Workaround rare deadlock with ristreceiver

Workaround a rare "deadlock" where a ristreceiver keeps requesting so much NACKS that the link becomes flooded, resulting in lots of late packets, buffers being full and never recovering.
Also fix a few segfaults, and lower loglevel of most per packet logs to DEBUG (saw 11 GB in 20min!!!)
5 jobs for workaround_flooded_deadlock in 1 minute and 15 seconds (queued for 2 seconds)