Multiple Udp sockets corrupt the w5500 device

[elpiel]

So, implementing some changes in #42 to the UDP and testing with multiple sockets at once makes me to believe that there is something wrong with the way the TX buffer (maybe even RX buffer) are used.

Pseudo code:

remote_addr = ...
socket_1 = w5500_device.socket().unwrap();
device.connect(socket_1, remote_addr)

socket_2 = w5500_device.socket().unwrap()
loop {
    socket_1.send(...);
   timeout(1/2 second)
   socket_2.send(...)
   timeout(1/2 second)
}

What ends up happening in this scenario is that the socket_2 gets into the infinite loop that waits for a SendOk or WriteTimeout and never receives one.

What I've managed to track so far is when I set the new TX write pointer and directly read it back before Command::Send, instead of having the newly set pointer it once again has 0.

Another (maybe unrelated to this one issue) is that at some point (even if I comment out the socket_2 code) I started to get WriteTimeout from the first socket_1.send().
This leads me to believe that it was left in an invalid state and flashing over and over didn't help with this issue either.
It did require a device reset (power down + up) to get it back working.

API improvement is needed

The main issues that were causing this problem were 2:

1. When Device.connect is not called, Udp socket is not being opened #46
2. When no destination is set and you call send - it should at least return an error if there's no destination remote address set

[ryan-summers]

@elpiel I think this issue is superceded by #46 and #59, is that correct? Multiple sockets works correctly after #57?

[elpiel]

Yes, I've tested it with 2 sockets