Port::read should take &mut self

[MikailBag]

Currently, this function takes port object through a shared reference.
To my mind, it is not good, because port internal state does mutate after read() (e.g. read is not idempotent, which is unusual for &-functions).
Since Port is Sync, user can by mistake read from port on multiple threads and get strange behaviour.
Finally, this is inconsistent with other similar APIs, like cpuio::Port::read, and std::io::Read::read.

[phil-opp]

Good point, I fully agree! Do you have time to open a PR for this?

[MikailBag]

Thank's for agreement!
Yes, I think I can implement it.

[phil-opp]

Great, thank you!