[mlir] Generalize OpFoldResult usage in ops with offsets, sizes and operands.

[nicolasvasilache]

Summary

This revision starts evolving the APIs to manipulate ops with offsets, sizes and operands towards a ValueOrAttr abstraction that is already used in folding under the name OpFoldResult.

The objective, in the future, is to allow such manipulations all the way to the level of ODS to avoid all the genuflexions involved in distinguishing between values and attributes for generic constant foldings.

Once this evolution is accepted, the next step will be a mechanical OpFoldResult -> ValueOrAttr.

[nicolasvasilache]

Fix test.

[jpienaar]

mlir/include/mlir/IR/OpDefinition.h

229

But: reduce lifeness scope:

if (Value ...)

[nicolasvasilache]

Address.

[nicolasvasilache]

Keep a ValueRange based builder around.

[pifon2a]

mlir/lib/Dialect/StandardOps/IR/Ops.cpp

45–52

nit:

if (auto v =  ofr.dyn_cast<Value>()) {
  dynamicVec.push_back(v);
  staticVec.push_back(sentinel);
  return;
}
APInt apInt = ofr.dyn_cast<Attribute>().cast<IntegerAttr>().getValue();
 taticVec.push_back(apInt.getSExtValue());
}

2144–2149

can it be initialized from (sizes.begin(), sizes.end())?

[nicolasvasilache]

mlir/lib/Dialect/StandardOps/IR/Ops.cpp

2144–2149

Unfortunately no, I get:

third_party/crosstool/v18/stable/toolchain/bin/../include/c++/v1/memory:2193:51: error: no matching constructor for initialization of 'value_type' (aka 'mlir::OpFoldResult')
            ::new ((void*)_VSTD::addressof(*__r)) value_type(*__f);
                                                  ^          ~~~~
third_party/llvm/llvm-project/llvm/include/llvm/ADT/SmallVector.h:490:10: note: in instantiation of function template specialization 'std::uninitialized_copy<const long *, mlir::OpFoldResult *>' requested here
    std::uninitialized_copy(I, E, Dest);
         ^
third_party/llvm/llvm-project/llvm/include/llvm/ADT/SmallVector.h:652:11: note: in instantiation of function template specialization 'llvm::SmallVectorTemplateBase<mlir::OpFoldResult, true>::uninitialized_copy<const long *, mlir::OpFoldResult *>' requested here
    this->uninitialized_copy(in_start, in_end, this->end());
          ^
third_party/llvm/llvm-project/llvm/include/llvm/ADT/SmallVector.h:1184:11: note: in instantiation of function template specialization 'llvm::SmallVectorImpl<mlir::OpFoldResult>::append<const long *, void>' requested here
    this->append(S, E);
          ^
third_party/llvm/llvm-project/mlir/lib/Dialect/StandardOps/IR/Ops.cpp:2127:29: note: in instantiation of function template specialization 'llvm::SmallVector<mlir::OpFoldResult, 6>::SmallVector<const long *, void>' requested here
  SmallVector<OpFoldResult> sizeValues(sizes.begin(), sizes.end());
                            ^
third_party/llvm/llvm-project/llvm/include/llvm/ADT/PointerUnion.h:104:5: note: candidate inherited constructor not viable: no known conversion from 'const long' to 'llvm::PointerIntPair<void *, 1, int, llvm::pointer_union_detail::PointerUnionUIntTraits<mlir::Attribute, mlir::Value>>' for 1st argument
    PointerUnionMembers(ValTy Val) : Val(Val) {}
    ^
third_party/llvm/llvm-project/mlir/include/mlir/IR/OpDefinition.h:220:41: note: constructor from base class 'PointerUnion<mlir::Attribute, mlir::Value>' inherited here
  using PointerUnion<Attribute, Value>::PointerUnion;
                                        ^
third_party/llvm/llvm-project/llvm/include/llvm/ADT/PointerUnion.h:117:5: note: candidate inherited constructor not viable: no known conversion from 'const long' to 'mlir::Value' for 1st argument
    PointerUnionMembers(Type V)
    ^
third_party/llvm/llvm-project/mlir/include/mlir/IR/OpDefinition.h:220:41: note: constructor from base class 'PointerUnion<mlir::Attribute, mlir::Value>' inherited here
  using PointerUnion<Attribute, Value>::PointerUnion;
                                        ^
third_party/llvm/llvm-project/llvm/include/llvm/ADT/PointerUnion.h:117:5: note: candidate inherited constructor not viable: no known conversion from 'const long' to 'mlir::Attribute' for 1st argument
    PointerUnionMembers(Type V)
    ^
third_party/llvm/llvm-project/mlir/include/mlir/IR/OpDefinition.h:220:41: note: constructor from base class 'PointerUnion<mlir::Attribute, mlir::Value>' inherited here
  using PointerUnion<Attribute, Value>::PointerUnion;
                                        ^
third_party/llvm/llvm-project/llvm/include/llvm/ADT/PointerUnion.h:167:3: note: candidate inherited constructor not viable: no known conversion from 'const long' to 'std::nullptr_t' (aka 'nullptr_t') for 1st argument
  PointerUnion(std::nullptr_t) : PointerUnion() {}
  ^
third_party/llvm/llvm-project/mlir/include/mlir/IR/OpDefinition.h:220:41: note: constructor from base class 'PointerUnion<mlir::Attribute, mlir::Value>' inherited here
  using PointerUnion<Attribute, Value>::PointerUnion;

We'll prob want a ValueOrAttrRange in the future, I'm not too hung up on the current impl.

[nicolasvasilache]

Address and rebase.

[nicolasvasilache]

Rebase.

mlir/include/mlir/Dialect/StandardOps/IR/Ops.td

let arguments = (ins
     I64ArrayAttr:$static_sizes,
     I64ArrayAttr:$static_strides
   );
   let results = (outs AnyMemRef:$result);
 
   let builders = [
     // Build a ReinterpretCastOp with mixed static and dynamic entries.
     OpBuilderDAG<(ins "MemRefType":$resultType, "Value":$source,
-      "int64_t":$staticOffset, "ArrayRef<int64_t>":$staticSizes,
-      "ArrayRef<int64_t>":$staticStrides, "ValueRange":$offset,
-      "ValueRange":$sizes, "ValueRange":$strides,
+      "OpFoldResult":$offset, "ArrayRef<OpFoldResult>":$sizes,
+      "ArrayRef<OpFoldResult>":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a ReinterpretCastOp with all dynamic entries.
+    // Build a ReinterpretCastOp with static entries.
     OpBuilderDAG<(ins "MemRefType":$resultType, "Value":$source,
-      "Value":$offset, "ValueRange":$sizes, "ValueRange":$strides,
+      "int64_t":$offset, "ArrayRef<int64_t>":$sizes,
+      "ArrayRef<int64_t>":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
+    // Build a ReinterpretCastOp with dynamic entries.
+    OpBuilderDAG<(ins "MemRefType":$resultType, "Value":$source,
+      "Value":$offset, "ValueRange":$sizes,
+      "ValueRange":$strides,
+      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>
   ];
 
   let extraClassDeclaration = extraBaseClassDeclaration # [{
     // The result of the op is always a ranked memref.
     MemRefType getType() { return getResult().getType().cast<MemRefType>(); }
     Value getViewSource() { return source(); }
 
     /// Return the rank of the source ShapedType.
let arguments = (ins
     Variadic<Index>:$strides,
     I64ArrayAttr:$static_offsets,
     I64ArrayAttr:$static_sizes,
     I64ArrayAttr:$static_strides
   );
   let results = (outs AnyMemRef:$result);
 
   let builders = [
-    // Build a SubViewOp with mixed static and dynamic entries.
-    OpBuilderDAG<(ins "Value":$source, "ArrayRef<int64_t>":$staticOffsets,
-      "ArrayRef<int64_t>":$staticSizes, "ArrayRef<int64_t>":$staticStrides,
-      "ValueRange":$offsets, "ValueRange":$sizes, "ValueRange":$strides,
+    // Build a SubViewOp with mixed static and dynamic entries and custom
+    // result type. If the type passed is nullptr, it is inferred.
+    OpBuilderDAG<(ins "Value":$source, "ArrayRef<OpFoldResult>":$offsets,
+      "ArrayRef<OpFoldResult>":$sizes, "ArrayRef<OpFoldResult>":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubViewOp with all dynamic entries.
-    OpBuilderDAG<(ins "Value":$source, "ValueRange":$offsets,
-      "ValueRange":$sizes, "ValueRange":$strides,
+    // Build a SubViewOp with mixed static and dynamic entries and inferred
+    // result type.
+    OpBuilderDAG<(ins "MemRefType":$resultType, "Value":$source,
+      "ArrayRef<OpFoldResult>":$offsets, "ArrayRef<OpFoldResult>":$sizes,
+      "ArrayRef<OpFoldResult>":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubViewOp with mixed static and dynamic entries
-    // and custom result type.
+    // Build a SubViewOp with static entries and custom result type. If the
+    // type passed is nullptr, it is inferred.
+    OpBuilderDAG<(ins "Value":$source, "ArrayRef<int64_t>":$offsets,
+      "ArrayRef<int64_t>":$sizes, "ArrayRef<int64_t>":$strides,
+      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
+    // Build a SubViewOp with static entries and inferred result type.
     OpBuilderDAG<(ins "MemRefType":$resultType, "Value":$source,
-      "ArrayRef<int64_t>":$staticOffsets, "ArrayRef<int64_t>":$staticSizes,
-      "ArrayRef<int64_t>":$staticStrides, "ValueRange":$offsets,
+      "ArrayRef<int64_t>":$offsets, "ArrayRef<int64_t>":$sizes,
+      "ArrayRef<int64_t>":$strides,
+      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
+    // Build a SubViewOp with dynamic entries and custom result type. If the
+    // type passed is nullptr, it is inferred.
+    OpBuilderDAG<(ins "Value":$source, "ValueRange":$offsets,
       "ValueRange":$sizes, "ValueRange":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubViewOp with all dynamic entries and custom result type.
+    // Build a SubViewOp with dynamic entries and inferred result type.
     OpBuilderDAG<(ins "MemRefType":$resultType, "Value":$source,
       "ValueRange":$offsets, "ValueRange":$sizes, "ValueRange":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>
   ];
 
   let extraClassDeclaration = extraBaseClassDeclaration # [{
     /// Returns the type of the base memref operand.
     MemRefType getSourceType() {
let arguments = (ins
     Variadic<Index>:$strides,
     I64ArrayAttr:$static_offsets,
     I64ArrayAttr:$static_sizes,
     I64ArrayAttr:$static_strides
   );
   let results = (outs AnyRankedTensor:$result);
 
   let builders = [
-    // Build a SubTensorOp with mixed static and dynamic entries.
-    OpBuilderDAG<(ins "Value":$source, "ArrayRef<int64_t>":$staticOffsets,
-      "ArrayRef<int64_t>":$staticSizes, "ArrayRef<int64_t>":$staticStrides,
-      "ValueRange":$offsets, "ValueRange":$sizes, "ValueRange":$strides,
-      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubTensorOp with all dynamic entries.
-    OpBuilderDAG<(ins "Value":$source, "ValueRange":$offsets,
-      "ValueRange":$sizes, "ValueRange":$strides,
-      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubTensorOp with mixed static and dynamic entries
-    // and custom result type.
-    OpBuilderDAG<(ins "RankedTensorType":$resultType, "Value":$source,
-      "ArrayRef<int64_t>":$staticOffsets, "ArrayRef<int64_t>":$staticSizes,
-      "ArrayRef<int64_t>":$staticStrides, "ValueRange":$offsets,
-      "ValueRange":$sizes, "ValueRange":$strides,
-      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubTensorOp with all dynamic entries and custom result type.
-    OpBuilderDAG<(ins "RankedTensorType":$resultType, "Value":$source,
-      "ValueRange":$offsets, "ValueRange":$sizes, "ValueRange":$strides,
-      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
     // Build a SubTensorOp with mixed static and dynamic entries and inferred
     // result type.
     OpBuilderDAG<(ins "Value":$source, "ArrayRef<OpFoldResult>":$offsets,
       "ArrayRef<OpFoldResult>":$sizes, "ArrayRef<OpFoldResult>":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
     // Build a SubTensorOp with mixed static and dynamic entries and custom
     // result type. If the type passed is nullptr, it is inferred.
     OpBuilderDAG<(ins "RankedTensorType":$resultType, "Value":$source,
       "ArrayRef<OpFoldResult>":$offsets, "ArrayRef<OpFoldResult>":$sizes,
       "ArrayRef<OpFoldResult>":$strides,
+      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
+    // Build a SubTensorOp with dynamic entries and custom result type. If the
+    // type passed is nullptr, it is inferred.
+    OpBuilderDAG<(ins "Value":$source, "ValueRange":$offsets,
+      "ValueRange":$sizes, "ValueRange":$strides,
+      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
+    // Build a SubTensorOp with dynamic entries and inferred result type.
+    OpBuilderDAG<(ins "RankedTensorType":$resultType, "Value":$source,
+      "ValueRange":$offsets, "ValueRange":$sizes, "ValueRange":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>
   ];
 
