)
326a1320dc98 [LSR][AArch64] Precommit tests showing lack of `mul vl` addressing (NFC) (#200149)
dbebcb02d3d8 [mlir][spirv] Remove unnecessary assertion (#200137)
eb1b57c63994 [LangRef] Specify that syncscopes can affect the monotonic modification order (#189017)
a554a1989f60 [DirectX] Generate shader debug file name part in llc (#199555)
b115dd117ac1 [AArch64] Fix definition of system register move instructions (#185709)
51d823197cb4 [TySan] Expose __tysan_set_type_unknown interface (#198800)
66ee43ce9d66 [LifetimeSafety] Improve diagnostics for use-after-scope (#200031)
172ea47e6d5b [AMDGPU] Use separate tables for VOPD3X and VOPDY in `getCanBeVOPD` (NFC) (#199072)
cee1c3ba576f [mlir][tosa] Limit consecutive concat rewrite to MAX_TENSOR_LIST_SIZE (#199051)
174eb79572b3 [compiler-rt][ARM] Optimized FP -> integer conversions (#179927)
86a74788dcc8 [X86] matchBinaryPermuteShuffle - match to X86ISD::SHLD funnel shift patterns (#200136)
1235c442a3c0 Fix Bazel build for 5dc8711 (#200142)
0e02ccdaf010 llvm: Fix most LLVM_C_ABI annotations (#199914)
4503872fe573 [IR][FunctionAttrs] Clarify memory effects of atomics (#193768)
8a9a82c53e82 [VPlan] Remove redundant vp_merge pattern. NFC (#200116)
0513c9cba2a0 [Bazel] Fixes 5dc8711 (#200129)
a63eacee2db2 [VPlan] Don't use Ingredient to get type in VPWidenMemoryRecipe (NFC) (#200049)
3a50239454f8 [GlobalISel] Replace `GIM_CheckFeatures` with `GIM_Try_CheckFeatures` (#198461)
597b1376bab3 [mlir][SPIR-V][complex] Convert complex.add/sub/mul/div to SPIR-V ops (#200123)
626f4061e7f2 [mlir][SPIRVToLLVM] Add conversions for more GL ops (#200115)
538140ff2cb7 [clang] Add missing BuiltinAnchorSources.def to module map (#200125)
af772866888e [mlir][spirv] Add SPIR-V NonSemantic.Graph.DebugInfo (#199519)
b51a8025bed1 Revert "[clang] remove lots of "innocuous" addrspacecasts" (#199685)
5dc8711b4e55 [mlir][mem2reg] Promote memory slots through transparent view operations (#196924)
5c95f6a85939 [LLD] [COFF] Fix handling of immediates in ARM64_SECREL_HIGH12A (#200060)
4306b4ab9c12 [libc++] Implement P3508R0: Wording for "constexpr for specialized memory algorithms" (#197313)
35bfc00c5c8d [InstCombine] Fix type mismatch in `foldBitmaskMul` (#199920)
91d765f6ab10 [FunctionAttrs][Attributor] Update nofree inference (#196266)
6ed05a0b53dc Revert "[Clang] prevent constexpr crash on invalid overrides" (#199895)
942292bf7516 [orc-rt] Remove NativeDylibManager unload operation. (#200119)
486370c33fbd [AMDGPU][Clang] refactor addrspace and scope checks [NFC] (#199175)
191aa295d9c9 [LangRef] Do not allow free via synchronization in nofree (#195658)
8059bc5944b7 [clang][bytecode] Fix a diagnostic difference in bitcasts (#197174)
690a25fd5b0b [clang] Don't optimize out no-op atomics in kernel mode (#193562)
9cd3c0b99e5d [HLSL] Codegen for handling global resource array initialization (#198891)
48a1ee7bc8e1 [AMDGPU] Remove redundant s_wait_xcnt after implicit XCNT drains (#198823)
5cee11e45226 merge main into amd-staging (#2697)
1da70ad6da64 [clang-tidy] Fix false positive of parentheses removal for overloaded operator (#192254)
846bfd77ddc4 [LifetimeSafety] Avoid assert on variadic placement new (#199588)
36ea901c3433 merge main into amd-staging
5acb952cb5d9 [HIP][Driver] Forward -fcoverage-mapping flags to device compiler (#198872)
da4894c874ed [LoopFusion] reject unsafe scalar flow dependences (#195895)
a20e85f74d7a [clang] NFC: add test cases from #111561 (#200105)
5b2ec64ecf02 merge main into amd-staging (#2695)
d5e97d77289c [Driver] Honor /Fo when deriving the split-dwarf .dwo path (#199613)
635e120fb873 [PGO][HIP] Stop pulling ROCm.o into every PGO host link (#200101)
2d5dac5b8167 [Bazel] Fixes 5db1364 (#200104)
f9185454a9a7 [clang][AMDGPU] Fix -ast-print crash on expanded predicate builtins (#199963)
412f238f2435 [AMDGPU] comgr: split non-stride64 DS 2-address ops for A0 (#2281)
6619c82f7294 [RISCV] Fix incorrect CM.MVSA01/QC_CM_MVSA01 generation with Zdinx (#200000)
5b17cdb0b582 [RISCV][P-ext] Split v4i16/v8i8 INSERT/EXTRACT_VECTOR_ELT on RV32. (#199917)
0381a09309c6 [SLP] Precommit tests for runtime strided stores (#200019)
dbe6800baeec [AMDGPU] This reverts patches to use fp16 inline constants for i16 (#200091)
d65897289e68 [alpha.webkit.UncountedLocalVarsChecker] Detect an assignment to a guardian argument (#198695)
