Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[mlir][llvm] Fix negative GEP crash in type consistency #74859

Merged
merged 2 commits into from Dec 11, 2023

Conversation

rikhuijzer
Copy link
Member

Fixes #74453.

The gepToByteOffset was implicitly casting an signed integer to an unsigned integer even though negative dimensions are valid for llvm.getelementptr.

@llvmbot
Copy link
Collaborator

llvmbot commented Dec 8, 2023

@llvm/pr-subscribers-mlir

@llvm/pr-subscribers-mlir-llvm

Author: Rik Huijzer (rikhuijzer)

Changes

Fixes #74453.

The gepToByteOffset was implicitly casting an signed integer to an unsigned integer even though negative dimensions are valid for llvm.getelementptr.


Full diff: https://github.com/llvm/llvm-project/pull/74859.diff

2 Files Affected:

  • (modified) mlir/lib/Dialect/LLVMIR/Transforms/TypeConsistency.cpp (+4-1)
  • (modified) mlir/test/Dialect/LLVMIR/type-consistency.mlir (+14)
diff --git a/mlir/lib/Dialect/LLVMIR/Transforms/TypeConsistency.cpp b/mlir/lib/Dialect/LLVMIR/Transforms/TypeConsistency.cpp
index b094c650ff1932..3e856827255149 100644
--- a/mlir/lib/Dialect/LLVMIR/Transforms/TypeConsistency.cpp
+++ b/mlir/lib/Dialect/LLVMIR/Transforms/TypeConsistency.cpp
@@ -161,7 +161,10 @@ static std::optional<uint64_t> gepToByteOffset(DataLayout &layout, GEPOp gep) {
     IntegerAttr indexInt = llvm::dyn_cast_if_present<IntegerAttr>(index);
     if (!indexInt)
       return std::nullopt;
-    indices.push_back(indexInt.getInt());
+    int32_t gepIndex = indexInt.getInt();
+    if (gepIndex < 0)
+      return std::nullopt;
+    indices.push_back((uint32_t)gepIndex);
   }
 
   uint64_t offset = indices[0] * layout.getTypeSize(gep.getElemType());
diff --git a/mlir/test/Dialect/LLVMIR/type-consistency.mlir b/mlir/test/Dialect/LLVMIR/type-consistency.mlir
index 1504a98e6f8cca..3a1ab924ebdacb 100644
--- a/mlir/test/Dialect/LLVMIR/type-consistency.mlir
+++ b/mlir/test/Dialect/LLVMIR/type-consistency.mlir
@@ -151,6 +151,20 @@ llvm.func @index_to_struct(%arg: i32) {
 
 // -----
 
+// CHECK-LABEL: llvm.func @no_crash_on_negative_gep_index
+llvm.func @no_crash_on_negative_gep_index() {
+  %0 = llvm.mlir.constant(1.000000e+00 : f16) : f16
+  %1 = llvm.mlir.constant(1 : i32) : i32
+  // CHECK: %[[ALLOCA:.*]] = llvm.alloca %{{.*}} x !llvm.struct<"foo", (i32, i32, i32)>
+  %2 = llvm.alloca %1 x !llvm.struct<"foo", (i32, i32, i32)> : (i32) -> !llvm.ptr
+  // CHECK: llvm.getelementptr %[[ALLOCA]][-1] : (!llvm.ptr) -> !llvm.ptr, f32
+  %3 = llvm.getelementptr %2[-1] : (!llvm.ptr) -> !llvm.ptr, f32
+  llvm.store %0, %3 : f16, !llvm.ptr
+  llvm.return
+}
+
+// -----
+
 // CHECK-LABEL: llvm.func @coalesced_store_ints
 // CHECK-SAME: %[[ARG:.*]]: i64
 llvm.func @coalesced_store_ints(%arg: i64) {

@gysit gysit requested a review from Moxinilian December 8, 2023 16:27
Copy link
Contributor

@gysit gysit left a comment

Choose a reason for hiding this comment

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

LGTM modulo the C style cast.

mlir/lib/Dialect/LLVMIR/Transforms/TypeConsistency.cpp Outdated Show resolved Hide resolved
Co-authored-by: Tobias Gysi <tobias.gysi@nextsilicon.com>
@rikhuijzer
Copy link
Member Author

Thanks for the review @gysit! I'll merge once CI on main is green again.

