Comment 4 for bug 1947588

I had a look at what's going on there. My understanding (with the caveat that the code of s_server is quite hard to follow, even within GDB) is that when the socket receives the packet, the server goes on and try to establish a connection, only to find out that it cannot because it has an inconsistent configuration (DTLS1 being disabled on seclevel 2 on Ubuntu), thus erroring out early, before it actually reads from the socket, thus triggering the loop all over again. This does not happen with TCP-based protocols, I assume because the underlying stream socket is closed (haven't checked the details though).

Fixing this cleanly would probably be a bit tricky (do we want to abort() the application? If not, what do we do with the incoming datagram?) but isn't very urgent either as it is an issue with the s_server code, which AIUI is a debugging tool.