   let extraClassDeclaration = extraBaseClassDeclaration # [{
     /// Returns the type of the base tensor operand.
     RankedTensorType getSourceType() {
       return source().getType().cast<RankedTensorType>();
     }
let arguments = (ins
     I64ArrayAttr:$static_sizes,
     I64ArrayAttr:$static_strides
   );
   let results = (outs AnyRankedTensor:$result);
 
   let builders = [
     // Build a SubTensorInsertOp with mixed static and dynamic entries.
     OpBuilderDAG<(ins "Value":$source, "Value":$dest,
-      "ArrayRef<int64_t>":$staticOffsets, "ArrayRef<int64_t>":$staticSizes,
-      "ArrayRef<int64_t>":$staticStrides, "ValueRange":$offsets,
-      "ValueRange":$sizes, "ValueRange":$strides,
-      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubTensorInsertOp with all dynamic entries.
-    OpBuilderDAG<(ins "Value":$source, "Value":$dest, "ValueRange":$offsets,
-      "ValueRange":$sizes, "ValueRange":$strides,
-      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
-    // Build a SubTensorInsertOp with mixed static and dynamic entries.
-    OpBuilderDAG<(ins "Value":$source, "Value":$dest,
       "ArrayRef<OpFoldResult>":$offsets, "ArrayRef<OpFoldResult>":$sizes,
       "ArrayRef<OpFoldResult>":$strides,
+      CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>,
+    // Build a SubTensorInsertOp with dynamic entries.
+    OpBuilderDAG<(ins "Value":$source, "Value":$dest,
+      "ValueRange":$offsets, "ValueRange":$sizes, "ValueRange":$strides,
       CArg<"ArrayRef<NamedAttribute>", "{}">:$attrs)>
   ];
 
   let extraClassDeclaration = extraBaseClassDeclaration # [{
     /// Returns the type of the base tensor operand.
     RankedTensorType getSourceType() {
       return source().getType().cast<RankedTensorType>();
     }

mlir/include/mlir/IR/OpDefinition.h

 // Allow comparing operators.
 inline bool operator==(OpState lhs, OpState rhs) {
   return lhs.getOperation() == rhs.getOperation();
 }
 inline bool operator!=(OpState lhs, OpState rhs) {
   return lhs.getOperation() != rhs.getOperation();
 }
 
+raw_ostream &operator<<(raw_ostream &os, OpFoldResult ofr);
+
 /// This class represents a single result from folding an operation.
 class OpFoldResult : public PointerUnion<Attribute, Value> {
   using PointerUnion<Attribute, Value>::PointerUnion;
+
+public:
+  void dump() { llvm::errs() << *this << "\n"; }
 };
 
 /// Allow printing to a stream.
+inline raw_ostream &operator<<(raw_ostream &os, OpFoldResult ofr) {
+  if (Value value = ofr.dyn_cast<Value>())
+    value.print(os);
+  else
+    ofr.dyn_cast<Attribute>().print(os);
+  return os;
+}
+
+/// Allow printing to a stream.
 inline raw_ostream &operator<<(raw_ostream &os, OpState &op) {
   op.print(os, OpPrintingFlags().useLocalScope());
   return os;
 }
 
 //===----------------------------------------------------------------------===//
 // Operation Trait Types
 //===----------------------------------------------------------------------===//

mlir/include/mlir/Interfaces/ViewLikeInterface.td

InterfaceMethod<
       /*args=*/(ins),
       /*methodBody=*/"",
       /*defaultImplementation=*/[{
         return $_op.sizes();
       }]
     >,
     InterfaceMethod<
       /*desc=*/[{
-        Return a vector of all the static or dynamic sizes of the op.
-      }],
-      /*retTy=*/"SmallVector<OpFoldResult, 4>",
-      /*methodName=*/"getMixedSizes",
-      /*args=*/(ins),
-      /*methodBody=*/"",
-      /*defaultImplementation=*/[{
-        SmallVector<OpFoldResult, 4> res;
-        std::array<unsigned, 3> ranks = $_op.getArrayAttrRanks();
-        unsigned numDynamic = 0;
-        unsigned count = ranks[getOffsetOperandGroupPosition()];
-        for (unsigned idx = 0; idx < count; ++idx) {
-          if (isDynamicSize(idx))
-            res.push_back($_op.sizes()[numDynamic++]);
-          else
-            res.push_back($_op.static_sizes()[idx]);
-        }
-        return res;
-      }]
-    >,
-    InterfaceMethod<
-      /*desc=*/[{
         Return the dynamic stride operands.
       }],
       /*retTy=*/"OperandRange",
       /*methodName=*/"strides",
       /*args=*/(ins),
       /*methodBody=*/"",
       /*defaultImplementation=*/[{
         return $_op.strides();
InterfaceMethod<
       /*retTy=*/"ArrayAttr",
       /*methodName=*/"static_strides",
       /*args=*/(ins),
       /*methodBody=*/"",
       /*defaultImplementation=*/[{
         return $_op.static_strides();
       }]
     >,
+    InterfaceMethod<
+      /*desc=*/[{
+        Return a vector of all the static or dynamic sizes of the op.
+      }],
+      /*retTy=*/"SmallVector<OpFoldResult, 4>",
+      /*methodName=*/"getMixedOffsets",
+      /*args=*/(ins),
+      /*methodBody=*/"",
+      /*defaultImplementation=*/[{
+        SmallVector<OpFoldResult, 4> res;
+        std::array<unsigned, 3> ranks = $_op.getArrayAttrRanks();
+        unsigned numDynamic = 0;
+        unsigned count = ranks[getOffsetOperandGroupPosition()];
+        for (unsigned idx = 0; idx < count; ++idx) {
+          if (isDynamicOffset(idx))
+            res.push_back($_op.offsets()[numDynamic++]);
+          else
+            res.push_back($_op.static_offsets()[idx]);
+        }
+        return res;
+      }]
+    >,
+    InterfaceMethod<
+      /*desc=*/[{
+        Return a vector of all the static or dynamic sizes of the op.
+      }],
+      /*retTy=*/"SmallVector<OpFoldResult, 4>",
+      /*methodName=*/"getMixedSizes",
+      /*args=*/(ins),
+      /*methodBody=*/"",
+      /*defaultImplementation=*/[{
+        SmallVector<OpFoldResult, 4> res;
+        std::array<unsigned, 3> ranks = $_op.getArrayAttrRanks();
+        unsigned numDynamic = 0;
+        unsigned count = ranks[getSizeOperandGroupPosition()];
+        for (unsigned idx = 0; idx < count; ++idx) {
+          if (isDynamicSize(idx))
+            res.push_back($_op.sizes()[numDynamic++]);
+          else
+            res.push_back($_op.static_sizes()[idx]);
+        }
+        return res;
+      }]
+    >,
+    InterfaceMethod<
+      /*desc=*/[{
+        Return a vector of all the static or dynamic strides of the op.
+      }],
+      /*retTy=*/"SmallVector<OpFoldResult, 4>",
+      /*methodName=*/"getMixedStrides",
+      /*args=*/(ins),
+      /*methodBody=*/"",
+      /*defaultImplementation=*/[{
+        SmallVector<OpFoldResult, 4> res;
+        std::array<unsigned, 3> ranks = $_op.getArrayAttrRanks();
+        unsigned numDynamic = 0;
+        unsigned count = ranks[getStrideOperandGroupPosition()];
+        for (unsigned idx = 0; idx < count; ++idx) {
+          if (isDynamicStride(idx))
+            res.push_back($_op.strides()[numDynamic++]);
+          else
+            res.push_back($_op.static_strides()[idx]);
+        }
+        return res;
+      }]
+    >,
 
     InterfaceMethod<
       /*desc=*/[{
         Return true if the offset `idx` is dynamic.
       }],
       /*retTy=*/"bool",
       /*methodName=*/"isDynamicOffset",
       /*args=*/(ins "unsigned":$idx),

mlir/lib/Dialect/Linalg/Transforms/Bufferize.cpp

matchAndRewrite(SubTensorOp op, ArrayRef<Value> operands,
     MemRefType subviewMemRefType =
         getTypeConverter()->convertType(op.getType()).cast<MemRefType>();
     // op.sizes() capture exactly the dynamic alloc operands matching the
     // subviewMemRefType thanks to subview/subtensor canonicalization and
     // verification.
     Value alloc =
         rewriter.create<AllocOp>(op.getLoc(), subviewMemRefType, op.sizes());
     Value subView = rewriter.create<SubViewOp>(
-        op.getLoc(), sourceMemref, extractFromI64ArrayAttr(op.static_offsets()),
-        extractFromI64ArrayAttr(op.static_sizes()),
-        extractFromI64ArrayAttr(op.static_strides()), op.offsets(), op.sizes(),
-        op.strides());
+        op.getLoc(), sourceMemref, op.getMixedOffsets(), op.getMixedSizes(),
+        op.getMixedStrides());
     rewriter.create<linalg::CopyOp>(op.getLoc(), subView, alloc);
     rewriter.replaceOp(op, alloc);
     return success();
   }
 };
 
 /// Convert `subtensor_insert %source into %dest [offsets][sizes][strides] ->
 /// %t` to an tensor_to_memref + subview + copy + tensor_load pattern.
matchAndRewrite(SubTensorInsertOp op, ArrayRef<Value> operands,
     // For now, be conservative and copy the converted input memref.
     // In general, the converted input memref here could be aliased or could
     // point into constant memory, so mutating it would lead to miscompilations.
     Value destMemRef = cloneMemref(op.getLoc(), adaptor.dest(), rewriter);
     assert(destMemRef.getType().isa<MemRefType>());
 
     // Take a subview to copy the small memref.
     Value subview = rewriter.create<SubViewOp>(
-        op.getLoc(), destMemRef, extractFromI64ArrayAttr(op.static_offsets()),
-        extractFromI64ArrayAttr(op.static_sizes()),
-        extractFromI64ArrayAttr(op.static_strides()), adaptor.offsets(),
-        adaptor.sizes(), adaptor.strides());
+        op.getLoc(), destMemRef, op.getMixedOffsets(), op.getMixedSizes(),
+        op.getMixedStrides());
     // Copy the small memref.
     rewriter.create<linalg::CopyOp>(op.getLoc(), sourceMemRef, subview);
     rewriter.replaceOp(op, destMemRef);
     return success();
   }
 };
 } // namespace
 

mlir/lib/Dialect/Linalg/Transforms/Fusion.cpp

 ///
 /// More advanced use cases, analyses as well as profitability heuristics are
 /// left for future work.
 
 // Fill `offset`, `sizes` and `strides` used to iterate over the shape indexed
 // by `permutationMap`.
 static void inferShapeComponents(AffineMap permutationMap,
                                  ArrayRef<Range> loopRanges,
-                                 SmallVectorImpl<Value> &offsets,
-                                 SmallVectorImpl<Value> &sizes,
-                                 SmallVectorImpl<Value> &strides) {
+                                 SmallVectorImpl<OpFoldResult> &offsets,
+                                 SmallVectorImpl<OpFoldResult> &sizes,
+                                 SmallVectorImpl<OpFoldResult> &strides) {
   assert(permutationMap.isProjectedPermutation() &&
          "expected some subset of a permutation map");
   SmallVector<Range, 4> shapeRanges(permutationMap.getNumResults());
   unsigned idx = 0;
   for (AffineExpr e : permutationMap.getResults()) {
     // loopToOperandRangesMaps are permutations-only, just swap indices.
     unsigned loopPos = e.cast<AffineDimExpr>().getPosition();
     shapeRanges[idx++] = loopRanges[loopPos];
static LinalgOp cloneWithLoopRanges(OpBuilder &b, Location loc, LinalgOp op,
 
