LLVM/project a8a1551mlir/include/mlir/Dialect/OpenACC OpenACCOps.td, mlir/lib/Dialect/OpenACC/IR OpenACC.cpp

[acc] RegionBranchOpInterface for acc regions (#171533)

Defining RegionBranchOpInterface for acc regions will help dataflow
analysis to propagate IN/OUT sets without losing information
DeltaFile
+50-37mlir/include/mlir/Dialect/OpenACC/OpenACCOps.td
+55-0mlir/lib/Dialect/OpenACC/IR/OpenACC.cpp
+105-372 files

LLVM/project 8b87edfllvm/lib/Analysis InstructionSimplify.cpp, llvm/test/Transforms/InstSimplify vp-reverse.ll

[InstSimplify] Ignore mask when combinining vp.reverse(vp.reverse). (#171542)

The mask doesn't really affect the reverse. It only poisons the masked
off elements in the results. It should be ok to ignore the mask if we
can eliminate the pair.

I don't have a specific use case for this, but it matches what I had
implemented in our downstream before the current upstream
implementation. Submitting upstream so I can remove the delta
in my downstream.
DeltaFile
+9-0llvm/test/Transforms/InstSimplify/vp-reverse.ll
+3-5llvm/lib/Analysis/InstructionSimplify.cpp
+12-52 files

LLVM/project c642fa0llvm/lib/Target/AArch64 AArch64AsmPrinter.cpp, llvm/test/CodeGen/AArch64 ptrauth-reloc.ll ptrauth-irelative.ll

AArch64: Relax restriction on discriminator when PAuth ifunc used.

When a PAuth ifunc is being used, we can represent any discriminator that
we want in the code and don't need to be restricted to 16 bits. For now we
only need this capability for address discriminated ptrauth expressions,
so keep the restriction in place for other discriminators.

Reviewers: atrosinenko, fmayer

Reviewed By: fmayer

Pull Request: https://github.com/llvm/llvm-project/pull/170945
DeltaFile
+14-9llvm/lib/Target/AArch64/AArch64AsmPrinter.cpp
+17-0llvm/test/CodeGen/AArch64/ptrauth-reloc.ll
+13-0llvm/test/CodeGen/AArch64/ptrauth-irelative.ll
+44-93 files

LLVM/project 390f17fllvm/lib/Target/AArch64 AArch64AsmPrinter.cpp, llvm/test/CodeGen/AArch64 ptrauth-irelative.ll

AArch64: Emit PAuth ifuncs into the same comdat as the containing global.

PAuth ifuncs contain a relocation pointing to the section they are
relocating (the place), so they need to be in the same comdat in order
to avoid relocations pointing to discarded sections.

Reviewers: atrosinenko, kovdan01, fmayer

Reviewed By: fmayer

Pull Request: https://github.com/llvm/llvm-project/pull/170944
DeltaFile
+8-2llvm/lib/Target/AArch64/AArch64AsmPrinter.cpp
+6-0llvm/test/CodeGen/AArch64/ptrauth-irelative.ll
+14-22 files

LLVM/project 01c4eb5clang-tools-extra/clangd GlobalCompilationDatabase.cpp GlobalCompilationDatabase.h, clang-tools-extra/clangd/tool ClangdMain.cpp Check.cpp

Revert "[clangd] Add a (currently hidden) --strong-workspace-mode flag (#155905)"

This reverts commit 2fa492726e774c050d6f21d57990c8bfbd7f1400.

This caused sanitizer bots to fail and sanitizer errors to show up in
our downstream testing:

```
[ RUN      ] LSPTest.DiagnosticsHeaderSaved
<<< initialize: {}
<-- initialize(0)
third_party/llvm/llvm-project/clang-tools-extra/clangd/ClangdLSPServer.cpp:557:14: runtime error: load of value 112, which is not a valid value for type 'bool'
```

