Subject: Re: kern/28375: kernel panic upon pppoe reconnect
To: Martin Husemann <martin@duskware.de>
From: None <root@hydra.kaervek.net>
List: netbsd-bugs
Date: 11/21/2004 14:19:49
>> #5  0xc0102e39 in calltrap ()
>> #6  0xc02b8cdc in pppoe_start (ifp=0xc0a24800)
>>     at ../../../../net/if_pppoe.c:1413
>
> This means it crashes in the first call to pppoe_output(). Could you try
> adding a few printfs to that function and see how far it gets there?

Of course :)
There seems to be a bogus pointer passed to pppoe_out, here's my output

pppoe_output( *sc, *m): sc: 0xc0a50800, m: 0xc0940300
814 I'm here :)
821 I'm here :)
823 I'm here :)
827 I'm here :)
832 I'm here :)
842 I'm here :)
sc->sc_eth_if: 0xc09fc038
sc->sc_eth_if->if_output: 0xdeadbeef
kernel: page faul trap, code=0
stopped at 0xdeadbeef: invalid address
db>


> If you have another NetBSD machine and a serial connection, you could set
> up kgdb and manually step through it (contact me off list if you need
> instructions).

Unfortunately the other machine is linux (doubt that's a problem
though) and this netbsd laptop only has usb/parall ports :(

Ralf