740e52bc9969 [DirectX] Drop debug labels (#197490)
e3ec3f62b457 merge main into amd-staging
830c8d625421 test(llvm-symbolizer): fix Wasm layering violation by using YAML (#200080)
1a5822e3c942 [coro] Use C calling convention for C++20 coroutines (#198943)
9cc6c9375faa [libclc] Optimize and vectorize signbit (#199497)
1670d39972a1 [DirectX] Handle undef the same way as null (#197507)
5db13643f4b7 [PGO][AMDGPU] Add basic HIP offload PGO support (#177665)
e69fb46030a0 [gsymutil] Disable readahead in `GsymReader::openFile()` (#199230)
f94c912511b6 merge main into amd-staging (#2693)
9d1644738134 [DirectX] Add an "offset" operand to llvm.dbg.value (#197478)
9409c07de637 [lldb][Darwin] Read Mach-O binaries out of memory more efficiently (#200072)
eb0cb0c69526 [LifetimeSafety] Propagate inner origins through std::move and related casts (#199600)
f13305bac8ea [VectorCombine] Fold deinterleave2 with smaller effective element size (#192121)
7f82efb39668 [Object] Add missing BBAddrMap.def to module map (#199961)
00b4695f1dc6 [CodeGenPrepare] Use recomputed split-branch weights. (#199822)
e8cc37e60d04 [clang][deps] Disable app extensions during scanning (#200041)
725425314070 merge main into amd-staging
e1dd984a809c [flang][FIRToMemRef] fix stride calculation for complex lowering (#200035)
8fc12f1e3f6a [flang-rt][cuda] Set RT_CUDA_THIN_IO=1 only for PTX object (#200063)
9fa9ab338f75 [RISCV] Remove unused Predicates for Zicfilp. NFC (#200028)
055a4ba0b277 [clang] Update C++ DR status page (#200053)
78987e85ae09 [lldb] Edits and clarifications to DataFileCache comments, NFC (#199787)
5ae73b942f44 [lldb] Keep addr for Memory Modules separate (#199810)
b6161b9f4c9e [libc] Add missing struct_mmsghdr dependency to sys_socket (#200051)
5a7e4a0cf7f5 [AMDGPU][MC] Replace shifted registers in CFI instructions (#183147)
4e39ea3d5e73 [AMDGPU] Implement -amdgpu-spill-cfi-saved-regs (#183149)
4fe4d9844862 [IR] Handle `expected` tag in switch branch weights. (#200025)
506ac5cde6f6 merge main into amd-staging
b0516c58736c [SeparateConstOffsetFromGEP] Set `inbounds` correctly. (#199304)
9ca1941d8ea2 [flang-rt][cuda] Move thin I/O to ptx only (#200054)
bf005a1227a4 [alpha.webkit.UnretainedCallArgsChecker] Emit a warning for a non-const RetainPtr member (#184243)
2f0815059c7b Reland: [lit] Move maxIndividualTestTime from global to test suite config (#199996)
f95f02c4b1ac [VPlan] Thread scalar types through VPReplicateRecipe. (NFC) (#199379)
c315c662cd2d [AMDGPU] Fix codesize estimate after #198005 (#200033)
c00484d99f36 [llvm-profgen] Fix -Wunused-variable (#200038)
5c0ff67e0c4a [CI] Add SPIRV-focused PR CI workflow for amd-staging (#2451)
b76503229dc3 [alpha.webkit.UncountedLocalVarsChecker] Detect a raw pointer/reference in a decomposition declaration. (#198582)
99bc986c527a [Support] Add MemoryBuffer::randomAccessIfMmap() (#199265)
bdb3232dfacb [AMDGPU] Implement CFI for CSR spills (#183150)
1286670a2156 Fix for noassert buildbot break in #183153 (#199781)
901ebeb3da80 [AMDGPU] Fix SuperReg to MCRegister conversion (#199993)
8dc8411a7570 [Bazel] Fixes 3d9d776 (#200029)
15113c3eedb1 [libc++] Update the base image hash to get GCC 16 (#199882)
59389b0ba9de [libc++] Add __is_transparently_comparable_v optimizations for set, multiset, and multimap functions [Issue 187105] (#189735)
80393ba1bdbe [llvm-profgen] Fix assertion condition for the top-level probes address range checks (#198674)
57b485f07d94 [AMDGPU] Use register pair for PC spill (#183146)
3a203a506691 [CUDA][HIP] Defer device diagnostics in implicit H+D explicit instantiations (#197214)
eeeb26f64efd [CIR] Emit globals for declarations that force externally visible defs
d0676f53ed5b [CIR] Handle throwing an exception from a cleanup scope (#199121)
aa1f119c6dcb [clang][lit] Add option to skip clang-repl checks (#199255)
8a55aca421d6 [SandboxVec][SeedCollection] Iterate over all seeds (#195964)
3d9d776b29c9 [bzl] Reduce the `deps` size of libc's shared_math_header library. (#200006)
0e84bfa79f11 [CIR] Handle the 'before case' block of a switch statement. (#199752)
2b3bc03b5ef0 [AMDGPU] Use shorter form for i16 operands (#198005)
0791b7d8e99a merge main into amd-staging
6c9225f1d1e2 Reapply "[clang] Use FileError in FileManager::getFileRef, getDirectoryRef" (#199759)
34d5523eaa8f [Driver][MSVC] Loosen regex for binary name in test (#200015)
