Re: NULL pointer dereference panic [resurrected]

Hello Robert!

Mon, Jun 19, 2006 at 09:29:25AM +0100 you wrote:

> Wojciech Koszek has been doing some work to debug a related pty/pts/devfs 
> problem, and mentioned to me a day or so ago that he may have identified a 
> workaround (and maybe a fix)?  If this panic is a result of the same or a 
> related problem, his work may be relevant.  I've CC'd him.

I've observed this same panic on a fresh(er) -CURRENT (and, certainly,
have a dump). This is on a UP system.

(kgdb) frame 14
#14 0xc06f5f24 in vop_stdgetwritemount (ap=0xd5c9aa2c)
    at /src/usr.src/sys/kern/vfs_default.c:356
356             vfs_ref(mp);
(kgdb) l
351              * will be set to NULL.  We only have to prevent this call from
352              * returning with a ref to an incorrect mountpoint.  It is not
353              * harmful to return with a ref to our previous mountpoint.
354              */
355             mp = ap->a_vp->v_mount;
356             vfs_ref(mp);
357             if (mp != ap->a_vp->v_mount) {
358                     vfs_rel(mp);
359                     mp = NULL;
360             }
(kgdb) print ap->a_vp->v_mount
$3 = (struct mount *) 0x0

Is there any progress in this area?

-- 
DoubleF
No virus detected in this message. Ehrm, wait a minute...
/kernel: pid 56921 (antivirus), uid 32000: exited on signal 9
Oh yes, no virus:)