With ASan at -O1.
DeltaFile
+9-32clang-tools-extra/clangd/GlobalCompilationDatabase.cpp
+5-21clang-tools-extra/clangd/GlobalCompilationDatabase.h
+0-14clang-tools-extra/clangd/unittests/GlobalCompilationDatabaseTests.cpp
+0-12clang-tools-extra/clangd/tool/ClangdMain.cpp
+2-6clang-tools-extra/clangd/tool/Check.cpp
+0-5clang-tools-extra/clangd/ClangdServer.h
+16-901 files not shown
+16-927 files

LLVM/project 468cffbllvm/test/CodeGen/RISCV/rvv vfadd-sdnode.ll

[NFC][RISCV] Make vfadd attribute list follow other tests (#170864)

DeltaFile
+147-23llvm/test/CodeGen/RISCV/rvv/vfadd-sdnode.ll
+147-231 files

LLVM/project b0e8c3cutils/bazel/llvm-project-overlay/mlir BUILD.bazel

Fix bazel build for 21147e7c95c03f554d4a7fb9b55b8e459357eb49
DeltaFile
+3-0utils/bazel/llvm-project-overlay/mlir/BUILD.bazel
+3-01 files

LLVM/project 09eb25fflang/docs OptionComparison.md

[flang][docs] Replace Flang to Classic Flang in old doc (#171558)

This document was written when Flang was known as F18 and Classic Flang
was known as Flang. The term "Flang" in this document refers to Classic
Flang, except in the first paragraph.

Also, a trivial HTML error is fixed.
DeltaFile
+15-15flang/docs/OptionComparison.md
+15-151 files

LLVM/project 17b13e9flang/docs Unsigned.md

[flang][docs] Correct the title in Unsigned.md (#171553)

Probablly the original author copied `Extensions.md` and forgot to
change the title.
DeltaFile
+1-1flang/docs/Unsigned.md
+1-11 files

LLVM/project f309fabllvm/lib/Target/RISCV/GISel RISCVLegalizerInfo.cpp, llvm/test/CodeGen/RISCV/GlobalISel/legalizer/rvv legalize-insert-subvector.mir

[RISCV][GISel] Fix legalizer of G_INSERT_SUBVECTOR (#171091)

Fix early exit condition and use the right type as mask type of
vslideup.
DeltaFile
+274-48llvm/test/CodeGen/RISCV/GlobalISel/legalizer/rvv/legalize-insert-subvector.mir
+2-2llvm/lib/Target/RISCV/GISel/RISCVLegalizerInfo.cpp
+276-502 files

LLVM/project 7533d56mlir/lib/Conversion/ArithAndMathToAPFloat ArithToAPFloat.cpp MathToAPFloat.cpp, mlir/lib/Conversion/ArithToAPFloat ArithToAPFloat.cpp

[mlir][math] Add FP software implementation lowering pass: math-to-apfloat
DeltaFile
+0-665mlir/lib/Conversion/ArithToAPFloat/ArithToAPFloat.cpp
+623-0mlir/lib/Conversion/ArithAndMathToAPFloat/ArithToAPFloat.cpp
+148-0mlir/lib/Conversion/ArithAndMathToAPFloat/MathToAPFloat.cpp
+49-0mlir/lib/Conversion/ArithAndMathToAPFloat/CMakeLists.txt
+47-1mlir/lib/ExecutionEngine/APFloatWrappers.cpp
+39-0mlir/lib/Dialect/Func/Utils/Utils.cpp
+906-66610 files not shown
+1,040-69116 files

LLVM/project 4582a80clang/lib/Format UnwrappedLineParser.cpp, clang/unittests/Format TokenAnnotatorTest.cpp

release/21.x: [clang-format] Fix a regression in annotating star before lambda (#170969)

Backport 4930e94011f6c62231de880273821d453dae0f14
DeltaFile
+4-4clang/lib/Format/UnwrappedLineParser.cpp
+6-0clang/unittests/Format/TokenAnnotatorTest.cpp
+10-42 files

LLVM/project a565880clang/lib/Format WhitespaceManager.cpp, clang/unittests/Format FormatTest.cpp

[clang-format] Fix a crash in AlignArrayOfStructures (#167099)