@rikhuijzer rikhuijzer merged commit 3764f5e into llvm:main Dec 11, 2023
4 checks passed
@rikhuijzer rikhuijzer deleted the rh/fix-neg-gep-dim branch December 11, 2023 11:29
fadlyas07 pushed a commit to greenforce-project/llvm-project that referenced this pull request Dec 16, 2023
* llvm-project/main:
  [clang][CGCUDANV] Unify PointerType members of CGNVCUDARuntime (NFC) (#75668)
  [mlir][linalg] Fix rank-reduced cases for extract/insert slice in DropUnitDims (#74723)
  Revert "Reapply "RegisterCoalescer: Add implicit-def of super register when coalescing SUBREG_TO_REG""
  [libc++][spaceship][NFC] Status page - added papers (#75043)
  [AArch64][SME2] Add FCLAMP, CNTP builtins for SME2 (#72487)
  [flang][nfc] Refactor linker invocation logic (#75648)
  [InstCombine] Canonicalize `icmp pred (X +/- C1), C2` into `icmp pred  X, C2 -/+ C1` with nowrap flag implied by with.overflow intrinsic (#75511)
  [InstCombine] Add test for missed opportunity to fold 'or' into 'mul' operand. NFC
  [NFC][CLANG] Rename duplicate loop attributes diagnostic functions (#75657)
  [RISCV] Make Zfh imply Zfhmin. (#75576)
  [llvm-readobj] Print the associated CUDA SM flags (#75664)
  [Sanitizers] Don't inline unpoisoning of small stacks when inlining disabled (#75555)
  [ASTReader] Fix readability-inconsistent-declaration-parameter-name. NFC
  [mlir][sparse] support loop range query using SparseTensorLevel. (#75670)
  [X86] Don't use rip-relative lea to get a function address in medium static mode (#75656)
  Revert "[NFC] [Serialization] Packing more bits and refactor AbbrevToUse"
  Revert "[NFC] Fix the warning Wcovered-switch-default"
  [PowerPC] Emit libcall to frexpl for calls to frexp(ppcDoublDouble) (#75226)
  [llvm-objcopy] Fix gap-fill/pad-to tests (#75631)
  [mlir][sparse] set up the skeleton for SparseTensorLevel abstraction. (#75645)
  Revert "[StackColoring] Delete dead stack slots (#75351)" (#75655)
  [llvm-exegesis] Adjust page size in unit tests to fix ppc failures
  [libc++] Fix constexpr initialization of std::array<T, 0> (#74667)
  [flang][openacc/mp] Do not read bounds on absent box (#75252)
  [X86][test] Update tagged-globals*.ll tests
  [VPlan] Unswitch cond in replaceUsesWithIf in optimizeInductions (NFC)
  [OpenMP] Use simple VLA implementation to replace uses of actual VLA
  [OpenMP] [OMPT] A pointer to HostOpId should be passed in EMI callbacks. (#75574)
  fixup! fixup! [GlobalISel] Always direct-call IFuncs and Aliases (#74902)
  [NVPTX] Add ELF flags for Nvidia cubin files (#75624)
  [MemProf][NFC] Clear each IndexedMemProfRecord after it is written (#75205)
  [MemProf][NFC] Free large data structures after last use (#75120)
  [mlir][math] Added `math.sinh` with expansions to `math.exp` (#75517)
  [flang] legacy branch target (#75628)
  [RISCV][InsertVSETVLI] Factor out isNonZeroLoadImmediate helper [nfc]
  fixup! [GlobalISel] Always direct-call IFuncs and Aliases (#74902)
  Revert "[flang][nfc] Refactor linker invocation logic (#75534)"
  [SystemZ] ABI support for single-element vector types
  [lldb][NFCI] Remove unused parameter from BreakpointResolver*::CreateFromStructuredData (#75374)
  [lld][WebAssembly] Add an `--initial-heap` option (#75594)
  [llvm-windres] Resolve the --preprocessor executable in $PATH (#75390)
  [llvm-windres] Pass user preprocessor arguments before the input filename (#75389)
  [LLD] [MinGW] Respect the -S/-s options when writing PDB files (#75181)
  [LLD] [COFF] Add /debug: options nodwarf and nosymtab (#75180)
  [LLD] [COFF] Parse all /debug: options, like /opt: (#75178)
  [LLD] [COFF] Rewrite handling of the /debug: option. NFC. (#75175)
  [LLD] [COFF] Rewrite the config flags for dwarf debug info or symtab. NFC. (#75172)
  [mlir][linalg] Use vector.shuffle to flatten conv filter (#75038)
  [Instrumentation][X86] Limit setting large section flag to medium/large code models (#75542)
  [flang][openacc/mp][NFC] Remove unused baseAddr argument (#75537)
  [llvm-exegesis] Validate that address annotations are aligned (#75554)
  [VPlan] Remove stale comment from optimizeInductions (NFC).
  [SLP] Fix OOB GEP index access for a no-op GEP
  [NFC][mlir][sparse] remove redundant parameter. (#75551)
  [RISCV] Prefer whole register loads and stores when VL=VLMAX (#75531)
  [llvm] Remove no-op ptr-to-ptr casts (NFC)
  [RISCV Add some vsetvli insertion test cases with vmv.s.x+reduction. NFC (#75544)
  [AArch64] Put legal action first for G_ATOMIC_CMPXCHG (#74613)
  [gn build] Port f7407411a1da
  [IR] Fix UB on Op<2> in ShuffleVector predicates (#75549)
  [AMDGPU] GFX12: select @llvm.prefetch intrinsic (#74576)
  [libc++] Optimize std::find for segmented iterators (#67224)
  [MLIR][bazel] Fix build files for 5e29112719df91b1c80916386e230252d56… (#75617)
  [flang][OpenMP] Move handling of OpenMP symbol flags to OpenMP.cpp (#75523)
  [MLIR][bazel] Fix build files for 681eacc1b670fd7137d8677fef6fc76c6e3… (#75615)
  [flang][driver][nfc] Rename one variable (res -> invoc) (#75535)
  [flang][nfc] Refactor linker invocation logic (#75534)
  [flang][OpenMP] Move nested eval conversion to OpenMP.cpp, NFC (#75502)
  [libc] Fix improper initialization of `StorageType` (#75610)
  [AMDGPU] - Add address space for strided buffers (#74471)
  [CodeGenPrepare] Remove unused TypePromotionTransaction::moveBefore to fix gcc Wunused-function warning. NFC.
  [Object] Drop unnecessary const qualifier to fix gcc Wignored-qualifiers warning. NFC.
  [mlir][mesh] Add verification and canonicalization for some collectives (#74905)
  Revert "[LinkerWrapper] Add 'Freestanding' config to the LTO pass" (#75528)
  [clang] Fix unexpected warnings after a01307a (#75591)
  [AMDGPU] CodeGen for GFX12 VFLAT, VSCRATCH and VGLOBAL instructions (#75493)
  Support for dynamic dimensions in 'tensor.splat' (#74626)
  [mlir][tensor] Fix tensor.concat reifyResultShapes for static result dims (#75558)
  [VPlan] Remove unneeded getNumUsers calls in replaceAllUsesWith (NFC).
  [clangd] Add header mapping for struct_rusage.h
  [clang][Interp][NFC] Remove outdated FIXME comment
  [OpenMP] Fix compile error caused by #74397 (#75419)
  [libc][NFC] Rename `MANTISSA_WIDTH` in `FRACTION_LEN` (#75489)
  Revert "[builtins][arm64] Build __init_cpu_features_resolver on Apple platforms (#73685)"
  [AMDGPU] CodeGen for GFX12 VBUFFER instructions (#75492)
  [X86][AVX10] Allow 64-bit mask register used without EVEX512 (#75571)
  [MLIR][transform][python] add sugared python abstractions for transform dialect (#75073)
  [AMDGPU] Fix -Wunused-variable in SIInsertWaitcnts.cpp (NFC)
  [lldb] Trying to fix windows buildbots after #74786 (#75566)
  [SystemZ] Support i128 as legal type in VRs (#74625)
  [Clang][AArch64]Add QCVTN builtin to SVE2.1 (#75454)
  [AMDGPU] CodeGen for GFX12 VIMAGE and VSAMPLE instructions (#75488)
  [mlir][vector] Add emulation patterns for vector masked load/store (#74834)
  [AMDGPU][SIInsertWaitcnts] Do not add s_waitcnt when the counters are known to be 0 already (#72830)
  AMDGPU: refactor phi lowering from SILowerI1Copies (NFCI) (#75349)
  [AMDGPU] CodeGen for SMEM instructions (#75579)
  [AArch64][GlobalISel] Look into array's element (#74109)
  [symbolizer] Support symbol+offset lookup (#75067)
  [AMDGPU] Update permlane test for GFX12 (#75572)
  [AMDGPU] Update s_get_barrier_state tests (#75575)
  [ValueTracking] isNonEqual Pointers with with a recursive GEP (#70459)
  [StackColoring] Delete dead stack slots (#75351)
  [CodeGen] Use `MachinePassKey` for machine passes (#75567)
