[Mlir-commits] [llvm] [mlir] [MLIR] Generalize expand_shape to take shape as explicit input (PR #69267)

[Gaurav Shukla]

================
@@ -778,6 +805,11 @@ fuseWithReshapeByExpansion(LinalgOp linalgOp, Operation *reshapeOp,
           expandedType.getShape(), collapsedType.getShape(), rewriter)))
     return std::nullopt;
 
+  // TODO: With the support of multiple dynamic dims expansion in
+  // tensor.expand_shape op, this case can be handled.
+  if (failed(validateDynamicDimExpansion(linalgOp, expansionInfo, rewriter)))
----------------
Shukla-Gaurav wrote:

Yes, this needs to be fixed anyway and it was easier to place this check here as we have expansionInfo.

https://github.com/llvm/llvm-project/pull/69267