   // Iterate over the shape operands in order.
   // Extract the subranges from the linearized ranges.
   for (auto en : llvm::enumerate(op.getShapedOperands())) {
     unsigned shapedOperandIdx = en.index();
     AffineMap map = op.getIndexingMap(shapedOperandIdx);
     LLVM_DEBUG(llvm::dbgs() << "shapedOperandIdx: " << shapedOperandIdx
                             << " with indexingMap: " << map << "\n");
-    SmallVector<Value, 4> offsets, sizes, strides;
+    SmallVector<OpFoldResult, 4> offsets, sizes, strides;
     inferShapeComponents(map, loopRanges, offsets, sizes, strides);
     Value shape = en.value();
     Value sub = shape.getType().isa<MemRefType>()
                     ? b.create<SubViewOp>(loc, shape, offsets, sizes, strides)
                           .getResult()
                     : b.create<SubTensorOp>(loc, shape, offsets, sizes, strides)
                           .getResult();
     clonedShapes.push_back(sub);

mlir/lib/Dialect/Linalg/Transforms/Promotion.cpp

 // boundary tiles. For now this is done with an unconditional `fill` op followed
 // by a partial `copy` op.
 Optional<PromotionInfo> mlir::linalg::promoteSubviewAsNewBuffer(
     OpBuilder &b, Location loc, SubViewOp subView,
     AllocBufferCallbackFn allocationFn, OperationFolder *folder) {
   ScopedContext scopedContext(b, loc);
   auto viewType = subView.getType();
   auto rank = viewType.getRank();
-  SmallVector<Value, 4> fullSizes, partialSizes;
+  SmallVector<Value, 4> fullSizes;
+  SmallVector<OpFoldResult> partialSizes;
   fullSizes.reserve(rank);
   partialSizes.reserve(rank);
   for (auto en : llvm::enumerate(subView.getOrCreateRanges(b, loc))) {
     auto rangeValue = en.value();
     // Try to extract a tight constant.
     LLVM_DEBUG(llvm::dbgs() << "Extract tightest: " << rangeValue.size << "\n");
     IntegerAttr sizeAttr = getSmallestBoundingIndex(rangeValue.size);
     Value size =
         (!sizeAttr) ? rangeValue.size : b.create<ConstantOp>(loc, sizeAttr);
     LLVM_DEBUG(llvm::dbgs() << "Extracted tightest: " << size << "\n");
     fullSizes.push_back(size);
-    partialSizes.push_back(folded_std_dim(folder, subView, en.index()));
+    partialSizes.push_back(folded_std_dim(folder, subView, en.index()).value);
   }
   SmallVector<int64_t, 4> dynSizes(fullSizes.size(), -1);
   // If a callback is not specified, then use the default implementation for
   // allocating the promoted buffer.
   Optional<Value> fullLocalView = allocationFn(b, subView, fullSizes, folder);
   if (!fullLocalView)
     return {};
-  auto zero = folded_std_constant_index(folder, 0);
-  auto one = folded_std_constant_index(folder, 1);
-  SmallVector<Value, 4> zeros(fullSizes.size(), zero);
-  SmallVector<Value, 4> ones(fullSizes.size(), one);
+  SmallVector<OpFoldResult, 4> zeros(fullSizes.size(), b.getIndexAttr(0));
+  SmallVector<OpFoldResult, 4> ones(fullSizes.size(), b.getIndexAttr(1));
   auto partialLocalView =
       folded_std_subview(folder, *fullLocalView, zeros, partialSizes, ones);
   return PromotionInfo{*fullLocalView, partialLocalView};
 }
 
 static Optional<MapVector<unsigned, PromotionInfo>>
 promoteSubViews(OpBuilder &b, Location loc,
                 LinalgOpInstancePromotionOptions options,

mlir/lib/Dialect/Linalg/Transforms/Tiling.cpp

for (auto en : llvm::enumerate(tiledOperands)) {
     AffineMap map = linalgOp.getIndexingMap(en.index());
     // If the shape is not tiled, we can use it as is.
     if (!isTiled(map, tileSizes)) {
       res.push_back(shapedOp);
       continue;
     }
 
     // Construct a new subview / subtensor for the tile.
-    SmallVector<Value, 4> offsets, sizes, strides;
+    SmallVector<OpFoldResult, 4> offsets, sizes, strides;
     offsets.reserve(rank);
     sizes.reserve(rank);
     strides.reserve(rank);
     for (unsigned r = 0; r < rank; ++r) {
       if (!isTiled(map.getSubMap({r}), tileSizes)) {
-        offsets.push_back(std_constant_index(0));
-        sizes.push_back(std_dim(shapedOp, r));
-        strides.push_back(std_constant_index(1));
+        offsets.push_back(b.getIndexAttr(0));
+        sizes.push_back(std_dim(shapedOp, r).value);
+        strides.push_back(b.getIndexAttr(1));
         continue;
       }
 
       // Tiling creates a new slice at the proper index, the slice step is 1
       // (i.e. the op does not subsample, stepping occurs in the loop).
       auto m = map.getSubMap({r});
       auto offset = applyMapToValues(b, loc, m, lbs).front();
       offsets.push_back(offset);
for (unsigned r = 0; r < rank; ++r) {
             b.getContext());
         auto d = std_dim(shapedOp, r);
         SmallVector<Value, 4> operands{size, d, offset};
         fullyComposeAffineMapAndOperands(&minMap, &operands);
         size = affine_min(b.getIndexType(), minMap, operands);
       }
 
       sizes.push_back(size);
-      strides.push_back(std_constant_index(1));
+      strides.push_back(b.getIndexAttr(1));
     }
 
     if (shapedType.isa<MemRefType>())
       res.push_back(
           b.create<SubViewOp>(loc, shapedOp, offsets, sizes, strides));
     else
       res.push_back(
           b.create<SubTensorOp>(loc, shapedOp, offsets, sizes, strides));

mlir/lib/Dialect/Linalg/Transforms/Transforms.cpp

static LogicalResult rewriteAsPaddedOp(PatternRewriter &rewriter,
   linalg::LinalgOp paddedOp =
       opToPad.clone(rewriter, loc, resultTensorTypes, operands);
 
   // Recover the subtensor out of the new static results. This keeps the
   // original linalg op around because it uses the dims of the original results.
   // This later folds away.
   SmallVector<Value> paddedSubviewResults;
   paddedSubviewResults.reserve(opToPad->getNumResults());
-  Value zero = rewriter.create<ConstantIndexOp>(loc, 0);
-  Value one = rewriter.create<ConstantIndexOp>(loc, 1);
   llvm::SetVector<Operation *> newUsersOfOpToPad;
   for (auto it : llvm::zip(opToPad->getResults(), paddedOp->getResults())) {
     auto rank = std::get<0>(it).getType().cast<RankedTensorType>().getRank();
-    SmallVector<Value> offsets(rank, zero);
-    auto sizes = llvm::to_vector<4>(
-        llvm::map_range(llvm::seq<unsigned>(0, rank), [&](unsigned d) -> Value {
+    SmallVector<OpFoldResult> offsets(rank, rewriter.getIndexAttr(0));
+    auto sizes = llvm::to_vector<4>(llvm::map_range(
+        llvm::seq<unsigned>(0, rank), [&](unsigned d) -> OpFoldResult {
           auto dimOp = rewriter.create<DimOp>(loc, std::get<0>(it), d);
           newUsersOfOpToPad.insert(dimOp);
-          return dimOp;
+          return dimOp.getResult();
         }));
-    SmallVector<Value> strides(rank, one);
+    SmallVector<OpFoldResult> strides(rank, rewriter.getIndexAttr(1));
     paddedSubviewResults.push_back(rewriter.create<SubTensorOp>(
         loc, std::get<1>(it), offsets, sizes, strides));
   }
   // Replace the transient `opToPad` locally, except for uses that we just
   // created for the purpose of extracting the dims.
   rewriter.replaceOpWithIf(opToPad, paddedSubviewResults, [&](OpOperand &opOp) {
     return !newUsersOfOpToPad.contains(opOp.getOwner());
   });

mlir/lib/Dialect/StandardOps/IR/Ops.cpp

 #include "llvm/Support/FormatVariadic.h"
 #include "llvm/Support/raw_ostream.h"
 
 // Pull in all enum type definitions and utility function declarations.
 #include "mlir/Dialect/StandardOps/IR/OpsEnums.cpp.inc"
 
 using namespace mlir;
 
+/// Helper function to dispatch an OpFoldResult into either the `dynamicVec` if
+/// it is a Value or into `staticVec` if it is an IntegerAttr.
+/// In the case of a Value, a copy of the `sentinel` value is also pushed to
+/// `staticVec`. This is useful to extract mixed static and dynamic entries that
+/// come from an AttrSizedOperandSegments trait.
+static void dispatchIndexOpFoldResult(OpFoldResult ofr,
+                                      SmallVectorImpl<Value> &dynamicVec,
+                                      SmallVectorImpl<int64_t> &staticVec,
+                                      int64_t sentinel) {
+  if (auto v = ofr.dyn_cast<Value>()) {
+    dynamicVec.push_back(v);
+    staticVec.push_back(sentinel);
+    return;
+  }
+  APInt apInt = ofr.dyn_cast<Attribute>().cast<IntegerAttr>().getValue();
+  staticVec.push_back(apInt.getSExtValue());
+}
+
+static void dispatchIndexOpFoldResults(ArrayRef<OpFoldResult> ofrs,
+                                       SmallVectorImpl<Value> &dynamicVec,
+                                       SmallVectorImpl<int64_t> &staticVec,
+                                       int64_t sentinel) {
+  for (auto ofr : ofrs)
+    dispatchIndexOpFoldResult(ofr, dynamicVec, staticVec, sentinel);
+}
+
 //===----------------------------------------------------------------------===//
 // StandardOpsDialect Interfaces
 //===----------------------------------------------------------------------===//
 namespace {
 /// This class defines the interface for handling inlining with standard
 /// operations.
 struct StdInlinerInterface : public DialectInlinerInterface {
   using DialectInlinerInterface::DialectInlinerInterface;

 OpFoldResult MemRefCastOp::fold(ArrayRef<Attribute> operands) {
   return succeeded(foldMemRefCast(*this)) ? getResult() : Value();
 }
 
 //===----------------------------------------------------------------------===//
 // MemRefReinterpretCastOp
 //===----------------------------------------------------------------------===//
 
-void mlir::MemRefReinterpretCastOp::build(
-    OpBuilder &b, OperationState &result, MemRefType resultType, Value source,
-    int64_t staticOffset, ArrayRef<int64_t> staticSizes,
-    ArrayRef<int64_t> staticStrides, ValueRange offset, ValueRange sizes,
-    ValueRange strides, ArrayRef<NamedAttribute> attrs) {
-  build(b, result, resultType, source, offset, sizes, strides,
-        b.getI64ArrayAttr(staticOffset), b.getI64ArrayAttr(staticSizes),
-        b.getI64ArrayAttr(staticStrides));
-  result.addAttributes(attrs);
-}
-
 /// Build a MemRefReinterpretCastOp with all dynamic entries: `staticOffsets`,
 /// `staticSizes` and `staticStrides` are automatically filled with
 /// source-memref-rank sentinel values that encode dynamic entries.
 void mlir::MemRefReinterpretCastOp::build(OpBuilder &b, OperationState &result,
                                           MemRefType resultType, Value source,
+                                          OpFoldResult offset,
+                                          ArrayRef<OpFoldResult> sizes,
+                                          ArrayRef<OpFoldResult> strides,
+                                          ArrayRef<NamedAttribute> attrs) {
+  SmallVector<int64_t> staticOffsets, staticSizes, staticStrides;
+  SmallVector<Value> dynamicOffsets, dynamicSizes, dynamicStrides;
+  dispatchIndexOpFoldResults(offset, dynamicOffsets, staticOffsets,
+                             ShapedType::kDynamicStrideOrOffset);
+  dispatchIndexOpFoldResults(sizes, dynamicSizes, staticSizes,
+                             ShapedType::kDynamicSize);
+  dispatchIndexOpFoldResults(strides, dynamicStrides, staticStrides,
+                             ShapedType::kDynamicStrideOrOffset);
+  build(b, result, resultType, source, dynamicOffsets, dynamicSizes,
+        dynamicStrides, b.getI64ArrayAttr(staticOffsets),
+        b.getI64ArrayAttr(staticSizes), b.getI64ArrayAttr(staticStrides));
+  result.addAttributes(attrs);
+}
+
+void mlir::MemRefReinterpretCastOp::build(OpBuilder &b, OperationState &result,
+                                          MemRefType resultType, Value source,
+                                          int64_t offset,
+                                          ArrayRef<int64_t> sizes,
+                                          ArrayRef<int64_t> strides,
+                                          ArrayRef<NamedAttribute> attrs) {
+  SmallVector<OpFoldResult> sizeValues =
+      llvm::to_vector<4>(llvm::map_range(sizes, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  build(b, result, resultType, source, b.getI64IntegerAttr(offset), sizeValues,
+        strideValues, attrs);
+}
+
+void mlir::MemRefReinterpretCastOp::build(OpBuilder &b, OperationState &result,
+                                          MemRefType resultType, Value source,
                                           Value offset, ValueRange sizes,
                                           ValueRange strides,
                                           ArrayRef<NamedAttribute> attrs) {
-  unsigned rank = resultType.getRank();
-  SmallVector<int64_t, 4> staticSizesVector(rank, ShapedType::kDynamicSize);
-  SmallVector<int64_t, 4> staticStridesVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  build(b, result, resultType, source,
-        /*staticOffset=*/ShapedType::kDynamicStrideOrOffset, staticSizesVector,
-        staticStridesVector, offset, sizes, strides, attrs);
+  SmallVector<OpFoldResult> sizeValues = llvm::to_vector<4>(
+      llvm::map_range(sizes, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [](Value v) -> OpFoldResult { return v; }));
+  build(b, result, resultType, source, offset, sizeValues, strideValues, attrs);
 }
 
 /// Print a memref_reinterpret_cast op of the form:
 /// ```
 ///   `memref_reinterpret_cast` ssa-name to
 ///       offset: `[` offset `]`
 ///       sizes: `[` size-list `]`
 ///       strides:`[` stride-list `]`
 ///   `:` any-memref-type to strided-memref-type
static ParseResult parseSubViewOp(OpAsmParser &parser, OperationState &result) {
 
   if (failed(parseOffsetsSizesAndStrides(parser, result,
                                          /*segmentSizes=*/{1}, // source memref
                                          preResolutionFn)))
     return failure();
   return parser.addTypeToList(dstType, result.types);
 }
 