  [AMDGPU] Add pseudo scalar trans instructions for GFX12 (#75204)
  [Flang] HLFIR maxloc intrinsic (#75450)
  [Flang] Remove kind from CountOp (#75466)
  [AMDGPU] Pre-commit test for #75573. NFC
  [AMDGPU][GFX12] Add new v_permlane16 variants (#75475)
  [CodeGen][NewPM] Add necessary codegen options (#70904)
  [libc][NFC] Split str_to_float tests and reduce type clutter (#75471)
  [clang] Report narrowing conversions with const references (#75332)
  [ValueTracking] Infer `X u<= X +nuw Y` for any Y (#75524)
  [InstCombine] Treat `lshr nneg` as `ashr` in `getBinOpsForFactorization` (#75521)
  [X86][MC] Support Enc/Dec for EGPR for promoted MOVDIR instruction (#74713)
  [AMDGPU][MC] Add GFX12 SMEM encoding (#75215)
  [AMDGPU][SIInsertWaitcnts] Set initial state for VS_CNT in non-kernel functions (#75436)
  [mlir][ArmSME] Fail instead of error in vector.outerproduct lowering (#75447)
  [RISCV] Support printing immediate of RISCV MCInst in hexadecimal format (#74053)
  [mlir][tosa] Add verifier for `tosa.transpose` (#75376)
  Revert "[RISCV] Support printing immediate of RISCV MCInst in hexadecimal format" (#75561)
  [RISCV] Remove 'experimental-' from extension name in diagnostic message.
  [CoverageMapping] Avoid use of pow() resulting in solaris build fail (#75559)
  [flang][openacc/mp][NFC] Fix order of template arguments (#75538)
  [RISCV] Update relax-per-target-feature.ll to use hexadecimal constants. NFC
  [AMDGPU][NFC] Check more autogenerated llc tests for COV5 (#75219)
  Reland "[Pass][CodeGen] Add some necessary passes for codegen" (#71783)
  [SelectionDAG] Use getVectorElementPointer in DAGCombiner::replaceStoreOfInsertLoad. (#74249)
  Reapply "RegisterCoalescer: Add implicit-def of super register when coalescing SUBREG_TO_REG"
  tests: temporarily add a X86 requirement on a test
  [Serialization] Fix -Wpessimizing-move after 9cdb825a4f1bf9e75829d03879620c6144d0b7bc
  [NFC] Fix the warning Wcovered-switch-default
  [clang][RISCV] Change default abi with f extension but without d extension (#73489)
  [NFC] [Serialization] Packing more bits and refactor AbbrevToUse
  [llvm][NVPTX] Inform that 'DYNAMIC_STACKALLOC' is unsupported (#74684)
  [RISCV] Support printing immediate of RISCV MCInst in hexadecimal format (#74053)
  Fix comparing in instrprof-binary-correlate.c
  [BOLT][DWARF] Fix handling .debug_str_offsets for type units (#75522)
  [mlir][SCF] Add support for peeling the first iteration out of the loop (#74015)
  [test][asan] Disable test on Android
  Revert "Use llvm-profdata to test indexed profile in instrprof-binary-correlate.c"
  [test][sanitizer] Add pthread_join in child
  [test][sanitizer] Don't use non-portable __WALL
  [scudo] simplify flag parser out of bounds logic (#72371)
  [Flang] fix ppc-vec intrinsics testcases on AIX (NFC) (#74347)
  [mlir] Handle simple commutative cases in CSE.
  [OptTable] Make new lines in help text respect indentation (#75366)
  [orc-rt] Add ORC_ENABLE_OSX to control whether to build the orc runtime (#75536)
  Add option to pass thread ID to thread select command (#73596)
  [test][android] Reduce buffer size in attempt to fix the test on Android
  [test][hwasan] Update test expectation for LAM
  [ThinLTO] Allow importing based on a workload definition (#74545)
  Revert "[clang] Substitute alias templates from correct context (#75069)"
  [LLD][COFF] add __buildid symbol. (#74652)
  [libc++][test] Enhance ADDITIONAL_COMPILE_FLAGS, use TEST_MEOW_DIAGNOSTIC_IGNORED sparingly (#75317)
  [llvm-objcopy] Add --gap-fill and --pad-to options (#65815)
  fixup! [builtins][arm64] Build __init_cpu_features_resolver on Apple platforms (#73685)
  Re-Reland [X86] Respect code models more when determining if a global reference can fit in 32 bits (#75386)
  [LLDB] Add more helper functions to CompilerType class (second try). (#73472)
  [LLD][COFF] Fix ARM64 EC chunks comparator. (#75495)
  [clang][modules] Deprecate module.map in favor of module.modulemap (#75142)
  [llvm][cmake] Use $<CONFIG> instead of ${CMAKE_CFG_INTDIR} for llvm-config (#75417)
  CodeGen: add a missing check for bit-slice overlap in CV (#75504)
  Use llvm-profdata to test indexed profile in instrprof-binary-correlate.c
  [GlobalISel] Always direct-call IFuncs and Aliases (#74902)
  [builtins][arm64] Build __init_cpu_features_resolver on Apple platforms (#73685)
  [clang] Function Multi Versioning supports IFunc lowerings on Darwin platforms (#73688)
  [clang] Support __attribute__((ifunc(...))) on Darwin platforms
  [llvm] Support IFuncs on Darwin platforms (#73686)
  [libc] fix msan failure in mempcpy_test (#75532)
  [libclang/python] Fix some minor typos (#74292)
  [AMDGPU] Add GFX12 s_sleep_var instruction and intrinsic (#75499)
  [flang][driver] Don't use -whole-archive on Darwin (#75393)
  [AMDGPU] Remove v_cmp_t_* and v_cmp_f_* for GFX12 (#75498)
  [AMDGPU] Remove s_cmpk_* for GFX12 (#75497)
  [AMDGPU] Remove s_cbranch_cdbg* for GFX12 (#75496)
  [flang] Adjust _FORTRAN_RUNTIME_IEEE_FENV_T_EXTENT for Solaris (#74590)
  Revert "Reland [X86] Respect code models more when determining if a global reference can fit in 32 bits (#75386)"
  [SPIR-V][NFC] Require asserts on 2 tests (#75087)
  [llvm-exegesis] Refactor MMAP platform-specific preprocessor directives (#75422)
  [Docs][llvm-exegesis] Add documentation on recently added options (#75408)
  [lldb-dap] Implement command directives (#74808)
  [openmp][wasm] Allow compiling OpenMP to WebAssembly  (#71297)
  [XCOFF] Fix UB 'left shift of negative value' after #72532
  [lldb] Remove PopPlan asserts and skip test on Arm/AArch64 Linux
  [Profile] Add binary profile correlation for code coverage. (#69493)
  [test][sanitizer] Disable new test failing on PPC
  Fix a crash from character type confusion interaction with libedit (#75388)
  [Sema] atomic_compare_exchange: check failure memory order (#74959)
  [mlir][gpu] Fix crash in `gpu-module-to-binary` (#75477)
  [mlir][sparse] comment cleanup in iteration graph sorter (#75508)
  [mlir][print]Add functions for printing memref f16/bf16/i16 (#75094)
  [AMDGPU][NFC] Regenerate .mir test
  Reland [X86] Respect code models more when determining if a global reference can fit in 32 bits (#75386)
  Revert "[ADT][StringMap] Add ability to precompute and reuse the string hash"
  Revert "lldb: Cache string hash during ConstString pool queries/insertions"
  [RISCV] Precommit test coverage for VLMAX encodable via vsetivli
  Revert "[X86] Respect code models more when determining if a global reference can fit in 32 bits" (#75500)
  [X86] combineLoad - allow constant loads to share matching 'lower constant bits' with larger VBROADCAST_LOAD/SUBV_BROADCAST_LOAD nodes
  [X86] Respect code models more when determining if a global reference can fit in 32 bits (#75386)
  [flang] Add genEval to the AbstractConverter (#75140)
  [flang][OpenMP] Avoid unnecessary init loop, use constructor instead,… (#75482)
  [clang][modules] Strip LLVM options (#75405)
  [mlir] Fix type transformation in DropUnitDimFromElementwiseOps (#75430)
  [StandardInstrumentations] add `unwrapIR` to simplify code NFCI (#75474)
  [riscv] Consolidate a set of load-add-store tests into one file
  [Libomptarget] Move ELF symbol extraction to the ELF utility (#74717)
  [riscv] Convert a set of tests to opaque pointers
  [lldb] Fixup PopPlan assert
  [lldb] Assert immediately prior to calling PopPlan
  [VPlan] Implement mayHaveSideEffects/mayWriteToMemory for VPInterleav… (#71360)
