[all-commits] [llvm/llvm-project] 7b744a: [MLIR][MLProgram] Fix crash in mlprogram-pipeline-...

[Andrei Elovikov via All-commits]

  Branch: refs/heads/users/eas/split-mem-widen-transform
  Home:   https://github.com/llvm/llvm-project
  Commit: 7b744a511e989df39f0a49f80528b0580348b863
      https://github.com/llvm/llvm-project/commit/7b744a511e989df39f0a49f80528b0580348b863
  Author: Mehdi Amini <joker.eph at gmail.com>
  Date:   2026-04-14 (Tue, 14 Apr 2026)

  Changed paths:
    M mlir/lib/Dialect/MLProgram/Transforms/PipelineGlobalOps.cpp
    M mlir/test/Dialect/MLProgram/pipeline-globals.mlir

  Log Message:
  -----------
  [MLIR][MLProgram] Fix crash in mlprogram-pipeline-globals on unresolvable callees (#189244)

The `MLProgramPipelineGlobals` pass crashed with a null pointer dereference
when a `CallOpInterface` operation referred to a callee symbol that could not
be resolved in the IR (e.g. an external function defined outside the module).

Instead  conservatively bail out when a callee symbol cannot be resolved, 
causing the pass to (preserving all loads/stores). This is consistent with
how Value-based callees are handled.

Fixes #109649

Assisted-by: Claude Code


  Commit: c47c3344dc01a86e34168623120820a29785a3d1
      https://github.com/llvm/llvm-project/commit/c47c3344dc01a86e34168623120820a29785a3d1
  Author: Keith Smiley <keithbsmiley at gmail.com>
  Date:   2026-04-14 (Tue, 14 Apr 2026)

  Changed paths:
    M mlir/include/mlir/Dialect/AMDGPU/IR/AMDGPUEnums.h

  Log Message:
  -----------
  [mlir][AMD] Add missing includes to AMDGPUEnums.h (NFC) (#191077)

This header assumed these had been imported


  Commit: 3ba18a675bf8bbfc4d52427e895a518a5685f19b
      https://github.com/llvm/llvm-project/commit/3ba18a675bf8bbfc4d52427e895a518a5685f19b
  Author: Justin Fargnoli <jfargnoli at nvidia.com>
  Date:   2026-04-14 (Tue, 14 Apr 2026)

  Changed paths:
    M llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
    M llvm/test/Transforms/LoopUnroll/debug.ll

  Log Message:
  -----------
  [LoopUnroll] Fix misleading runtime unroll debug message (#190709)

Avoid the confusing `Runtime unrolling with count: 0` `LLVM_DEBUG`
statement.


  Commit: 142d3c2a445365498e8e7d8a2926595ac0508615
      https://github.com/llvm/llvm-project/commit/142d3c2a445365498e8e7d8a2926595ac0508615
  Author: Chi-Chun, Chen <chichun.chen at hpe.com>
  Date:   2026-04-14 (Tue, 14 Apr 2026)

  Changed paths:
    M llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
    M llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp

  Log Message:
  -----------
  [NFC][OMPIRBuilder][OpenMP] Pass AffinityData by reference instead of… (#191863)

… value


  Commit: 8a69fb096cb3f3f99ee9bf25247f4b0bc442ca84
      https://github.com/llvm/llvm-project/commit/8a69fb096cb3f3f99ee9bf25247f4b0bc442ca84
  Author: hanbeom <kese111 at gmail.com>
  Date:   2026-04-15 (Wed, 15 Apr 2026)

  Changed paths:
    M llvm/lib/Transforms/InstCombine/InstCombineAddSub.cpp
    M llvm/test/Transforms/InstCombine/add.ll

  Log Message:
  -----------
  [InstCombine] Fold (X + C) + (Y & ~C) to X + (Y | C) (#191334)

Add an InstCombine fold for masked overwrite patterns where the add
constant matches the cleared bits in the mask:

  (X + C) + (Y & ~C) -> X + (Y | C)

Since `Y & ~C` clears all bits set in C, adding C cannot generate carry
through those bits and is equivalent to setting them with `or`.

Proof: https://alive2.llvm.org/ce/z/277UFK
Fixed: https://github.com/llvm/llvm-project/issues/191171


  Commit: 1859c66da433cfc80f6fddcc10760bbd01f4df2a
      https://github.com/llvm/llvm-project/commit/1859c66da433cfc80f6fddcc10760bbd01f4df2a
  Author: Andrei Elovikov <andrei.elovikov at sifive.com>
  Date:   2026-04-14 (Tue, 14 Apr 2026)

  Changed paths:
    M llvm/include/llvm/Frontend/OpenMP/OMPIRBuilder.h
    M llvm/lib/Frontend/OpenMP/OMPIRBuilder.cpp
    M llvm/lib/Transforms/InstCombine/InstCombineAddSub.cpp
    M llvm/lib/Transforms/Scalar/LoopUnrollPass.cpp
    M llvm/test/Transforms/InstCombine/add.ll
    M llvm/test/Transforms/LoopUnroll/debug.ll
    M mlir/include/mlir/Dialect/AMDGPU/IR/AMDGPUEnums.h
    M mlir/lib/Dialect/MLProgram/Transforms/PipelineGlobalOps.cpp
    M mlir/test/Dialect/MLProgram/pipeline-globals.mlir

  Log Message:
  -----------
  Merge branch 'main' into users/eas/split-mem-widen-transform


Compare: https://github.com/llvm/llvm-project/compare/cb11242f00d0...1859c66da433

To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications