[Mlir-commits] [mlir] [mlir][bufferization] Introduce reconcileBufferTypeMismatchFn hook (PR #202667)

Andrei Golubev llvmlistbot at llvm.org
Thu Jun 11 07:55:47 PDT 2026


================
@@ -83,28 +81,35 @@ struct OpWithUnstructuredControlFlowBufferizableOpInterfaceExternalModel
       if (bufferType == callerType)
         continue;
 
-      // If the computed buffer type does not match the computed buffer type
-      // of the earlier forwarded operands, fall back to a buffer type with a
-      // fully dynamic layout map.
+      // If the computed buffer type does not match the computed buffer type of
+      // the earlier forwarded operands, fall back to a reconciled buffer type.
 #ifndef NDEBUG
-      if (auto rankedTensorType = dyn_cast<RankedTensorType>(tensorType)) {
-        assert(bufferType.hasRank() && callerType.hasRank() &&
-               "expected ranked memrefs");
-        assert(llvm::all_equal({bufferType.getShape(), callerType.getShape(),
-                                rankedTensorType.getShape()}) &&
-               "expected same shape");
-      } else {
-        assert(!bufferType.hasRank() && !callerType.hasRank() &&
-               "expected unranked memrefs");
+      auto tensorType = value.getType();
----------------
andrey-golubev wrote:

thanks for asking, i remembered we have a verifier function inside `TensorLikeType`, i believe we can use it here.
i left the `NDEBUG` and also made it through `if (dyn_cast<TensorLikeType>())` to guard against already-bufferized cases - probably good enough so long as it's under `NDEBUG`.

https://github.com/llvm/llvm-project/pull/202667


More information about the Mlir-commits mailing list