[PATCH] D72022: [mlir][Linalg] Extend generic ops to allow tensors

[Nicolas Vasilache via Phabricator via llvm-commits]

nicolasvasilache added inline comments.


================
Comment at: mlir/include/mlir/Dialect/Linalg/Transforms/LinalgTransforms.h:83
+LogicalResult vectorizeGenericLinalgOpPrecondition(Operation *op);
+SmallVector<Value, 0> vectorizeGenericLinalgOp(PatternRewriter &rewriter,
+                                               Operation *op);
----------------
rriddle wrote:
> Why SmallVector<..., 0>?
This needs a type compatible with `SmallVector<Value, x>` to be used as a NativeCall with for an op with Variadic returns.
Do you have a better suggestion?


================
Comment at: mlir/lib/Dialect/VectorOps/VectorTransforms.cpp:500
   // Unroll 'op' with 'iterationBounds' to 'targetShape'.
-  return unrollSingleResultStructuredOp(op, iterationBounds, vectors,
-                                        resultIndex, targetShape, builder);
+  return SmallVector<Value, 1>{unrollSingleResultStructuredOp(
+      op, iterationBounds, vectors, resultIndex, targetShape, builder)};
----------------
ftynse wrote:
> Nit: would llvm::to_vector<1> work?
Nope, this is just constructing a 1-element vector from a single value.


================
Comment at: mlir/tools/mlir-tblgen/RewriterGen.cpp:577
+      if (resultTree.isNativeCodeCall()) {
+        os << symbolInfoMap.getAllRangeUse(
+            val, "SmallVector<Value, 4> values{0};", "\n");
----------------
rriddle wrote:
> It isn't immediately clear to me why this is necessary. What return value is convertible to small vector, but not ArrayRef for example?
Reworked and added a comment.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D72022/new/

https://reviews.llvm.org/D72022