Hi,

I keep on running into a strange problem with the bridge-netfilter
0.0.8 patch against 2.4.19, applied against 2.4.20 (it applies with
a few hunks, but it applies) .

The machine is a sun4m (sparc10) with two lance eth cards. The
2.4.20 kernel it is running as well as in particular the ethernet
bridge works quite well (and, as it appears, as stable as with
2.4.19) .

However with the bridge-nf filter applied the machine keeps oopsing
(Unable to handle kernel paging request in mna handler<1> at virtual
address ...) as soon as the bridge gets some traffic and the machine
starts swapping because some user space process starts taking memory
The machine has only 32mb ram and it's perfectly reproducible for me
when putting traffic on the bridge (outside traffic, just forwarded)
and running say 'dpkg -l' or anything else causing swapping.

The call trace resolved with ksymoops does not point to any
particular code unfortunately, or rather: the place where the
exception handler gets called is random, so I suspect it to be the
result of memory corruption (I wish there would be valgrind for the
kernel :) .

But given that the problem is perfectly reproducible when the
bridge-nf patch is applied (and not otherwise) I suspect the bug to
be in there.

As a sidenote it is reproducible independent of any iptables rules
defined.

Now I know, this isn't a very helpful bugreport due to the lack of
call trace pointing right into the bridging code :} . But I wonder
if maybe anyone else is experiencing similar problems or can
reproduce the problem?


Simon
_______________________________________________
Bridge mailing list
[EMAIL PROTECTED]
http://www.math.leidenuniv.nl/mailman/listinfo/bridge

Reply via email to