add dealloc alias check to bufferization

[kushanam]

Summary

Traverse the cloneOp for aliases to find the alloc op

[kushanam]

add dealloc alias check to bufferization
Traverse the cloneOp for aliases to find the alloc op

[frgossen]

Great! Thank's for fixing this.

mlir/lib/Dialect/Bufferization/IR/BufferizationOps.cpp

454

Can we call this canonicalSource?

457

Dont think this is needed. The while loop will already check this, no?

[kushanam]

add dealloc alias check to bufferization
Traverse the cloneOp for aliases to find the alloc op

[bondhugula]

LGTM

mlir/lib/Dialect/Bufferization/IR/BufferizationOps.cpp

452–453

Find the dealloc op for the canonical source...?

[frgossen]

Nice, exactly what we want.
Can you address Uday's comment and the failing CI? I believe this is just a formatting issue:

ERROR   git-clang-format returned an non-zero exit code 1

[kushanam]

add dealloc alias check to bufferization
Traverse the cloneOp for aliases to find the alloc op

[frgossen]

Thanks. Will land this when the CI passes

[frgossen]

It's still failing the linter
You can likely just run git-clang-format HEAD~1 to fix this.

[kushanam]

add dealloc alias check to bufferization
Traverse the cloneOp for aliases to find the alloc op

[kushanam]

In D131797#3730232, @frgossen wrote:

It's still failing the linter
You can likely just run git-clang-format HEAD~1 to fix this.

Thanks Frederik, I ran the clang formatter and resubmitted.

[frgossen]

The Ci is running but I do not see any changed. This is strange.

[kushanam]

This comment was removed by kushanam.

[kushanam]

add dealloc alias check to bufferization
Traverse the cloneOp for aliases to find the alloc op

[kushanam]

Sorry, the issue was with my amend. Resubmitted the patch..

mlir/lib/Dialect/Bufferization/IR/BufferizationOps.cpp

struct SimplifyClones : public OpRewritePattern<CloneOp> {
   LogicalResult matchAndRewrite(CloneOp cloneOp,
                                 PatternRewriter &rewriter) const override {
     if (cloneOp.use_empty()) {
       rewriter.eraseOp(cloneOp);
       return success();
     }
 
     Value source = cloneOp.getInput();
+    // We aim to check dealloc op for aliases below, e.g. ReshapeOp
+    // which otherwise could prevent removal of unnecessary allocs.
+    Value canonicalSource = source;
+    while (auto iface =
+               dyn_cast_or_null<ViewLikeOpInterface>(canonicalSource.getDefiningOp()))
+      canonicalSource = iface.getViewSource();
 
-    // This only finds dealloc operations for the immediate value. It should
-    // also consider aliases. That would also make the safety check below
-    // redundant.
     llvm::Optional<Operation *> maybeCloneDeallocOp =
         memref::findDealloc(cloneOp.getOutput());
     // Skip if either of them has > 1 deallocate operations.
     if (!maybeCloneDeallocOp.has_value())
       return failure();
     llvm::Optional<Operation *> maybeSourceDeallocOp =
-        memref::findDealloc(source);
+        memref::findDealloc(canonicalSource);
     if (!maybeSourceDeallocOp.has_value())
       return failure();
     Operation *cloneDeallocOp = *maybeCloneDeallocOp;
     Operation *sourceDeallocOp = *maybeSourceDeallocOp;
 
     // If both are deallocated in the same block, their in-block lifetimes
     // might not fully overlap, so we cannot decide which one to drop.
     if (cloneDeallocOp && sourceDeallocOp &&

mlir/test/Dialect/Bufferization/canonicalize.mlir

func.func @alloc_tensor_canonicalize() -> (tensor<4x5x?xf32>) {
   %c6 = arith.constant 6 : index
   %0 = bufferization.alloc_tensor(%c6) : tensor<4x5x?xf32>
   return %0 : tensor<4x5x?xf32>
 }
 // CHECK: func @alloc_tensor_canonicalize
 // CHECK:   %[[T0:.+]] = bufferization.alloc_tensor() : tensor<4x5x6xf32>
 // CHECK:   %[[T1:.+]] = tensor.cast %[[T0]] : tensor<4x5x6xf32> to tensor<4x5x?xf32>
 // CHECK:   return %[[T1]]
+
+// -----
+
+func.func @dealloc_canonicalize_clone_removal(%arg0: memref<?xindex>) -> memref<*xf32> {
+  %c1 = arith.constant 1 : index
+  %0 = memref.alloc(%c1) : memref<?xf32>
+  %1 = memref.reshape %0(%arg0) : (memref<?xf32>, memref<?xindex>) -> memref<*xf32>
+  %2 = bufferization.clone %1 : memref<*xf32> to memref<*xf32>
+  memref.dealloc %0 : memref<?xf32>
+  return %2 : memref<*xf32>
+}
+// CHECK-LABEL: @dealloc_canonicalize_clone_removal
+//   CHECK-NOT:   bufferization.clone
+//   CHECK-NOT:   memref.dealloc
+//       CHECK:   return {{.*}}