Re: A minor problem...|
Karl Kleinpaste <karl,AT,charcoal,DOT,com>|
Thu, 16 Aug 2001 14:28:02 +0200|
Robert Davidson <puttputt,AT,ebbs,DOT,com,DOT,au> writes:
> Another thing I get on boot-up when CIPE fires up is cipcb0: route
> loop or something similar to that. Exactly what does that mean? I
> don't think there is a loop in my routing anywhere, I havn't spotted
> an obvious one yet anyway.
The diagnostic is "looped route", and it comes out of the cipe_xmit()
routine in output.c. It's an indication that the attempt at
forwarding the packet from the original cipcbN interface to the
carrier interface (eth0 or somesuch) has not resulted in finding such
a device at all -- the "new" device is still cipcbN.
This is indeed a routing problem. Now, perhaps there's some problem
with this in 2.4.8, but I'm using 2.4.3 successfully so unless there's
been some damage done between 2.4.3 and 2.4.8, I suspect you really do
have a simple routing problem, where the carrier interface addresses
aren't chosen correctly.
Take a close look at your routing table, both before and after you
(try to) start CIPE. Make sure that the "me" and "peer" carrier
addresses are properly routable as expected, and prove to yourself,
with ping or somesuch, that you really can reach the remote endpoint.