8f03570e7d04 [AMDGPU] Add regression test for wave.reduce constant folding. (#198673)
c8efcf55ee47 [flang][PPC] Improve vector type names in expression diagnostics (NFC) (#199383)
f828c0e39fca [CIR] Report NYI for STDC FENV_ACCESS under -fclangir
a3d935402420 [CodeGen] Compare MMO atomic ordering and syncscope. (#199892)
0d8ba697df74 [bazel] Add config for hermetic clang toolchain (#192528)
3a8b5e3bb1ae [lldb-dap] Use MainLoop instead of a background thread in OutputRedirector. (#199970)
d62792499205 [mlir][SliceAnalysis] Fix visited set to avoid infinite recursion  (#200008)
f8bf8af12c81 [WebAssembly] Add f16x8.demote_f32x4_zero to wasm_simd128.h. (#199795)
5dc633bc765a [AArch64][GlobalISel] Add BF16 fabs and fneg (#198655)
f16c0ecf2a0b workflows/issue-release-workflow: Remove template expansion of login names (#199772)
a8e1f5c475ab [flang-rt][cuda] Use a thinner I/O in CUDA build (#199769)
5b38edd3723a workflows/pr-code-lint: Pin container image (#199767)
ed918c1496d5 [AtomicExpand] Preserve volatile in widenPartwordAtomicRMW. (#199722)
82755077b739 [ProfCheck] Fix #199174 (#200013)
f6af6cedfd2f workflows/libclang-abi-tests: Remove template expansion (#199792)
7a1a5421b1a1 workflows/release-documentation: Validate input and remove template expansion (#199760)
077210a3b1a5 [gn build] Port commits (#200011)
a1cba5a2f8e9 [flang][OpenMP] Optionally get final symbol in Get(Argument|Object)Sy… (#196816)
00b13536e3da [Flang][OpenMP] Support declare reduction without initializer (#196211)
d98fd416e0ee [ExpandIRInsts] Support llvm.fpto{u,s}i.sat (#199174)
aac106b7146e [AMDGPU] comgr: Add DS_ADDTID trampoline patch for A0 16-bit M0 truncation (#2302)
7b993d207fee [InstCombine] Use sadd.sat for chained ldexp fold (#199274)
860e4b803e91 [X86][AvoidStoreForwardingBlocks] Skip volatile/atomic accesses. (#199698)
e1e52c9ccef3 [win][x64] Updated `llvm-objdump` and `llvm-readobj` to be able to dump Windows x64 Unwind v3 information. (#199120)
2713d9441cfd Reapply "[SandboxIR][Tracker] Implement accept(/*AcceptAll*/) and revert(/*RevertAll*/)" (#199776) (#199805)
db9b595ae3b3 [X86][APX] Add CodeView register IDs and mapping for APX EGPR (#199586)
ea3cb0fef970 Fix for noassert buildbot break in #183153 (#199781)
3768e1329807 [X86] Add test coverage for #145276 (#200004)
95f08b1b743d [RISCV][P-ext] Make the direction argument for RVPPairShift* classes required. NFC (#199799)
38555dbd81ad [RISCV][P-ext] Replace some custom isel code with tablegen patterns. NFC (#199881)
358f5e770361 [RISCV][P-ext] Add missing let Inst{31} = 0b0 to RVPPairShift_rr. (#199885)
7c11b9f4aeb8 [Comgr] Fix race-condition on compilation/execution of Comgr's lit tests (#2689)
a3acd8083c3b [MLIR][XeGPU] Propagate layout onto loop-carried iter_arg entry edges (#198862)
c3fc506e29ba [AMDGPU] Remove explicit PartialThreshold setting in loop unrolling (#198901)
261dbebd7b01 [NFC][Comgr] Remove dead variables from tests (#2688)
97b9c24cb9f2 Add missing annotations for Apple platforms (#198864)
dd2ce3d8b558 [LLVM] Add per-target runtime directory to rpath (#199755)
b65a71d4942a [libc][bazel] Add rules for __support/threads tests. (#199871)
06a6cbb18846 [AMDGPU] Fix SuperReg to MCRegister conversion (#199993)
e5ad7f7e700e [flang][OpenMP] Remove ompFlagsRequireMark from symbol resolution (#198591)
bf8dae18aa31 [InstCombine] Narrow llvm.abs through trunc. (#199643)
3e8c217c6829 [AMDGPU] Fix ShiftAmt32Imm to use unsigned comparison (#199052)
dcf50fe35fd0 [SystemZ] Don't fold memops after SSA if tied regs don't match. (#197475)
6cb00b6c344c [Hexagon] Fix up vector predicate before compressing it for bitcast (#199283)
c0a302d416aa [AMDGPU] Refactor insertRelease into insertWriteback + insertWait (NFC) (#199486)
6170eeb1fa81 [flang][OpenMP] Fix copyprivate crash with unlimited polymorphic pointer (#199768)
9757718b591f [Bazel] Fixes 81c523c (#199988)
d25ec4b84487 [llvm][RISCV] Support batched dot product extensions MC layer (#196467)
9e279e4757b5 [X86] vector-shuffle-combining-avx512vbmi2.ll - add VLX/NOVLX test coverage (#199984)
f55acff22c41 [lld][ELF] Exclude SHT_NOBITS sections from LMA overlap checks (#196423)
7664fc742fa6 [Offload] Fix missing SONAME version on offload libraries (#199975)
f56774213bf4 [clang] fix finding class template instantiation pattern for member specializations (#199979)
58156c21585a [lldb-dap] Fix typo in StepInTarget serialization (#199907)