  [InstrRef][nfc] Remove usage of unique_ptrs of arrays (#74203)
  [llvm][llvm-exegesis] Fix unused private field warning on Windows
  Use StringRef::{starts,ends}_with (NFC)
  [gn build] Port fd6e19cdc399
  [RISCV][InsertSETVTLI] Handle large immediates in backwards walk (#75409)
  [X86] combineLoad - improve constant pool matches by ignoring undef elements
  [DAG] visitTRUNCATE - format (truncate (load x)) fold code.
  [InstCombine] Explicitly fold `~(~X >>u Y)` into `X >>s Y` (#75473)
  [SystemZ][z/OS] yaml2obj for header and end records (#73859)
  [SystemZ] Improve shouldCoalesce() for i128. (#74942)
  [libc++] Remove anchors for agent queues from the BuildKite pipeline configuration (#75359)
  [SLP] Pessimistically handle unknown vector entries in SLP vectorizer (#75438)
  [ARM] Do not emit unwind tables when saving LR around outlined call (#69611)
  [clang][Interp] Start implementing binary operators for complex types
  [InstCombine] Fix insertion point
  [gn] port d7aee33029f2d029c (ws2_32.lib in llvm-config test)
  [clang][Interp] Allow evaluating standalone complex expressions
  [libc][NFC] Remove MantissaWidth traits (#75458)
  [SystemZ][z/OS] Fix STDOUT to STDERR
  [mlir][llvm] Add invariant intrinsics (#75354)
  [AMDGPU][MC] Add GFX12 VFLAT, VSCRATCH and VGLOBAL encodings (#75193)
  [AArch64][SME2] Add REQUIRES to new test
  [InstCombine] Change (add x, c) to (xor x, c) (#75129)
  [VPlan] Add a test for testing unused interleave recipes (#75026)
  [AMDGPU] Enable GCNRewritePartialRegUses pass by default. (#72975)
  [GlobalIsel][NFC] Harden MachineIRBuilder (#75465)
  [InstCombine] Improve `foldICmpWithDominatingICmp` with DomConditionCache (#75370)
  [clang][Interp] ComplexFloatingToBoolean casts
  [clang][Interp] IntegralComplexToBoolean casts
  [X86] broadcast-elm-cross-splat-vec.ll - drop constant pool check
  [AMDGPU][MC] Add GFX12 VBUFFER encoding (#75195)
  [clang][Interp] Support empty initlist initializers for complex types
  [flang][MLIR][OpenMP] Add support for `target update` directive. (#75047)
  [AArch64][SME2] Add SQRSHRN, UQRSHRN, SQRSHRUN builtins for SME2, SVE2p1 (#75325)
  [llvm][Support] Fix missing field in WSADATA warning
  [AArch64] ORRWrs is copy instruction when there's no implicit def of the X register (#75184)
  Revert "[gn] port 2e45326b088b (arm streaming attrs)"
  [GlobalISel] IRTranslator::translateGetElementPtr - don't assume a gep constant offset is representable as i64
  [clang][Interp] Start supporting complex types
  Reland "Add a test for evicting unreachable modules from the global module cache (#74894)"
  [AArch64][SME2] Add builtins for SQDMULH (#75326)
  [mlir][ArmSME][NFC] Move conversion tests (#75446)
  [mlir][ArmSME][NFC] Remove arm_sme::populateVectorTransferLoweringPatterns decl (#75442)
  [Support] Canonicalise ws2_32 casing, fix MinGW cross compilation
  Revert rG2047ab00eaf0a17e71ce5e8a5b27a8c90f034c3d "[VPlan] Add a test for testing unused interleave recipes (#75026)"
  [MachinePass] Run instrumentation before/after module pass (#70919)
  [AMDGPU][MC] Add GFX12 VDS encoding (#75316)
  Revert "[bazel] Port 2e45326b088b3b2f5c8327f6d5e61bdd2845bbbe"
  [llvm][Bazel] Add missing library to linkopts on Windows.
  [X86] Rename VPERMI2/VPERMT2 to VPERMI2*Z/VPERMT2*Z (#75192)
  [clang][NFC] Inline some lambdas to their only call site
  [M68k] Fix ODR violation in GISel code (#72797)
  [VPlan] Add a test for testing unused interleave recipes (#75026)
  Revert "[AArch64][SME] Warn when using a streaming builtin from a non-streaming function" (#75449)
  [IR] Avoid redundant map lookup [NFC]
  [libc][NFC] Remove ExponentWidth traits (#75362)
  [MLIR][SCFToOpenMP] Add num-threads option (#74854)
  [IR] Add dead_on_unwind attribute (#74289)
  [docs] remove some out-of-date content in LLVM Programmer's Manual (#74989)
  [DebugInfo] Pass string ownership to MarkupFilter (#75403)
  [flang] Use StringRef::{starts,ends}_with (NFC)
  [BOLT] Use StringRef::{starts,ends}_with (NFC)
  [clangd] Use StringRef::{starts,ends}_with (NFC)
  [clang-tidy] Use StringRef::{starts,ends}_with (NFC)
  [mlir] Use StringRef::{starts,ends}_with (NFC)
  [llvm] Use StringRef::{starts,ends}_with (NFC)
  [mlir] [tensor] Fix typo in tensor.pack documentation (#74922)
  [test][llvm-config] Allow extra libraries in test on Windows
  [CommandLine] Remove unused variable 'NearestHandler' (fixup for #74811)
  [gn build] Port ef3f476097c7
  [clang] Parse attribute [[gnu::no_stack_protector]] (#75289)
  [gn] port a5ffabce98a4 (unix sockets -> ws2_32.lib dep everywhere O_o)
  [X86][test] Merge encoding tests for avx512dq_vl
  [mlir] Fix loop pipelining when the operand of `yield` is not defined in the loop body (#75423)
  [Support] Fix WS2_32 casing
  [gn] port 2e45326b088b (arm streaming attrs)
  [LLVM] Add file magic detection for SPIR-V files. (#75363)
  [riscv] Fix build due to missing test update
  [attributes][analyzer] Implement [[clang::suppress]] - suppress static analysis warnings.
  [LLVM][Support] Fixed the compile error caused by #73603 (#75418)
  [Bazel] Add arm_vector_types.h (fixup #73258)
  [RISCV] When VLEN is exactly known, prefer VLMAX encoding for vsetvli (#75412)
  [llvm][Windows] Don't run socket tests on old versions of Windows
  [CommandLine] Better report unknown subcommands (#74811)
  [Bazel] Define BUILTIN_THREAD_POINTER where appropriate on x86-64 linux (#74574)
  [bazel] Port 2e45326b088b3b2f5c8327f6d5e61bdd2845bbbe
  [lldb] Skip 2 newly introduced tests from running on DWARF2 and clang 11 (#75406)
  [AArch64][SME] Warn when using a streaming builtin from a non-streaming function (#74064)
  [mlir][affine] Allow `memref.cast` in `isDimOpValidSymbol` (#74401)
  [dsymutil] Improve missing symbol warning message (#75378)
  [msan] Use `pthread_atfork` instead of interceptor (#75398)
  [lldb] Return index of element in ValueObject path instead of the element's value (#74413)
  [mlir][sparse] minor cleanup of transform/utils (#75396)
  [libc++] Add CI job for testing macOS C++03 (#75355)
  [hwasan] Improve support of forking with threads (#75291)
  [NFC][msan] Clang-format some includes
  [test][hwasan] Implement sanitizer_specific for HWASAN (#75280)
  [mlir][tensor] Fold consumer linalg transpose with producer tensor pack (#74206)
  [NFC][GlobalISel] Fix case of local variable
  [GlobalISel] Fix misaligned read after #74429
  [clang-format] Fix a bug in git-clang-format.bat (#75268)
  Revert "[MLIR] Fuse locations of merged constants (#74670)"
  [X86][FastISel] Support medium code model in more places (#75375)
  [asan] Install `pthread_atfork` (#75290)
  [RISCV] Extract a utility for computing bounds on VLMAX [nfc]
  [mlir,vector] Fix -Wunused-variable
  Revert "[clangd] Add test for GH75115 (#75116)"
  [libc] add mempcpy to bazel overlay (#75383)
  Reland "[Coverage][llvm-cov] Enable MC/DC Support in LLVM Source-based Code Coverage (2/3)"
  [lsan] Install `pthread_atfork` (#75281)
  [X86][test] Use separate check prefix in code-model-elf.ll
  Revert "[VPlan] Mark Select VPInstructions as not having sideeffects."
  [LV] Add test case for #75298.
  [RISCV] Add test coverage for profitable vsetvli a0, zero, <vtype> cases
  workflows/release-binaries: Add schedule to run job once per month (#73812)
  [mlir][vector] Add pattern to drop unit dim from elementwise(a, b)) (#74817)
  [clang-format] Fix parsing of `operator<() {}` (#75144)
  [runtimes] Add missing test dependencies to check-all (#75090)
  [hwasan] Add `__hwasan_get_tag_from_pointer` (#75267)
