[Mlir-commits] [mlir] [MLIR] support dynamic indexing of `vector.maskedload` in `VectorEmulateNarrowTypes` (PR #115070)

[Han-Chung Wang]

================
@@ -549,29 +571,30 @@ struct ConvertVectorMaskedLoad final
             ? getConstantIntValue(linearizedInfo.intraDataOffset)
             : 0;
 
-    if (!foldedIntraVectorOffset) {
-      // unimplemented case for dynamic intra vector offset
-      return failure();
-    }
-
-    FailureOr<Operation *> newMask =
-        getCompressedMaskOp(rewriter, loc, op.getMask(), origElements, scale,
-                            *foldedIntraVectorOffset);
+    auto maxIntraDataOffset = foldedIntraVectorOffset.value_or(scale - 1);
----------------
hanhanW wrote:

nit: spell the type out. The `auto` does not make the code more readable, IMHO.

https://llvm.org/docs/CodingStandards.html#use-auto-type-deduction-to-make-code-more-readable

https://github.com/llvm/llvm-project/pull/115070