opt -passes="aggressive-instcombine" output depends on debug info #69925
Labels
Comments
|
This starts happening with https://reviews.llvm.org/D127392 |
mikaelholmen
added a commit
to mikaelholmen/llvm-project
that referenced
this issue
Oct 26, 2023
We get different results with/without debug info present.
zahiraam
pushed a commit
to zahiraam/llvm-project
that referenced
this issue
Oct 26, 2023
We get different results with/without debug info present.
zahiraam
pushed a commit
to zahiraam/llvm-project
that referenced
this issue
Oct 26, 2023
…llvm#70200) We previously included debug instructions when counting instructions when looking for loads to combine. This meant that the presence of debug instructions could affect optimization, as shown in the updated testcase. This fixes llvm#69925.
fadlyas07
pushed a commit
to greenforce-project/llvm-project
that referenced
this issue
Oct 28, 2023
* llvm-project/main: [clang-format] Change LLVM style to BreakAfterAttributes == ABS_Leave (llvm#70360) [MLIR][Presburger] Add simplify function (llvm#69107) [compiler-rt] Fix src_rep_t_clz and clz_in_sig_frac Revert "Revert "[InstCombine] Add oneuse checks to shr + cmp constant folds."" [LoopDist] Update the pragma info of loop distribute, NFC (llvm#69825) [JITLink][RISCV] Implement eh_frame handling (llvm#68253) [SimplifyCFG] Reland transform for redirecting phis between unmergeable BB and SuccBB (llvm#68473) [analyzer] Extend EnumCastOutOfRange diagnostics (llvm#68191) [mlir][tosa] Add verifier for tosa.reverse (llvm#70500) TypeSize: remove unused deprecated methods [mlir][bytecode] Fix D155919 and enable backward-compatibility and back-deployment between version 5 and version 6 of the bytecode encoding (llvm#70498) [BasicBlockSections] Apply path cloning with -basic-block-sections. (llvm#68860) [mlir][Interfaces] `LoopLikeOpInterface`: Add helpers to query tied inits/iter_args (llvm#70408) [OpenMP] Fix building for 32-bit DragonFly, NetBSD, OpenBSD (llvm#70527) [Driver][NFC] Fix a typo in the function name [Flang][OpenMP] skip CompilerDirective on ResolveOmpTopLevelParts. (llvm#69580) [DataFlowSanitizer] Remove no-op ptr-to-ptr bitcasts (NFC) [mlir][sparse] remove obsoleted output methods from runtime (llvm#70523) [mlir][sparse] unify sparse_tensor.out rewriting rules (llvm#70518) [OpenMP] record-replay use static-cast (llvm#70516) [lldb] Remove some declarations without definitions (llvm#70514) [InstCombine] Add folds for (icmp eq/ne (and (add/sub/xor A, P2), P2), 0/P2) [InstCombine] Add tests for folding (icmp eq/ne (and (add/sub/xor A, P2), P2), 0/P2); NFC [AMDGPU] make v32i16/v32f16 legal (llvm#70484) [mlir][sparse] minor edit of map reinterpretation pass doc (llvm#70513) [lldb] Add logging to ObjectFileMachO::ParseSymtab (llvm#70490) [libc++] Disable test which timeouts on ARM with sanitizers Revert "[InstCombine] Add oneuse checks to shr + cmp constant folds." [clang] Language to String function (llvm#69487) [mlir][sparse] add ReinterpretMapScopeOption for the pass (llvm#70486) [mlgo][regalloc] Fix tests post 9a091de [AArch64] Clarify that Anyext is OK for MOPS instructions. NFC [Test] NFC. Add missing "REQUIRES: x86-registered-target" to LoopStrengthReduce/scev-incorrect-nuw-inference.ll [RISCV] Reduce the number of parameters to copyPhysRegVector. NFC (llvm#70502) [RISCV] Use RISCVInstrInfo::movImm to implement most of RISCVPostRAExpandPseudo::expandMovImm (llvm#70389) [RISCV] Refactor RISCVPostRAExpandPseudo::expandMovImm and RISCVInstrInfo::movImm to prepare for merging. [AMDGPU] Remove unneeded implicit-def from shrink-i32-kimm.mir. NFC. (llvm#70489) [flang][openacc] Convert rhs expr to the lhs type on atomic read/write (llvm#70377) [Test] NFC. Add a test exposing a SCEV bug causing an LSR miscompile [clang] Remove no-op ptr-to-ptr bitcasts (NFC) [OpenMP][AMDGPU] Do not include 'ockl' implementations in OpenMP (llvm#70462) [libc] Optimize the RPC memory copy for the AMDGPU target (llvm#70467) [Driver][NFC] Make some derived classes of Tool final (llvm#70416) [X86, Peephole] Enable FoldImmediate for X86 [RISCV] Add copyPhysRegVector to extract common vector code out of copyPhysRegVector. (llvm#70497) [RISCV] Separate FPR and VR copyPhysReg implementation. (llvm#70492) [clang-tidy][test] Add more bugprone-string-constructor tests (llvm#70456) TosaToLinalgNamed: add option to prefer HWCF kernel layout for Conv2D ops. (llvm#70482) [lldb] Make ScriptedInterface::CreateObjectPlugin less strict SwiftCallingConv: Fix the splitVectorEntry function (llvm#69953) [AArch64][TTI] Improve `LegalVF` when gather loads are scalarized [libc] Fix incorrect printing for alt mode ints (llvm#70252) [Headers] Add \returns to _rdpid_u32 (llvm#70481) [AMDGPU] Set Size to 4 for V_MOV_B64_PSEUDO and S_MOV_B64_IMM_PSEUDO (llvm#70376) [RISCV][GISel] Fix failing test case for G_BSWAP [RISCV][GISEL] Add legalizer for G_BSWAP (llvm#70226) [mlir][Vector] Update v.contract -> v.outerproduct tests (NFC) (llvm#70449) [mlir][ml_program] Add asm names for load (llvm#70409) [openmp][flang] Add tests for map clause (llvm#70394) TypeSize: remove unused deprecated methods [Flang][OpenMP] Port openmp threadprivate tests to HLFIR flow (llvm#70440) [PassTimingInfo] Handle nested timers in passes (llvm#70165) [clang-tidy]Fix PreferMemberInitializer false positive for reassignment (llvm#70316) [SVE] Fix incorrect offset calculation when rewriting an instruction's frame index. (llvm#70315) [clang]set invalid for lambda which missing capture `this` (llvm#70432) [ARM] Add a test for incorrect demand bits / undef fold. NFC [lldb][lldb-dap] Remove extra spaces at the end of --help examples [flang] Consider bind(c) when lowering calls to intrinsic module procedures (llvm#70386) [libc++] Fix incorrect length check in std::basic_filebuf [libc++] Make sure we implement and test LWG2280 properly (llvm#67670) Revert "[CodeGen] Improve ExpandMemCmp for more efficient non-register aligned sizes handling (llvm#69942)" [clang] use relative paths for builtin headers during module compilation (llvm#68023) [mlir][SVE] Add more e2e test for vector.contract (llvm#70367) [gn build] Port b0b8864 [libc++] Fix flakiness in `atomic_notify_all.pass.cpp` (llvm#70436) [AMDGPU] Try to fix the block prologs broken by RA inserted instructions (llvm#69924) [VPlan] Add initial anlysis to