maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[mlir][amdgpu] Fix signed/unsigned comparison for abid/cbsz comparison 

Unsigned/signed comparison failure due to implicit signed value.

Reviewed By: stella.stamenova

Differential Revision: https://reviews.llvm.org/D133061
This commit is contained in:
Rob Suderman 2022-08-31 15:29:27 -07:00
parent 69193c6cd7
commit 7e52e0fc72
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
mlir/lib/Dialect/AMDGPU/IR/AMDGPUDialect.cpp
View File

@ -111,7 +111,7 @@ LogicalResult MFMAOp::verify() {
if (destElem.isF64() && getCbsz() != 0)
return emitOpError(
"double-precision ops do not support permuting lanes of A");
if (getAbid() >= (1 << getCbsz()))
if (getAbid() >= (1u << getCbsz()))
return emitOpError(
"block ID for permuting A (abid) must be below 2 ** cbsz");