Adopt Properties to store operations inherent Attributes in the Vector dialect

[mehdi_amini]

Summary

This is part of an on-going migration to adopt Properties inside MLIR.

[mehdi_amini]

clang-format

[mehdi_amini]

Rebase

[mehdi_amini]

Rebase after improving diagnostic on invalid IR

[rriddle]

mlir/lib/Dialect/Vector/IR/VectorOps.cpp

1157–1158

Would be nice if there was a way to just extract it from the op, without the explicit casting. This API gets a bit nastier, which is slightly sad.

[mehdi_amini]

Rebase

mlir/include/mlir/Dialect/Vector/IR/VectorOps.td

 include "mlir/Interfaces/ViewLikeInterface.td"
 
 def Vector_Dialect : Dialect {
   let name = "vector";
   let cppNamespace = "::mlir::vector";
 
   let useDefaultAttributePrinterParser = 1;
   let hasConstantMaterializer = 1;
+  let usePropertiesForAttributes = 1;
   let dependentDialects = ["arith::ArithDialect"];
 }
 
 // Base class for Vector dialect ops.
 class Vector_Op<string mnemonic, list<Trait> traits = []> :
     Op<Vector_Dialect, mnemonic, traits>;
 
 // The "kind" of combining function for contractions and reductions.

mlir/lib/Dialect/Vector/IR/VectorOps.cpp

void vector::ExtractOp::build(OpBuilder &builder, OperationState &result,
   build(builder, result, source, positionConstants);
 }
 
 LogicalResult
 ExtractOp::inferReturnTypes(MLIRContext *, std::optional<Location>,
                             ValueRange operands, DictionaryAttr attributes,
                             OpaqueProperties properties, RegionRange,
                             SmallVectorImpl<Type> &inferredReturnTypes) {
-  ExtractOp::Adaptor op(operands, attributes);
+  ExtractOp::Adaptor op(operands, attributes, *properties.as<ExtractOp::Properties *>());
   auto vectorType = op.getVector().getType().cast<VectorType>();
   if (static_cast<int64_t>(op.getPosition().size()) == vectorType.getRank()) {
     inferredReturnTypes.push_back(vectorType.getElementType());
   } else {
     auto n =
         std::min<size_t>(op.getPosition().size(), vectorType.getRank() - 1);
     inferredReturnTypes.push_back(VectorType::get(
         vectorType.getShape().drop_front(n), vectorType.getElementType()));
   }
LogicalResult ShuffleOp::verify() {
   return success();
 }
 
 LogicalResult
 ShuffleOp::inferReturnTypes(MLIRContext *, std::optional<Location>,
                             ValueRange operands, DictionaryAttr attributes,
                             OpaqueProperties properties, RegionRange,
                             SmallVectorImpl<Type> &inferredReturnTypes) {
-  ShuffleOp::Adaptor op(operands, attributes);
+  ShuffleOp::Adaptor op(operands, attributes, *properties.as<ShuffleOp::Properties *>());
   auto v1Type = op.getV1().getType().cast<VectorType>();
   auto v1Rank = v1Type.getRank();
   // Construct resulting type: leading dimension matches mask
   // length, all trailing dimensions match the operands.
   SmallVector<int64_t, 4> shape;
   shape.reserve(v1Rank);
   shape.push_back(std::max<size_t>(1, op.getMask().size()));
   // In the 0-D case there is no trailing shape to append.