-void mlir::SubViewOp::build(OpBuilder &b, OperationState &result, Value source,
-                            ArrayRef<int64_t> staticOffsets,
-                            ArrayRef<int64_t> staticSizes,
-                            ArrayRef<int64_t> staticStrides, ValueRange offsets,
-                            ValueRange sizes, ValueRange strides,
+// Build a SubViewOp with mixed static and dynamic entries and custom result
+// type. If the type passed is nullptr, it is inferred.
+void mlir::SubViewOp::build(OpBuilder &b, OperationState &result,
+                            MemRefType resultType, Value source,
+                            ArrayRef<OpFoldResult> offsets,
+                            ArrayRef<OpFoldResult> sizes,
+                            ArrayRef<OpFoldResult> strides,
                             ArrayRef<NamedAttribute> attrs) {
+  SmallVector<int64_t> staticOffsets, staticSizes, staticStrides;
+  SmallVector<Value> dynamicOffsets, dynamicSizes, dynamicStrides;
+  dispatchIndexOpFoldResults(offsets, dynamicOffsets, staticOffsets,
+                             ShapedType::kDynamicStrideOrOffset);
+  dispatchIndexOpFoldResults(sizes, dynamicSizes, staticSizes,
+                             ShapedType::kDynamicSize);
+  dispatchIndexOpFoldResults(strides, dynamicStrides, staticStrides,
+                             ShapedType::kDynamicStrideOrOffset);
   auto sourceMemRefType = source.getType().cast<MemRefType>();
-  auto resultType = inferResultType(sourceMemRefType, staticOffsets,
-                                    staticSizes, staticStrides);
-  build(b, result, resultType, source, offsets, sizes, strides,
-        b.getI64ArrayAttr(staticOffsets), b.getI64ArrayAttr(staticSizes),
-        b.getI64ArrayAttr(staticStrides));
+  // Structuring implementation this way avoids duplication between builders.
+  if (!resultType) {
+    resultType = SubViewOp::inferResultType(sourceMemRefType, staticOffsets,
+                                            staticSizes, staticStrides)
+                     .cast<MemRefType>();
+  }
+  build(b, result, resultType, source, dynamicOffsets, dynamicSizes,
+        dynamicStrides, b.getI64ArrayAttr(staticOffsets),
+        b.getI64ArrayAttr(staticSizes), b.getI64ArrayAttr(staticStrides));
   result.addAttributes(attrs);
 }
 
-/// Build a SubViewOp with all dynamic entries: `staticOffsets`, `staticSizes`
-/// and `staticStrides` are automatically filled with source-memref-rank
-/// sentinel values that encode dynamic entries.
+// Build a SubViewOp with mixed static and dynamic entries and inferred result
+// type.
 void mlir::SubViewOp::build(OpBuilder &b, OperationState &result, Value source,
-                            ValueRange offsets, ValueRange sizes,
-                            ValueRange strides,
+                            ArrayRef<OpFoldResult> offsets,
+                            ArrayRef<OpFoldResult> sizes,
+                            ArrayRef<OpFoldResult> strides,
                             ArrayRef<NamedAttribute> attrs) {
-  auto sourceMemRefType = source.getType().cast<MemRefType>();
-  unsigned rank = sourceMemRefType.getRank();
-  SmallVector<int64_t, 4> staticOffsetsVector;
-  staticOffsetsVector.assign(rank, ShapedType::kDynamicStrideOrOffset);
-  SmallVector<int64_t, 4> staticSizesVector;
-  staticSizesVector.assign(rank, ShapedType::kDynamicSize);
-  SmallVector<int64_t, 4> staticStridesVector;
-  staticStridesVector.assign(rank, ShapedType::kDynamicStrideOrOffset);
-  build(b, result, source, staticOffsetsVector, staticSizesVector,
-        staticStridesVector, offsets, sizes, strides, attrs);
+  build(b, result, MemRefType(), source, offsets, sizes, strides, attrs);
 }
 
-/// Build a SubViewOp as above but with custom result type.
+// Build a SubViewOp with static entries and inferred result type.
+void mlir::SubViewOp::build(OpBuilder &b, OperationState &result, Value source,
+                            ArrayRef<int64_t> offsets, ArrayRef<int64_t> sizes,
+                            ArrayRef<int64_t> strides,
+                            ArrayRef<NamedAttribute> attrs) {
+  SmallVector<OpFoldResult> offsetValues = llvm::to_vector<4>(
+      llvm::map_range(offsets, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  SmallVector<OpFoldResult> sizeValues =
+      llvm::to_vector<4>(llvm::map_range(sizes, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  build(b, result, source, offsetValues, sizeValues, strideValues, attrs);
+}
+
+// Build a SubViewOp with dynamic entries and custom result type. If the
+// type passed is nullptr, it is inferred.
 void mlir::SubViewOp::build(OpBuilder &b, OperationState &result,
                             MemRefType resultType, Value source,
-                            ArrayRef<int64_t> staticOffsets,
-                            ArrayRef<int64_t> staticSizes,
-                            ArrayRef<int64_t> staticStrides, ValueRange offsets,
-                            ValueRange sizes, ValueRange strides,
+                            ArrayRef<int64_t> offsets, ArrayRef<int64_t> sizes,
+                            ArrayRef<int64_t> strides,
                             ArrayRef<NamedAttribute> attrs) {
-  build(b, result, resultType, source, offsets, sizes, strides,
-        b.getI64ArrayAttr(staticOffsets), b.getI64ArrayAttr(staticSizes),
-        b.getI64ArrayAttr(staticStrides));
-  result.addAttributes(attrs);
+  SmallVector<OpFoldResult> offsetValues = llvm::to_vector<4>(
+      llvm::map_range(offsets, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  SmallVector<OpFoldResult> sizeValues =
+      llvm::to_vector<4>(llvm::map_range(sizes, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [&](int64_t v) -> OpFoldResult {
+        return b.getI64IntegerAttr(v);
+      }));
+  build(b, result, resultType, source, offsetValues, sizeValues, strideValues,
+        attrs);
 }
 
-/// Build a SubViewOp as above but with custom result type.
+// Build a SubViewOp with dynamic entries and custom result type. If the type
+// passed is nullptr, it is inferred.
 void mlir::SubViewOp::build(OpBuilder &b, OperationState &result,
                             MemRefType resultType, Value source,
                             ValueRange offsets, ValueRange sizes,
                             ValueRange strides,
                             ArrayRef<NamedAttribute> attrs) {
-  auto sourceMemRefType = source.getType().cast<MemRefType>();
-  unsigned rank = sourceMemRefType.getRank();
-  SmallVector<int64_t, 4> staticOffsetsVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  SmallVector<int64_t, 4> staticSizesVector(rank, ShapedType::kDynamicSize);
-  SmallVector<int64_t, 4> staticStridesVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  build(b, result, resultType, source, staticOffsetsVector, staticSizesVector,
-        staticStridesVector, offsets, sizes, strides, attrs);
+  SmallVector<OpFoldResult> offsetValues = llvm::to_vector<4>(
+      llvm::map_range(offsets, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> sizeValues = llvm::to_vector<4>(
+      llvm::map_range(sizes, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [](Value v) -> OpFoldResult { return v; }));
+  build(b, result, resultType, source, offsetValues, sizeValues, strideValues);
+}
+
+// Build a SubViewOp with dynamic entries and inferred result type.
+void mlir::SubViewOp::build(OpBuilder &b, OperationState &result, Value source,
+                            ValueRange offsets, ValueRange sizes,
+                            ValueRange strides,
+                            ArrayRef<NamedAttribute> attrs) {
+  build(b, result, MemRefType(), source, offsets, sizes, strides, attrs);
 }
 
 /// For ViewLikeOpInterface.
 Value SubViewOp::getViewSource() { return source(); }
 
 llvm::Optional<SmallVector<bool, 4>>
 mlir::computeRankReductionMask(ArrayRef<int64_t> originalShape,
                                ArrayRef<int64_t> reducedShape) {
Value stride =
             : b.create<ConstantIndexOp>(loc, op.getStaticStride(idx));
     res.emplace_back(Range{offset, size, stride});
   }
   return res;
 }
 
 namespace {
 
-/// Take a list of `values` with potential new constant to extract and a list
-/// of `constantValues` with`values.size()` sentinel that evaluate to true by
-/// applying `isDynamic`.
 /// Detects the `values` produced by a ConstantIndexOp and places the new
 /// constant in place of the corresponding sentinel value.
-void canonicalizeSubViewPart(SmallVectorImpl<Value> &values,
-                             SmallVectorImpl<int64_t> &constantValues,
+void canonicalizeSubViewPart(SmallVectorImpl<OpFoldResult> &values,
                              llvm::function_ref<bool(int64_t)> isDynamic) {
-  bool hasNewStaticValue = llvm::any_of(
-      values, [](Value val) { return matchPattern(val, m_ConstantIndex()); });
-  if (hasNewStaticValue) {
-    for (unsigned cstIdx = 0, valIdx = 0, e = constantValues.size();
-         cstIdx != e; ++cstIdx) {
-      // Was already static, skip.
-      if (!isDynamic(constantValues[cstIdx]))
-        continue;
-      // Newly static, move from Value to constant.
-      if (matchPattern(values[valIdx], m_ConstantIndex())) {
-        constantValues[cstIdx] =
-            cast<ConstantIndexOp>(values[valIdx].getDefiningOp()).getValue();
-        // Erase for impl. simplicity. Reverse iterator if we really must.
-        values.erase(std::next(values.begin(), valIdx));
-        continue;
-      }
-      // Remains dynamic move to next value.
-      ++valIdx;
-    }
+  for (OpFoldResult &ofr : values) {
+    if (ofr.is<Attribute>())
+    if (auto cstOp = ofr.dyn_cast<Value>().getDefiningOp<ConstantIndexOp>())
+      ofr = OpBuilder(cstOp).getIndexAttr(cstOp.getValue());
   }
 }
 
 static void replaceWithNewOp(PatternRewriter &rewriter, SubViewOp op,
                              SubViewOp newOp) {
   rewriter.replaceOpWithNewOp<MemRefCastOp>(op, newOp, op.getType());
 }
 
LogicalResult matchAndRewrite(OpType op,
     // No constant operand, just return;
     if (llvm::none_of(op.getOperands(), [](Value operand) {
           return matchPattern(operand, m_ConstantIndex());
         }))
       return failure();
 
     // At least one of offsets/sizes/strides is a new constant.
     // Form the new list of operands and constant attributes from the existing.
-    SmallVector<Value, 8> newOffsets(op.offsets());
-    SmallVector<int64_t, 8> newStaticOffsets =
-        extractFromI64ArrayAttr(op.static_offsets());
-    std::array<unsigned, 3> ranks = op.getArrayAttrRanks();
-    (void)ranks;
-    assert(newStaticOffsets.size() == ranks[0]);
-    canonicalizeSubViewPart(newOffsets, newStaticOffsets,
-                            ShapedType::isDynamicStrideOrOffset);
-
-    SmallVector<Value, 8> newSizes(op.sizes());
-    SmallVector<int64_t, 8> newStaticSizes =
-        extractFromI64ArrayAttr(op.static_sizes());
-    assert(newStaticSizes.size() == ranks[1]);
-    canonicalizeSubViewPart(newSizes, newStaticSizes, ShapedType::isDynamic);
-
-    SmallVector<Value, 8> newStrides(op.strides());
-    SmallVector<int64_t, 8> newStaticStrides =
-        extractFromI64ArrayAttr(op.static_strides());
-    assert(newStaticStrides.size() == ranks[2]);
-    canonicalizeSubViewPart(newStrides, newStaticStrides,
-                            ShapedType::isDynamicStrideOrOffset);
+    SmallVector<OpFoldResult> mixedOffsets(op.getMixedOffsets());
+    SmallVector<OpFoldResult> mixedSizes(op.getMixedSizes());
+    SmallVector<OpFoldResult> mixedStrides(op.getMixedStrides());
+    canonicalizeSubViewPart(mixedOffsets, ShapedType::isDynamicStrideOrOffset);
+    canonicalizeSubViewPart(mixedSizes, ShapedType::isDynamic);
+    canonicalizeSubViewPart(mixedStrides, ShapedType::isDynamicStrideOrOffset);
 
     // Create the new op in canonical form.
-    auto newOp = rewriter.create<OpType>(
-        op.getLoc(), op.source(), newStaticOffsets, newStaticSizes,
-        newStaticStrides, newOffsets, newSizes, newStrides);
+    auto newOp = rewriter.create<OpType>(op.getLoc(), op.source(), mixedOffsets,
+                                         mixedSizes, mixedStrides);
 
     replaceWithNewOp(rewriter, op, newOp);
 
     return success();
   }
 };
 
 } // end anonymous namespace
assert(staticOffsets.size() == rank &&
          "unexpected staticOffsets size mismatch");
   assert(staticSizes.size() == rank && "unexpected staticSizes size mismatch");
   assert(staticStrides.size() == rank &&
          "unexpected staticStrides size mismatch");
   return RankedTensorType::get(staticSizes,
                                sourceRankedTensorType.getElementType());
 }
 
