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

[llvmlistbot at llvm.org]

================
@@ -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:

updated `shape` to `maskShape` to better represent the meaning.

Currently limiting it to only 1 d because the `arith.constant` op needs different treatment from the other two ops.  

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