81c523c71c92 Reapply "[clang][ssaf][NFC] Rework how the Force linker anchors are defined and used" (#194693)
1e1a1ff8a6c6 workflows/release-binaries-all: Validate input and remove template expansion (#199427)
78f5f7717a2e [PowerPC] Drop invalid range metadata when lowering i64 load to fp in INT_TO_FP (#198705)
f0ae26c925ef workflows/issue-release-workflow: Fix bug when specifying multiple commits (#199973)
7de3f46cf771 [InstSimplify] Fold fshl/fshr of complementary shifts to identity (#196887)
ef1b6d232216 Add Pre-Optimization FunctionPropertiesStatistics Pass - All Properties (#199293)
c09041a33a8a [DebugInfo] Introduce DW_AT_LLVM_language_dialect (#194898)
cef3e0f1280d [libc++][C++03] Fix array/size_and_alignment.compile.pass.cpp (#198822)
89f024609b13 [RISCV][CodeGen] Use vzip.vv for e64 interleave shuffles with Zvzip (#199923)
ed11d7a52a50 [compiler-rt][ARM] Optimized FP double <-> single conversion (#179926)
8ec87744a09c merge main into amd-staging (#2681)
bdaf3cfda6d6 [SLP] Improve InsertElement scalarization cost modeling
5bfcf13e48c9 [VPlan] Construct VPlan1 once, share across buildVPlans calls. (#197276)
61341994bb45 [SLP][NFC]Add some more tests with phi external uses, NFC
2245dd779f8c [clangd] Prefer .hpp files over .h with header source switch (#198152)
e6fe3f4f55ef [lldb][test] Require Python for a few more tests (#199913)
c71f9f0aec3a [LV] Handle loop.dependence.mask in verifyLastActiveLaneRecipe() (#199897)
b63787a1c5d7 [MLIR][AMDGPU] Add permlane16.var and permlanex16.var intrinsic ops (#199501)
2a574828c62f [SelectionDAGBuilder] Replace asserts inside LLVM_DEBUG (#199748)
3c16e92da529 [AArch64][TTI][EarlyCSE] Add support for ld1xN and st1xN intrinsics (#198765)
090df8f0c306 Revert "[lit] Move maxIndividualTestTime from global to test suite config" (#199886)
0d6aac78ad96 [libc++] Remove workarounds for __{add,remove}_pointer on AppleClang (#199821)
3060f65b84f2 Revert "[RISCV][CodeGen] Use vzip.vv for e64 interleave shuffles with Zvzip" (#199899)
4def779282d9 [libc++][NFC] Remove lit annotations for older AppleClang versions (#199817)
99e6632cf619 [CIR][AArch64] Upstream vector-shift-right-and-insert NEON builtins (#196776)
c94e5f327f47 AMDGPU/GlobalISel: Move executeInWaterfallLoop call from lower (#199701)
698d44bf9ff6 [clang] Add builtin to clear padding bytes (prework for P0528R3) (#75371)
d63f99627b9b [BOLT][merge-fdata] Preserve event names
3e0c818e6f80 [CIR][AMDGPU] Adds lowering for amdgcn image load/store builtins (#198184)
4c9296d94c95 [BOLT][NFC] Unify perf data setting
b926cf7da091 [mlir][tosa][spirv] Lower TOSA elementwise ops to SPIR-V TOSA (#199505)
992387147411 [BOLT][NFC] Split out DataAggregator::parseInput
66bafd43ff2b [MLIR] Improve TypeID anonymous namespace check to handle GCC's __PRETTY_FUNCTION__ format (#199634)
f5bda2b94460 [LifetimeSafety] Change new tests' warnings to new ones  (#199887)
a4b1361f3313 [RISCV][CodeGen] Use vzip.vv for e64 interleave shuffles with Zvzip (#199512)
6de9a33debfd Revert "[llvm-debuginfo-analyzer] Add support for LLVM IR format. (#135440)" (#199890)
263494977277 [Bazel] Fixes 532940b (#199874)
77d3da64d2bb [Clang] Accept gnu vectors in __builtin_masked* (#198248)
2b286794010f [MLIR] Add a non-const ActionHandler getter to MLIRContext (#199652)
6bbbf743ae49 [llvm-debuginfo-analyzer] Add support for LLVM IR format. (#135440)
a3adb545b6a3 [LifetimeSafety] Add details for `-Wlifetime-safety-return-stack-addr` diagnostic (#199432)
505a611530cc Auto-label LifetimeSafety issues with `clang:temporal-safety` (#199522)
4f54064a6e19 [IR] Introduce an appendTags() idiom to set MMRA metadata [NFC] (#199621)
4ee7e900f321 [lit] Move maxIndividualTestTime from global to test suite config (#198192)
532940bdee66 [BBAddrMap] Drive Features and Metadata bits from BBAddrMap.def (#196906)
18ddec75ba82 [lldb] Use private stop for breakpoint-delaying decision (#199639)
c85c8af5426e [compiler-rt][ARM] Add missing SUPERSEDES for optimized FP comparison sources (#199604)
3913c828ab91 [MC] Diagnose unfinished CFI frame from an earlier section (#196775)
80490b8cfa34 [clang][diagnostics] Reject embedded NUL characters in inline asm (#196462)
22ba468002d5 [clang][bytecode] Fix non-defaulted union copy/move ctors (#199394)
5a616ce490ab [libc++] Update the GCC head version to 17 (#199823)
