[Mlir-commits] [mlir] [mlir][vector] Prevent masked transfer read/write identity folding (PR #192966)

[llvmlistbot at llvm.org]

https://github.com/zackc6 updated https://github.com/llvm/llvm-project/pull/192966

>From b7f120c4c5778d13ee9692fdab3006d21049b556 Mon Sep 17 00:00:00 2001
From: zack <zackchen666 at gmail.com>
Date: Mon, 20 Apr 2026 21:00:07 +0800
Subject: [PATCH] [mlir][vector] Prevent masked transfer read/write identity
 folding

Avoid folding transfer_read + transfer_write to the base tensor when either op is masked, since masked transfer semantics are not identity. Add regression coverage for masked read-only, masked write-only, and masked read+write cases.
---
 mlir/lib/Dialect/Vector/IR/VectorOps.cpp   |  3 ++
 mlir/test/Dialect/Vector/canonicalize.mlir | 55 ++++++++++++++++++++++
 2 files changed, 58 insertions(+)

diff --git a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
index 3d3e49134363f..f78b3c37832c2 100644
--- a/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
+++ b/mlir/lib/Dialect/Vector/IR/VectorOps.cpp
@@ -5850,6 +5850,9 @@ static LogicalResult foldReadInitWrite(TransferWriteOp write,
   // Bail on potential out-of-bounds accesses.
   if (read.hasOutOfBoundsDim() || write.hasOutOfBoundsDim())
     return failure();
+  // Masked transfers have padding/select semantics and are not identity folds.
+  if (read.getMask() || write.getMask())
+    return failure();
   // Tensor types must be the same.
   if (read.getBase().getType() != rankedTensorType)
     return failure();
diff --git a/mlir/test/Dialect/Vector/canonicalize.mlir b/mlir/test/Dialect/Vector/canonicalize.mlir
index de32ab45aa7ea..6aa92ab79a0dd 100644
--- a/mlir/test/Dialect/Vector/canonicalize.mlir
+++ b/mlir/test/Dialect/Vector/canonicalize.mlir
@@ -1816,6 +1816,61 @@ func.func @transfer_folding_1(%t0: tensor<2x3x4xf32>, %t1: tensor<2x3x4xf32>)
 
 // -----
 
+// CHECK-LABEL: func @negative_transfer_folding_masked_read
+//       CHECK:   vector.transfer_read {{.*}}, {{.*}}, %[[MASK:.*]]
+//       CHECK:   %[[R:.*]] = vector.transfer_write
+//       CHECK:   return %[[R]]
+func.func @negative_transfer_folding_masked_read(
+    %t0: tensor<2x3x4xf32>, %t1: tensor<2x3x4xf32>,
+    %mask: vector<2x3x4xi1>) -> tensor<2x3x4xf32> {
+  %c0 = arith.constant 0 : index
+  %pad = arith.constant 0.0 : f32
+  %v = vector.transfer_read %t0[%c0, %c0, %c0], %pad, %mask {in_bounds = [true, true, true]} :
+    tensor<2x3x4xf32>, vector<2x3x4xf32>
+  %r = vector.transfer_write %v, %t1[%c0, %c0, %c0] {in_bounds = [true, true, true]} :
+    vector<2x3x4xf32>, tensor<2x3x4xf32>
+  return %r : tensor<2x3x4xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func @negative_transfer_folding_masked_write
+//       CHECK:   vector.transfer_read
+//       CHECK:   %[[R:.*]] = vector.transfer_write {{.*}}, {{.*}}, %[[MASK:.*]]
+//       CHECK:   return %[[R]]
+func.func @negative_transfer_folding_masked_write(
+    %t0: tensor<2x3x4xf32>, %t1: tensor<2x3x4xf32>,
+    %mask: vector<2x3x4xi1>) -> tensor<2x3x4xf32> {
+  %c0 = arith.constant 0 : index
+  %pad = arith.constant 0.0 : f32
+  %v = vector.transfer_read %t0[%c0, %c0, %c0], %pad {in_bounds = [true, true, true]} :
+    tensor<2x3x4xf32>, vector<2x3x4xf32>
+  %r = vector.transfer_write %v, %t1[%c0, %c0, %c0], %mask {in_bounds = [true, true, true]} :
+    vector<2x3x4xf32>, tensor<2x3x4xf32>
+  return %r : tensor<2x3x4xf32>
+}
+
+// -----
+
+// CHECK-LABEL: func @negative_transfer_folding_masked_read_and_write
+//  CHECK-SAME:   %[[MASK:[0-9a-zA-Z_]+]]: vector<2x3x4xi1>
+//       CHECK:   %[[V:.*]] = vector.transfer_read {{.*}}, {{.*}}, %[[MASK]]
+//       CHECK:   %[[R:.*]] = vector.transfer_write %[[V]], {{.*}}, %[[MASK]]
+//       CHECK:   return %[[R]]
+func.func @negative_transfer_folding_masked_read_and_write(
+    %t0: tensor<2x3x4xf32>, %t1: tensor<2x3x4xf32>,
+    %mask: vector<2x3x4xi1>) -> tensor<2x3x4xf32> {
+  %c0 = arith.constant 0 : index
+  %pad = arith.constant 0.0 : f32
+  %v = vector.transfer_read %t0[%c0, %c0, %c0], %pad, %mask {in_bounds = [true, true, true]} :
+    tensor<2x3x4xf32>, vector<2x3x4xf32>
+  %r = vector.transfer_write %v, %t1[%c0, %c0, %c0], %mask {in_bounds = [true, true, true]} :
+    vector<2x3x4xf32>, tensor<2x3x4xf32>
+  return %r : tensor<2x3x4xf32>
+}
+
+// -----
+
 // CHECK-LABEL: func @store_after_load_tensor
 //  CHECK-SAME: (%[[ARG:.*]]: tensor<4x4xf32>)
 //   CHECK-NOT:   vector.transfer_read