  [gn build] Port a5ffabce98a4
  [test][sanitizer] Disable test on Darwin
  [libc++] Fix `take_view::__sentinel`'s `operator==` (#74655)
  [runtimes] Don't link against compiler-rt explicitly when we use -nostdlib++ (#75089)
  [XCOFF] Display branch-absolute targets in hex. (#72532)
  [AMDGPU] Fix no waitcnt produced between LDS DMA and ds_read on gfx10 (#75245)
  [llvm][Support] Add UNIX socket support (#73603)
  Remove the builtin_headers_in_system_modules feature (#75262)
  [flang] module namelist IO with renaming (#75264)
  [SystemZ][z/OS] Add missing strnlen function for z/OS to fix build failures (#75339)
  [libc] fix unit tests (#75361)
  [libc++abi] Fix test on Android (#74753)
  [gn build] Port 6892c175c565
  [RISCV] Remove setJumpIsExpensive(). (#74647)
  [gn build] Manually port ed2d497291f
  [ConstantHoisting] Add a TTI hook to prevent hoisting. (#69004)
  [MLIR][Presburger] Fix IntegerRelation::swapVar not swapping identifiers (#74407)
  Move status badge as suggested in post-commit review
  [libc][NFC] Shorten type names in tests (#75358)
  [clang] Use StringRef::{starts,ends}_with (NFC) (#75149)
  [libc] fix -Wmacro-redefined (#75261)
  [clang][Driver] Support -fms-volatile as equivalent to /volatile:ms (#74790)
  [Transforms] Fix a warning
  [SVE2.1][Clang][LLVM]Int/FP reduce builtin in Clang and LLVM intrinsic (#69926)
  AMDGPU/GlobalISel: add AMDGPUGlobalISelDivergenceLowering pass (#75340)
  [CMake] Include opt-viewer in Fuchsia toolchain (#75296)
  [InstCombine] Fix uninitialized variable usage
  [libc++] Fix incomplete user-defined ctype specialization in test (#74630)
  [ConstraintElim] Refactor `checkCondition`. NFC. (#75319)
  [RISCV] Eliminate dead li after emitting VSETVLIs (#65934)
  [libc++][NFC] Add a few explicit 'inline' keywords, mostly in <chrono> (#75234)
  [BOLT] Fix some dwarf tests affected by 75095 (#75327)
  [X86][test] Rename some encoding tests
  [DebugInfo][RemoveDIs] Switch some insertion routines to use iterators (#75330)
  [AMDGPU] GFX12: Add Split Workgroup Barrier (#74836)
  [mlir][ArmSME] Add sve streaming compatible attribute (#75222)
  Revert "[StackColoring] Delete dead stack slots (#72633)"
  [AMDGPU] Min/max changes for GFX12 (#75214)
  [NFC] Remove dead code (#75336)
  [bazel] Port ab380c287a42c0701cd86ae2932c0cb125b9a294
  [AMDGPU] Update IEEE and DX10_CLAMP for GFX12 (#75030)
  [RemoveDIs] Support DPValue dbg.declares in MemoryOpRemark (#74108)
  [lldb][PDB] TypeQuery parameter should be ConstString
  [SystemZ][z/OS] Complete EBCDIC I/O support (#75212)
  [RemoveDIs] Update Coroutine passes to handle DPValues (#74480)
  [X86][test] Merge the decoding tests for avx512vp2intersectvl and unify the names
  [X86][test] Remove duplicated tests for avx512_vp2intersect
  [StackColoring] Delete dead stack slots (#72633)
  [X86][test] Add missing encoding/decoding tests for avx512vp2intersect and unify the names
  [lldb][PDB] Fix more issues with PDB tests
  [MLIR][NFC] Add fast path to fused loc flattening. (#75312)
  Revert "Add a test for evicting unreachable modules from the global module cache (#74894)"
  [X86][test] Add missing encoding/decoding tests for avx512vbmi and unify the names
  Add cosh op to the math dialect. (#75153)
  [CostModel] Mark ssa_copy as free (#75294)
  [X86][MC] Pre-commit test for 74713 (#75288)
  [reland][libc][NFC] Implement `FPBits` in terms of `FloatProperties` to reduce clutter (#75196) (#75318)
  [RemoveDIs] Enable conversion from dbg.declare to DPValue (#74090)
  [LoopVersioningLICM] add comment regarding dubious check (NFC)
  [gn] port 9ed20568e7de
  [lldb][PDB] Attempt to fix tests on Windows
  [AST] Switch to MemoryLocation in add method (NFC)
  [clang][Interp] Don't diagnose undefined functions when checking... (#75051)
  [X86][test] Move the x86 tests to correct folder
  [LICM] Add test show missed promotion due to AAInfo merging (NFC)
  [AArch64][ELF][llvm-readobj] Support the GCS .note.gnu.property bit (#75065)
  [Clang][AArch64] Add REQUIRES to new test
  [lld][COFF] Merge .00cfg section into .rdata. (#75207)
  [libc++] `views::split` and `views::lazy_split` shouldn't be range adaptor closures (#75266)
  [mlir][ArmSME][test] Use `only-if-required-by-ops` rather than `enable_arm_streaming_ignore` (NFC) (#75209)
  [mlir][ArmSME] Add missing dependencies in ArmSME transforms (#75269)
  [AMDGPU] Extend clang-format directive in SIDefines.h
  [clang][Sema][NFC] Add a boolean parameter comment
  [llvm][TableGen][Docs] Add tools/resources links
  [CVP] Add tests for using block values in conditions (NFC)
  [MLIR][LLVM] Support nameless and scopeless global constants (#75307)
  [X86][test] Merge the decoding tests for avx512_bf16 and unify the names
  [MLIR][LLVM] Remove disallowlist from LLVM inliner (#75303)
  [InstCombine] simplify `(X * C0) / (X * C1)` into `C0 / C1`. (#73204)
  [MachineSink] Clear kill flags of sunk addressing mode registers (#75072)
  [X86][test] Merge the decoding tests for avx512_bf16_vl and unify the names
  [lldb] Fix buildbots after PR 74786 (#75272)
  [Clang][AArch64] Add  fix vector types to header into SVE (#73258)
  [MC] Reduce size of MCLEBFragment (NFC) (#75050)
  Revert "[libc][NFC] Implement `FPBits` in terms of `FloatProperties` to reduce clutter" (#75304)
  [X86][test] Add missing encoding/decoding tests for avx512bitalg
  [X86][test] Merge the decoding tests for avx512_bf16 and unify the names
  [MLIR][Presburger] Add Gram-Schmidt (#70843)
  [libc][NFC] Implement `FPBits` in terms of `FloatProperties` to reduce clutter (#75196)
  [CodeGen] Port `ExpandMemCmp` to new pass manager (#74050)
  RegisterCoalescer: Fix implicit operand handling during rematerialize (#75271)
  [libc] [riscv] support build with `scudo` on `riscv64` (#74951)
  [CodeGen] Port `IndirectBrExpand` to new pass manager (#75287)
  [TableGen][GlobalISel] Add specialized opcodes (#74823)
  RegisterCoalescer: Add undef flags in removePartialRedundancy (#75152)
  [bazel] Port 27259f17e9d273147c648331e92000a48677f489
  [X86][test] Add missing encoding/decoding tests for avx512dq_vl
  [CodeGen] Add analyses to help for porting GC passes (#74972)
  [clang] Substitute alias templates from correct context (#75069)
  [GlobalISel] Change MatchTable entries to 1 byte each (#74429)
  [clang] Fix false positive -Wmissing-field-initializer for anonymous unions (#70829)
  [CostModel] Add some ssa.copy costmodel tests. NFC
  [test][hwasan] Add `__hwasan_enable_allocator_tagging`
  [RISCV] Bump zicfilp to 0.4 (#75134)
  [test][sanitizer] Convert test from C++ to C
  [ELF,test] Improve PROVIDE tests
  [test][sanitizer] Add `fork` include to the test
  [InstCombine] Fold `binop (select cond, a, b), (select cond, b, a)` to `binop a, b` (#74953)
  Add libc++ main builder status badge (#75283)
  Disable PGO instrumentation on naked function (#75224)
  [test][sanitizer] Remove unnececary `printf` and `gettid`
  [BPF] improve error handling by custom lowering & fail() (#75088)
  Revert "[coroutines] Use DILocation from new storage for hoisted dbg.declare" (#75282)
  [ASan][libc++] std::basic_string annotations (#72677)
  [NFC][lsan] clang-format includes
  [clangd] Add test for GH75115 (#75116)
  [CodeGen] Port `InterleavedLoadCombine` to new pass manager (#75164)
  [test][sanitizer] Allow fork_threaded test on Msan, Tsan, Ubsan (#75260)
  llvm-reduce: Handle disjoint flag
  llvm-reduce: Handle nneg flag
  Revert "[TextAPI] Add missing link to libObject" and "[TextAPI] Add DylibReader (#75006)"
