Re: New possible exploit for 2.0.33 (kfree_skb error)

[alan () LXORGUK UKUU ORG UK (Alan Cox)]

> This could be the sign of a new exploit. I have also managed to find a few
> others on IRC that have had the same exact error message. If anyone knows
> what exactly causes it and how to fix it please let us know.   The
> "syndrop" program generates the kfree error but it does not crash the box
> (at least the one that I have tested)

I've been given 2 copies of a syndrop program, and right now I can make
neither break my boxes. Which is proving in this case troublesome to fixing
things. If its directly syndrop related then firewalling packets from
your source address incoming via external interfaces should block it.
Remember also to cover 127.*

> 0286a554 00000000 077a8958 00000000
> Apr 14 23:51:36 web1 kernel:        077a8958 03bd8e18 00000040 00000040
> 0286a528 0286a554 00142661 077a8958
> Apr 14 23:51:36 web1 kernel:        0000002e 0000cb0c 0000cb3c 001b0008
> 00000000 03bd8e18 0000002c 00000014

Below this there should be a function trace, although your stack trace
looks a bit weird in itself. The call trace is important.

Im looking into these but the more info I get the better

Alan