Replying to my onw messages.. Hmm.. Oh well..
> OK I was happy, until I noticed that the TX and RX lights on my break-out box
> are burning brightly after I send some traffic across the line. Routing
> problem? (loop?).. Checked that out. Doesn't seem to be that way. Our side has
> one static network-route for the customers class C address and the customer
> has a default route (using 0.0.0.0) to our PM. Sounds reasonable..
Found out what the problem is...
The cisco software (hacked comOS) _does_ look at the netmask of a PPP (or SLIP)
link, unlike the PM. So when I assigned adresses to both ends of the link from
the customer's address space the cisco sent every packet for the ethernet to the
PM as well because it assumes a 255.255.255.0 netmask on that link. The PM then
happily routed the packet back and the loop started all over again...
Aha! Now I've set the link up with two adresses from a separate address block
and now it's working!
This is the final solution:
customer address block --- cisco --- intermediate addresses --- PM --- our address block
The intermediate adresses are not taken from either network.
Perhaps this helps people out who run into the same kind of problem..
Bye Arno.