infer scalar type of VPValues. (llvm#69013) [X86] Add test case for Issue llvm#66150 [LV] Add test case to scalarize ptrtoint instructions. [InstSimplify] Fold (a != 0) ? abs(a) : 0 (llvm#70305) [Flang][OpenMP] Move the host-ir-flag test to the Driver directory (llvm#70441) [VPlan] Handle VPValues without underlying values in getTypeForVPValue. [mlir][vector] Update v.contract -> v.outerproduct tests (1/N) (llvm#70379) [AMDGPU] Add pseudo instructions for SGPR spill to VGPR (llvm#69923) [CodeGen] Improve ExpandMemCmp for more efficient non-register aligned sizes handling (llvm#69942) [lldb] Return unimplemented error from ScriptedProcessInterface::CreatePluginObject [RISCV] Keep same SEW/LMUL ratio if possible in forward transfer (llvm#69788) [LLVM][AArch64][Assembly]: Add FAMINMAX assembly/disasse… (llvm#70115) [Driver][test][NFC] Simplify some of the OpenBSD tests [clang-format][NFC] Remove extraneous newlines in remaining unit tests [AMDGPU] Give some of the VI V_PK_* instructions a subtarget predicate. (llvm#70334) [mlir][SVE] Add an e2e test for vectorization of linalg.matmul (llvm#70372) [InstCombine] Remove redundant cast of GEP fold (NFC) [mlir][Vector] Fix integration test for vector.maskedload narrow type… (llvm#70431) [clang-format][NFC] Remove more extraneous newlines in unit tests Reland [mlir][tools] Introduce tblgen-to-irdl tool (llvm#70121) [libc++] Fix _CopySegment helper in ranges::copy(join_view, out) when called in a static assertion context (llvm#69593) [C++20] [Modules] Chose BMI from for module m with the last -fmodule-file=<module-name>= option [test][compiler-rt] Mark several tests as UNSUPPORTED on LoongArch (llvm#69699) [mlir][Vector] Add narrow type emulation pattern for vector.maskedload (llvm#68443) [clang][Interp][NFC] Rename a parameter [clang][Interp][NFC] Use delegate() address-of operators [Driver] Link Flang runtime on FreeBSD, NetBSD, OpenBSD, DragonFly and Haiku (llvm#69817) AMDGPU: Add infinite looping testcase after subrange spilling change [mlir][vector] Fix off-by-one error in `getTransferChunkAccessed` (llvm#70292) [mlir][vector] Add scalable vectors to tests for vector.contract (llvm#70039) [flang] Implement legacy %VAL and %REF actual arguments (llvm#70343) [InstSimplify] Add tests for llvm#69050 and llvm#69091 (NFC) [docs] Improve README: point to office hours and online sync-ups (llvm#69323) TypeSize: remove redundant getFixed and getScalable [CMake] Address the issue introduced in llvm#69869 Replace TypeSize::{getFixed,getScalable} with canonical TypeSize::{Fixed,Scalable}. NFC [X86][EVEX512] Do not allow 512-bit memcpy without EVEX512 (llvm#70420) [flang] Move whole allocatable assignment implicit conversion to lowering (llvm#70317) Revert "[libc++][tests] Fix a few remaining instances of outdated static assertion regexes in our test suite" Apply clang-tidy fixes for misc-include-cleaner in DenseAnalysis.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in DeadCodeAnalysis.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in ConstantPropagationAnalysis.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in CallGraph.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in LocalAliasAnalysis.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in AliasAnalysis.cpp (NFC) [RISCV] Correct copyPhysReg for GPRPF64. (llvm#70419) [mlir][IR] Make `OpOperand` comparable (llvm#70410) Fix typo in c07903a [RISCV] Fix incorrect use of Zfa fli instruction for negative minimum value. (llvm#70411) [RISCV] Add test cases showing fli being used for negative min normalized value. [clang][ExprConst] Fix crash on uninitialized array subobject (llvm#67817) [libc++abi] Disable test with a leaks [clang-format][NFC] Remove extraneous newlines in some unit test files [lldb/test] Fix failures following ec456ba [llvm] Add myself as codeowner for TextAPI (llvm#70399) [Driver][NFC] Make use of auto (llvm#70400) [Workflow] Expand code-format-helper.py error reporting (llvm#69686) [DominanceFrontier] make iterating dereferenced DominanceFrontierBase::find deterministic (llvm#69711) [CodeLayout] Changed option names cds to cdsort (llvm#69668) [mlir][sparse] add boilterplate code for a new reintepret map pass (llvm#70393) Add target REQUIRES to 3 more tests Revert "Revert "[OpenMP][NFC] Add min/max threads/teams count into the KernelEnvironment (llvm#70257)"" Revert "[OpenMP][NFC] Add min/max threads/teams count into the KernelEnvironment (llvm#70257)" [mlir][www] Add missing include for ARM SVE passes [mlir][sparse] implements sparse_tensor.reinterpret_map (llvm#70388) [RISCV][GISel] Rename XLenVT sXLen to be consistent with other LLTs. (llvm#70288) [mlir][sparse] Return actual identity map instead of null map (llvm#70365) Always use pthread_rwlock on Apple platforms (llvm#70151) [OpenMP] Repair mlir test [mlir][www] Fix generation of the Mesh dialect doc [scudo] Pass the max number of blocks to popBlocks (llvm#70243) [lldb] Add OperatingSystem base class to the lldb python module [lldb] Introduce OperatingSystem{,Python}Interface and make use it [mlir][sparse] introduce sparse_tensor.reinterpret_map operation. (llvm#70378) Revert "[OpenMP] Patch for Support to loop bind clause : Checking Parent Region" [NFC] In InstrProf, generalize helper functions to take 'GlobalObject'. They currently take 'Functions' as function parameters or have 'Func' in the name. (llvm#70287) [OpenMP][NFC] Add min/max threads/teams count into the KernelEnvironment (llvm#70257) Add another missing REQUIRES: line [Flang][OpenMP] Port copyin and commonblock privatisation tests to HLFIR flow [OpenMP] Pass min/max thread and team count to the OMPIRBuilder (llvm#70247) [OpenMP][NFC] Fix test (remove wrong autogen header) [OpenMP][NFC] Remove untested code emitting no-op call [OpenMP] Remove SPMD specific handling during globalization [OpenMP][NFC] Clang format some tests [OpenMP][FIX] Move workaround code to avoid races Fix opt/invalid-target.ll on Windows bots Add missing REQUIRES lines to unbreak buildbots [Driver] Reject unsupported -mcmodel= (llvm#70262) [libc++] <experimental/simd> Add assignment operator of simd reference (llvm#70020) [clang] Document -Wglobal-constructors behavior (llvm#68084) [clang-tidy] Improve bugprone-unused-return-value check (llvm#66573) [AArch64][GlobalISel] Add support for extending indexed