The rapid growth of io_uring

> Unix historically didn't have asynchronous disk I/O, and this effected the architecture of subsystems and drivers. Linux still doesn't have asynchronous buffered disk I/O. All you're doing is offloading the operation to a thread pool in the kernel, which performs the operation synchronously, rather than a thread pool in user land.
That's not quite true, the BIO layer is already asynchronous (with its own scheduler and everything). And that's where the most of the waiting goes.

And it's not like the VFS layer is impossible to implement asynchronously, it's just that before now it was not needed at all.