Fixes #157405

(cherry picked from commit 836919bb34493333767fc1734e402d3ebf989acb)
DeltaFile
+13-0clang/unittests/Format/FormatTest.cpp
+4-1clang/lib/Format/WhitespaceManager.cpp
+17-12 files

LLVM/project b3a5ad1mlir/include/mlir-c/Dialect IRDL.h, mlir/lib/CAPI/Dialect IRDL.cpp

[MLIR][IRDL] Add C API for IRDL Variadicity attributes. (#171076)

This add the basic APIs to create VariadicityAttr and
VariadicityArrayAttr attributes from the C API. This is necessary for
C API users that want to create IRDL dialect declarations.
DeltaFile
+34-0mlir/test/CAPI/irdl.c
+27-0mlir/lib/CAPI/Dialect/IRDL.cpp
+14-0mlir/include/mlir-c/Dialect/IRDL.h
+75-03 files

LLVM/project 94c40c6clang/lib/Format QualifierAlignmentFixer.cpp, clang/unittests/Format QualifierFixerTest.cpp

[clang-format] Don't swap `(const override)` with QAS_Right (#167191)

Fixes #154846

(cherry picked from commit dfe9838f9c790aa632bed0a1b67976c2a7e95f76)
DeltaFile
+13-4clang/lib/Format/QualifierAlignmentFixer.cpp
+2-0clang/unittests/Format/QualifierFixerTest.cpp
+15-42 files

LLVM/project 2d0c14cutils/bazel/llvm-project-overlay/mlir BUILD.bazel

[bazel] Port 21147e7c95c03f554d4a7fb9b55b8e459357eb49 (#171545)

Adds a couple additional deps to OpenACCUtils.
DeltaFile
+3-0utils/bazel/llvm-project-overlay/mlir/BUILD.bazel
+3-01 files

LLVM/project c155124mlir/lib/Dialect/OpenACC/Utils OpenACCUtilsTiling.cpp

[mlir][acc] Fix build error for tiling API return value (#171546)

The build error looks like:
error: could not convert 'newLoops' from 'SmallVector<[...],3>' to
'SmallVector<[...],6>'
  310 |   return newLoops;

The fix is to remove the explicit size in the local declaration for the
SmallVector being returned.
DeltaFile
+1-1mlir/lib/Dialect/OpenACC/Utils/OpenACCUtilsTiling.cpp
+1-11 files

LLVM/project eab9394clang/include/clang/DependencyScanning DependencyScannerImpl.h, clang/lib/DependencyScanning DependencyScannerImpl.cpp DependencyScanningWorker.cpp

[clang][DependencyScanning] Fix spelling mistake for DiagnosticsEngineWithDiagOpts (#171535)

DeltaFile
+5-5clang/include/clang/DependencyScanning/DependencyScannerImpl.h
+2-2clang/lib/DependencyScanning/DependencyScannerImpl.cpp
+1-1clang/lib/DependencyScanning/DependencyScanningWorker.cpp
+8-83 files

LLVM/project 53cd4abllvm/lib/Target/AArch64 AArch64InstrInfo.cpp AArch64RedundantCondBranchPass.cpp, llvm/test/CodeGen/AArch64 arm64-shrink-wrapping.ll pr164181.ll

Revert "[AArch64] Run optimizeTerminators earlier too." (#171505)

Reverts llvm/llvm-project#170907

Causes crashes, see
https://github.com/llvm/llvm-project/pull/170907#issuecomment-3634271414
DeltaFile
+52-22llvm/test/CodeGen/AArch64/arm64-shrink-wrapping.ll
+37-29llvm/test/CodeGen/AArch64/pr164181.ll
+0-47llvm/lib/Target/AArch64/AArch64InstrInfo.cpp
+45-1llvm/lib/Target/AArch64/AArch64RedundantCondBranchPass.cpp
+24-12llvm/test/CodeGen/AArch64/block-placement-optimize-branches.ll
+16-9llvm/test/tools/UpdateTestChecks/update_llc_test_checks/Inputs/aarch64_generated_funcs.ll.generated.expected
+174-1206 files not shown
+203-14812 files