853d532794be Fix llvm-symbolizer test broken by #199739 (#199801)
0d5b7522dc72 [RISCV][NFC] Remove SegInstSEW for unused function (#199598)
05e1af73ba2d [RISCV][P-ext] Remove duplicate hasSideEffects=0, mayLoad=0, mayStore=0. NFC (#199798)
7259dd668674 [RISCV][P-ext] Add DefVXSAT argument to tablegen classes. NFC (#199797)
73de4c73dc9b [NFC][AMDGPU] Improve the predicate uses for WMMAs (#199807)
1b19eccf1751 [BOLT][NFC] Split out function marking from profile parsing
40602b6346cc [LLD][COFF] Gate second-dot section-name stripping on MinGW (#199625)
7e6f337e1dcf [llvm-mca] Fix total execution count in Average Wait times (#199500)
d139f653019e [SLP][NFC]Add another test for external phi user, NFC
7938535a54b5 Compute GUIDs once and store in metadata (#184065)
e0ef143823ce [lld][WebAssembly] Only include __stack_pointer when needed (#199739)
e6d8a8f9384b [Clang] Emit prefix map normalization before generating hashes for the unique linkage names. (#198667)
2bd872b03fe1 [LV] Add support for partial alias masking with tail folding (#182457)
9eb0d424db2d [lit][NFC] remove future statements for mandatory features in Python 3 (#199786)
a005c2562f30 merge main into amd-staging
e918a5ada8a9 [lit][NFC] remove new-style class opt-ins (#199784)
577e9a7cc82f [WebAssembly] WASIP3 Library Call Thread Context Support (#175800)
fc60e08b4574 [libc] Use containers for overlay precommit CIs. (#199294)
bf420f0d1856 [AArch64] Fix hasNearbyPairedStore to handle non-inbounds GEPs (#199137)
38326d65239e [Flang][OpenMP][Offload] remove umt speciifc fix from runtime
187dfef80ee9 [flang] Enabled pulling of rebox into array_coor. (#199161)
f3c0f26a3961 [flang][FIRToMemRef] Get strides from descriptor for some array_coor cases. (#199158)
f263446abbdd clang/AMDGPU: Report all runtimeless sanitizers as available (#199642)
2d8a2861ab00 merge main into amd-staging (#2672)
e9e5d4ee914a clang/AMDGPU: Remove unnecessary fallback to check -march (#199780)
7e98d1960e98 [LoopFusion] Do not fuse loops with different guards (#199724)
e42046a747f5 [MLIR][XeGPU] Fix pass name in RUN command (#199766)
27abffa3efe2 [lldb] New expedited register specfication for unavailable regs (#193894)
a7aceff0b1e5 Revert "[SandboxIR][Tracker] Implement accept(/*AcceptAll*/) and revert(/*RevertAll*/)" (#199776)
a14d084bbb1a Reland "[libc] Enable baremetal float printf using modular format" (#199758)
f5a3f1d4bf10 [SLP][NFC]Add a test iwhtthre vectorization regression, NFC
0e0127e5b21d [lldb] Fix vtable support on arm64e (#199116)
8a64511b59a2 [LV] Add tests with pointers based on URem expressions (NFC). (#199763)
0eb28e6de265 [SLP] Propagate through instrinsics in BoUpSLP::getVectorElementSize() (#199129)
2c336879888a [AMDGPU] Add regression test for extract of vector binop scalarization (#198825)
a4c8cfdac226 [AMDGPU] Fix ELF note emission to include null terminator (#199720)
632cadcd2ced [AMDGPU]comgr: Added CLOCK_MONOTONIC_RAW define to allow successful complications and usage on FreeBSD (#2666)
c052a26f2d8c Revert "[clang][driver][darwin] Hold onto full triples in Darwin SDKP… (#199756)
f78a233ac89d [AMDGPU] Implement CFI for non-kernel functions (#183153)
4ef6ca40f53f [Clang][AArch64] Fix crash with large arguments to MTE built-ins (#197620)
fdd3b84773f1 [SLP] Fix FMA regression in FMA-candidate retry
5157be7bd5c1 [SandboxIR][Tracker] Implement accept(/*AcceptAll*/) and revert(/*RevertAll*/) (#197289)
3dd3b6fe758c [AMDGPU][docs][NFC] Fix some instruction names in gfx950 doc (#199094)
3212caa942de [bazel] Use `additional_compiler_inputs` to handle include scanning for TargetPassRegistry.inc (#199201)
3ce7b405579d Revert "[clang] Use FileError in FileManager::getFileRef, getDirectoryRef" (#199721)
593a23887939 [AMDGPU] Diagnose unsupported fma_legacy/sudot4/sudot8 intrinsics on some subtargets (#198464)
13f9f468dcf8 [AMDGPU] Emit entry function Dwarf CFI (#183152)
7a66e99b12ee [AMDGPU][True16] Add regbank combiner cases to fix regression around G_SEXTLOAD (#198671)
6519c04eb459 [1/3][RegAlloc][LiveRegMatrix] Fix inconsistency in HoistSpillHelper delegates (#197773)
08e83a533e47 workflows/release-tasks: Remove template expansion (#199444)
ef59dbea76d8 [flang][cuda] Lower c_devptr value arguments in bind(c) like c_ptr (#199316)
584b596d8120 workflows/release-doxygen: Remove template expansions (#199456)
88fbc0614e82 [DSE] Restrict partial-overlap store merging to matching orderings. (#199728)