  [sanitizer] Pre-commit disabled test for fork (#75257)
  [TextAPI] Add missing link to libObject
  [TextAPI] Add DylibReader (#75006)
  [gn] port 27259f17e9d2
  [RISCV] Reduce the size of the index used for RVV intrinsics. NFC (#74906)
  [lldb] Make only one function that needs to be implemented when searching for types (#74786)
  [CodeGen] Port `CFGuard` to new pass manager (#75146)
  Revert "[X86] Set SHF_X86_64_LARGE for globals with explicit well-known large section name (#74381)"
  [mlir][sparse] refactor utilities into transform/utils dir (#75250)
  Add missing dep on MLIRToLLVMIRTranslationRegistration to mlir-opt. (#75111)
  [asan] Switch initialization to "double-checked locking"
  [RISCV] Add missing break to last case in switch. NFC
  [libc] Add bind function (#74014)
  [bazel] Port 4b3446771f745bb5169354ad9027c0a1c9fca394
  [lld][WebAssembly] Don't set importUndefined when -shared is used. NFC (#75241)
  [libc++][module] Fixes std::string UDL. (#75000)
  [MLIR] Flatten fused locations when merging constants. (#75218)
  [OpenMP][NFC] Move mapping related code into OpenMP/Mapping.cpp (#75239)
  [SystemZ][z/OS] Fix build errors on z/OS in the Unix .inc files (#74758)
  [X86][FastISel] Bail out on large objects when materializing a GlobalValue
  [mlir][sparse] cleanup of CodegenEnv reduction API (#75243)
  Add `std::basic_string` test cases (#74830)
  [libc] fix issues around stack protector (#74567)
  [AMDGPU] Precommit test for LDS DMA waitcounts. NFC. (#75240)
  Add a test for evicting unreachable modules from the global module cache (#74894)
  [NFC][InstrProf] Rename internal `InstrProfiling` to `InstrLowerer` (#75139)
  [mlir][sparse]Make isBlockSparsity more robust (#75113)
  [X86] avx512-vbroadcast.ll - fix orphan check prefixes
  [bazel,unittest] Export llvm_orc_registerJITLoaderGDBWrapper
  [mlir][mesh] Add endomorphism simplification for all-reduce (#73150)
  [mlir][vector] Allow vector distribution with multiple written elements (#75122)
  [ELF] Don't create copy relocation/canonical PLT entry for a defined symbol (#75095)
  [libc++][chrono] Fixes year_month year wrapping. (#74938)
  [DebugInfo] Fix duplicate DIFile when main file is preprocessed (#75022)
  [libc++][doc] Updates module information. (#75003)
  [coroutines] Use DILocation from new storage for hoisted dbg.declare (#75104)
  [AMDGPU] CodeGen for GFX12 64-bit scalar add/sub (#75070)
  [ORC][MachO] For convenience, make MachOPlatform ref available to subclasses.
  [gn build] Port 6a6646749900
  [analyzer] Fix broken testcase (#75216)
  [RegAllocFast] NFC cleanups (#74860)
  [clang][Sema] Always clear UndefinedButUsed (#73955)
  [Transforms] Fix a warning
  [readtapi] Cleanup printing command line options (#75106)
  [flang] Fix compilation error due to variable no being used (#75210)
  [C API] Add getters and setters for fast-math flags on relevant instructions (#75123)
  [libc++][CI] Tests the no RTTI configuration. (#65518)
  [RemoveDIs] Fold variable into assert, it's only used once. NFC
  [RemoveDI] Handle DPValues in SROA (#74089)
  [AArch64][GlobalISel] Test Pre-Commit for Look into array's element
  [mlir][tensor] Fix bug in `tensor.extract(tensor.from_elements)` folder (#75109)
  [analyzer] Move alpha checker EnumCastOutOfRange to optin (#67157)
  [RemoveDIs] Handle DPValues in replaceDbgDeclare (#73507)
  [SHT_LLVM_BB_ADDR_MAP] Implements PGOAnalysisMap in Object and ObjectYAML with tests.
  [X86][GlobalISel] Add instruction selection for G_SELECT (#70753)
  [AMDGPU] Remove unused function splitScalar64BitAddSub
  [LLVM][DWARF] Add compilation directory and dwo name to TU in dwo section (#74909)
  [clang][Interp] Implement __builtin_ffs (#72988)
  [RemoveDIs] Update ConvertDebugDeclareToDebugValue after #72276 (#73508)
  [libc][NFC] Reuse `FloatProperties` constant instead of creating new ones (#75187)
  [RemoveDIs] Fix removeRedundantDdbgInstrs utils for dbg.declares (#74102)
  Reapply "[RemoveDIs][NFC] Find DPValues using findDbgDeclares  (#73500)"
  [GitHub] Remove author association print from new-prs workflow
  [libc++] tests with picolibc: mark fenv tests as unsupported (#74610)
  [clang][NFC] Remove unused parameter
  [clang][Interp][NFC] Fix a comment typo
  [libc++] P2770R0: Stashing stashing iterators for proper flattening (#66033)
  Revert "[RemoveDIs][NFC] Find DPValues using findDbgDeclares  (#73500)"
  [AMDGPU] Turn off clang-format in moveToVALU (#75188)
  [libc++] Fix incorrectly-placed attribute
  [LVI] Don't return optional from getEdgeValueLocal() (NFC)
  [RemoveDIs] Handle DPValues in LowerDbgDeclare (#73504)
  [LVI] Move bulk of getConstantRangeAtUse() implementation into Impl (NFC)
  [AMDGPU] Miscellaneous clang-format changes (#75186)
  [NFC][CLANG] Fix static analyzer bugs about large copy by values (#75060)
  [NFC][MLIR][OpenMP] Add test to check lowering omp.wsloop for GPU (#74857)
  [lld][NFC] Silence -Wuninitialized GCC 11 warnings. (#75183)
  [LVI] Reuse LatticeValue to ConstantRange conversion more
  [RemoveDIs][NFC] Find DPValues using findDbgDeclares  (#73500)
  [LVI] Don't require DataLayout in getConstantRangeOrFull() (NFC)
  [X86] Set MaxAtomicSizeInBitsSupported. (#75112)
  [Clang][SME2] Add multi-vector unpack builtins (#75075)
  [libc][NFC] Remove last use of BitsType in FloatProperties (#75174)
  [lldb] Correctly check and report error states in StackFrame.cpp (#74414)
  [TableGen] Use getSizeInBits (#75157)
  [Flang] Add a HLFIR Minloc intrinsic (#74436)
  [libc] Add mask functions to math_extras (#75169)
  [flang] Lower VALUE derived types in BIND(C) interface (#74847)
  [LVI] Drop bitcast handling (NFCI)
  [MLIR][Linalg] improve silenceable failure msg for `lower_pack` (NFC) (#75053)
  [libc][NFC] Remove remaining specializations in FloatProperties (#75165)
  [libc][NFC] Make EXPONENT_BIAS int32_t (#75046)
  [LVI] Require UndefAllowed argument to getConstantRangeAtUse() (NFC)
  [DebugInfo][RemoveDIs] Handle dbg.declares in SelectionDAGISel (#73496)
  [Clang][docs] Update  extension documentation (#75150)
  [RISCV][NFC] Use AddTargetFeature to add fast-unaligned-access (#74280)
  [NFC][DSE] Fix typo comment in eliminateDeadStores (#75166)
  [CVP] Don't use undef range for LHS of div/rem transforms
  [CVP] Add test for invalid use of undef range in urem transform (NFC)
  [X86] canonicalizeShuffleWithBinOps - generalize to handle some unary ops
  [libc] Make the bit header compatible with uint128 types (#75161)
  [CVP] Don't use undef ranges in willNotOverflow()
  [CVP] Add test for invalid use of undef range for saturating insts (NFC)
  [flang] Add struct passing target rewrite hooks and partial X86-64 impl (#74829)
  [libc][NFC] Swap ifdef logic for UInt128 (#75160)
  [LVI] Make UndefAllowed argument of getConstantRange() required
  [AMDGPU] Update VOP instructions for GFX12 (#74853)
  [X86] X86FixupVectorConstants - create f32/f64 broadcast constants if the source constant data was f32/f64
  [CVP] Don't allow undef range when inferring nowrap flags
  [CVP] Add test for incorrect use of range with undef (NFC)
  [CVP] Regenerate test checks (NFC)
  [TableGen][NFC] Remove leading spaces
  [TableGen][NFC] Format parts of DAGISelMatcher.h/DAGISelMatcherGen.cpp
  [clang][Interp] Reject static lambdas with captures (#74718)
  [clang][Interp] Decay arrays to the first element (#72660)
  [LVI] Switch getValueFromCondition() to use recursion
  [GitHub] Try a workaround to get the new contributor greeting to work (#75036)