loads. (llvm#70373) [clang-format][NFC] Remove some extraneous newlines at end of test cases [AArch64] Fix st2 check for nearby store with debug info. [clang][OpenMP] Fix target data if/logical expression assert fail (llvm#70268) [mlir][SVE] Add an e2e test for vector.contract (llvm#69845) Update llvm/docs/MyFirstTypoFix.rst for post-Phabricator / Pull-requests world (llvm#70310) [mlir][sparse] merger cleanup (llvm#70371) [Tosa] Add tosa-to-linalg-pipeline for testing (llvm#69997) Fix typo in swift section name [RISCV] Disable lax vector conversions between RVVBuiltin types and RVVFixedLengthDataVector. [opt] Infer DataLayout from triple if not specified Remove unnecessary newline from error message [clang-tidy] Add StrictMode to cppcoreguidelines-pro-type-static-cast-downcast (llvm#69529) [Driver] Clean up unused architecture related bits for *BSD's (llvm#69809) [AArch64][GlobalISel] Fold variable into assert [Driver] Fix -r handling on Solaris (llvm#70322) [InstCombine] Add oneuse checks to shr + cmp constant folds. [AST] Only dump desugared type when visibly different (llvm#65214) [Github] Fix libc docs build (llvm#70363) [mlir][sparse] remove reshape dot test (llvm#70359) clarify NaN propagation in fptrunc (llvm#68554) [mlir] Remove no-op ptr-to-ptr bitcasts (NFC) [RISCV][GISel] Allow G_AND/G_OR/G_XOR to have s32 types on RV64. [gn build] Port 88d00a6 [mlir][sparse] cleanup conversion test (llvm#70356) Reland [dsymutil] Add support for mergeable libraries (llvm#70256) [MLIR] Add SyclRuntimeWrapper (llvm#69648) [AArch64][GlobalISel] Add support for pre-indexed loads/stores. (llvm#70185) Fix comment in wasm unreachable test (llvm#70340) [libc++][NFC] Remove unused typedefs in filesystem::path helpers (llvm#70331) [libc++] Optimize string operator[] for known large inputs (llvm#69500) [MLIR][Arith] Fix arith::AtomicRMWKind::maximumf's identity value (llvm#70312) [clang] Prioritze decl comments from macro expansion site (llvm#65481) [lldb][test][Windows] XFAIL IR memory map test [Flang][OpenMP] Port a few parallel tests to HLFIR flow [mlir][sparse] cleanup merger test, add header (llvm#70279) Diagnose use of VLAs in a coroutine (llvm#70341) [mlir][linalg] Replace CopyOp from memref to linalg in linalg PromoteOp (llvm#69154) [SLP]Improve isGatherShuffledEntry by trying per-register shuffle. [gn build] Port a349092 Revert "[SLP]Improve isGatherShuffledEntry by trying per-register shuffle." [Libomptarget] Do not pass 'nogpulib' to the non-LTO Nvidia tests (llvm#70327) [libc++][tests] Fix a few remaining instances of outdated static assertion regexes in our test suite [libc++] Make sure std::declval() produces an error when ODR-used [libc++] Add __small_buffer [clang][NFC] Refactor enums that hold size of `Type` and `DeclContext` bit-fields (llvm#70296) [Clang] Fix several `-Wdocumentation` warnings (NFC) [llvm][AArch64][Assembly]: Add FP8 instructions assembly and disassembly. (llvm#69632) [X86] Fix gcc warning about mix of enumeral and non-enumeral types. NFC [AMDGPU] Correct assert that incorrectly chained multiple == operators. (llvm#70291) [AMDGPU] Use const reference in SIInstrInfo::buildExtractSubReg. NFC. [AMDGPU] Assert that we can find subregs in copyPhysReg. NFC. (llvm#70332) [Security Group] add github names of security group members. (llvm#69304) [lldb] Try to fix build after d1556e5 [AArch64] Regenerate addr-of-ret-addr.ll [AArch64] Regenerate tests to show missing constant comments [clang][Interp] Only diagnose null field access in constant contexts (llvm#69223) [AMDGPU] Use alloca address space in rewrite-out-arguments.ll (llvm#70269) [mlir][emitc] Add a structured for operation (llvm#68206) [SLP][X86] Regenerate hadd/hsub tests with full set of check-prefixes [clang][Interp] Add explicit dummy descriptors (llvm#68888) [analyzer] Loop should contain CXXForRangeStmt (llvm#70190) [clang][Interp][NFC] Fix a doc comment [libc++] Encode additional ODR-affecting properties in the ABI tag (llvm#69669) [SLP]Improve isGatherShuffledEntry by trying per-register shuffle. [clang][Interp] Correctly emit destructors for multi-dimensional arrays (llvm#69140) Revert "[mlir][SVE] Add an e2e test for vectorization of linalg.matmul (llvm#69592)" [flang] Regularize TODO messages for coarray intrinsics (llvm#70281) [VP] Check if VP ops with functional intrinsics are speculatable (llvm#69504) Let M68kMCCodeEmitter set Scratch size. (llvm#69898) Revert "[SimplifyCFG] Delete the unnecessary range check for small mask operation (llvm#70324) [InstSimplify] Remove redundant pointer icmp fold (NFCI) [flang][runtime] Avoid dependency on libc++ for `std::__libcpp_verbose_abort` [AMDGPU] precommit lit test for PR 69924. Revert "[AMDGPU] Use `S_CSELECT` for uniform i1 ext (llvm#69703)" [mlir][SVE] Add an e2e test for vectorization of linalg.matmul (llvm#69592) [clang][Interp] Handle unknown-size arrays better (llvm#68868) [LoongArch][test] Add some ABI regression tests for empty struct. NFC Add `isBatchVecmat` utilities for `linalg.batch_vecmat` (llvm#70284) [Bazel] Fixes for 96e040a. [libc] memmove optimizations (llvm#70043) Revert "[AMDGPU] Cleanup hasUnwantedEffectsWhenEXECEmpty function (llvm#70206)" [mlir][ArmSVE] Add `-arm-sve-legalize-vector-storage` pass (llvm#68794) [clang][Interp][NFC] Make another Descriptor param const [InstCombine] Drop exact flag instead of increasing demanded bits (llvm#70311) [SimplifyCFG] Delete the unnecessary range check for small mask operation [SimplifyCFG] Precommit tests for PR65835 [OpenMP] Patch for Support to loop bind clause : Checking Parent Region [Inline Spiller] Consider bundles when marking defs as dead [Inline Spiller] Pre-commit test [clang][Interp][NFC] Make Block::Desc const [clang][NFC] Move a variable into the closest scope [clang][Interp][NFC] Make InlineDescriptor::Desc const [DAG] SimplifyDemandedBits - ensure we drop NSW/NUW flags when we simplify a SHL node's input [DAG] Update test case for Issue llvm#69965 [AMDGPU] Rematerialize scalar loads (llvm#68778) [mlir] Remove `printCString()` from RunnerUtils (llvm#70197) [libc] Add --sweep-min-size flag for benchmarks (llvm#70302) [InstCombine] Add additional demanded bits tests for shifts (NFC) Revert "[AArch64] Move SLS later in pass pipeline" [scev-aa] Make TypeSize -> uint64_t conversion explicit [Bazel] Fixes for