[Mlir-commits] [mlir] [mlir][bufferize] Make drop-equivalent-buffer-results support mult blocks (PR #163388)

Matthias Springer llvmlistbot at llvm.org
Fri Oct 17 01:58:50 PDT 2025


================
@@ -490,3 +490,32 @@ func.func @collapse_shape_regression(
   tensor.collapse_shape %0[[0, 1]] : tensor<5x6xf32> into tensor<30xf32>
   return
 }
+
+// -----
+
+// CHECK-LABEL: func private @mult_return_callee(
+//  CHECK-SAME:   %[[T:.*]]: memref<?xf32, strided<[?], offset: ?>>, %[[COND:.*]]: i1,
+//  CHECK-SAME:   %[[A:.*]]: index, %[[B:.*]]: index) -> index {
+func.func private @mult_return_callee(%t: tensor<?xf32>,  %cond:i1, %a: index, %b: index) -> (tensor<10xf32>, index) {
+  %casted = tensor.cast %t : tensor<?xf32> to tensor<10xf32>
+  // CHECK: cf.cond_br %[[COND]], ^bb1, ^bb2
+  // CHECK:  ^bb1:
+  // CHECK:    return %[[A]] : index
+  // CHECK:  ^bb2:
+  // CHECK:    return %[[B]] : index
+  cf.cond_br %cond,^a, ^b
+  ^a:
----------------
matthias-springer wrote:

nit: `^a:` and `^b:` should not be indented. (Same level as `func.func`.)

https://github.com/llvm/llvm-project/pull/163388


More information about the Mlir-commits mailing list