-void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
-                              Value source, ArrayRef<int64_t> staticOffsets,
-                              ArrayRef<int64_t> staticSizes,
-                              ArrayRef<int64_t> staticStrides,
-                              ValueRange offsets, ValueRange sizes,
-                              ValueRange strides,
-                              ArrayRef<NamedAttribute> attrs) {
-  auto sourceRankedTensorType = source.getType().cast<RankedTensorType>();
-  auto resultType = inferResultType(sourceRankedTensorType, staticOffsets,
-                                    staticSizes, staticStrides);
-  build(b, result, resultType, source, offsets, sizes, strides,
-        b.getI64ArrayAttr(staticOffsets), b.getI64ArrayAttr(staticSizes),
-        b.getI64ArrayAttr(staticStrides));
-  result.addAttributes(attrs);
-}
-
-/// Build a SubTensorOp with all dynamic entries: `staticOffsets`, `staticSizes`
-/// and `staticStrides` are automatically filled with sentinel values that
-/// encode dynamic entries.
-void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
-                              Value source, ValueRange offsets,
-                              ValueRange sizes, ValueRange strides,
-                              ArrayRef<NamedAttribute> attrs) {
-  auto sourceRankedTensorType = source.getType().cast<RankedTensorType>();
-  unsigned rank = sourceRankedTensorType.getRank();
-  SmallVector<int64_t, 4> staticOffsetsVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  SmallVector<int64_t, 4> staticSizesVector(rank, ShapedType::kDynamicSize);
-  SmallVector<int64_t, 4> staticStridesVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  build(b, result, source, staticOffsetsVector, staticSizesVector,
-        staticStridesVector, offsets, sizes, strides, attrs);
-}
-
-/// Build a SubTensorOp as above but with custom result type.
-void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
-                              RankedTensorType resultType, Value source,
-                              ArrayRef<int64_t> staticOffsets,
-                              ArrayRef<int64_t> staticSizes,
-                              ArrayRef<int64_t> staticStrides,
-                              ValueRange offsets, ValueRange sizes,
-                              ValueRange strides,
-                              ArrayRef<NamedAttribute> attrs) {
-  build(b, result, resultType, source, offsets, sizes, strides,
-        b.getI64ArrayAttr(staticOffsets), b.getI64ArrayAttr(staticSizes),
-        b.getI64ArrayAttr(staticStrides));
-  result.addAttributes(attrs);
-}
-
-/// Build a SubTensorOp as above but with custom result type.
-void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
-                              RankedTensorType resultType, Value source,
-                              ValueRange offsets, ValueRange sizes,
-                              ValueRange strides,
-                              ArrayRef<NamedAttribute> attrs) {
-  auto sourceRankedTensorType = source.getType().cast<RankedTensorType>();
-  unsigned rank = sourceRankedTensorType.getRank();
-  SmallVector<int64_t, 4> staticOffsetsVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  SmallVector<int64_t, 4> staticSizesVector(rank, ShapedType::kDynamicSize);
-  SmallVector<int64_t, 4> staticStridesVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  build(b, result, resultType, source, staticOffsetsVector, staticSizesVector,
-        staticStridesVector, offsets, sizes, strides, attrs);
-}
-
-/// Dispatch `ofr` into either `dynamicVec` if it is a Value or into `staticVec`
-/// otherwise.  In the dynamic case, `sentinel` is appended to `staticVec` to
-/// represent the dynamic value `?`.
-static void unpackOpFoldResult(OpFoldResult ofr,
-                               SmallVectorImpl<Value> &dynamicVec,
-                               SmallVectorImpl<int64_t> &staticVec,
-                               int64_t sentinel) {
-  Value v = ofr.dyn_cast<Value>();
-  if (v) {
-    dynamicVec.push_back(v);
-    staticVec.push_back(sentinel);
-  } else {
-    APInt apInt = ofr.dyn_cast<Attribute>().cast<IntegerAttr>().getValue();
-    staticVec.push_back(apInt.getSExtValue());
-  }
-}
-
-static void unpackOpFoldResults(ArrayRef<OpFoldResult> ofrs,
-                                SmallVector<Value> &dynamicVec,
-                                SmallVector<int64_t> &staticVec,
-                                int64_t sentinel) {
-  for (auto ofr : ofrs)
-    unpackOpFoldResult(ofr, dynamicVec, staticVec, sentinel);
-}
-
 // Build a SubTensorOp with mixed static and dynamic entries and custom result
 // type. If the type passed is nullptr, it is inferred.
 void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
                               RankedTensorType resultType, Value source,
                               ArrayRef<OpFoldResult> offsets,
                               ArrayRef<OpFoldResult> sizes,
                               ArrayRef<OpFoldResult> strides,
                               ArrayRef<NamedAttribute> attrs) {
   SmallVector<int64_t> staticOffsets, staticSizes, staticStrides;
   SmallVector<Value> dynamicOffsets, dynamicSizes, dynamicStrides;
-  unpackOpFoldResults(offsets, dynamicOffsets, staticOffsets,
+  dispatchIndexOpFoldResults(offsets, dynamicOffsets, staticOffsets,
                              ShapedType::kDynamicStrideOrOffset);
-  unpackOpFoldResults(sizes, dynamicSizes, staticSizes,
+  dispatchIndexOpFoldResults(sizes, dynamicSizes, staticSizes,
                              ShapedType::kDynamicSize);
-  unpackOpFoldResults(strides, dynamicStrides, staticStrides,
+  dispatchIndexOpFoldResults(strides, dynamicStrides, staticStrides,
                              ShapedType::kDynamicStrideOrOffset);
   auto sourceRankedTensorType = source.getType().cast<RankedTensorType>();
   // Structuring implementation this way avoids duplication between builders.
   if (!resultType) {
     resultType =
         SubTensorOp::inferResultType(sourceRankedTensorType, staticOffsets,
                                      staticSizes, staticStrides)
             .cast<RankedTensorType>();
   }
-  build(b, result, resultType, source, staticOffsets, staticSizes,
-        staticStrides, dynamicOffsets, dynamicSizes, dynamicStrides, attrs);
+  build(b, result, resultType, source, dynamicOffsets, dynamicSizes,
+        dynamicStrides, b.getI64ArrayAttr(staticOffsets),
+        b.getI64ArrayAttr(staticSizes), b.getI64ArrayAttr(staticStrides));
+  result.addAttributes(attrs);
 }
 
 // Build a SubTensorOp with mixed static and dynamic entries and inferred result
 // type.
 void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
                               Value source, ArrayRef<OpFoldResult> offsets,
                               ArrayRef<OpFoldResult> sizes,
                               ArrayRef<OpFoldResult> strides,
                               ArrayRef<NamedAttribute> attrs) {
   build(b, result, RankedTensorType(), source, offsets, sizes, strides, attrs);
 }
 
+// Build a SubTensorOp with dynamic entries and custom result type. If the type
+// passed is nullptr, it is inferred.
+void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
+                              RankedTensorType resultType, Value source,
+                              ValueRange offsets, ValueRange sizes,
+                              ValueRange strides,
+                              ArrayRef<NamedAttribute> attrs) {
+  SmallVector<OpFoldResult> offsetValues = llvm::to_vector<4>(
+      llvm::map_range(offsets, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> sizeValues = llvm::to_vector<4>(
+      llvm::map_range(sizes, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [](Value v) -> OpFoldResult { return v; }));
+  build(b, result, resultType, source, offsetValues, sizeValues, strideValues);
+}
+
+// Build a SubTensorOp with dynamic entries and inferred result type.
+void mlir::SubTensorOp::build(OpBuilder &b, OperationState &result,
+                              Value source, ValueRange offsets,
+                              ValueRange sizes, ValueRange strides,
+                              ArrayRef<NamedAttribute> attrs) {
+  build(b, result, RankedTensorType(), source, offsets, sizes, strides, attrs);
+}
+
 /// Verifier for SubTensorOp.
 static LogicalResult verify(SubTensorOp op) {
   // Verify result type against inferred type.
   auto expectedType = SubTensorOp::inferResultType(
       op.getSourceType(), extractFromI64ArrayAttr(op.static_offsets()),
       extractFromI64ArrayAttr(op.static_sizes()),
       extractFromI64ArrayAttr(op.static_strides()));
   auto result = isRankReducedType(expectedType, op.getType());
static ParseResult parseSubTensorInsertOp(OpAsmParser &parser,
   if (failed(parseOffsetsSizesAndStrides(
           parser, result,
           /*segmentSizes=*/{1, 1}, // source tensor, destination tensor
           preResolutionFn)))
     return failure();
   return parser.addTypeToList(dstType, result.types);
 }
 
-void mlir::SubTensorInsertOp::build(
-    OpBuilder &b, OperationState &result, Value source, Value dest,
-    ArrayRef<int64_t> staticOffsets, ArrayRef<int64_t> staticSizes,
-    ArrayRef<int64_t> staticStrides, ValueRange offsets, ValueRange sizes,
-    ValueRange strides, ArrayRef<NamedAttribute> attrs) {
-  build(b, result, dest.getType(), source, dest, offsets, sizes, strides,
-        b.getI64ArrayAttr(staticOffsets), b.getI64ArrayAttr(staticSizes),
-        b.getI64ArrayAttr(staticStrides));
-  result.addAttributes(attrs);
-}
-
-/// Build a SubViewOp with all dynamic entries: `staticOffsets`, `staticSizes`
-/// and `staticStrides` are automatically filled with source-memref-rank
-/// sentinel values that encode dynamic entries.
-void mlir::SubTensorInsertOp::build(OpBuilder &b, OperationState &result,
-                                    Value source, Value dest,
-                                    ValueRange offsets, ValueRange sizes,
-                                    ValueRange strides,
-                                    ArrayRef<NamedAttribute> attrs) {
-  auto destRankedTensorType = dest.getType().cast<RankedTensorType>();
-  unsigned rank = destRankedTensorType.getRank();
-  SmallVector<int64_t, 4> staticOffsetsVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  SmallVector<int64_t, 4> staticSizesVector(rank, ShapedType::kDynamicSize);
-  SmallVector<int64_t, 4> staticStridesVector(
-      rank, ShapedType::kDynamicStrideOrOffset);
-  build(b, result, source, dest, staticOffsetsVector, staticSizesVector,
-        staticStridesVector, offsets, sizes, strides, attrs);
-}
-
 // Build a SubTensorInsertOp with mixed static and dynamic entries.
 void mlir::SubTensorInsertOp::build(OpBuilder &b, OperationState &result,
                                     Value source, Value dest,
                                     ArrayRef<OpFoldResult> offsets,
                                     ArrayRef<OpFoldResult> sizes,
                                     ArrayRef<OpFoldResult> strides,
                                     ArrayRef<NamedAttribute> attrs) {
   SmallVector<int64_t> staticOffsets, staticSizes, staticStrides;
   SmallVector<Value> dynamicOffsets, dynamicSizes, dynamicStrides;
-  unpackOpFoldResults(offsets, dynamicOffsets, staticOffsets,
+  dispatchIndexOpFoldResults(offsets, dynamicOffsets, staticOffsets,
                              ShapedType::kDynamicStrideOrOffset);
-  unpackOpFoldResults(sizes, dynamicSizes, staticSizes,
+  dispatchIndexOpFoldResults(sizes, dynamicSizes, staticSizes,
                              ShapedType::kDynamicSize);
-  unpackOpFoldResults(strides, dynamicStrides, staticStrides,
+  dispatchIndexOpFoldResults(strides, dynamicStrides, staticStrides,
                              ShapedType::kDynamicStrideOrOffset);
-  build(b, result, source, dest, staticOffsets, staticSizes, staticStrides,
-        dynamicOffsets, dynamicSizes, dynamicStrides, attrs);
+  build(b, result, dest.getType(), source, dest, dynamicOffsets, dynamicSizes,
+        dynamicStrides, b.getI64ArrayAttr(staticOffsets),
+        b.getI64ArrayAttr(staticSizes), b.getI64ArrayAttr(staticStrides));
+  result.addAttributes(attrs);
+}
+
+// Build a SubTensorInsertOp with dynamic entries.
+void mlir::SubTensorInsertOp::build(OpBuilder &b, OperationState &result,
+                                    Value source, Value dest,
+                                    ValueRange offsets, ValueRange sizes,
+                                    ValueRange strides,
+                                    ArrayRef<NamedAttribute> attrs) {
+  SmallVector<OpFoldResult> offsetValues = llvm::to_vector<4>(
+      llvm::map_range(offsets, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> sizeValues = llvm::to_vector<4>(
+      llvm::map_range(sizes, [](Value v) -> OpFoldResult { return v; }));
+  SmallVector<OpFoldResult> strideValues = llvm::to_vector<4>(
+      llvm::map_range(strides, [](Value v) -> OpFoldResult { return v; }));
+  build(b, result, source, dest, offsetValues, sizeValues, strideValues);
 }
 
 /// Verifier for SubViewOp.
 static LogicalResult verify(SubTensorInsertOp op) {
   if (op.getType() != op.dest().getType())
     return op.emitError("expected result type to be ") << op.dest().getType();
   return success();
 }

mlir/lib/Dialect/StandardOps/Transforms/ExpandOps.cpp

public:
 
