Re: Panic in nd6_slowtimo()

On Friday 13 August 2004 07:41, George V. Neville-Neil wrote:
> At Fri, 13 Aug 2004 07:32:16 +0200,
>
> sebastian ssmoller wrote:
> > hi,
> > if i remember correctly i got this panic from time to time on my
> > workstation. what i did as a "work-around" i guess was to disable ipv6
> > (option INET6).
> >
> > the system runs stable now but i had to configure X11 to -nolisten tcp
> > which makes it difficult to start X11 progs from another user :)
> >
> > have a look at:
> > http://lists.freebsd.org/pipermail/freebsd-current/2004-January/019524.ht
> >ml
>
> I would be interested to know if this still happens.  We've been

Obviously it does. The original report mentions current as of yesterday (Aug, 
12th).

> trying to fix up the IPv6 code.  I am running IPv6 on several machines
> (CURRENT and STABLE) and have yet to see this.  I am also using
> neighbor discovers (nd) extensively.

Could you also load pf in one of these boxes, just to check if it triggers the 
panic?
Anyhow, the timeout is initialized unconditionally and it runs unconditionally 
over all interfaces present. At least that's my reading so far. The only 
problem I see is the mutex-trickery in if_attachdomain1. But as the original 
report indicates if_afdata_initialized was set to 1, so the last culprit can 
be in the domain.dom_ifattach() function.

-- 
/"\  Best regards,			| mlaier_at_freebsd.org
\ /  Max Laier				| ICQ #67774661
 X   http://pf4freebsd.love2party.net/	| mlaier_at_EFnet
/ \  ASCII Ribbon Campaign		| Against HTML Mail and News