Re: panic: tcp_input: TCPS_LISTEN in netinet/tcp_input.c:2370

Hey Mohan,

On Tue, 21 Dec 2004 09:57:16 -0800 (PST), Mohan Srinivasan
<mohan_srinivasan_at_yahoo.com> wrote:
> Arjan,
> 
> This panic is different from the one you experienced on stable.
> This panic is from the tcp_input() path.
> 
> The panic you saw was from the tcp_output()->m_copym() path,
> which is likely the result of attempting to pull more data out
> of the snd buffer than is present there. I am pretty sure the
> panic you saw is TCP SACK related, but without a core dump
> (a dump of the SACK state), it would be impossible to fix the
> problem you saw.

Hmm... I'm not sure. See
http://www.piwebs.com/freebsd/pagefault-network.jpg - there's
ip_input, but no tcp_output. It's probably still a different panic
though :). I'll resume my hunt for a coredump on this machine when I'm
able to replace it with another one, I hope next week.

Arjan

> 
> mohan
> 
> --- Arjan Van Leeuwen <avleeuwen_at_gmail.com> wrote:
> 
> > On Mon, 20 Dec 2004 22:52:31 +0100, Peter Holm <peter_at_holm.cc> wrote:
> > > On Mon, Dec 20, 2004 at 10:16:00PM +0100, Andre Oppermann wrote:
> > > > Peter Holm wrote:
> > > > >On Mon, Dec 20, 2004 at 09:11:46PM +0100, Andre Oppermann wrote:
> > > > >
> > > > >>Peter Holm wrote:
> > > > >>
> > > > >>>During stress test with GENERIC HEAD from Dec 20 12:18 UTC I got:
> > > > >>>
> > > > >>>panic(c08374d0,8,c08f46e0,c1523170,3e0) at panic+0x190
> > > > >>>tcp_input(c2877900,14,2,c082b363,246) at tcp_input+0x2689
> > > > >>>ip_input(c2877900,18,c091a0b8,cbc7fcf4,c0681867) at ip_input+0xd6
> > > > >>>netisr_processqueue(c154a080,c154c180,c1523170,cbc7fd1c,c05ffa66)
> > > > >>>   at netisr_processqueue+0xf
> > > > >>>swi_net(0,0,0,c1554bd0,0) at swi_net+0x8b
> > > > >>>ithread_loop(c154c180,cbc7fd48,c154c180,c05ff8c8,0) at
> > > > >>>ithread_loop+0x19e
> > > > >>>fork_exit(c05ff8c8,c154c180,cbc7fd48) at fork_exit+0x7e
> > > > >>>fork_trampoline() at fork_trampoline+0x8
> > > > >>>
> > > > >>>http://www.holm.cc/stress/log/cons96.html
> > > > >>
> > > > >>Interesting.  Can you specify what kind of test was running at the
> > > > >>time of panic?
> > > > >>
> > > > >
> > > > >The tests from http://www.holm.cc/stress/src/stress.tgz
> > > >
> > > > Can you find out which of those test was doing the connect and listen?
> > > >
> > >
> > > That would be the "net" test.
> > >
> > > > >Let me know if you need any more info or gdb output.
> > > >
> > > > Yes, please give *th, *inp and *so as well.
> > > >
> > >
> > > I've updated http://www.holm.cc/stress/log/cons96.html
> >
> > This panic was also mentioned several times on stable_at_ (with
> > 5.3-RELEASE). I was unable to get a kernel dump when I posted to
> > stable, but I hope it can be fixed now that Peter has all the info!
> >
> > Best regards,
> >
> > Arjan
> >
> > >
> > > --
> > > Peter Holm
> > > _______________________________________________
> > > freebsd-current_at_freebsd.org mailing list
> > > http://lists.freebsd.org/mailman/listinfo/freebsd-current
> > > To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
> > >
> > _______________________________________________
> > freebsd-current_at_freebsd.org mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-current
> > To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
> >
> 
>