   LogicalResult matchAndRewrite(MemRefReshapeOp op,
                                 PatternRewriter &rewriter) const final {
     auto shapeType = op.shape().getType().cast<MemRefType>();
     if (!shapeType.hasStaticShape())
       return failure();
 
     int64_t rank = shapeType.cast<MemRefType>().getDimSize(0);
-    SmallVector<Value, 4> sizes, strides;
+    SmallVector<OpFoldResult, 4> sizes, strides;
     sizes.resize(rank);
     strides.resize(rank);
 
     Location loc = op.getLoc();
     Value stride = rewriter.create<ConstantIndexOp>(loc, 1);
     for (int i = rank - 1; i >= 0; --i) {
       Value index = rewriter.create<ConstantIndexOp>(loc, i);
       Value size = rewriter.create<LoadOp>(loc, op.shape(), index);
       if (!size.getType().isa<IndexType>())
         size = rewriter.create<IndexCastOp>(loc, size, rewriter.getIndexType());
       sizes[i] = size;
       strides[i] = stride;
       if (i > 0)
         stride = rewriter.create<MulIOp>(loc, stride, size);
     }
-    SmallVector<int64_t, 2> staticSizes(rank, ShapedType::kDynamicSize);
-    SmallVector<int64_t, 2> staticStrides(rank,
-                                          ShapedType::kDynamicStrideOrOffset);
     rewriter.replaceOpWithNewOp<MemRefReinterpretCastOp>(
-        op, op.getType(), op.source(), /*staticOffset = */ 0, staticSizes,
-        staticStrides, /*offset=*/llvm::None, sizes, strides);
+        op, op.getType(), op.source(), /*offset=*/rewriter.getIndexAttr(0),
+        sizes, strides);
     return success();
   }
 };
 
 /// Expands SignedCeilDivIOP (n, m) into
 ///   1) x = (m > 0) ? -1 : 1
 ///   2) (n*m>0) ? ((n+x) / m) + 1 : - (-n / m)
 struct SignedCeilDivIOpConverter : public OpRewritePattern<SignedCeilDivIOp> {

mlir/lib/Dialect/Vector/VectorTransforms.cpp

 // TODO: view intersection/union/differences should be a proper std op.
 static Value createScopedSubViewIntersection(VectorTransferOpInterface xferOp,
                                              Value alloc) {
   using namespace edsc::intrinsics;
   int64_t memrefRank = xferOp.getShapedType().getRank();
   // TODO: relax this precondition, will require rank-reducing subviews.
   assert(memrefRank == alloc.getType().cast<MemRefType>().getRank() &&
          "Expected memref rank to match the alloc rank");
-  Value one = std_constant_index(1);
   ValueRange leadingIndices =
       xferOp.indices().take_front(xferOp.getLeadingShapedRank());
-  SmallVector<Value, 4> sizes;
+  SmallVector<OpFoldResult, 4> sizes;
   sizes.append(leadingIndices.begin(), leadingIndices.end());
   xferOp.zipResultAndIndexing([&](int64_t resultIdx, int64_t indicesIdx) {
     using MapList = ArrayRef<ArrayRef<AffineExpr>>;
     Value dimMemRef = std_dim(xferOp.source(), indicesIdx);
     Value dimAlloc = std_dim(alloc, resultIdx);
     Value index = xferOp.indices()[indicesIdx];
     AffineExpr i, j, k;
     bindDims(xferOp.getContext(), i, j, k);
     SmallVector<AffineMap, 4> maps =
         AffineMap::inferFromExprList(MapList{{i - j, k}});
     // affine_min(%dimMemRef - %index, %dimAlloc)
     Value affineMin = affine_min(index.getType(), maps[0],
                                  ValueRange{dimMemRef, index, dimAlloc});
     sizes.push_back(affineMin);
   });
-  return std_sub_view(xferOp.source(), xferOp.indices(), sizes,
-                      SmallVector<Value, 4>(memrefRank, one));
+
+  SmallVector<OpFoldResult, 4> indices = llvm::to_vector<4>(llvm::map_range(
+      xferOp.indices(), [](Value idx) -> OpFoldResult { return idx; }));
+  return std_sub_view(
+      xferOp.source(), indices, sizes,
+      SmallVector<OpFoldResult>(memrefRank, OpBuilder(xferOp).getIndexAttr(1)));
 }
 
 /// Given an `xferOp` for which:
 ///   1. `inBoundsCond` and a `compatibleMemRefType` have been computed.
 ///   2. a memref of single vector `alloc` has been allocated.
 /// Produce IR resembling:
 /// ```
 ///    %1:3 = scf.if (%inBounds) {

mlir/test/Dialect/Linalg/fusion-sequence.mlir

 //  CHECK-SAME:         ) outs(%[[STARG2]] : tensor<?x?xf32>)
 //       CHECK:       %[[T1:.+]] = linalg.matmul
 //  CHECK-SAME:         ins(%[[T0]], %[[STARG3]]
 //  CHECK-SAME:         ) outs(%[[STARG4]] : tensor<?x?xf32>)
 //       CHECK:       %[[T2:.+]] = linalg.matmul
 //  CHECK-SAME:         ins(%[[T1]], %[[ARG5]]
 //  CHECK-SAME:         ) outs(%[[STARG6]] : tensor<?x?xf32>)
 //       CHECK:       %[[R1:.+]] = subtensor_insert %[[T2]]
-//  CHECK-SAME:         into %[[ARG8]][%[[IV0]], %[[C0]]]
+//  CHECK-SAME:         into %[[ARG8]][%[[IV0]], 0]
 //       CHECK:       scf.yield %[[R1]]
 //       CHECK:     }
 //       CHECK:     return %[[R0]]
 //       CHECK:   }

mlir/test/Dialect/Linalg/promote.mlir

 // RUN: mlir-opt %s -linalg-promote-subviews | FileCheck %s
 // RUN: mlir-opt %s -linalg-promote-subviews="test-promote-dynamic" | FileCheck %s --check-prefix=DYNAMIC
 // RUN: mlir-opt %s -linalg-promote-subviews="test-use-alloca" | FileCheck %s --check-prefix=ALLOCA
 
 #map1 = affine_map<(d0) -> (d0 + 2)>
 #map2 = affine_map<(d0) -> (d0 + 4)>
 #map3 = affine_map<(d0) -> (d0 + 3)>
 
 // CHECK-DAG: #[[$strided2D:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
-// CHECK-DAG: #[[$strided2D_dynamic:.*]] = affine_map<(d0, d1)[s0, s1, s2] -> (d0 * s1 + s0 + d1 * s2)>
 
 func @matmul_f32(%A: memref<?xi8>, %M: index, %N: index, %K: index) {
   %c4 = constant 4 : index
   %c3 = constant 3 : index
   %c2 = constant 2 : index
   %c0 = constant 0 : index
   %c1 = constant 1 : index
   %3 = view %A[%c0][%M, %K] : memref<?xi8> to memref<?x?xf32>

 //       CHECK:         %[[vA:.*]] = subview {{.*}} : memref<?x?xf32>
 //       CHECK:         %[[vB:.*]] = subview {{.*}} : memref<?x?xf32>
 //       CHECK:         %[[vC:.*]] = subview {{.*}} : memref<?x?xf32>
 ///
 //       CHECK:         %[[tmpA:.*]] = alloc() : memref<32xi8>
 //      ALLOCA:         %[[tmpA:.*]] = alloca() : memref<32xi8>
 //       CHECK:         %[[fullA:.*]] = std.view %[[tmpA]][{{.*}}][{{.*}}] : memref<32xi8> to memref<?x?xf32>
 //     DYNAMIC:         std.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECK:         %[[partialA:.*]] = subview %[[fullA]]{{.*}} : memref<?x?xf32> to memref<?x?xf32, #[[$strided2D_dynamic]]>
+//       CHECK:         %[[partialA:.*]] = subview %[[fullA]]{{.*}} : memref<?x?xf32> to memref<?x?xf32, #[[$strided2D]]>
 ///
 //       CHECK:         %[[tmpB:.*]] = alloc() : memref<48xi8>
 //      ALLOCA:         %[[tmpB:.*]] = alloca() : memref<48xi8>
 //       CHECK:         %[[fullB:.*]] = std.view %[[tmpB]][{{.*}}][{{.*}}] : memref<48xi8> to memref<?x?xf32>
 //     DYNAMIC:         std.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECK:         %[[partialB:.*]] = subview %[[fullB]]{{.*}} : memref<?x?xf32> to memref<?x?xf32, #[[$strided2D_dynamic]]>
+//       CHECK:         %[[partialB:.*]] = subview %[[fullB]]{{.*}} : memref<?x?xf32> to memref<?x?xf32, #[[$strided2D]]>
 ///
 //       CHECK:         %[[tmpC:.*]] = alloc() : memref<24xi8>
 //      ALLOCA:         %[[tmpC:.*]] = alloca() : memref<24xi8>
 //       CHECK:         %[[fullC:.*]] = std.view %[[tmpC]][{{.*}}][{{.*}}] : memref<24xi8> to memref<?x?xf32>
 //     DYNAMIC:         std.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?x?xf32>
-//       CHECK:         %[[partialC:.*]] = subview %[[fullC]]{{.*}} : memref<?x?xf32> to memref<?x?xf32, #[[$strided2D_dynamic]]>
+//       CHECK:         %[[partialC:.*]] = subview %[[fullC]]{{.*}} : memref<?x?xf32> to memref<?x?xf32, #[[$strided2D]]>
 
-//       CHECK:         linalg.copy(%[[vA]], %[[partialA]]) : memref<?x?xf32, #[[$strided2D]]>, memref<?x?xf32, #[[$strided2D_dynamic]]>
-//       CHECK:         linalg.copy(%[[vB]], %[[partialB]]) : memref<?x?xf32, #[[$strided2D]]>, memref<?x?xf32, #[[$strided2D_dynamic]]>
-//       CHECK:         linalg.copy(%[[vC]], %[[partialC]]) : memref<?x?xf32, #[[$strided2D]]>, memref<?x?xf32, #[[$strided2D_dynamic]]>
+//       CHECK:         linalg.copy(%[[vA]], %[[partialA]]) : memref<?x?xf32, #[[$strided2D]]>, memref<?x?xf32, #[[$strided2D]]>
+//       CHECK:         linalg.copy(%[[vB]], %[[partialB]]) : memref<?x?xf32, #[[$strided2D]]>, memref<?x?xf32, #[[$strided2D]]>
+//       CHECK:         linalg.copy(%[[vC]], %[[partialC]]) : memref<?x?xf32, #[[$strided2D]]>, memref<?x?xf32, #[[$strided2D]]>
 //
 //       CHECK:         linalg.matmul ins(%[[partialA]], %[[partialB]]{{.*}} outs(%[[partialC]]
 //
 //       CHECK:         linalg.copy(%[[partialC]], %[[vC]]) :
-//       CHECK:           memref<?x?xf32, #[[$strided2D_dynamic]]>,
+//       CHECK:           memref<?x?xf32, #[[$strided2D]]>,
 //       CHECK:           memref<?x?xf32, #[[$strided2D]]>
 //
 //       CHECK:         dealloc %[[tmpA]] : memref<32xi8>
 //       CHECK:         dealloc %[[tmpB]] : memref<48xi8>
 //       CHECK:         dealloc %[[tmpC]] : memref<24xi8>
 //  ALLOCA-NOT:         dealloc %[[tmpA]] : memref<32xi8>
 //  ALLOCA-NOT:         dealloc %[[tmpB]] : memref<48xi8>
 //  ALLOCA-NOT:         dealloc %[[tmpC]] : memref<24xi8>

 //       CHECK:       scf.for %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} {
 //       CHECK:         %[[vA_f64:.*]] = subview {{.*}} : memref<?x?xf64>
 //       CHECK:         %[[vB_f64:.*]] = subview {{.*}} : memref<?x?xf64>
 //       CHECK:         %[[vC_f64:.*]] = subview {{.*}} : memref<?x?xf64>
 ///
 //       CHECK:         %[[tmpA_f64:.*]] = alloc() : memref<64xi8>
 //       CHECK:         %[[fullA_f64:.*]] = std.view %[[tmpA_f64]][{{.*}}][{{.*}}] : memref<64xi8> to memref<?x?xf64>
 //     DYNAMIC:         std.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?x?xf64>
-//       CHECK:         %[[partialA_f64:.*]] = subview %[[fullA_f64]][%{{.*}}, %{{.*}}] : memref<?x?xf64> to memref<?x?xf64, #[[$strided2D_dynamic]]>
+//       CHECK:         %[[partialA_f64:.*]] = subview %[[fullA_f64]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf64> to memref<?x?xf64, #[[$strided2D]]>
 ///
 //       CHECK:         %[[tmpB_f64:.*]] = alloc() : memref<96xi8>
 //       CHECK:         %[[fullB_f64:.*]] = std.view %[[tmpB_f64]][{{.*}}][{{.*}}] : memref<96xi8> to memref<?x?xf64>
 //     DYNAMIC:         std.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?x?xf64>
-//       CHECK:         %[[partialB_f64:.*]] = subview %[[fullB_f64]][%{{.*}}, %{{.*}}] : memref<?x?xf64> to memref<?x?xf64, #[[$strided2D_dynamic]]>
+//       CHECK:         %[[partialB_f64:.*]] = subview %[[fullB_f64]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf64> to memref<?x?xf64, #[[$strided2D]]>
 ///
 //       CHECK:         %[[tmpC_f64:.*]] = alloc() : memref<48xi8>
 //       CHECK:         %[[fullC_f64:.*]] = std.view %[[tmpC_f64]][{{.*}}][{{.*}}] : memref<48xi8> to memref<?x?xf64>
 //     DYNAMIC:         std.view %{{.*}}[{{.*}}][{{.*}}] : memref<?xi8> to memref<?x?xf64>
-//       CHECK:         %[[partialC_f64:.*]] = subview %[[fullC_f64]][%{{.*}}, %{{.*}}] : memref<?x?xf64> to memref<?x?xf64, #[[$strided2D_dynamic]]>
+//       CHECK:         %[[partialC_f64:.*]] = subview %[[fullC_f64]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf64> to memref<?x?xf64, #[[$strided2D]]>
 