LLVM/project 687986ellvm/docs CMake.rst

[Docs] Add documentation for LLVM_ENABLE_CURL (#170928)

DeltaFile
+6-0llvm/docs/CMake.rst
+6-01 files

LLVM/project 21147e7mlir/include/mlir/Dialect/OpenACC OpenACCUtilsTiling.h, mlir/lib/Dialect/OpenACC/Utils OpenACCUtilsTiling.cpp CMakeLists.txt

[mlir][acc] Add loop tiling utilities for OpenACC (#171490)

Add utilities in OpenACCUtilsTiling.h/.cpp to support tiling
transformations on acc.loop operations:

- uncollapseLoops: Expand collapsed loops with multiple IVs into nested
loop structures when tile count exceeds collapse count
- tileACCLoops: Transform loop nests into tile and element loops based
on provided tile sizes, with automatic resolution of unknown tile sizes
(tile(*) represented as -1)

These utilities prepare for the ACCLoopTiling pass which handles the
OpenACC loop tile directive.

---------

Co-authored-by: Vijay Kandiah <vkandiah at nvidia.com>
DeltaFile
+348-0mlir/unittests/Dialect/OpenACC/OpenACCUtilsTilingTest.cpp
+311-0mlir/lib/Dialect/OpenACC/Utils/OpenACCUtilsTiling.cpp
+83-0mlir/include/mlir/Dialect/OpenACC/OpenACCUtilsTiling.h
+3-0mlir/lib/Dialect/OpenACC/Utils/CMakeLists.txt
+1-0mlir/unittests/Dialect/OpenACC/CMakeLists.txt
+746-05 files

LLVM/project 29760celldb/source/Interpreter CommandInterpreter.cpp, lldb/test/API/functionalities/abbreviation TestAbbreviations.py

[lldb] Fix capitalization in ambiguous command error (#171519)

We follow LLVM's style guide for diagnostics, which instructs to start
the first sentence with a lowercase letter, and finish the last sentence
without a period, if it would end in one otherwise.
DeltaFile
+2-4lldb/source/Interpreter/CommandInterpreter.cpp
+1-1lldb/test/API/functionalities/abbreviation/TestAbbreviations.py
+1-1lldb/test/API/functionalities/ambigous_commands/TestAmbiguousCommands.py
+1-1lldb/test/API/functionalities/wrong_commands/TestWrongCommands.py
+5-74 files

LLVM/project 2da1699flang-rt/lib/runtime cudadevice.f90 __ppc_intrinsics.f90, flang/module cudadevice.f90 __ppc_intrinsics.f90

Merge branch 'main' into revert-170907-gh-a64-cbzwzrearly
DeltaFile
+0-2,242flang-rt/lib/runtime/cudadevice.f90
+2,242-0flang/module/cudadevice.f90
+1,911-0flang/module/__ppc_intrinsics.f90
+0-1,911flang-rt/lib/runtime/__ppc_intrinsics.f90
+0-1,122flang-rt/lib/runtime/mma.f90
+1,122-0flang/module/mma.f90
+5,275-5,275153 files not shown
+9,009-9,087159 files

LLVM/project dda715dbolt/lib/Core BinaryContext.cpp, bolt/test dwarf5-missing-dwo.c

[BOLT][DWARF] Improve reporting on missing DWOs (#171506)

List all required missing DWO files and report a summary with
recommendations on how to proceed.
DeltaFile
+20-5bolt/lib/Core/BinaryContext.cpp
+18-0bolt/test/dwarf5-missing-dwo.c
+38-52 files

LLVM/project 782f507lldb/source/Plugins/ObjectFile/wasm ObjectFileWasm.cpp ObjectFileWasm.h, lldb/test/Shell/Symtab symtab-wasm.test

[lldb][Wasm] Handle imports when parsing Wasm name sections (#170960)

LLDB can use the wasm name section to populate its symbol table and get
names for functions. However the index space used in the name section is
the "function index space" which includes imported as well as locally
defined functions.
DeltaFile
+74-10lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.cpp
+47-27lldb/test/Shell/Symtab/Inputs/simple.wasm.yaml
+6-4lldb/test/Shell/Symtab/symtab-wasm.test
+1-0lldb/source/Plugins/ObjectFile/wasm/ObjectFileWasm.h
+128-414 files

