fix: Udp implementation and add more methods

[elpiel]

Split out from #42

• Fixes the UDP implementation and uses the socket's ring buffers accordingly for better performance.
• Adds set_port and get_port for setting and getting respectively the socket port

[ryan-summers]

Looking really nice! Thanks for the changes. Some minor comments/cleanup requests, but it looks really good otherwise.

[elpiel]

There's definitely a lot of room for improvements, however, this UDP implementation now works and for me that was more important.

You're welcome to improve on the current working implementation.

[ryan-summers]

@elpiel It's not clear to me what this PR is actually fixing in the UDP API outside of adding the new functions and refactoring the code, but you claim this has resolved your issues. What issues does this PR fix? From my review, I would guess #59, but I don't see any changes that would address #46 or #44

[elpiel]

Yes @ryan-summers there are a few things that this PR fixes and improves:

• Instead of always writing from the start of the buffer it uses it in it's entirety moving the pointers. UDP receive incorrectly updates RX pointer #59
• If you haven't sent a destination but you try to send data it will return an error - When Device.connect is not called, Udp socket is not being opened #46
• Status is checked to ensure When Device.connect is not called, Udp socket is not being opened #46 gives you feedback if the socket is not opened.
• It also removes socketn::Command::Open when sending which caused issues with the sockets.

[ryan-summers]

Thanks for the explanation on what this PR is solving, it helped me understand the purpose of this restructering. After taking a closer look, everything looks to be in order and makes sense to me. Thanks for taking the time to help me understand a bit more, and I appreciate the contribution here a lot!

[ryan-summers]

@elpiel Would you be able to resolve the merge conflicts? I am happy to merge afterwards :)

[elpiel]

Will do.
Thanks for taking the time to look into it!

Hopefully we can improve the crate not to block while waiting for ARP response, make it async and allow usage of the interrupt pin.