  [SelectionDAG] Add OPC_MoveSibling (#73643)
  [clang][Interp] Implement __builtin_rotate{right,left} (#72984)
  [SelectionDAG] Add space-optimized forms of OPC_EmitNode/OPC_MorphNodeTo (#73502)
  [clang][Interp] Implement inc/dec for IntegralAP (#69597)
  [NFC] Change FindDbgDeclareUsers interface to match findDbgUsers/values (#73498)
  [SelectionDAG] Add instantiated OPC_CheckChildType (#73297)
  [JITLink][AArch32] Drop anonymous namespaces around FixupInfo helper classes (NFC)
  [llvm-c] Mitigate debug support test for systems that still default to RuntimeDyld
  [SelectionDAG] Add space-optimized forms of OPC_EmitCopyToReg (#73293)
  [NFC] Fix typo in ConstraintElimination assertion. (#75151)
  [SelectionDAG] Add space-optimized forms of OPC_EmitConvertToTarget (#73286)
  [SelectionDAG] Add instantiated OPC_CheckType (#73283)
  [LoopVectorize] Improve algorithm for hoisting runtime checks (#73515)
  [AMDGPU][NFC] Test autogenerated llc tests for COV5 (#74339)
  [mlir][Bazel] Adjust for d5fb4c0f118b47db74233af2d99ae075e1dbe148
  [RISCV] Remove forward declaration and unused argument. NFC
  [SCEV] Use loop guards when checking that RHS >= Start (#75039)
  [RISCV] Move test to RVV directory. NFC
  [RISCV] Don't set AVL if only zeroness is demanded (#74049)
  [RISCV] Remove unecessary early exit in transferBefore (#74040)
  [BPF][GlobalISel] select non-PreISelGenericOpcode (#75034)
  [CodeGen] Port `SjLjEHPrepare` to new pass manager (#75023)
  [mlir][GPU] Apply ClangTidy fixes
  [llvm-exegesis] Use explicit error classes for different snippet crashes (#74210)
  [Preprocessor] Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_16 for AArch64 (#74954)
  [M68k][MC] Add fmove.l from / to fp control registers
  Revert "[RISCV] Update the interface of sifive vqmaccqoq (#74284)"
  [clang] Crash when referencing capture in static lambda (#74661)
  [AArch64]: Added code for generating XAR instruction (#75085)
  [Github] Set start rev to merge base in code format action (#75132)
  [LV] Added pre-commit tests for changing loop interleaving count computation (#74689)
  [RISCV] Update the interface of sifive vqmaccqoq (#74284)
  Link NVVM translation in the to LLVMIR registration library
  [llvm] Use StringRef::{starts,ends}_with (NFC) (#74956)
  [MLIR][NVVM] Enable nvvm intrinsics import to LLVMIR   (#68843)
  [MacroFusion] Support multiple predicators (#72219)
  [CodeGen] Port `SelectOptimize` to new pass manager (#74920)
  [MLIR] Fuse locations of merged constants (#74670)
  Revert "[MacroFusion] Support multiple predicators (#72219)"
  [MacroFusion] Support multiple predicators (#72219)
  [X86] Handle unsized types in TargetMachine::isLargeGlobalObject() (#74952)
  [MachinePipeliner] Fix store-store dependences (#72575)
  [X86] Handle ifuncs in TargetMachine::isLargeGlobalObject() (#74911)
  [X86] Allow constant pool references under medium code model in X86InstrInfo::foldMemoryOperandImpl() (#75011)
  [gn] port 6747fc07d1aa9 (made necessary by 54397f9ac128)
  [ORC][LLJIT] Expose ORCPlatformSupport to aid custom native Platform setup.
  [OpenMP][NFC] Improve profiling for the offload runtime
  [OpenMP][NFCI] Remove effectively unused mutex
  [OpenMP][NFC] Move rtl.cpp to OffloadRTL.cpp
  [OpenMP][NFC] Move api.cpp to OpenMP/API.cpp
  [CodeGen][GC] Remove `GCInfoPrinter` (#75033)
  [CodeGen] Port `JMCInstrumenter` to new pass manager (#75049)
  [scudo] Add utilization percentages for stats. (#75101)
  Add missing paren
  [mlir][python] update type stubs (#75099)
  lldb: Cache string hash during ConstString pool queries/insertions
  [ADT][StringMap] Add ability to precompute and reuse the string hash
  [mlir][SparseTensor][NFC] Use `tensor.empty` for dense tensors (#74804)
  [BOLT][DWARF] Fix handling of debug_str_offsets (#75100)
  [mlir][bufferization] Fix `regionOperatesOnMemrefValues` (#75016)
  [NFC][InstrProf] Move `InstrProfiling` to the .cpp file (#75018)
  [test] Change llc -march to -mtriple
  [clang] Remove stale release note
  [flang] Modify lto testcase for AIX (#74496)
  [SpecialCaseList] Use glob by default (#74809)
  [-Wunsafe-buffer-usage] Add FixableGadget for AddAssign in UnspecifiedUntypedContext (#71862)
  [AArch64] Set MaxAtomicSizeInBitsSupported. (#74385)
  [lldb] Disable new TestLocationListLookup when clang version is <= 11 (#75102)
  [ORC][MachO] Expose SimpleMachOHeaderMU and related utilities.
  [bazel] Port #73257 54397f9ac128568838f2ac7bfc8e1f94b3eb264d
  [RISCV] Macro-fusion support for veyron-v1 CPU. (#70012)
  [readtapi] Swap anon namespaces in favor of annotating `static`, NFC (#75078)
  Revert "[mlir][vector] Make `TransposeOpLowering` configurable (#73915)" (#75062)
  [bazel] Port #73257 54397f9ac128568838f2ac7bfc8e1f94b3eb264d
  [libc++][test] disable all atomic<long double> tests (#74201)
  [flang] Handle continuation line edge case (#74751)
  [flang][runtime] Fix fixed-width field internal wide character input (#74683)
  [flang][runtime] Fix octal input of REAL(10) (#74658)
  [flang][runtime] Support READ after WRITE w/o positioning (#74650)
  [BOLT] Fix warnings
  [Profile] Add missing COMPILER_RT_VISIBILITY to __llvm_profile_disable_continuous_mode.
  [RISCV] Correct the SEW=64 MUL diagnostic to refer to V as an extension.
  [RISCV] Shorten diagnostic a bit.
  [flang][runtime] Don't use endfile record number for EOF detection on… (#74640)
  [flang][runtime] Delete CharacterAssign() (#74621)
  [flang][runtime] Clear last record in internal WRITE even if nothing … (#74528)
  [ELF][test] Improve copy relocation/canonical PLT entry tests
  [BOLT] Provide backwards compatibility for YAML profile with std::hash (#74253)
  [flang][runtime] Terminate last partial record after non-advancing write (#74524)
  [RISCV] Simplify checking whether SEW=64 multiply builtins require V. NFC
  [gn build] Port 54397f9ac128
  [lldb-dap] Emit declarations along with variables (#74865)
  [CUDA] Add support for CUDA-12.3 and sm_90a (#74895)
  [flang][runtime] Handle Fw.0 case that needs to round up to 1.0 (#74384)
  [NVPTX] Custom lower integer<->bf16 conversions for sm_80 (#74827)
  [flang] Correct definability checking for INTENT(IN) pointers (#74158)
  [RISCV] Use getBuiltinVectorTypeInfo to simplify code. NFC
  [flang][runtime] Make defined formatted I/O process format elementally (#74150)
  [llvm-c] Expose debug support for LLJIT in Orc C-API bindings (#73257)
  [lldb-dap] Include [opt] in the frame name only if a custom frame format is not specified. (#74861)
  [flang][runtime] Crash more informatively in defined I/O error case (#74134)
  [github] Enable assertions on test workflow (#74849)
  [AArch64] Correctly mark Neoverse N2 as an Armv9.0a core (#75055)
  [libc][NFC] fix uint128 init in float properties (#75084)
  [OpenMP] Add extra flags to libomptarget and plugin builds (#74520)
  Fix to msvc::no_unique_address causing assert when used with __declspec(empty_bases) (#74776)
  [mlir][sparse] fix bug in custom reduction scalarization code (#74898)
  [mlir][spirv] Add folding for Bitwise[Or|And|Xor] (#74193)
  [ORC][MachO] Enable customization of MachO-headers produced by MachOPlatform.
  [llvm][SanitizerCoverage] Remove no-op 'ptr addrspace(0)' to 'ptr addrspace(0)' pointercast (NFC)
  [flang] update ppc-vec-store-elem-order.f90 after #74709 (NFC) (#75064)
  [llvm-readobj] --needed-libs: support --elf-output-style=JSON (#75028)
  [AArch64][SME2] Add PEXT, PSEL builtins for SME2 (#72827)
