strange logs (0.1.8)

[Adam Kumiszcza]

On Sat, Sep 18, 1999 at 02:47:03PM +0930, Paul Rusty Russell wrote:
> In message <19990917063947.A875@3net.net.pl> you write:
> > I got this in my logs:
> > 
> > Sep 17 06:33:22 qman kernel: nf_iterate: NF_STOLEN for 00000000.
> > Sep 17 06:33:22 qman kernel: nf_iterate: NF_STOLEN for 00000000.
> > Sep 17 06:33:22 qman kernel: ip_finish_output: bad unowned skb =
> > c39bd620: LOCAL_OUT POST_ROUTING
> > Sep 17 06:33:22 qman kernel: skb: pf=2 (unowned) dev=lo len=8133
> > Sep 17 06:33:22 qman kernel: PROTO=17 127.0.0.1:1027 127.0.0.1:1031
> > L=8133 S=0x00 I=3537 F=0x0000 T=64
> 
> Connection tracking (NAT) and local fragments: I need to fixup the
> nf_debug headers for this case... if it's not a problem for you, you
> can ignore it.

I didn't notice any problems, so it's probably harmless. BTW:
cat /var/log/syslog|grep "LOCAL_OUT POST_ROUTING"|wc -l
     32

so it's not so uncommon... 

> 
> It's a very large UDP packet going from port 1031 to 1027; I'd say
> maybe NFS, except the ports are a bit wierd.
>

I just checked and found some leftouts of the
testsuite (btw, it failed -- I'll check again and let you know what
exactly was the problem later today) -- tap0, tap1 and ip_conntrack.
Maybe it produced some unwanted udp? I don't use NAT in that moment at
all, and, as for NFS, I have the kernel server installed (with userspace
support), but I don't use it (/etc/exports is empty).

Strange thing is that all of these 32 appearances of "LOCAL_OUT..." in
my logs (except the first two) happend just after (in a matter of
several seconds) after the connection appeared. And, to support the
theory of testsuite leftouts, I checked that the first two logs happened
just after starting the testsuite (many nf_register... before).

Adam Kumiszcza
<qman@from.pl>