ec6da06. [IR] Require index width to be ule pointer width (llvm#70015) NFC. Move out and expose affine expression simplification utility out of AffineOps lib (llvm#69813) [mlir] Fixed typo in type (128x64 -> 64x128) in TMA load test (llvm#70022) [mlir][cuda] Avoid driver call to check max shared memory (llvm#70021) [AggressiveInstCombine] Ignore debug instructions when load combining (llvm#70200) [test][AggressiveInstCombine] Precommit testcase for llvm#69925 [LLD] [COFF] Recognize Itanium vtables for ICF (llvm#70196) [AMDGPU] Use `S_CSELECT` for uniform i1 ext (llvm#69703) [PowerPC] Fix use of FPSCR builtins in smmintrin.h (llvm#67299) [clang] [Gnu] Improve GCCVersion parsing to match versions such as "10-win32" (llvm#69079) [AArch64][SVE2] Use rshrnb for masked stores (llvm#70026) [lldb][lldb-server] Enable sending RegisterFlags as XML (llvm#69951) [AMDGPU] Shrink to SOPK with 32-bit signed literals (llvm#70263) [gn build] Port 897cc8a [RISCV] Support predefined macro __riscv_misaligned_[fast,avoid]. (llvm#65756) [RecursiveASTVisitor] Fix RecursiveASTVisitor (RAV) fails to visit the initializer of a bitfield (llvm#69557) [RISCV] Add tests for vmadd for VP intrinsics. NFC (llvm#70042) [Support] Better error msg when cache dir can't be created. (llvm#69575) [AArch64] Prevent argument promotion of vector with size > 128 bits (llvm#70034) [clang-tidy][NFC]refactor PreferMemberInitializerCheck for readability Apply clang-tidy fixes for misc-include-cleaner in MLIR examples [clang-tidy] Improved cppcoreguidelines-narrowing-conversions.IgnoreConversionFromTypes (llvm#69242) [clang-tidy] Improved cppcoreguidelines-pro-type-const-cast (llvm#69501) [clang-tidy] Support functional cast in bugprone-dangling-handle (llvm#69067) [RISCV] Separate addend from FMA operands to support cascade FMA. NFC. (llvm#70241) [AMDGPU] Lower __builtin_amdgcn_read_exec_hi to use amdgcn_ballot (llvm#69567) [Github] Add lld to docs CI (llvm#69821) [Driver] Remove some misused NoXarchOption [mlir][sparse] avoid excessive macro magic (llvm#70276) [mlir][doc] Improve Destination-passing-style documentation (llvm#70283) [Driver] Use StringSet::contains (NFC) [OpenMP] [OMPD] Fix CMake install command [TableGen][NFC] Format CompressInstEmitter (llvm#68564) [mlir][doc] Slightly clarify bufferization documentation (llvm#70212) [clang-tidy]fix misc-unused-using-decls false positive false for using in elaborated type (llvm#70230) [lldb] Fix assertions caused by un-checked errors in ScriptedProcess [gn build] Port 109aa58 [RISCV] Add an experimental pseudoinstruction to represent a rematerializable constant materialization sequence. (llvm#69983) [clang-format] Fix a JavaScript import order bug [llvm] Remove no-op ptr-to-ptr bitcasts (NFC) Fix log format strings [lldb] Add test dependency on the `runtimes` instead of the `cxx` target [lldb] Part 1 of 2 - Refactor `CommandObject::Execute(...)` return `void` (not `bool`) (llvm#69989) [AMDGPU] Fix gcc -Wparentheses warning. NFC (llvm#70239) [OpenMP 5.2] Deprecate old syntax of linear clause (llvm#70152) Improve CI output. (llvm#70236) [LLDB][NFC] Remove DWARFASTParserClang as friend from SymbolFileDWARF (llvm#70157) [clang][deps] Fix `__has_include` behavior with umbrella headers (llvm#70144) [gn build] Manually port 078ae8c [LowerMemIntrinsics] Remove no-op ptr-to-ptr bitcasts (NFC) [mlir][sparse] test for linalg tensor semantics (llvm#70254) [flang][NFC] Update comment to be generic [mlir][sparse] add verification of absent value in sparse_tensor.unary (llvm#70248) [Driver][test] Improve mcmodel.c [RISCV] Use a switch instead of a series of if-clauses [nfc] (try 2) [RISCV][GISel] Add instruction selection for G_FADD/G_FSUB/G_FMUL/G_FDIV with F/D extensions. (llvm#69808) Revert "[RISCV] Use a switch instead of a series of if-clauses [nfc]" [X86][GlobalISel] Add legalization of 64-bit G_ICMP for i686 (llvm#69478) [RISCV] Use a switch instead of a series of if-clauses [nfc] [lldb] Refactor InstrumentationRuntimeAsan and add a new plugin (llvm#69388) [RISCV][GISel] Add regbank selection for G_FADD/G_FSUB/G_FMUL/G_FDIV with F/D extensions. (llvm#69805) [RISCV][GISel] Add missing using LegalityPredicates. [RISCV][GISel] Add legalizer support for G_FADD/G_FSUB/G_FMUL/G_FDIV with F/D extensions. (llvm#69804) [LLVM[NFC] Refactor to allow debug_names entries to conatain DIE offset (llvm#69399) [RISCV][GISel] Add FP calling convention support (llvm#69138) [Libomptarget] Add a wavefront sync builtin for the AMDGPU implementation (llvm#70228) [libc] Compile the GPU functions with '-fconvergent-functions' (llvm#70229) [CMake] Correctly handle LLVM_ENABLE_RUNTIMES in targets (llvm#69869) [clang] Fix trailing whitespace in DiagnosticParseKinds.td [MLIR] Modify lowering of gpu.alloc op to llvm (llvm#69969) [mlir][nvvm] Fix mov.u32 to mov.pred (llvm#70027) [OpenMP][Obvious] Fix incorrect variant selector in test Improve debug names index fetching global variables performance (llvm#70231) [clang-tidy] Ignore deleted functions in cppcoreguidelines-rvalue-reference-param-not-moved (llvm#69514) Update m68k tests to new block placement [CodeLayout] Update to resolve Wdangling warning. [RISCV] Remove RISCVISD opcodes for LGA, LA_TLS_IE, and LA_TLS_GD. (llvm#70137) [Mips][GISel] Fix a couple issues with passing f64 in 32-bit GPRs. (llvm#69131) [libclang/python] Add missing concept declaration CursorKind (llvm#69125) [mlgo][regalloc] Fix reference file post e3cf80c [RISCV][GISel] Add widenScalarToNextPow2 to G_SEXTLOAD/G_ZEXTLOAD legalization. Update stdckdint.h and make it available in pre-C23 modes. (llvm#69649) [AMDGPU] Set SchedRW = Write64Bit on V_MOV_B64 (llvm#70135) [lldb][libc++] Adds chrono data formatters. [lldb] Fix build failure introduced by f22d82c [Libomptarget] Bump up PTX version from +ptx61 to +ptx63 (llvm#70227) [RISCV][GISel] Add clampScalar G_ZEXTLOAD/G_SEXTLOAD legalization rules. [Clang] Ensure zero-init is not overridden when initializing a base class in a constant expression context (llvm#70150) [X86] Regenerate select-mmx.ll [lldb/Interpreter] Make ScriptedInterface Object creation more generic (llvm#68052) [AMDGPU] Cleanup hasUnwantedEffectsWhenEXECEmpty function (llvm#70206) [MLIR][pass registry] show overlapping pass pipeline name (llvm#70167) Scalarizer: add negative test for lrint, llrint (llvm#70203) SLP/RISCV: add negative test for llrint, increase coverage (llvm#69940) [RISCV][GISel] Falling back to SDISel for scalable vector type values (llvm#70133) [Sema] -Wzero-as-null-pointer-constant: don't warn for __null (llvm#69126) [DAG] WidenVectorOperand - add basic handling for *_EXTEND_VECTOR_INREG nodes [llvm] Followup fix for "Use XMACROS for MachO platforms" (llvm#70140) [OpenMP][Obvious] Fix incompatbile function prototype causing failures [LinkerWrapper][Obvious] Fix dangling stringref on Xcuda-ptxas options Add missing `linalg.batch_vecmat` named op (llvm#70218) [clang] Remove redundant ptr-to-ptr bitcasts (NFC) [mlir][sparse] update block24 example (llvm#70145) [MLIR][OpenMP] Use opaque pointers in OpenMP translation tests 2/2 (llvm#70138) [InstCombine] Rename some variables (NFC) [SelectionDAG] Salvage debug info for non-constant ADDs (2nd try) (llvm#68981) [CodeLayout] Faster basic block reordering, ext-tsp (llvm#68617) [bazel] Add missing dependency [bazel] Port 078ae8c [OpenMP][Obvious] Fix function prototype when used in C mode [flang][openacc] Add implicit copy for reduction in combined construct (llvm#70148) [clang]improve diagnosing redefined defaulted constructor with different exception specs (llvm#69688) [AArch64][GlobalISel] Legalize NEON smin,smax,umin,umax,fmin,fmax intrinsics [clang][dataflow] Add `Environment::allows()`. (llvm#70046) [clang][analyzer][NFC] Combine similar methods of StreamChecker (llvm#70170) [NFC][Clang] Make GetSDLFromOffloadArchive return void [NFC][Clang] Make read-only arguments of GetSDLFromOffloadArchive and SDLSearch read-only references instead of copying them [X86] combineArithReduction - use PACKUSWB directly for PSADBW(TRUNCATE(v8i16 X)) reduction patterns Revert "[SelectionDAG] Salvage debug info for non-constant ADDs (llvm#68981)" [mlir][ArmSME] Add tile slice layout attr to vector <-> tile ops (llvm#69186) [AArch64] Add test for llvm#70207 (NFC) [clang][NFC] Refactor `Selector` to use `PointerIntPair` inside (llvm#69916) [tooling/include-mapping] Add missing localtime_r symbols (llvm#66091) [NFC][Clang] Make GetSDLFromOffloadArchive and SDLSearch static (llvm#70201) [Offloading][NFC] Move creation of offloading entries from OpenMP (llvm#70116) [lldb][docs][AArch64] Update example in SME docs [SystemZ][NFC] Fix a couple of style issues (llvm#69958) [clang-format] Fix the version for a newly added option [Flang][OpenMP] NFC: Copy target test to HLFIR flow [X86] Regenerate pr38539.ll [AMDGPU][NFCI] Decouple actual register encodings from HWEncoding values. (llvm#69452) [clang-format] AllowShortCompoundRequirementOnASingleLine [C23] Use thread_local semantics (llvm#70107) [LLD] [MinGW] Hook up --icf=safe to -opt:safeicf (llvm#70037) [libc] Fix a constexpr violation from b4e5529 [OpenMP][Archer] Do not check for column numbers in backtraces (llvm#70075) [SelectionDAG] Salvage debug info for non-constant ADDs (llvm#68981) [clang-format] Don't align comments over scopes [flang] add TBAA tags to global and direct variables [lldb][AArch64] Simplify AArch64SMEFA64 check Revert "[AMDGPU] Accept arbitrary sized sources in CalculateByteProvider" [AArch64] Disable by default MachineSink sink-and-fold (llvm#70101) [RISCV] Add missing break in switch-case in convertToThreeAddress function. NFC. [AArch64] Move SLS later in pass pipeline [mlir][LLVM] Verify too many indices in GEP verifier (llvm#70174) [Clang][DebugInfo] Clang generates an extra spurious unnamed 'dbg.declare' (llvm#69681) [InstCombine] Remove false commutativity from processUMulZExtIdiom() (NFCI) [AArch64] Add test showing incorrect code-gen [InstCombine] Remove unnecessary eq/ne handling from processUMulZExtIdiom() (NFCI) [mlir][test] Update tests to use vector.print str (NFC) (llvm#68973) [mlir][tosa] fix a crash when sliceOp has invalid attribute (llvm#68486) (llvm#70063) [lldb] On POSIX, check for duplicate interpreter modules without loading them (llvm#69932) [lldb][AArch64] Add isAArch64SMEFA64 check to SME testing (llvm#68094) [lldb][AArch64] Read mte_ctrl register from core files (llvm#69689) [InstCombine] Test extra and use in processUMulZExtIdiom() fold (NFC) [mlir][NFC] Move `foldAttributesIntoMap` to `IR` build unit (llvm#70155) [llvm][TableGen] Add a README to the main TableGen folder (llvm#69943) [InstCombine] Remove unnecessary handling of non-canonical predicates (NFCI) Reland "[lldb][AArch64] Invalidate SVG prior to reconfiguring ZA regdef (llvm#66768)"" [InstCombine] Remove unnecessary typed pointer handling (NFC) [InstCombine] Remove redundant icmp gep fold (NFCI) Revert "[lldb][AArch64] Invalidate SVG prior to reconfiguring ZA regdef (llvm#66768)" [RISCV] Run mem2reg to simplify Zbc tests (llvm#70169) [lldb][AArch64] Invalidate SVG prior to reconfiguring ZA regdef (llvm#66768) [lldb][AArch64] Correct type of 32 bit GPR RegisterValues when using core files (llvm#70054) [mlir][transform] Handle multiple library preloading passes. (llvm#69705) [Sema] Fixed faulty shift count warning (llvm#69521) [flang][codegen] Update FIR codegen to use mlir.llvm opaque pointers (llvm#69692) [MLIR] Properly add operations to blocks during `createOrFold` (llvm#70010) [Driver] Add `--` to some test clang-cl command lines. (llvm#70055) [flang] Allow lowering of sub-expressions to be overridden (llvm#69944) [RISCV][GISel] Add G_SEXTLOAD to legalizer and regbank select. Add instruction selection tests. [RISCV][GISel] Fix failure to legalize non-power of 2 shifts between i32 and i64 on RV64. [ADT] Rename llvm::erase_value to llvm::erase (NFC) (llvm#70156) [Flang][OpenMP] Replace fir.store with hlfir.assign in OpenMP single … (llvm#70164) [mlir][python] Clear PyOperations instead of invalidating them. (llvm#70044) [ValueTracking] Analyze `Select` in `isKnownNonEqual`. (llvm#68427) [LoongArch] Set some operations action for LSX and LASX [Flang][OpenMP] Port OpenMP single tests to hlfir flow (llvm#69877) [lldb-vscode] Allow specifying a custom escape prefix for LLDB commands (llvm#69238) [ADT] Remove an extraneous ternary operator (NFC) BlockFrequencyInfoImpl: Avoid big numbers, increase precision for small spreads [RISCV][NFC] Fix comments in foldMemoryOperandImpl (llvm#70033) [llvm][support] Show name of overlapping cl option (llvm#70108) [LowerSwitch] Don't let pass manager handle the dependency (llvm#68662) [builtins] Revert accidental change to PPC implementation in 05a4212 [builtins] Avoid using long double in generic sources (llvm#69754) [libc] Disable -NaN test on float128 systems (llvm#70146) [mlir][tensor][NFC] Simplify `SubsetInsertionOpInterface` implementation (llvm#69999) [SLP] Improve gather tree nodes matching when users are PHIs. (llvm#70111) [RISCV][GISel] Add ISel support for SHXADD_UW and SLLI.UW (llvm#69972) [GISel] Make assignValueToReg take CCValAssign by const reference. (llvm#70086) [libc] Fix printf long double inf, bitcast in msan (llvm#70067) [OpenMP 5.2] Initial parsing and semantic analysis suppport for 'step' modifier on 'linear' clause [SLP][NFC]Remove unused variables, NFC. [clangd] Support `-stdlib` flags in SystemIncludeExtractor. (llvm#69283) [clangd] Do not offer RawStringLiteral code action in C and C++98 (llvm#69775) [RISCV] Mark V0 regclasses as larger superclasses of non-V0 classes (llvm#70109) [OpenMP][MLIR] Use opaque pointers in OpenMP translation tests 1/2 (llvm#70057) [AArch64][GlobalISel] Add support for post-indexed loads/stores. (llvm#69532) Reapply "[compiler-rt] Allow Fuchsia to use 64-bit allocator for RISCV (llvm#68343)" [mlir] Fix nvvm integration tests build error (llvm#70113) [mlir][sparse] implement sparse_tensor.lvl operation. (llvm#69993) [mlir][sparse] add COO to python tests (llvm#70090) [NFC][SLP] Add test case for issue llvm#69670. (llvm#70088) [DAGCombiner] Fix misuse of getZeroExtendInReg in SimplifySelectCC. (llvm#70066) [BOLT] Fix incorrect basic block output addresses (llvm#70000) [Github] Add support for building libc docs in Github actions (llvm#69824) [libc++] Update status of P2770R0, P2441R2 and P2711R1 (llvm#70045) [RISCV][InsertVSETVLI] Add Subtarget variable to class [nfc] Revert "[dsymutil] Add support for mergeable libraries" [BOLT] Fix build issues after llvm#69836 (llvm#70087) [Clang] Mark declarators invalid in the presence of ill-formed explicit parameters. (llvm#70018) [flang][openacc] Support single array element in data clause (llvm#70065) [RISCV] Rename hasFixedResult to willVLBeAVL [nfc] [mlir][sparse] hoists alloca outside the outermost loop. (llvm#70085) Reland "[clang][Sema] Use original template pattern when declaring implicit deduction guides for nested template classes" (llvm#69676) Expose DWARFDIE::GetDeclContext() in lldb_private::Function. (llvm#69981) [mlir][tosa] Add verifiers to ReduceOps, fix shape inference crash (llvm#69843) Revert "[clang] Support fixed point types in C++ (llvm#67750)" (llvm#69963) [OpenMP][Tests] Sync struct DEP with the runtime (llvm#69982) [OpenMP] Provide big-endian bitfield definitions (llvm#69995) [dsymutil] Add support for mergeable libraries [ADT] Support appending multiple values (llvm#69891) [Analysis] Add Scalable field in MemoryLocation.h (llvm#69716) [OpenMP 5.2] Deprecate syntax of map modifiers without comma separators (llvm#69534) [builtins] Fix floattitf.c etc. compilation on Solaris/SPARC (llvm#70058) [GISel] Pass MPO and VA to assignValueToAddress by const reference. NFC (llvm#69810) [lldb][gardening] Remove full name of "DWARFDIE" type in GetTypeForDIE (llvm#70062) [builtins] Support building the 128-bit float functions on ld80 platforms (llvm#68132) [NFC][Clang] Remove unused variable "AOBFileNames" Reintroduce accidentally deleted "protected" keyword in SymbolFileDWARF (llvm#69990) [ModuleInliner] Remove an extraneous pair of std::push_heap and std::pop_heap (NFC) (llvm#69672) [SLP]Fix PR70004: Do not change insert point for reduction gather nodes. [flang][openacc] Allow acc routine at the top level (llvm#69936) [FunctionAttrs] Improve handling of alias-preserving intrinsic calls (llvm#68453) [Clang][LTO][GISel] Propagate `-fglobal-siel` to LTO (llvm#69747) Make -frewrite-includes handle -include correctly [NVPTX] Add a test to verify the .version with sm_90(a) [clang]Transform uninstantiated ExceptionSpec in TemplateInstantiator (llvm#68878) [libunwind][AIX] static_cast the value from getLR() to avoid the warning from -Wconversion (llvm#69767) [Reland] Detect against invalid variant index for LibStdC++ std::variant data formatters (llvm#69614) Revert "[mlgo] Fix tests post 760e7d0" [InstCombine] Remove unnecessary removeBitcastsFromLoadStoreOnMinMax() fold (NFCI) [SLP]Fix PR69246: do not treat resizing maskas identity. [InstCombine] Remove redundant add+and fold (NFCI) [InstCombine] Remove redundant folds in foldCastedBitwiseLogic() (NFCI) [InstCombine] Remove redundant fold in foldUnsignedUnderflowCheck() (NFCI) [RISCV][GISel][NFC] Correct the test case in constant32.mir (llvm#70003) [mlir] Change default NVVM compilation to `fatbin` from `bin` (llvm#70052) [bazel] Add missing dependencies for ba8ae98 [clang][Sema] Correct end for the `CastOperation.OpRange` (llvm#69480) [InstCombine] Remove unnecessary typed pointer fold (NFCI) Reland "Add Documentation for Execution Results Handling in Clang-Repl (llvm#65650)" [mlir][transform] Improve error when merging of modules fails. (llvm#69331) [InstCombine] Remove redundant logical select fold (NFCI) [InstCombine] Remove redundant fold in foldSelectExtConst() (NFCI) [InstCombine] Remove some redundant select folds (NFCI) [X86] Add scalar isel test coverage for AND/OR/XOR types [mlir] use transform-interpreter in test passes (llvm#70040) [libc] Speed up memmove overlapping check (llvm#70017) [VectorCombine] Add special handling for truncating shuffles (llvm#70013) [SLP]Fix PR69976: Check for multi-node uses during node building. [MLIR] Fixes NVGPU Integration Test Passes Ordering (llvm#69934) [MLIR][CFGToSCF] Fix exit latch location preservation (llvm#70032) [CXXNameMangler] Correct the mangling of SVE ACLE types within function names. (llvm#69460) [DAG] Add test case for Issue llvm#69965 [flang][OpenMP] Fix LASTPRIVATE for iteration variables (llvm#69773) [mlir][Vector] Remove duplicated test cases (llvm#70029) [gn build] Port 5d7f346 [IPSCCP] Don't crash on ptrtoint [InstCombine] Extend Phi-Icmp use to include or (llvm#67682) [libc++] Remove the Buildkite job that builds documentation (llvm#69992) [libc++][NFC] Remove the list of `nodiscard` extensions from the docs. (llvm#69546) [Driver] Ignore non-clang pch files when -include a.h probes a.h.gch (llvm#69204) [libc++][PSTL] Implement std::rotate_copy [analyzer][clangsa] Add