[all-commits] [llvm/llvm-project] 51eeda: AMDGPU: Set the addrspacecast nonnull flag instead...

Matt Arsenault via All-commits all-commits at lists.llvm.org
Tue Sep 8 05:16:37 PDT 2026


  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 51eedac31577e2394d240aa9267374183a42b231
      https://github.com/llvm/llvm-project/commit/51eedac31577e2394d240aa9267374183a42b231
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2026-09-08 (Tue, 08 Sep 2026)

  Changed paths:
    M clang/test/CodeGenCUDA/fp-contract.cu
    M llvm/lib/Target/AMDGPU/AMDGPUCodeGenPrepare.cpp
    M llvm/test/CodeGen/AMDGPU/codegen-prepare-addrspacecast-non-null.ll

  Log Message:
  -----------
  AMDGPU: Set the addrspacecast nonnull flag instead of the intrinsic (#220927)

AMDGPUCodeGenPrepare proved the source of certain flat<->local/private
addrspacecasts non-null and rewrote them to
llvm.amdgcn.addrspacecast.nonnull. Now that the flag is honored in
codegen, set it in place on the existing instruction instead.

Co-authored-by: Claude (Claude-Opus-4.8) <noreply at anthropic.com>



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


More information about the All-commits mailing list