[PATCH] D70418: AMDGPU: Handle waitcnt overflow

[Matt Arsenault via Phabricator via llvm-commits]

arsenm added inline comments.


================
Comment at: llvm/test/CodeGen/AMDGPU/waitcnt-overflow.mir:1
+# RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -run-pass si-insert-waitcnts -o - %s | FileCheck -check-prefixes=GFX9,GFX9_10 %s
+# RUN: llc -march=amdgcn -mcpu=gfx1010 -verify-machineinstrs -run-pass si-insert-waitcnts -o - %s | FileCheck -check-prefixes=GFX10,GFX9_10 %s
----------------
These look like generated checks, but the update_mir_test_checks comment is missing?


================
Comment at: llvm/test/CodeGen/AMDGPU/waitcnt-overflow.mir:8
+---
+name:            max-counter-lgkmcnt
+body:             |
----------------
Should also add an exp test


================
Comment at: llvm/test/CodeGen/AMDGPU/waitcnt-overflow.mir:24
+    ; GFX10-LABEL: name: max-counter-lgkmcnt
+    ; GFX10: $vgpr34_vgpr35 = DS_READ2_B32 renamable $vgpr99, 34, 35, 0, implicit $m0, implicit $exec
+    ; GFX10-NEXT: S_WAITCNT 53631
----------------
gfx9 and 10 should both be using the _gfx9 variant without the implicit m0


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70418/new/

https://reviews.llvm.org/D70418