[PATCH] D21764: AMDGPU: Remove superfluous string attributes from tests

Nicolai Hähnle via llvm-commits llvm-commits at lists.llvm.org
Wed Jul 6 01:33:58 PDT 2016


nhaehnle added a subscriber: nhaehnle.
nhaehnle accepted this revision.
nhaehnle added a reviewer: nhaehnle.
nhaehnle added a comment.
This revision is now accepted and ready to land.

Apart from some minor comments, LGTM.


================
Comment at: test/CodeGen/AMDGPU/fmul.ll:1
@@ -1,2 +1,2 @@
-; RUN: llc -march=amdgcn -mcpu=SI -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
-; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=SI -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
+; RUN: llc -march=amdgcn -mcpu=tonga -verify-machineinstrs < %s | FileCheck -check-prefix=GCN -check-prefix=FUNC %s
----------------
Isn't it better to be explicit about -mcpu? Not that I care strongly either way...

================
Comment at: test/CodeGen/AMDGPU/llvm.SI.sendmsg-m0.ll:17-19
@@ -16,5 +16,4 @@
 ; Function Attrs: nounwind
 declare void @llvm.SI.sendmsg(i32, i32) #1
 
-attributes #0 = { "unsafe-fp-math"="true" }
-attributes #1 = { nounwind }
+attributes #0 = { nounwind }
----------------
Now #1 is missing in the sendmsg declare, isn't it?

================
Comment at: test/CodeGen/AMDGPU/sgpr-copy.ll:398-400
@@ -398,4 +397,5 @@
+attributes #0 = { nounwind }
 attributes #1 = { nounwind readnone }
-attributes #2 = { readonly }
-attributes #3 = { readnone }
+attributes #2 = { nounwind readonly }
+attributes #3 = { nounwind readnone }
 
----------------
Duplicate #1 and #3.


http://reviews.llvm.org/D21764





More information about the llvm-commits mailing list