new option to alpha.security.cert.InvalidPtrChecker (llvm#67663) [Tooling/Inclusion] Fix MSVC build [CVP] Flip signedness icmp predicate in use level (llvm#69948) [Tooling/Inclusion] Refactor to use tables for compile time [lldb][AArch64][NFC] Fix typo in get trap handler function [wasm] Don't crash on non-simple value types during shuffle combine [BasicAA] Update comment (NFC) [clang][Interp] Handle variadic functions (llvm#67814) [polly] Reformat due to d68826d [clang][Interp] Fix ArrayInitLoop common expr life time [clang][Interp] Don't explicitly call InterpState destructor() [clang-format] Don't break between string literal operands of << (llvm#69871) [lldb] Add test for reserved XML chars in register field names Revert "[clang-format] Fix align consecutive declarations over function pointers" [LoopIdimo] Use tryZExtValue() instead of getZExtValue() [CodeGen] Update a comment from NoSSA to IsSSA [IPSCCP] Fix a mistake in b796eac so the test actually passes [AMDGPU] Fix subreg check in the SIFixSGPRCopies (llvm#70007) [analyzer][NFC] Simplifications in ArrayBoundV2 (llvm#67572) [IPSCCP] Silence sign compare warnings in test [mlir][VectorOps] Support string literals in `vector.print` (llvm#68695) [flang][driver] support -dumpversion and -dumpmachine (llvm#68896) [mlir][ods] Emit bytecode readers with generic Attribute/Type (NFC) [IR] Require that ptrmask mask matches pointer index size (llvm#69343) [clang][Interp] Only emit function_param_value_unknown in C++11 (llvm#67990) [clang][Interp] Fix scalar inits of void type (llvm#69868) Revert 24633ea and 760e7d0 "Enable FoldImmediate for X86" [clang][Interp] Call destructors of local variables [clang][dataflow] Remove `declToLocConsistent()` assertion. (llvm#69819) [TableGen][NFC] Add record kind to `Record` class (llvm#69919) [clang] Add triple to builtin-memcpy test [AMDGPU] Regenerate udiv.ll [DAG] Constant Folding for U/SMUL_LOHI (llvm#69437) [IPSCCP] Variable not visible at Og. (llvm#66745) [Github] Fetch before files changed action on push in docs action Revert "[clang][Interp] Call destructors of local variables" [clang][ExprConst] Handle 0 type size in builtin_memcpy etc. [Github] Use API to fetch PR diff for docs action (llvm#70001) [mlir][sparse] extend sparse output test (llvm#69986) [clang][Interp] Call destructors of local variables [clang][Diagnostics] Provide source range to constexpr function diags (llvm#69721) [Github] Add libcxx docs to CI (llvm#69828) [Github] Add libunwind to docs CI (llvm#69830) [AArch64] Lower mathlib call ldexp into fscale when sve is enabled (llvm#67552) [clangd] Adapt Inlay Hint support for Deducing This (llvm#68177) -fstack-usage: fix filename for functions in an included file (llvm#69896) [flang][openacc] Fix post_alloc declare function ordering (llvm#69980) [AMDGPU] Fix -Wunused-variable in SIISelLowering.cpp (NFC) [clang][NFC] Remove trailing whitespace in release note [mlir][Interfaces] `LoopLikeOpInterface`: Expose mutable inits/yielded values (llvm#69137) [lldb][NFC] Move DWARFASTParserClang::GetTypeForDIE to DWARFASTParser (llvm#69764) [X86][GlobalISel] Reorganize shift scalar tests (NFC) (llvm#68232) [lldb][NFC] Remote reference to Swift in DWARFASTParser::Kind enum (llvm#69984) [lldb][NFC] Implement llvm-style RTTI for DWARFASTParser (llvm#69762) [AMDGPU] Accept arbitrary sized sources in CalculateByteProvider [amdgpu] D2D memcpy via streams and HSA (llvm#69977) [mlir][sparse] introduce sparse_tensor.lvl operation. (llvm#69978) [mlir][sparse] support 2:4 structured sparsity and loose compressed (llvm#69968) [lldb] Move comments and add missing periods (NFC) Fix test clang/test/Driver/cl-offload.cu [RISCV] Only check for scalar VT at depth 0 in hasAllNBitUsers. Revert tests missed in b997ff4 [ThinLTOBitcodeWriter] Do not crash on a typed declaration (llvm#69564) [MLIR][scf.parallel] Don't allow a tile size of 0 (llvm#68762) [flang][openacc] Support fixed form sentinel format (llvm#69970) [MLIR,Flang,OpenMP] Remove usage of getElementType in OpenMPTranslation (llvm#69772) [libc++][Android] Enable Android testing in BuildKite CI (llvm#69275) [RFC][flang] Trampolines for internal procedures. (llvm#66157) [NVPTX] Add lowering for bitcasts float<->v4i8 (llvm#69960) [clang-format] Don't alter SpacesRequiredBefore of finalized tokens (llvm#69897) [LLDB] Update docs on building documentation (llvm#69858) [LV] Use variable instead of value number in vplan-dot-printing.ll test. [compiler-rt][lsan][Fuchsia] Adjust lsan allocator settings (llvm#69401) [NFC][lsan] Extract and rename SizeClassMap type from AP64 (llvm#69526) Revert "[RISCV][InsertVSETVLI] Make VL preserving vsetvli emission more explicit [nfc]" [mlir][sparse] Parser cleanup (llvm#69792) [clang] Add clang::preferred_type attribute for bitfields (llvm#69104) [AMDGPU] Force the third source operand of the MAI instructions to VGPR if no AGPRs are used. (llvm#69720) [RISCV] Allow swapped operands in reduction formation (llvm#68634) [lldb][docs] Update contributing links (llvm#69726) [libc] Enable remaining string functions on the GPU (llvm#68346) [lldb] Add a single bit constructor for RegisterFlags::Field (llvm#69315) [AArch64] Don't generate st2 for 64bit store that can use stp (llvm#69901) Fixed Windows build warnings (llvm#68978) [LowerGlobalDtors] Skip __cxa_atexit call completely when arg0 is unused (llvm#68758) Update call-long1.ll [opt] Properly report errors when loading pass plugins (llvm#69745) [lldb] Move template python files to separate directory [lldb] Move ScriptInterpreter Interfaces to subdirectory (NFC) [mlir][Transform] NFC - Add a C++ builder for NamedSequenceOp Revert "[Sema] Add check for bitfield assignments to integral types (llvm#69049)" [NFC] use `StringSet::insert(iter, iter)` instead for loop to insert (llvm#69851) Revert [clang] Handle templated operators with reversed arguments and [STLExtras] Undo C++20 hack (llvm#69937) [lldb][docs] Add strace example to Debugging doc [clang][Interp] Only evaluate the source array initialization of an `ArrayInitLoopExpr` once (llvm#68039) [Flang][HLFIR][OpenMP] Fix offloading tests broken by HLFIR (llvm#69457) [AMDGPU] Remove unnecessary conditionality on atomic CSUB pseudo-ops (llvm#69914) [mlir] Verify TestBuiltinAttributeInterfaces eltype (llvm#69878) [llvm][Release] Add note about binaries to Github release description (llvm#69698) [GVN] Add additional captured before tests (NFC) [RISCV][GISel] Select G_SELECT (G_ICMP, A, B) (llvm#68247) Update my contact email in CodeOwners.rst [Clang] Handle real and imaginary parts of complex lvalue in `APValue::printPretty` (llvm#69252) [DSE] Return std::optional from getPointerSize() (NFC) [ScalarizeMaskedMemIntrin] Add missing lit.local.cfg (NFC) [BasicAA] Return std::optional from getObjectSize() (NFC) [Clang] Report an error and crash on source location exhaustion in macros (llvm#69908) [LegacyPM] Clean up includes in LowerExpectIntrinsicPass Use the correct namespace for looking up matching operator!