mrloop
mrloop copied to clipboard
`mr_writev` calls `io_uring_submit` in batches?

Maybe the aim is improving performance, but what if no much writev in a short-time? Every writev will wait for a long time to be submitted to kernel?