Quoting Stefano Corti ([EMAIL PROTECTED]):
> Yes, that patch seems to solve the 'oops' generated by downing an ax25
> interface,
> *BUT*
> every time an ax25 connection is cleared (at user level) a stale definition
> in the kernel keeps the ax25 connection in a 'disconnected' state. So you
> can't connect that station anymore, and the system needs to be rebooted.
Oops.
Ok, I must admit that I did not really knew how the kernel-ax25 works,
it just looked reasonable because the other timers are stopped in this
function, too. Ok, I think ax25_stop_heartbeat() has to be called, but not
in ax25_disconnect() which is called at normal disconnects too, but only
when the interface is shut down. perhaps in ax25_kill_by_device in
ax25_dev.c? Can people who have some more knowledge than me comment
on that please?
Sorry, hp