= (llvm#68922) Fix Clang sphinx build Correct behavior of VLA extension diagnostic in C89 mode [AArch64] Allow SVE code generation for fixed-width vectors (llvm#67122) [mlir][VectorOps] Add fold `ExtractOp(CreateMask) -> CreateMask` (llvm#69456) [Sema] Change order of displayed overloads in diagnostics [STLExtras] Undo C++20 hack [InstCombine] Remove scalable vector extracts to and from the same type (llvm#69702) [libc++] <experimental/simd> Add operator value_type() of simd reference (llvm#68960) [mlir][SCF] Minor fixes in documentation examples (llvm#69802) [mlir][tosa] Check for unranked tensors during validation (llvm#68509) [clang] Fix designated initializers inside templates (llvm#69712) Typos: 'maxium', 'minium' [flang] Place MIN/MAX A1/A2 first in semantic analysis (llvm#69722) [BOLT][RISCV] Set minimum function alignment to 2 for RVC (llvm#69837) [MemDep] Use EarliestEscapeInfo (llvm#69727) [flang] Do not stop on mismatched DATA substring length (llvm#69336) [mlir][linalg] Enable CollapseLinalgDimensions to collapse linalg::CopyOp (llvm#68526) [BOLT][RISCV] Use target features from object file (llvm#69836) [clang][dataflow] Remove `DataflowAnalysisContext::flowConditionIsTautology()`. (llvm#69601) [clang-format] Add a new style for the clang-format source code (llvm#69814) [mlir][Bazel] Add missing dependency. [mlir][SCF] Use getResult() instead of static_cast<Value>(). [RISCV] Disable hasAllNBitUsers for vector types. [TextAPI] Use std::remove_reference_t (NFC) [lldb] Use llvm::is_contained (NFC) Use llvm::count (NFC) [TableGen][NFC] Remove MultiClass argument and Scoper in QualifyName (llvm#69297) [Clang][RISCV] Support CSRs in clobbered registers of inline assembly (llvm#67646) [C++20] [Modules] [Driver] Don't enable -fdelayed-template-parsing by default on windows with C++20 (llvm#69431) [OpenMP][FIX] Enlarge thread state array, improve test and add second [mlir][bufferization] Ownership-based deallocation: Allow manual (de)allocs (llvm#68648) [llvm-profgen] More tweaks to warnings (llvm#68608) [CodeLayout] cache-directed sort: limit max chain size (llvm#69039) [LegacyPM] Remove LowerExpectIntrinsicPass [LLDB] Update breakpoint-command.test to use string instead of number. (llvm#69796) [llvm] Stop including llvm/ADT/iterator_range.h (NFC) [mlir][DeadCodeAnalysis] Don't Require `RegionBranchTerminatorOpInterface` in `visitRegionTerminator()` (llvm#69043) [JITLink] Allow multiple relocations at same offset in EHFrameEdgeFixer (llvm#68252) [OpenMP][FIX] Fix memset oversight to partially unblock test Apply clang-tidy fixes for misc-include-cleaner in AST.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in AST.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in MLIRGen.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in Dialect.cpp (NFC) Apply clang-tidy fixes for misc-include-cleaner in toyc.cpp (NFC) [clang-format][NFC] Simplify the logic in a return statement [OpenMP][FIX] Ensure test runs correct with (at least) 2 threads [compiler-rt] Use std::clamp (NFC) [llvm] Stop including llvm/ADT/DepthFirstIterator.h (NFC) [lldb] Remove an unused using decl (NFC) [llvm] Stop including llvm/ADT/StringMap.h (NFC) [CodeGen][Remarks] Add the function name to the stack size remark (llvm#69346) [Utils] Use std::remove_pointer_t (NFC) [llvm] Use llvm::any_of (NFC) [mlir] Remove an extraneous typename (NFC) [llvm] Stop including llvm/ADT/SmallString.h (NFC) [mlir][minimal-opt] Fix typo [clangd] Show alignment for records and fields decls (llvm#67213) [VPlan] Make ExpandedSCEVs argument const (NFC). [VPlan] Simplify redundant trunc (zext A) pairs to A. [gn build] Port d2a46e6 [libc++][PSTL] Implement std::move [AMDGPU] Set size to all SOP pseudos (llvm#69756) [Clang][OHOS] Keep ARM ABI selection logic in sync between Clang and LLVM (llvm#68656) [llvm] Stop including Endian.h (NFC) [LIT] Print discovered tests and percentages (llvm#66057) (llvm#69831) [docs] Fix suggested darker command in coding standards (llvm#69860) Reland: "[mlir][index][spirv] Add conversion for index to spirv" (llvm#69790) Apply clang-tidy fixes for llvm-qualified-auto in IRNumbering.cpp (NFC) Apply clang-tidy fixes for llvm-qualified-auto in CallGraph.cpp (NFC) Apply clang-tidy fixes for llvm-qualified-auto in LowerToLLVM.cpp (NFC) Apply clang-tidy fixes for llvm-qualified-auto in LowerToLLVM.cpp (NFC) [lldb] improve dwo path in missing dwo error when relative (llvm#69783) [Github] Fetch an additional commit for docs CI on PRs [mlir][doc] Add basic doc for extraTraitClassDeclaration. [mlir][doc] Include ml_program passes in passes doc [Github] Add lldb docs step to Github docs action (llvm#69832) [Github] Add clang-tools-extra docs to CI (llvm#69827) [OpenMP] Rewrite test to check the correct (CPU) result [OpenMP] Basic BumpAllocator for (AMD)GPUs (llvm#69806) [OpenMP][FIX] Ensure thread states do not crash on the GPU [LV] Enforce order of reductions with intermediate stores in VPlan (NFC) Reland [LLD] [COFF] Don't try to detect MSVC installations in mingw mode [Attributor] Ignore different kernels for kernel lifetime objects [Attributor][NFC] Precommit test [lldb] Update qRegisterInfo docs to recommend target.xml (llvm#69853) [RISCV] Replace RISCV -> RISC-V in comments. NFC Change-Id: I5672108f77731b793c3011a911ab626d49e3ce6c 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
llvm commit: 324d1bb
Reproduce with:
With the dbg.value present aggressive-instcombine does nothing and leaves the two loads from @e be.
If we remove the dbg.value call aggressive-instcombine optimizes the function to
bbi-87915.ll.gz
The text was updated successfully, but these errors were encountered: