I agree that io_uring is a fundamentally more efficient approach, but I think the performance limits you're currently measuring with mmap() aren't the fundament

I agree that io_uring is a fundamentally more efficient approach, but I think the performance limits you're currently measuring with mmap() aren't the fundamental ones imposed by the mmap() API, and I think that's what you're saying too?