[all-commits] [llvm/llvm-project] cf765a: [AMDGPU] Treat GEP offsets as signed in AMDGPUProm...

[Fabian Ritter via All-commits]

  Branch: refs/heads/users/ritter-x2a/09-09-_amdgpu_treat_gep_offsets_as_signed_in_amdgpupromotealloca
  Home:   https://github.com/llvm/llvm-project
  Commit: cf765a07cf544e07d1cb6d0ae245642cf1d36890
      https://github.com/llvm/llvm-project/commit/cf765a07cf544e07d1cb6d0ae245642cf1d36890
  Author: Fabian Ritter <fabian.ritter at amd.com>
  Date:   2025-09-09 (Tue, 09 Sep 2025)

  Changed paths:
    A llvm/test/CodeGen/AMDGPU/promote-alloca-negative-index.ll

  Log Message:
  -----------
  [AMDGPU] Treat GEP offsets as signed in AMDGPUPromoteAlloca

AMDGPUPromoteAlloca can transform i32 GEP offsets that operate on
allocas into i64 extractelement indices. Before this patch, negative GEP
offsets would be zero-extended, leading to wrong extractelement indices
with values around (2**32-1).

This fixes failing LlvmLibcCharacterConverterUTF32To8Test tests for
AMDGPU.


  Commit: b32f5d1d26671d219c4cecf35d0c84a46190629e
      https://github.com/llvm/llvm-project/commit/b32f5d1d26671d219c4cecf35d0c84a46190629e
  Author: Fabian Ritter <fabian.ritter at amd.com>
  Date:   2025-09-09 (Tue, 09 Sep 2025)

  Changed paths:
    M llvm/lib/Target/AMDGPU/AMDGPUPromoteAlloca.cpp
    M llvm/test/CodeGen/AMDGPU/promote-alloca-negative-index.ll

  Log Message:
  -----------
  Add the actual implementation, which fixes the test.


Compare: https://github.com/llvm/llvm-project/compare/cf765a07cf54%5E...b32f5d1d2667

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