5df91f6c83ca workflows/sycl-tests: Pin container image reference (#199466)
9d751a2985c1 [llvm][Object] Add COFF support to extractOffloadBundleFatBinary (#199574)
010faf1e36c7 [TableGen] Add missing grammar comment for !cond(NFC) (#199663)
fafc2b32ffff [TableGen] Fix wrong op name in a grammar comment(NFC) (#199661)
3aa913fe6474 [libc][math] Temporarily disable exception tests for NextAfter and NextToward tests on Windows. (#199740)
282e90796bb9 [libc][ci] Improve full build precommit CIs caching keys. (#199742)
95a6a4bb8e73 [comgr] Adding comgr's multi-threaded test back in (#2646)
92cfd0a3f0ac [clang-doc] Add option for compact JSON (#190822)
8d2f190609b6 [clang][SemaOpenACC] Reject VLA reduction (#199178)
ff243860d793 [lld][MachO] Fix SIGBUS crash in saveOrHardlinkBuffer (#198381)
95cb4d4b536a AMDGPU/GlobalISel: RegBankLegalize rules for wave_shuffle (#196412)
6d99fcb5c8a2 [libc++] Mark string.capacity/over_max_size.pass.cpp as UNSUPPORTED on old dylibs (#199682)
631d16eaaa88 [SSAF][WPA] Bounds propagation graph is a supergraph of the pointer-flow graph (#198889)
461f816d4420 Reland "[HLSL][DirectX] Emit convergence control tokens when targeting DirectX" (#194452)
769ffe95aac0 [lldb-dap][NFC] Use GetStringValue helper (#199672)
85717da6350c [clang] Add -fno-debug-record-sysroot (#192541)
2a80a9d52ee1 [-Wunsafe-buffer-usage] Move warning-only analysis back to function-based (#198006)
09709d7f5357 [Github] Pin container image reference in libc-shared-tests (#199737)
31111955fc5c [lldb] Set SO_NOSIGPIPE on platforms that support it (#198044)
9ec4c66daffd [AMDGPU] Lay groundwork for stridemark type / addrspace(9) via sgep (#198930)
da57dc996221 [AMDGPU][True16] Legalize extloads into 16-bit registers (#198670)
196c7018fc11 [mlir][MemRef] Move FlattenMemRefs to interfaces (#198625)
041c5e0e06c9 [VPlan] Add missing maybe_unsused to OpTy (NFC). (#199734)
ae0dc3104bf5 [flang][OpenMP] Make `iv` const-reference to original parser::Name (#199727)
e925643b1e7f [flang][OpenMP] Warn that declare simd in an interface body has no effect (#199248)
4bc1cd92e41f workflows/release-binaries: Fix yaml error (#199735)
460556ae66e4 [ORC] Shutdown the socket FD before closing it in FDSimpleRemoteEPCTransport (#196835)
1869d9925f5d Update `memset.inline` third argument documentation (#199725)
bacd87613449 [lld][MachO] Preserve __eh_frame ordering during BP section sorting (#191412)
b395ca789219 [SPIR-V] Fix legalized load of single-element vector from array (#198330)
80a5207d940b [VPlan] Thread scalar types through VPInstruction and VPPhi. (NFC) (#199378)
990e35ea9d80 [Bazel] Fixes 44da860 (#199723)
a917c00ab13f [libc++] Use __in_out_result in the remaining appropriate algorithms (#198156)
e526b2003977 [SLP] Add new tests for instrinsics in getVectorElementSize() (#199612)
7146087fb494 [AArch64][DAG] Use getSetCCResultType in PromoteSETCC (#198320)
6e94fa069fa9 [CIR][CUDA] Emit global var registration (#199270)
68ebd09d5e19 [lld][macho] Restructure thunk generation algorithm (#193367)
4b7ec0c71539 [GlobalsModRef] Don't erase while iterating (#199656)
eca8638e1303 [ORC] Avoid iterator invalidation when erasing image info symbols (#199655)
44da8601d71e [MLIR][XeGPU] Clone trivial operations with multiple consumers in layout propagation to avoid layout conflict (#197514)
2d3fea0916c8 merge main into amd-staging (#2670)
d685a3b0e30f Use symbol matching for types in llvm-mode.el (#199250)
8310cc5522f7 [psdb] trigger psdb on PR reopened event
551ff98cc3e6 [SLP][TTI][AMDGPU] Add TTI hook preferSLPInstCountCheck for per-target opt-out (#199696)
29fa3cc89864 [psdb] enable psdb trigger on reopened event
26f19eed998d [clang] propagate constexpr/constinit to binding VarDecl (#195860)
6fdfe1ead81f llvm: Fix most LLVM_ABI annotations in Transforms (#199042)
4acfc32f0075 Revert "[JumpThreading] Freeze undef/poison select condition in unfoldSelectInstr" (#199713)
d1b31664ea1c [X86] getFauxShuffleMask - add handling for X86ISD::VSHLD/VSHRD funnel shifts (#199707)
3718ab29aa88 [gn build] Port c437052c (#199709)
c927d44cb119 Revert "Revert "[libc] Port 2b2a63819f9f26d661bad5c269a03077d22ff6b4"" (#199681)
fe4c2bb1b3d5 [mlir][xegpu] Deprecate XeGPUSubgroupDistribute and rename XeGPUSgToWiDistributeExperimental to XeGPUSgToLaneDistribute (#198027)
