Re: panic and fs corruption in -current

On 8/7/07, Kris Kennaway <kris_at_obsecurity.org> wrote:
>
> On Tue, Aug 07, 2007 at 09:11:46PM +0200, A. Rios wrote:
> > On 8/5/07, A. Rios <cosasvarias_at_gmail.com> wrote:
> > >
> > > I compiled yesterday's kernel without problems, and it restarted fine.
> I
> > > applied jeff roberson's ulehtt.diff (P4 with hyperthreading and 2GB
> RAM
> > > here)
> > > and disabled all kernel debugging (I was testing performance)
> > >
> > > Earlier this morning I was portupgrading and everything start to fail.
> One
> > > of the
> > > (random) panics was:
> > >
> > >
> > Ok, now I have a kernel with debug symbols and I've found a way to
> reproduce
> > the panic. It occurs when installing a port and after the message
> "SHA256
> > checksum OK for port XXXXX", at this point the system freezes and
> sometimes
> > it dumps the memory and reboot itself, and other times it simply freezes
> and
> > I have to restart the computer. I think the problem came from somewhere
> of
> > the gcc/make toolchain but I don't know where to start.
>
> I would guess that your filesystem remains corrupted from a previous
> panic and you need to offline and fsck -f it.
>
> Kris
>

Ok, thanks a lot, that was the problem. I didn't have a filesystem, I had a
hell of unreferenced files and nodes... I don't know how even it started.