[Mlir-commits] [mlir] [MLIR] Fix VectorEmulateNarrowType constant op mask bug (PR #116064)

llvmlistbot at llvm.org llvmlistbot at llvm.org
Thu Nov 14 14:33:20 PST 2024


================
@@ -129,6 +132,45 @@ static FailureOr<Operation *> getCompressedMaskOp(OpBuilder &rewriter,
       auto denseAttr = DenseElementsAttr::get(newMaskType, newMaskValues);
       newMask = rewriter.create<arith::ConstantOp>(loc, newMaskType, denseAttr);
     }
+  } else if (constantOp) {
+    assert(shape.size() == 1 && "expected 1-D mask");
----------------
lialan wrote:

I honestly don't know if there is any > 1-d masks being generated, but `arith.constant` does allow arbitrary dimension constants being generated in this way. Is there any mechanism guarding it off from any such cases before this pass?

cc @hanhanW @dcaballe 

https://github.com/llvm/llvm-project/pull/116064


More information about the Mlir-commits mailing list