6b17cf8d11ce [RISCV][P-ext] Support packed SSHLSAT with non-constant splat shift amount. (#198937)
52aac75c0054 [RISCV][GISel] Use sXLen member variable instead of creating it locally. NFC (#199594)
bd9ce0ab1edd [clang][analyzer] Make CallAndMessage:ArgPointeeInitializedness released (NFC) (#199184)
721cddfd8f06 [SLP][NFC]Add a test with inst count heuristic for AMDHSA, NFC
e8d5037b400d [clang] NFC: readd test cases reverted in 79f4d8f014 (#199676)
3a25f7b645fa [clang] preserve exact redeclaration for getTemplateInstantiationPattern (#199473)
fbbaa545ce1a [mlir][async] Lazily create the coroutine destroy-cleanup block (#199583)
ca9884dd0659 [clang] Add missing type source info of `decltype` specifier in destructor call to AST (#197680)
cb96a3edf7a8 [clang][test] Move offload tools tests to clang/test/OffloadTools (#198434)
8b2751198db3 [X86] Add test coverage showing failure to recognise VPSHLD/VPSHRD nodes as potential faux shuffles (#199673)
471244d79075 [InferAlignment] Propagate alignment information across blocks (#198114)
e5f4f31b3508 [JumpThreading] Freeze undef/poison select condition in unfoldSelectInstr (#199408)
095faea999a2 [SLP] Fix assertion "Deleting out-of-tree value" for struct-typed vectorized operands
3eb9aacb7c47 [clang][CodeGen][NFC] Fixed typos in comments. (#197052)
f2ce225f0f77 Exclude documentation files from 80 columns rule in Coding Standard (#197666)
573138d54127 Handle more cases in DebugInfoFinder (#194684)
f72be679b604 [ProfCheck] Add test from #197698 to xfail list (#199650)
70321b9c3b26 [Support] Support runtime override for LLVM_WINDOWS_PREFER_FORWARD_SLASH (#199210)
ded7aa0dcf49 [z/OS] Add --ignore-case to FileCheck on output from od. (#196396)
05a237293730 [RISCV][GlobalISel] Lower i8 bitreverse using brev8 with Zbkb (#199469)
a903b9512750  [CloneModule] Clone undefined ifuncs (#197353)
028153a165d6 [VPlan] Make TransformState::get BCast-logic robust (#197589)
79337832f7a3 [OpenCL] Add cl_intel_split_work_group_barrier builtins (#199424)
10c0de887256 [LinkerWrapper] Fix temps being dumped to CWD instead of output path (#198679)
dfa8f5cd3549 merge main into amd-staging
c437052c0a47 Reland "[APFloat] Add exp functions for single and double using exp/expf implementations from LLVM libc." (#197440) (#199570)
0f1d083eff4b [Clang][test] check-clang-format not created with LLVM_ENABLE_IDE (#199638)
adcad45cbdf8 [CIR] Vector saturating rounding shift right and narrow intrinsics  (#198947)
495e6c54420f [Coroutines] Allow rematerialization of unary operators and selected intrinsics (#197698)
b2634fc4f8b1 [clang][bytecode] Fix a crash in __builtin_subcb (#199400)
3398f4e66558 [mlir][mem2reg] fix assert for indirect blocking uses inside regions (#199193)
5ab7435196ce [LV] Add support for partial reduction chains with fsubs. (#197114)
322ff9f317bb [RISCV] Remove TargetLowering arg from getContainerForFixedLengthVector. NFC (#199629)
73f141f75caf [NFC] Add null terminator assert to CodeViewRecordIO::mapStringZ (#199624)
19e915fc5c91 [LoongArch] Fix musttail with indirect arguments by forwarding incoming pointers (#198965)
4448a1a7c095 [X86] LowerBUILD_VECTORvXi1 - scalarize the bool masks if we insert a single non-const value (#199523)
a975b7cd3538 [RISCV] Fix fixed-length masked.{u,s}{div,rem} lowering not converting operands (#197913)
970ecaef1db3 [Clang][Sema] Fix crash in __builtin_dump_struct with immediate callables (#192880)
4ab4d2d9bb4a [VPlan] Fold canonical IV recipe creation into createLoopRegion. (#198383)
fda6309919a0 [AArch64][GlobalISel] Add tablegen pattern for uaddo combine (#198724)
d06d3a7bdca1 [clang][ASTImporter] Fix of crash at ConstraintSatisfaction import (#197407)
a69ce910db2d [Clang][AMDGPU] clean up auto-generated CHECK lines in a test (NFC) (#199614)
49e2295cf52c [GVN] Properly combine AA metadata if available load is hoisted (#197948)
a225aafbd1a4 [DenseMap] Invalidate iterators on erase (#199369)
be39e065d325 [IR] Inline remove_if in PMDataManager::removeNotPreservedAnalysis (#199571)
377af85dad4e [InferAttrs] Annotate math and basic string libcalls with `nosync` (#197761)
b86512ce648f [RISCV][P-ext] Split v4i16/v8i8 vselect on RV32. (#198898)
c618e112145f [MLIR][NVVM] Add `nvvm.divf` Op (#198744)
fb8243f32c95 [RISCV] Use append TableGen feature in RISCVInstrInfoXqci.td (#199603)
de188bbad3df [DirectX][ObjectYAML] Fix CI build issue in DXContainerInfo.cpp (#199605)
