[PATCH] D130537: AMDGPU: Take care of "tied" operand when removeOperand

Thu Jul 28 16:47:56 PDT 2022

arsenm accepted this revision.
arsenm added inline comments.
This revision is now accepted and ready to land.

================
Comment at: llvm/test/CodeGen/AMDGPU/frame-index-elimination-tied-operand.mir:2
+# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
+# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 -run-pass=prologepilog -o - %s | FileCheck -check-prefix=GCN %s
+...
----------------
Should add gfx11 to the test name?

================
Comment at: llvm/test/CodeGen/AMDGPU/frame-index-elimination.ll:4
 ; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx900 -mattr=-promote-alloca,+enable-flat-scratch -amdgpu-sroa=0 -verify-machineinstrs < %s | FileCheck -enable-var-scope -check-prefixes=GCN,GFX9,GFX9-FLATSCR %s
+; RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 < %s | FileCheck --check-prefixes=GFX11 %s

----------------
Would be better to have full checks for all targets on all functions

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D130537/new/

https://reviews.llvm.org/D130537