ppenzin wrote: > It remains unclear whether the final upstream kernel will support this scenario of asymmetric vlen heterogeneity. Is there a thread on kernel mailing list about that? https://github.com/llvm/llvm-project/pull/174052