Investigate tokio-uring usage

[mempirate]

Context

A large part of CPU cycles in the benchmarks are spent on syscalls and the associated copying of data. We could (potentially) improve on this by leveraging io_uring through tokio-uring.

This should be gated by a feature and only possible on Linux platforms (if kernel is up-to-date).

For more information about tokio-uring check out the design document

[merklefruit]

A good starting point here would be to run strace or get a flamegraph going to understand what syscalls we're doing most and how much time is spent on them. my intuition is that for network I/O the default epoll implementation will be hard to beat with io_uring, but I'm curious as well!