LLVM/project 93404e0mlir/lib/Conversion/XeGPUToXeVM XeGPUToXeVM.cpp, mlir/test/Conversion/XeGPUToXeVM loadstore_matrix.mlir loadstoreprefetch.mlir

add more tests
DeltaFile
+159-90mlir/test/Conversion/XeGPUToXeVM/loadstore_matrix.mlir
+25-0mlir/test/Conversion/XeGPUToXeVM/loadstoreprefetch.mlir
+1-1mlir/lib/Conversion/XeGPUToXeVM/XeGPUToXeVM.cpp
+185-913 files

LLVM/project a9bcedfllvm/lib/Target/SPIRV SPIRVLegalizerInfo.cpp SPIRVPreLegalizer.cpp

[NFC][SPIRV] Fix breakage introduced by #170798 (#171513)

Adding support for i128 missed a few quirks of legalisation, which were
masked previously by early erroring out on bitwidth > 64. i128 uses
should be legal, we decide whether or not the resulting module is viable
(i.e. if the required extensions are present) in the ModuleAnalysis
pass.
DeltaFile
+9-8llvm/lib/Target/SPIRV/SPIRVLegalizerInfo.cpp
+2-2llvm/lib/Target/SPIRV/SPIRVPreLegalizer.cpp
+11-102 files

LLVM/project 7062cd6clang/lib/Driver/ToolChains Clang.cpp, clang/lib/Frontend CompilerInvocation.cpp

[Matrix] Add a row\col major toggle in the clang driver (#167628)

fixes #167621

- define the new options in `Options.td` limit the naming to row-major
or column-major.
- In `ToolChains/Clang.cpp` limit the opt usage to only when
`-fenable-matrix` is used.

---------

Co-authored-by: Florian Hahn <flo at fhahn.com>
DeltaFile
+49-0clang/test/Driver/fmatrix-memory-layout.c
+36-0clang/unittests/Frontend/CompilerInvocationTest.cpp
+30-0clang/lib/Frontend/CompilerInvocation.cpp
+19-0clang/lib/Sema/SemaChecking.cpp
+13-0clang/test/Sema/matrix-col-major-builtin-disable.c
+12-0clang/lib/Driver/ToolChains/Clang.cpp
+159-08 files not shown
+190-014 files

LLVM/project 2765113llvm/lib/Target/AMDGPU BUFInstructions.td, llvm/test/Analysis/UniformityAnalysis/AMDGPU atomics.ll

AMDGPU: Drop and upgrade llvm.amdgcn.atomic.csub/cond.sub to atomicrmw (#105553)

These both perform conditional subtraction, returning the minuend and
zero respectively, if the difference is negative.
DeltaFile
+169-125llvm/test/CodeGen/AMDGPU/atomics_cond_sub.ll
+0-270llvm/test/CodeGen/AMDGPU/GlobalISel/llvm.amdgcn.global.atomic.csub.ll
+0-219llvm/test/CodeGen/AMDGPU/llvm.amdgcn.atomic.cond.sub.ll
+147-0llvm/test/Bitcode/amdgcn-atomic.ll
+0-57llvm/test/Analysis/UniformityAnalysis/AMDGPU/atomics.ll
+18-36llvm/lib/Target/AMDGPU/BUFInstructions.td
+334-70710 files not shown
+346-76216 files

LLVM/project 4af81d6llvm/lib/Transforms/IPO GlobalOpt.cpp, llvm/test/Transforms/GlobalOpt disable-globals-aa.ll

[GlobalOpt][profcheck] Mark as `unknown` the branch weights of global shrunk to boolean
DeltaFile
+24-4llvm/test/Transforms/GlobalOpt/disable-globals-aa.ll
+0-6llvm/utils/profcheck-xfail.txt
+4-1llvm/lib/Transforms/IPO/GlobalOpt.cpp
+28-113 files