-//       CHECK:         linalg.copy(%[[vA_f64]], %[[partialA_f64]]) : memref<?x?xf64, #[[$strided2D]]>, memref<?x?xf64, #[[$strided2D_dynamic]]>
-//       CHECK:         linalg.copy(%[[vB_f64]], %[[partialB_f64]]) : memref<?x?xf64, #[[$strided2D]]>, memref<?x?xf64, #[[$strided2D_dynamic]]>
-//       CHECK:         linalg.copy(%[[vC_f64]], %[[partialC_f64]]) : memref<?x?xf64, #[[$strided2D]]>, memref<?x?xf64, #[[$strided2D_dynamic]]>
+//       CHECK:         linalg.copy(%[[vA_f64]], %[[partialA_f64]]) : memref<?x?xf64, #[[$strided2D]]>, memref<?x?xf64, #[[$strided2D]]>
+//       CHECK:         linalg.copy(%[[vB_f64]], %[[partialB_f64]]) : memref<?x?xf64, #[[$strided2D]]>, memref<?x?xf64, #[[$strided2D]]>
+//       CHECK:         linalg.copy(%[[vC_f64]], %[[partialC_f64]]) : memref<?x?xf64, #[[$strided2D]]>, memref<?x?xf64, #[[$strided2D]]>
 //
 //       CHECK:         linalg.matmul ins(%[[partialA_f64]], %[[partialB_f64]]{{.*}} outs(%[[partialC_f64]]
 //
 //       CHECK:         linalg.copy(%[[partialC_f64]], %[[vC_f64]]) :
-//       CHECK:           memref<?x?xf64, #[[$strided2D_dynamic]]>,
+//       CHECK:           memref<?x?xf64, #[[$strided2D]]>,
 //       CHECK:           memref<?x?xf64, #[[$strided2D]]>
 //
 //       CHECK:         dealloc %[[tmpA_f64]] : memref<64xi8>
 //       CHECK:         dealloc %[[tmpB_f64]] : memref<96xi8>
 //       CHECK:         dealloc %[[tmpC_f64]] : memref<48xi8>

mlir/test/Dialect/Linalg/transform-patterns.mlir

 // RUN: mlir-opt %s -test-linalg-transform-patterns=test-patterns | FileCheck %s
 
-// CHECK-DAG: #[[$STRIDED_1D:.*]] = affine_map<(d0)[s0, s1] -> (d0 * s1  + s0)>
+// CHECK-DAG: #[[$STRIDED_1D:.*]] = affine_map<(d0)[s0] -> (d0 + s0)>
 // Map corresponding to a 2D memory access where the stride along the last dim is known to be 1.
 // CHECK-DAG: #[[$STRIDED_2D_u_1:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
 // Map corresponding to a 2D memory access where the stride along all dims are unknown.
 // CHECK-DAG: #[[$STRIDED_2D:.*]] = affine_map<(d0, d1)[s0, s1, s2] -> (d0 * s1 + s0 + d1 * s2)>
 // CHECK-DAG: #[[$kn:.*]] = affine_map<(d0, d1, d2) -> (d2, d1)>
 // CHECK-DAG: #[[$nm:.*]] = affine_map<(d0, d1, d2) -> (d1, d0)>
 // CHECK-DAG: #[[$km:.*]] = affine_map<(d0, d1, d2) -> (d2, d0)>
 

 }
 // CHECK-LABEL: func @matvec
 // CHECK-DAG:     %[[c0:.*]] = constant 0 : index
 // CHECK-DAG:     %[[c5:.*]] = constant 5 : index
 // CHECK-DAG:     %[[c6:.*]] = constant 6 : index
 // CHECK:         scf.parallel {{.*}} step (%[[c5]])
 // CHECK:           scf.for {{.*}} step %[[c6]]
 // CHECK:             linalg.matvec
-// CHECK:               ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D]]>, memref<?xf32, #[[$STRIDED_1D]]>)
+// CHECK:               ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D_u_1]]>, memref<?xf32, #[[$STRIDED_1D]]>)
 // CHECK:              outs({{.*}}: memref<?xf32, #[[$STRIDED_1D]]>)
 
 func @matmul(%A: memref<?x?xf32, offset: ?, strides: [?, 1]>,
              %B: memref<?x?xf32, offset: ?, strides: [?, 1]>,
              %C: memref<?x?xf32, offset: ?, strides: [?, 1]>) {
   linalg.matmul { __internal_linalg_transform__ = "MEM" }
     ins(%A, %B: memref<?x?xf32, offset: ?, strides: [?, 1]>,
                 memref<?x?xf32, offset: ?, strides: [?, 1]>)

 // CHECK:                   scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c400]] {
 // CHECK:                     scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c20]] {
 // CHECK:                       scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c30]] {
 // CHECK:                         scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c40]] {
 // CHECK:                           scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c2]] {
 // CHECK:                             scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c3]] {
 // CHECK:                               scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c4]] {
 // CHECK:                                 linalg.matmul
-// CHECK:                                   ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D]]>, memref<?x?xf32, #[[$STRIDED_2D]]>)
-// CHECK:                                  outs({{.*}}: memref<?x?xf32, #[[$STRIDED_2D]]>)
+// CHECK:                                   ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D_u_1]]>, memref<?x?xf32, #[[$STRIDED_2D_u_1]]>)
+// CHECK:                                  outs({{.*}}: memref<?x?xf32, #[[$STRIDED_2D_u_1]]>)
 
 #matmul_accesses = [
   affine_map<(m, n, k) -> (m, k)>,
   affine_map<(m, n, k) -> (k, n)>,
   affine_map<(m, n, k) -> (m, n)>
 ]
 #generic_matmul_trait = {
   args_in = 2,

 }
 // CHECK-LABEL: func @matvec_perm
 // CHECK-DAG:     %[[c0:.*]] = constant 0 : index
 // CHECK-DAG:     %[[c5:.*]] = constant 5 : index
 // CHECK-DAG:     %[[c6:.*]] = constant 6 : index
 // CHECK:         scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c6]]
 // CHECK:           scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c5]]
 // CHECK:             linalg.matvec
-// CHECK:               ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D]]>, memref<?xf32, #[[$STRIDED_1D]]>)
+// CHECK:               ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D_u_1]]>, memref<?xf32, #[[$STRIDED_1D]]>)
 // CHECK:              outs({{.*}}: memref<?xf32, #[[$STRIDED_1D]]>)
 
 func @matmul_perm(%A: memref<?x?xf32, offset: ?, strides: [?, 1]>,
              %B: memref<?x?xf32, offset: ?, strides: [?, 1]>,
              %C: memref<?x?xf32, offset: ?, strides: [?, 1]>) {
   linalg.matmul {__internal_linalg_transform__ = "__with_perm__"}
     ins(%A, %B: memref<?x?xf32, offset: ?, strides: [?, 1]>,
                 memref<?x?xf32, offset: ?, strides: [?, 1]>)

 // CHECK:             scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c2000]] {
 // CHECK:               scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c300]] {
 // CHECK:                 scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c200]] {
 // CHECK:                   scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c400]] {
 // CHECK:                     scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c20]] {
 // CHECK:                       scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c30]] {
 // CHECK:                         scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c40]] {
 // CHECK:                                 linalg.matmul
-// CHECK:                                  ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D]]>, memref<?x?xf32, #[[$STRIDED_2D]]>)
-// CHECK:                                   outs({{.*}}: memref<?x?xf32, #[[$STRIDED_2D]]>)
+// CHECK:                                  ins({{.*}}, {{.*}}: memref<?x?xf32, #[[$STRIDED_2D_u_1]]>, memref<?x?xf32, #[[$STRIDED_2D_u_1]]>)
+// CHECK:                                   outs({{.*}}: memref<?x?xf32, #[[$STRIDED_2D_u_1]]>)
 
 func @promote_subview_matmul(%arg0: memref<?x?xf32, offset: ?, strides: [?, 1]>,
                              %arg1: memref<?x?xf32, offset: ?, strides: [?, 1]>,
                              %arg2: memref<?x?xf32, offset: ?, strides: [?, 1]>) {
   %c2000 = constant 2000 : index
   %c3000 = constant 3000 : index
   %c4000 = constant 4000 : index
   %c0 = constant 0 : index

 // CHECK:         scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c2000]] {
 // CHECK:           scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c3000]] {
 // CHECK:             scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c4000]] {
 // CHECK:               %[[s0:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:               %[[s1:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:               %[[s2:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:               %[[a0:.*]] = alloc({{%.*}}) : memref<?xi8>
 // CHECK:               %[[v0:.*]] = std.view %[[a0]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK:               %[[l0:.*]] = subview %[[v0]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK:               %[[l0:.*]] = subview %[[v0]][0, 0] [%{{.*}}, %{{.*}}] [1, 1]
+// CHECK-SAME:            memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK:               %[[a1:.*]] = alloc({{%.*}}) : memref<?xi8>
 // CHECK:               %[[v1:.*]] = std.view %[[a1]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK:               %[[l1:.*]] = subview %[[v1]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK:               %[[l1:.*]] = subview %[[v1]][0, 0] [%{{.*}}, %{{.*}}] [1, 1]
+// CHECK-SAME:            memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK:               %[[a2:.*]] = alloc({{%.*}}) : memref<?xi8>
 // CHECK:               %[[v2:.*]] = std.view %[[a2]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK:               %[[l2:.*]] = subview %[[v2]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK:               %[[l2:.*]] = subview %[[v2]][0, 0] [%{{.*}}, %{{.*}}] [1, 1]
+// CHECK-SAME:            memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK:               linalg.copy(%[[s0]], %[[l0]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>
 // CHECK:               linalg.copy(%[[s1]], %[[l1]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>
 // CHECK:               linalg.copy(%[[s2]], %[[l2]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>
 // CHECK:               linalg.matmul
 // CHECK-SAME:                 ins(%[[v0]], %[[v1]] : memref<?x?xf32>, memref<?x?xf32>)
 // CHECK-SAME:                outs(%[[v2]] : memref<?x?xf32>)
 
 func @promote_first_subview_matmul(%arg0: memref<?x?xf32, offset: ?, strides: [?, 1]>,

