Re: lockdebug panic when running fsck_ffs

[Sverre Froyen <sverre%viewmark.com@localhost>]

On Thu June 11 2009 11:36:55 Tonnerre LOMBARD wrote:
> Salut,
>
> On Wed, Jun 10, 2009 at 10:49:25PM +0100, Patrick Welche wrote:
> > Do you have NFSSERVER in your kernel config?
>
> Yes, removing it removed the panic when invoking fsck. Odd.
>
>                               Tonnerre

My apologies for following up to this message, but I wanted to point out that 
I see the error too (current amd64).  Without LOCKDEBUG defined.  It would 
good if the problem could be fixed.  Christoph Egger has proposed a fix:

http://mail-index.netbsd.org/tech-kern/2009/06/14/msg005238.html

which is correct if no method called via

        vfs_hooks_reexport(mp, path, data);

assumes that the mp->mnt_updating mutex is already held by the caller.  Since 
I don't know if this is true, I'm worried about simply adding the patch to my 
local tree.

Comments?

Regards,
Sverre