[Mlir-commits] [llvm] [mlir] [mlir][bufferization] Add a ValueBoundsOpInterface model for to_tensor (PR #215357)

[Victor Perez]

================
@@ -0,0 +1,48 @@
+//===- ValueBoundsOpInterfaceImpl.cpp - Impl. of ValueBoundsOpInterface ---===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "mlir/Dialect/Bufferization/IR/ValueBoundsOpInterfaceImpl.h"
+
+#include "mlir/Dialect/Bufferization/IR/Bufferization.h"
+#include "mlir/Interfaces/ValueBoundsOpInterface.h"
+
+using namespace mlir;
+
+namespace mlir {
+namespace bufferization {
+namespace {
+
+struct ToTensorOpInterface
+    : public ValueBoundsOpInterface::ExternalModel<ToTensorOpInterface,
+                                                   ToTensorOp> {
+  void populateBoundsForShapedValueDim(Operation *op, Value value, int64_t dim,
+                                       ValueBoundsConstraintSet &cstr) const {
+    auto toTensorOp = cast<ToTensorOp>(op);
+    assert(value == toTensorOp.getResult() && "invalid value");
+
+    // The tensor and the buffer describe the same memory, so they have the same
+    // size in every dimension. The op also accepts types that are not shaped,
+    // for which no bound can be computed.
+    if (isa<ShapedType>(toTensorOp.getResult().getType()) &&
+        isa<ShapedType>(toTensorOp.getBuffer().getType()))
+      cstr.bound(value)[dim] == cstr.getExpr(toTensorOp.getBuffer(), dim);
----------------
victor-eds wrote:

I'm pretty sure that the convention in LLVM is not adding braces here. I'll keep it as is to follow [the convention](https://llvm.org/docs/CodingStandards.html#don-t-use-braces-on-simple-single-statement-bodies-of-if-else-loop-statements).

https://github.com/llvm/llvm-project/pull/215357