[etherlab-users] UNMATCHED and SKIPPED datagrams

[Erwin Burgstaller]

* Wed, Jan 14, 2009 - 12:01:46 +0100, Florian Pose - fp at igh-essen.com:
> > [57061.710532] EtherCAT DEBUG: skipping datagram f4d85640.
                                                     ^^^^^^^^
> 
> skipped datagrams mean, that the datagram is sent again, although the
> answer to the last sending hasn't been received yet. This usually means
> that the realtime cycle was executed too short after the last one (maybe
> because the last one was delayed for some reason).

It seems the problem is there are other rtai tasks running. As they have
the same priority, it might happen that my module is delayed too long,
so I'll try to lower the other tasks priority.

But, what's strange, if that happens once, the system never gets right
anymore, even if the other tasks are removed.

Is it correct, that the same datagram (f4d85640 in that very case) is
skipped for several hours? Or could there be a problem in the masters
datagram queue?


Thanks,
  Erwin

-- 
Erwin Burgstaller