[all-commits] [llvm/llvm-project] ed1ee9: AtomicExpand: Stop using report_fatal_error (#147300)

[Matt Arsenault via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: ed1ee9a9bf6deb5ec147e79a156e2f8d9465616d
      https://github.com/llvm/llvm-project/commit/ed1ee9a9bf6deb5ec147e79a156e2f8d9465616d
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2025-07-09 (Wed, 09 Jul 2025)

  Changed paths:
    M llvm/include/llvm/IR/RuntimeLibcalls.td
    M llvm/lib/CodeGen/AtomicExpandPass.cpp
    R llvm/test/CodeGen/AMDGPU/atomic-oversize.ll
    A llvm/test/CodeGen/AMDGPU/unsupported-atomics.ll
    A llvm/test/CodeGen/NVPTX/atomicrmw-expand.err.ll
    M llvm/test/CodeGen/NVPTX/atomicrmw-expand.ll
    R llvm/test/Transforms/AtomicExpand/AMDGPU/expand-atomic-fp128.ll
    R llvm/test/Transforms/AtomicExpand/AMDGPU/expand-atomic-i128.ll
    M llvm/test/Transforms/AtomicExpand/AMDGPU/expand-atomic-mmra.ll
    M llvm/test/Transforms/AtomicExpand/AMDGPU/expand-atomicrmw-fp-vector.ll
    M llvm/test/Transforms/AtomicExpand/AMDGPU/unaligned-atomic.ll

  Log Message:
  -----------
  AtomicExpand: Stop using report_fatal_error (#147300)

Emit a context error and delete the instruction. This
allows removing the AMDGPU hack where some atomic libcalls
are falsely added. NVPTX also later copied the same hack,
so remove it there too.

For now just emit the generic error, which is not good. It's
missing any useful context information (despite taking the instruction).
It's also confusing in the failed atomicrmw case, since it's reporting
failure at the intermediate failed cmpxchg instead of the original
atomicrmw.



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