Re: pfctl -f causes fatal trap 12

On Sun, Feb 13, 2005 at 08:40:20AM +0100, Divacky Roman wrote:
> On Sun, Feb 13, 2005 at 11:51:44AM +1100, Emil Mikulic wrote:
> > On Sat, Feb 12, 2005, Divacky Roman top-posted:
> > > same here.... I have pf as module and see this panic on boot
> > > 
> > > when I reboot after the panic, fsck takes place and then it doesn
> > > panic (obviously the fsck changes something which causes pf not to
> > > panic)
> > 
> > Really?  I hadn't noticed that.  You should send your message to the
> > mailing list in case it can help one of the developers track down the
> > problem.
>  
> yes.... and oh, I forgot to cc: current (and I am doing it now)
> 
> > >From the kgdb output:
> > > > (kgdb) frame 11
> > > > #11 0xc0627fdc in softclock (dummy=0x0) at \
> > > > /usr/src/sys/kern/kern_timeout.c:315
> > > > 315                                             mtx_unlock(c_mtx);
> > > > (kgdb) print c_mtx
> > > > $1 = (struct mtx *) 0x0
> > 
> > It looks to me like that part of the code is trying to unlock an
> > uninitialized mutex.
> > 
> > I guess running fsck somehow causes the mutex to be initialized before
> > loading the pf rules breaks stuff.

This should be fixed now -- callout_init() was not being called.

-- 
Brian Fundakowski Feldman                           \'[ FreeBSD ]''''''''''\
  <> green_at_FreeBSD.org                               \  The Power to Serve! \
 Opinions expressed are my own.                       \,,,,,,,,,,,,,,,,,,,,,,\