[mlir] Extend tile_using_for verifier to fix a crash (#98366)

This patch adds a check for the correct number of `loops` results of the
`transform.structured.tile_using_for` Op to the verifier, fixing a
crash.

Fix #98008

Author: ubfx

mlir/lib/Dialect/Linalg/TransformOps/LinalgTransformOps.cpp

@@ -2888,8 +2888,14 @@ void transform::TileUsingForOp::build(
 LogicalResult transform::TileUsingForOp::verify() {
   if (getMixedSizes().size() != getScalableSizes().size())
     return emitOpError("expected same number of sizes (")
-           << getMixedSizes().size() << ") and scalable sizes ()"
+           << getMixedSizes().size() << ") and scalable sizes ("
            << getScalableSizes().size() << ")";
+  ArrayRef<int64_t> staticSizes = getStaticSizes();
+  unsigned numExpectedLoops = staticSizes.size() - llvm::count(staticSizes, 0);
+  if (getLoops().size() != numExpectedLoops)
+    return emitOpError("expected number of loops to tile (")
+           << numExpectedLoops << ") to match number of `loops` results ("
+           << getLoops().size() << ")";
   return success();
 }
 

mlir/test/Dialect/Linalg/transform-op-tile.mlir

@@ -253,3 +253,23 @@ module attributes {transform.with_named_sequence} {
     transform.yield
   }
 }
+
+// -----
+
+module attributes {transform.with_named_sequence} {
+  transform.named_sequence @__transform_main(%arg1: !transform.any_op {transform.readonly}) {
+    %0 = transform.structured.match ops{["linalg.matmul"]} in %arg1 : (!transform.any_op) -> !transform.any_op
+    // expected-error @below {{op expected number of loops to tile (3) to match number of `loops` results (1)}}
+    %1, %loops = transform.structured.tile_using_for %0 tile_sizes [4, 4, 4] : (!transform.any_op) -> (!transform.any_op, !transform.any_op)
+    transform.yield
+  }
+}
+
+func.func @tile_linalg_matmul(
+  %arg0: tensor<128x128xf32>, %arg1: tensor<128x128xf32>, %arg2: tensor<128x128xf32>)
+    -> tensor<128x128xf32> {
+  %0 = linalg.matmul  ins(%arg0, %arg1: tensor<128x128xf32>, tensor<128x128xf32>)
+                     outs(%arg2: tensor<128x128xf32>)
+    -> tensor<128x128xf32>
+  return %0 : tensor<128x128xf32>
+}