[all-commits] [llvm/llvm-project] 837096: AMDGPU: Fix implicit vcc def to vcc_lo on wave32 t...

[Matt Arsenault via All-commits]

  Branch: refs/heads/users/arsenm/si-shrink-instructions-vcc-lo-fix
  Home:   https://github.com/llvm/llvm-project
  Commit: 83709620763a6ac4017d412d8bbfaf8ec999cca1
      https://github.com/llvm/llvm-project/commit/83709620763a6ac4017d412d8bbfaf8ec999cca1
  Author: Matt Arsenault <Matthew.Arsenault at amd.com>
  Date:   2024-09-21 (Sat, 21 Sep 2024)

  Changed paths:
    M llvm/lib/Target/AMDGPU/SIInstrInfo.cpp
    M llvm/test/CodeGen/AMDGPU/GlobalISel/divergence-divergent-i1-phis-no-lane-mask-merging.ll
    M llvm/test/CodeGen/AMDGPU/GlobalISel/divergence-divergent-i1-used-outside-loop.ll
    M llvm/test/CodeGen/AMDGPU/atomic_optimizations_global_pointer.ll
    M llvm/test/CodeGen/AMDGPU/atomic_optimizations_local_pointer.ll
    M llvm/test/CodeGen/AMDGPU/global_atomics_scan_fadd.ll
    M llvm/test/CodeGen/AMDGPU/global_atomics_scan_fmax.ll
    M llvm/test/CodeGen/AMDGPU/global_atomics_scan_fmin.ll
    M llvm/test/CodeGen/AMDGPU/global_atomics_scan_fsub.ll
    M llvm/test/CodeGen/AMDGPU/llvm.amdgcn.wqm.demote.ll
    M llvm/test/CodeGen/AMDGPU/set-inactive-wwm-overwrite.ll
    M llvm/test/CodeGen/AMDGPU/should-not-hoist-set-inactive.ll
    M llvm/test/CodeGen/AMDGPU/shrink-true16.mir
    A llvm/test/CodeGen/AMDGPU/shrink-v-cmp-wave32-dead-vcc-lo.mir
    M llvm/test/CodeGen/AMDGPU/skip-if-dead.ll
    M llvm/test/CodeGen/AMDGPU/vgpr-liverange.ll
    M llvm/test/CodeGen/AMDGPU/wave32.ll

  Log Message:
  -----------
  AMDGPU: Fix implicit vcc def to vcc_lo on wave32 targets

Fixes a crash when shrinking wave32 compares with dead defs

The shrinking code was not replacing the vcc implicit-def operand
with vcc_lo on wave32. If the operand was dead, the code trying
to preserve the dead flag would crash.



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