 // CHECK:   scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c2000]] {
 // CHECK:     scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c3000]] {
 // CHECK:       scf.for {{.*}} = %[[c0]] to {{.*}} step %[[c4000]] {
 // CHECK:         %[[s0:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:         %[[s1:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:         %[[s2:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:         %[[a0:.*]] = alloc({{%.*}}) : memref<?xi8>
 // CHECK:         %[[v0:.*]] = std.view %[[a0]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK:         %[[l0:.*]] = subview %[[v0]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK:         %[[l0:.*]] = subview %[[v0]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK-NOT:     %[[a1:.*]] = alloc({{%.*}}) : memref<?xi8>
 // CHECK-NOT:     %[[v1:.*]] = std.view %[[a1]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK-NOT:     %[[l0:.*]] = subview %[[v1]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK-NOT:     %[[l0:.*]] = subview %[[v1]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK-NOT:     %[[a2:.*]] = alloc({{%.*}}) : memref<?xi8>
 // CHECK-NOT:     %[[v2:.*]] = std.view %[[a2]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK-NOT:     %[[l0:.*]] = subview %[[v2]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK-NOT:     %[[l0:.*]] = subview %[[v2]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK:         linalg.copy(%[[s0]], %[[l0]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>
 // CHECK-NOT:     linalg.copy(%[[s1]], %[[l1]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>
 // CHECK-NOT:     linalg.copy(%[[s2]], %[[l2]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>^
 // CHECK:         linalg.matmul
 // CHECK-SAME:           ins(%[[v0]], %[[s1]] : memref<?x?xf32>, memref<?x?xf32, #[[$STRIDED_2D]]>)
 // CHECK-SAME:          outs(%[[s2]] : memref<?x?xf32, #[[$STRIDED_2D]]>)
 
 func @aligned_promote_fill(%arg0: memref<?x?xf32, offset: ?, strides: [?, 1]>) {
   %c2000 = constant 2000 : index
   %c4000 = constant 4000 : index
   %c0 = constant 0 : index
   %c1 = constant 1 : index
   %cf = constant 1.0 : f32
   %3 = std.subview %arg0[%c0, %c0][%c2000, %c4000][%c1, %c1] :
    memref<?x?xf32, offset: ?, strides: [?, 1]> to memref<?x?xf32, offset: ?, strides: [?, ?]>
   linalg.fill(%3, %cf) { __internal_linalg_transform__ = "_promote_views_aligned_"}
     :  memref<?x?xf32, offset: ?, strides: [?, ?]>, f32
   return
 }
 // CHECK-LABEL: func @aligned_promote_fill
 // CHECK:   %[[cf:.*]] = constant {{.*}} : f32
 // CHECK:         %[[s0:.*]] = subview {{%.*}}[{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32, #map{{.*}}> to memref<?x?xf32, #map{{.*}}>
 // CHECK:         %[[a0:.*]] = alloc({{%.*}}) {alignment = 32 : i64} : memref<?xi8>
 // CHECK:         %[[v0:.*]] = std.view %[[a0]][{{.*}}][{{%.*}}, {{%.*}}] : memref<?xi8> to memref<?x?xf32>
-// CHECK:         %[[l0:.*]] = subview %[[v0]][{{%.*}}, {{%.*}}] [{{%.*}}, {{%.*}}] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D]]>
+// CHECK:         %[[l0:.*]] = subview %[[v0]][0, 0] [%{{.*}}, %{{.*}}] [1, 1] : memref<?x?xf32> to memref<?x?xf32, #[[$STRIDED_2D_u_1]]>
 // CHECK:         linalg.fill(%[[v0]], {{%.*}}) : memref<?x?xf32>, f32
 // CHECK:         linalg.copy(%[[s0]], %[[l0]]) : memref<?x?xf32, #map{{.*}}>, memref<?x?xf32, #map{{.*}}>
 // CHECK:         linalg.fill(%[[v0]], %[[cf]]) : memref<?x?xf32>, f32
 
 func @tile_permute_parallel_loop(%arg0: memref<?x?xf32>,
                                  %arg1: memref<?x?xf32>,
                                  %arg2: memref<?x?xf32>) {
   linalg.matmul {__internal_linalg_transform__ = "par__with_perm__"}

mlir/test/Dialect/Vector/vector-transfer-full-partial-split.mlir

 // RUN: mlir-opt %s -test-vector-transfer-full-partial-split | FileCheck %s
 // RUN: mlir-opt %s -test-vector-transfer-full-partial-split=use-linalg-copy | FileCheck %s --check-prefix=LINALG
 
 // CHECK-DAG: #[[$map_p4:.*]] = affine_map<()[s0] -> (s0 + 4)>
 // CHECK-DAG: #[[$map_p8:.*]] = affine_map<()[s0] -> (s0 + 8)>
 // CHECK-DAG: #[[$map_2d_stride_1:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
 
 // LINALG-DAG: #[[$map_p4:.*]] = affine_map<()[s0] -> (s0 + 4)>
 // LINALG-DAG: #[[$map_p8:.*]] = affine_map<()[s0] -> (s0 + 8)>
 // LINALG-DAG: #[[$map_2d_stride_1:.*]] = affine_map<(d0, d1)[s0, s1] -> (d0 * s1 + s0 + d1)>
-// LINALG-DAG: #[[$map_2d_dynamic:.*]] = affine_map<(d0, d1)[s0, s1, s2] -> (d0 * s1 + s0 + d1 * s2)>
+// LINALG-DAG: #[[$map_2d_stride_8x1:.*]] = affine_map<(d0, d1)[s0] -> (d0 * 8 + s0 + d1)>
 // LINALG-DAG: #[[$bounds_map_4:.*]] = affine_map<(d0, d1, d2) -> (d0 - d1, 4)>
 // LINALG-DAG: #[[$bounds_map_8:.*]] = affine_map<(d0, d1, d2) -> (d0 - d1, 8)>
 
 // CHECK-LABEL: split_vector_transfer_read_2d(
 //  CHECK-SAME: %[[A:[a-zA-Z0-9]*]]: memref
 //  CHECK-SAME: %[[i:[a-zA-Z0-9]*]]: index
 //  CHECK-SAME: %[[j:[a-zA-Z0-9]*]]: index
 
func @split_vector_transfer_read_2d(%A: memref<?x8xf32>, %i: index, %j: index) -> vector<4x8xf32> {
   // CHECK-SAME:     memref<4x8xf32> to memref<?x8xf32>
   //      CHECK:   scf.yield %[[yielded]], %[[c0]], %[[c0]] :
   // CHECK-SAME:     memref<?x8xf32>, index, index
   //      CHECK: }
   //      CHECK: %[[res:.*]] = vector.transfer_read %[[ifres]]#0[%[[ifres]]#1, %[[ifres]]#2], %[[cst]]
   // CHECK_SAME:   {masked = [false, false]} : memref<?x8xf32>, vector<4x8xf32>
 
   //  LINALG-DAG: %[[c0:.*]] = constant 0 : index
-  //  LINALG-DAG: %[[c1:.*]] = constant 1 : index
   //  LINALG-DAG: %[[c4:.*]] = constant 4 : index
   //  LINALG-DAG: %[[c8:.*]] = constant 8 : index
   //  LINALG-DAG: %[[cst:.*]] = constant 0.000000e+00 : f32
   // alloca for boundary full tile
   //      LINALG: %[[alloc:.*]] = alloca() {alignment = 32 : i64} : memref<4x8xf32>
   // %i + 4 <= dim(%A, 0)
   //      LINALG: %[[idx0:.*]] = affine.apply #[[$map_p4]]()[%[[i]]]
   //      LINALG: %[[d0:.*]] = dim %[[A]], %[[c0]] : memref<?x8xf32>
   //      LINALG: %[[cmp0:.*]] = cmpi sle, %[[idx0]], %[[d0]] : index
   // %j + 8 <= dim(%A, 1)
   //      LINALG: %[[idx1:.*]] = affine.apply #[[$map_p8]]()[%[[j]]]
   //      LINALG: %[[cmp1:.*]] = cmpi sle, %[[idx1]], %[[c8]] : index
   // are both conds true
   //      LINALG: %[[cond:.*]] = and %[[cmp0]], %[[cmp1]] : i1
   //      LINALG: %[[ifres:.*]]:3 = scf.if %[[cond]] -> (memref<?x8xf32>, index, index) {
   //               inBounds, just yield %A
   //      LINALG:   scf.yield %[[A]], %[[i]], %[[j]] : memref<?x8xf32>, index, index
   //      LINALG: } else {
   //               slow path, fill tmp alloc and yield a memref_casted version of it
   //      LINALG:   linalg.fill(%[[alloc]], %[[cst]]) : memref<4x8xf32>, f32
   //      LINALG:   %[[d0:.*]] = dim %[[A]], %[[c0]] : memref<?x8xf32>
   //      LINALG:   %[[sv0:.*]] = affine.min #[[$bounds_map_4]](%[[d0]], %[[i]], %[[c4]])
   //      LINALG:   %[[sv1:.*]] = affine.min #[[$bounds_map_8]](%[[c8]], %[[j]], %[[c8]])
-  //      LINALG:   %[[sv:.*]] = subview %[[A]][%[[i]], %[[j]]] [%[[sv0]], %[[sv1]]] [%[[c1]], %[[c1]]]
-  // LINALG-SAME:     memref<?x8xf32> to memref<?x?xf32, #[[$map_2d_dynamic]]>
-  //      LINALG:   linalg.copy(%[[sv]], %[[alloc]]) : memref<?x?xf32, #[[$map_2d_dynamic]]>, memref<4x8xf32>
+  //      LINALG:   %[[sv:.*]] = subview %[[A]][%[[i]], %[[j]]] [%[[sv0]], %[[sv1]]] [1, 1]
+  // LINALG-SAME:     memref<?x8xf32> to memref<?x?xf32, #[[$map_2d_stride_8x1]]>
+  //      LINALG:   linalg.copy(%[[sv]], %[[alloc]]) : memref<?x?xf32, #[[$map_2d_stride_8x1]]>, memref<4x8xf32>
   //      LINALG:   %[[yielded:.*]] = memref_cast %[[alloc]] :
   // LINALG-SAME:     memref<4x8xf32> to memref<?x8xf32>
   //      LINALG:   scf.yield %[[yielded]], %[[c0]], %[[c0]] :
   // LINALG-SAME:     memref<?x8xf32>, index, index
   //      LINALG: }
   //      LINALG: %[[res:.*]] = vector.transfer_read %[[ifres]]#0[%[[ifres]]#1, %[[ifres]]#2], %[[cst]]
   // LINALG_SAME:   {masked = [false, false]} : memref<?x8xf32>, vector<4x8xf32>
   %1 = vector.transfer_read %A[%i, %j], %f0 : memref<?x8xf32>, vector<4x8xf32>
func @split_vector_transfer_read_strided_2d(
   // CHECK-SAME:     memref<4x8xf32> to memref<?x8xf32, #[[$map_2d_stride_1]]>
   //      CHECK:   scf.yield %[[yielded]], %[[c0]], %[[c0]] :
   // CHECK-SAME:     memref<?x8xf32, #[[$map_2d_stride_1]]>, index, index
   //      CHECK: }
   //      CHECK: %[[res:.*]] = vector.transfer_read {{.*}} {masked = [false, false]} :
   // CHECK-SAME:   memref<?x8xf32, #[[$map_2d_stride_1]]>, vector<4x8xf32>
 
   //  LINALG-DAG: %[[c0:.*]] = constant 0 : index
-  //  LINALG-DAG: %[[c1:.*]] = constant 1 : index
   //  LINALG-DAG: %[[c4:.*]] = constant 4 : index
   //  LINALG-DAG: %[[c7:.*]] = constant 7 : index
   //  LINALG-DAG: %[[c8:.*]] = constant 8 : index
   //  LINALG-DAG: %[[cst:.*]] = constant 0.000000e+00 : f32
   // alloca for boundary full tile
   //      LINALG: %[[alloc:.*]] = alloca() {alignment = 32 : i64} : memref<4x8xf32>
   // %i + 4 <= dim(%A, 0)
   //      LINALG: %[[idx0:.*]] = affine.apply #[[$map_p4]]()[%[[i]]]
func @split_vector_transfer_read_strided_2d(
   // LINALG-SAME:     memref<7x8xf32, #[[$map_2d_stride_1]]> to memref<?x8xf32, #[[$map_2d_stride_1]]>
   //      LINALG:   scf.yield %[[casted]], %[[i]], %[[j]] :
   // LINALG-SAME:     memref<?x8xf32, #[[$map_2d_stride_1]]>, index, index
   //      LINALG: } else {
   //               slow path, fill tmp alloc and yield a memref_casted version of it
   //      LINALG:   linalg.fill(%[[alloc]], %[[cst]]) : memref<4x8xf32>, f32
   //      LINALG:   %[[sv0:.*]] = affine.min #[[$bounds_map_4]](%[[c7]], %[[i]], %[[c4]])
   //      LINALG:   %[[sv1:.*]] = affine.min #[[$bounds_map_8]](%[[c8]], %[[j]], %[[c8]])
-  //      LINALG:   %[[sv:.*]] = subview %[[A]][%[[i]], %[[j]]] [%[[sv0]], %[[sv1]]] [%[[c1]], %[[c1]]]
-  // LINALG-SAME:     memref<7x8xf32, #[[$map_2d_stride_1]]> to memref<?x?xf32, #[[$map_2d_dynamic]]>
-  //      LINALG:   linalg.copy(%[[sv]], %[[alloc]]) : memref<?x?xf32, #[[$map_2d_dynamic]]>, memref<4x8xf32>
+  //      LINALG:   %[[sv:.*]] = subview %[[A]][%[[i]], %[[j]]] [%[[sv0]], %[[sv1]]] [1, 1]
+  // LINALG-SAME:     memref<7x8xf32, #[[$map_2d_stride_1]]> to memref<?x?xf32, #[[$map_2d_stride_1]]>
+  //      LINALG:   linalg.copy(%[[sv]], %[[alloc]]) : memref<?x?xf32, #[[$map_2d_stride_1]]>, memref<4x8xf32>
   //      LINALG:   %[[yielded:.*]] = memref_cast %[[alloc]] :
   // LINALG-SAME:     memref<4x8xf32> to memref<?x8xf32, #[[$map_2d_stride_1]]>
   //      LINALG:   scf.yield %[[yielded]], %[[c0]], %[[c0]] :
   // LINALG-SAME:     memref<?x8xf32, #[[$map_2d_stride_1]]>, index, index
   //      LINALG: }
   //      LINALG: %[[res:.*]] = vector.transfer_read {{.*}} {masked = [false, false]} :
   // LINALG-SAME:   memref<?x8xf32, #[[$map_2d_stride_1]]>, vector<4x8xf32>
   %1 = vector.transfer_read %A[%i, %j], %f0 :
     memref<7x8xf32, offset:?, strides:[?, 1]>, vector<4x8xf32>
 
   // CHECK: return %[[res]] : vector<4x8xf32>
   return %1 : vector<4x8xf32>
 }