[PATCH] D101115: [AMDGPU] Move code sinking before structurizer

Fri Apr 23 05:54:32 PDT 2021

arsenm added a comment.

In D101115#2711462 <https://reviews.llvm.org/D101115#2711462>, @piotr wrote:

> I know it is a legacy pass, but I am convinced of its usefulness in our flow - both in real-world content and lit testing (e.g., no_skip_no_successors in skip-if-dead.ll).

I'm not worried about legacy, but whether it's redundant since it should have already run at this point

================
Comment at: llvm/test/CodeGen/AMDGPU/multilevel-break.ll:198
 ; GCN-NEXT:    ; =>This Inner Loop Header: Depth=1
-; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], 0
+; GCN-NEXT:    buffer_load_dword v1, off, s[0:3], 0 glc
 ; GCN-NEXT:    s_waitcnt vmcnt(0)
----------------
Did this drop the memory operand or something? This looks like a regression

================
Comment at: llvm/test/CodeGen/AMDGPU/sink-image-sample.ll:3
+; RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx1010 -verify-machineinstrs < %s | FileCheck -check-prefixes=GCN %s
+
+; GCN-LABEL: {{^}}sinking_img_sample:
----------------
Maybe add a comment explaining what this is for?

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D101115/new/

https://reviews.llvm.org/D101115