  [RISCV] Enable target attribute when invoked through clang driver (#74889)
  [libc][NFC] Simplify FloatProperties implementation (#74821)
  [flang] Support -mvscale-min and mvscale-max on all targets (#74633)
  [SystemZ]  Use LCGR/AGHI for i64 XOR with -1 (#74882)
  [clang][Interp] Fix float->int casts overflowing (#72658)
  [X86] X86FixupVectorConstants - create f32/f64 broadcast constants if the source constant data was ANY floating point type
  [X86] combineConcatVectorOps - constant fold vector load concatenation directly into a new load.
  [libc][NFC] Simplify FloatProperties implementation (#74481)
  [libc][NFC] Remove custom leading_zeroes, factor in frequent typenames (#74825)
  [libc][NFC] Fix mixed up biased/unbiased exponent (#75037)
  [flang] fix ppc test broken after #74709 (#74826)
  [clang][Interp][NFC] Handle body-less functions like implicit ones
  Fix test rocm-detect.hip (#74872)
  [X86] combineConcatVectorOps - pull out repeated DAG.getContext calls. NFC.
  [Dexter] Set ShouldBuild=false for Visual Studio solutions (#75045)
  [MLIR][Linalg] (NFC) Drop `verify-diagnostics` from `transpose-conv2d.mlir`
  [Profile] Disable continuous mode when reset to default.profraw due to malformed LLVM_PROFILE_FILE. (#74879)
  [clang] Add support for -fcx-limited-range, #pragma CX_LIMITED_RANGE and -fcx-fortran-rules.  (#70244)
  [OpenMP] Change check for OS to check for defined for a macro (#75012)
  [Clang][Sema] Diagnose friend function specialization definitions (#72863)
  [llvm][Attributor] Strip AddressSpaceCast from 'constructPointer' (#74742)
  [lldb][NFC] Simplify DWARRFDeclContext::GetQualifiedName (#74788)
  [Linkerwrapper] Make -Xoffload-linker pass directly to `clang`
  [InstCombine] Preserve inalloca tag when transforming alloca
  [clang][dataflow] Convert `SpecialBoolAnalysis` to synthetic fields. (#74706)
  [bazel] Port 2460bf2facd1c0208ecda68c427a256710246dbb
  [gn build] Port 2460bf2facd1
  [flang][driver] Add -fno-fortran-main (link time) option to remove Fortran_main from link line (#74139)
  [VPlan] Mark Select VPInstructions as not having sideeffects.
  [AssignmentTracking] Skip large types in redundant debug info pruning (#74329)
  [InstCombine] Don't create unnecessary zero-index GEP (NFCI)
  [BPF][GlobalISel] add initial gisel support for BPF (#74999)
  [Bitcode] Check validity of fcmp/icmp predicate
  [X86] Rename VBROADCASTF128/VBROADCASTI128 to VBROADCASTF128rm/VBROADCASTI128rm (#75040)
  [Bitcode] Check type of alloca size argument
  [MemCpyOpt] Don't perform call slot opt if alloc type is scalable (#75027)
  [Bitcode] Add missing getValue() return value checks
  [NFC][RemoveDIs] Add LocationType parameter to DPValue ctor (#74091)
  [mlir][CAPI] Add mlirOpOperandGetValue (#75032)
  [mlir][llvm] Fix negative GEP crash in type consistency (#74859)
  [mlir][ArmSME] Remove `vector.print` legality from ArmSMEToSCF (NFC) (#74875)
  [VFABI] Improve VFABI unit tests (#73907)
  [SCEV] Add test for unnecessary umax in BECount (NFC)
  [X86] evex-to-vex-compress.mir - strip trailing whitespace
  [mlir][vector] Fix crash on invalid `permutation_map` (#74925)
  [GlobalISel] Add G_PREFETCH (#74863)
  [SymbolFileDWARF][NFC] Remove duplicated code checking for type tags (#74773)
  [AMDGPU] Form V_MAD_U64_U32 from mul24 (#72393)
  [Serialization] Use packed bits to initialize UserDefinedLiteral
  [NFC][AMDGPU] Unify AMDGPU address space enum (#73944)
  [mlir][Python] Apply ClangTidy findings.
  [MLIR][IntegerRangeAnalysis] Avoid crash reached when loop bound is uninitialized (#74832)
  [BPF] use target triple for pattern predicates (#74998)
  [AMDGPU] Update sendmsg types for GFX12 (#74888)
  [NFC] Missing argument for the PR issue number in the code format helper
  Recommit [NFC] [Serialization] Packing more bits
  [LLVM-C] Support operand bundles (#73914)
  code-format: Improve the code-format-helper to be able to run as a git hook (#73957)
  [NFC] Modify test to use autogenerated assertions
  [MLIR][SCF] Add support for pipelining dynamic loops (#74350)
  Revert "[clang] Remove unused variable 'ExprDependenceBits' in ASTWriterDecl.cpp (NFC)"
  Revert "[NFC] [Serialization] Packing more bits"
  [CodeGen] Rename `winehprepare` -> `win-eh-prepare` (#75024)
  [CodeGen][MachineScheduler][NFC]Update some comments of scheduler (#74705)
  [clangd] Initialize HighlightingsBuilder::Resolver (#74971)
  [NFC][VPlan] Simplify VPValue::removeUser (#74708)
  [clang] Remove unused variable 'ExprDependenceBits' in ASTWriterDecl.cpp (NFC)
  Fix uninitialized field post PR #74970
  [BOLT] Fix local out-of-range stub issue in LongJmp (#73918)
  [LoongArch] Add codegen support for [X]VF{MSUB/NMADD/NMSUB}.{S/D} instructions (#74819)
  [NFC] [Serialization] Packing more bits
  [NFC][InstrProf] Refactor InstrProfiling lowering pass (#74970)
  [CodeGen] Update DwarfEHPreparePass references in `CodeGenPassBuilder.h` (#74068)
  [libc++][test] Fix more MSVC and Clang warnings (#74965)
  [libc++][test] Fix MSVC warnings with `static_cast`s (#74962)
  [NFC][AArch64][ELF][PAC] Update AUTH relocation IDs (#74986)
  [libc][math] Ensure fputil::nextafter is called with correct type args (#75009)
  [DirectoryWatcher] Include limits.h instead of MathExtras.h (NFC)
  [ARM] Include bit.h instead of MathExtras.h (NFC)
  [ADT] Include bit.h instead of MathExtras.h (NFC)
  [NFC][test][llvm-readobj] Use single `#` for FileCheck directives (#74985)
  [Target] Remove unused forward declarations (NFC)
  [Transforms] Remove unused forward declarations (NFC)
  [llvm][docs] Fix typo in `CoverageMappingFormat.rst` (#71189)
  [libc++][test] Cleanup `LIBCPP_ONLY(meow_assert(...))` to `LIBCPP_MEOW_ASSERT(...)` (#74967)
  [libc++][test] Consistently use `TEST_SHORT_WCHAR` (#74958)
  [libc++][test] `concat_macros.h`: Fix `TEST_HAS_NO_WIDE_CHARACTERS` syntax damage (#74987)
  [LAA] Add tests with dependencies may preventing st-to-ld forwarding.
  [libc++][test] Fix `MaybePOCCAAllocator` to finally meet the allocator requirements (#74960)
  [AArch64] Fix case of 0 dynamic alloc when stack probing (#74877)
  [CodeGen] Port `InterleavedAccess` to new pass manager (#74904)
  [libc++] LWG-4021 "`mdspan::is_always_meow()` should be `noexcept`", use `LIBCPP_STATIC_ASSERT` for `noexcept` strengthening (#74254)
  [GISel][TableGen] Fix accidental operand name clashes in patterns (#74492)
  [GlobalISel] Add a missing include
  [Support] Remove unnecessary includes (NFC)
  [llvm] Adjust inlucdes of ArrayRef.h (NFC)
  [docs] Add missing quotation mark after #73774
  [Target] Include bitset (NFC)
  [Transforms] Remove unnecessary includes (NFC)
  [X86] Rearrange a few atomics tests. NFC.
  [clang-tools-extra] Use llvm::to_underlying (NFC)
  [clang] Use llvm::to_underlying (NFC)
  [RISCV] Remove unnecessary call to isSupportedExtensionFeature.
  [ADT] Rename SmallString::{starts,ends}with to {starts,ends}_with (#74916)
  [RISCV] Use Triple::isRISCV64(). NFC
  [LAA] Remove duplicated test.
  [ADT] Remove StringRef::{starts,ends}with_insensitive (#74918)
  [test][llvm-objdump][AArch64] Add tests for ELF AUTH constants (#74298)
  [HLSL] Define RasterizerOrderedBuffer resource
  [BOLT] Determine address size from binary (#74870)
  [RISCV] Remove Name and OverloadedName from RVVIntrinsicDef. NFC (#74907)
  [DirectX] Move ROV info into HLSL metadata. NFC

Change-Id: I7f909a9bb731070213995ae08d1bb3c6032a7b7d
Signed-off-by: greenforce-auto-merge <greenforce-auto-merge@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
3 participants