44583eca36b6 [mlir][x86] Fix - multiple issues / F8 support for AMX dot-product lowering (#196984)
fe3d6b051b31 [NFC] [clangd] [C++20] [Modules] Fix false duplicate module warning for equivalent paths (#199343)
44d60bad70c7 sanitizer_common: Fix build on MIPS with _TIME_BITS=64 (#199590)
d4550cc82efa [OpenCL] Add cl_intel_bfloat16_conversions builtins (#199422)
23b754f5745a [RISCV] Use the new "let append" TableGen feature to reduce duplication. NFC (#199596)
d7856fed0268 [RISCV][GISel] Add exact flag to G_LSHR in RISCVLegalizerInfo::legalizeVScale (#199595)
386e356b5f0a [AArch64] Add support for MSVC-style mangling for SVE (#196738)
03c8a834d6a9 [libc++] Address most LLVM23 TODOs (#199397)
50a110ace755 [CodeGenPrepare] splitMergedValStore: don't split atomic stores. (#199592)
d1902d34606e [RISCV][P-ext] Fold bitcast of v4i8/v2i16 const splat to scalar on RV64 (#199513)
5723dd03231f [RISCV][P-ext] Split v4i16/v8i8 MUL on RV32. (#199504)
78f0576901ad merge main into amd-staging
b929dbabe54b [X86] FixupInstTuning: ProcessShiftLeftToAdd should return true after mutating. (#199589)
7dabf946d908 [SelectionDAG] Handle CSE in PromoteIntOp_VP_STRIDED. (#199562)
b938d8b562d6 [DA] Consolidate accumulating GCD functions (NFCI) (#197936)
669e19ba4195 [Flang][OpenMP][NFC] Remove duplicate code, and default to upstream version
97e7ee26b2b3 [InstrProf] Do not emit metadata for zero values with zero counts (#199380)
18bb5400d6f8 merge main into amd-staging (#2667)
269719b1f63c Revert "[compiler-rt][ASan] Add function copying annotatio…
@djramic djramic requested a review from causten as a code owner June 1, 2026 13:50
@umangyadav umangyadav requested a review from Copilot June 3, 2026 12:48
@umangyadav umangyadav added the claude-review Trigger automated PR review by claude[bot]; auto-removed after the run. label Jun 3, 2026

Copilot AI left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copilot wasn't able to review this pull request because it exceeds the maximum number of files (300). Try reducing the number of changed files and requesting a review from Copilot again.

@rocmlir-pr-reviewer rocmlir-pr-reviewer Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Verdict: COMMENT  ·  Findings: 0 (0 Critical, 0 Major, 0 Minor)


Scope

LLVM upstream merge (~6000 files under external/llvm-project/). The only rocMLIR-side change in the diff is a regeneration of mlir/tools/rocmlir-lib/librockcompiler_deps.cmake, which the merge checklist requires when dependencies shift. Per the PR description, three upstream files needed manual conflict resolution: AMDGPUAsmPrinter.cpp, AMDGPUToROCDL.cpp, and EmulateNarrowType.cpp.

Findings

No blocking rocMLIR-side issues identified. The vast majority of the diff is verbatim upstream LLVM/MLIR content carried in by the merge and is out of scope for line-by-line rocMLIR review; the conflict resolutions described in the PR body (kept rocMLIR's commented-out occupancy-warning block, adopted upstream's InstPrefSize removal, picked up GlobalTransposeLoadOpLowering, adopted upstream ConvertMemRefCast/assumeAligned) look consistent with the surrounding upstream changes and the stated intent.

Notes

  • The "external changes must be separate, prefixed [EXTERNAL]" rule from the checklist is not applied here because the PR's purpose is the upstream merge itself; the lone rocMLIR-side touch (librockcompiler_deps.cmake) is a mechanical regeneration tied to the merge and would be awkward to split off.
  • Worth double-checking before merging that all three Test Plan boxes (PR / Nightly / Weekly) are exercised, since the Technical Details only call out check-llvm / check-mlir as passing.

CI status

  • Jenkins: one run reports FAILURE (a second Jenkins entry is SUCCESS, but the failing one needs investigation).
  • ml-ci-internal.amd.com: ERROR.

These red signals should be triaged before merge; with such a large surface area, a CI regression here is very likely real and not just infrastructure noise.

@rocmlir-pr-reviewer rocmlir-pr-reviewer Bot removed the claude-review Trigger automated PR review by claude[bot]; auto-removed after the run. label Jun 3, 2026

@justinrosner justinrosner left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LLVM and MLIR diffs look good!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants