[Mlir-commits] [mlir] [mlir] Add direct vectorization lowering for `tensor.pack` ops (PR #78660)

[Han-Chung Wang]

================
@@ -1585,6 +1727,32 @@ vectorizeLinalgOpPrecondition(LinalgOp linalgOp,
   return success();
 }
 
+/// TODO: Use a matcher to check for a constant padding value.
+static LogicalResult
+vectorizePackOpPrecondition(tensor::PackOp packOp,
+                            ArrayRef<int64_t> inputVectorSizes) {
+  auto padValue = packOp.getPaddingValue();
+  if (padValue && !padValue.getDefiningOp<arith::ConstantOp>()) {
+    LDBG("pad value is not constant: " << packOp << "\n");
+    return failure();
+  }
+
----------------
hanhanW wrote:

good catch, I think the current implementation is okay, which checks if they are from arith.constant.

https://github.com/llvm/llvm-project/pull/78660