Subject: Re: PR/32462 CVS commit: src/sys
To: None <yamt@netbsd.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org,>
From: Reinoud Zandijk <reinoud@netbsd.org>
List: netbsd-bugs
Date: 01/06/2006 15:00:04
The following reply was made to PR kern/32462; it has been noted by GNATS.
From: Reinoud Zandijk <reinoud@netbsd.org>
To: YAMAMOTO Takashi <yamt@mwd.biglobe.ne.jp>
Cc: gnats-bugs@netbsd.org, gnats-admin@netbsd.org,
netbsd-bugs@netbsd.org
Subject: Re: PR/32462 CVS commit: src/sys
Date: Fri, 6 Jan 2006 15:59:22 +0100
Dear Takashi,
On Fri, Jan 06, 2006 at 10:32:50PM +0900, YAMAMOTO Takashi wrote:
> > Running with DIAGNOSTIC+DEBUG+LOCKDEBUG, the kernel immediately crashes
> > when reading when it starts syslogd:
>
> what do you mean "when reading when it starts syslogd"?
>
> both of panics were with lfs used on the system?
no, this last panic was when booting from an FFS partion (all is FFS here,
only got a vnd with LFS). During the startup of the system when it just
printed `starting syslogd' it panic'd.
I'll have to excuse you... it turns out that the kernel propably is
somewhere due to my patch. A bog-standard kernel seems to run fine. I'm
looking into this now. It runs fine though when DIAGNOSTICS is on it
somehow failes a check.
> > also the `bp->b_vp = NULL;' would help on other places.
> other places?
I could imagine other places would have this hidden dependency on bp->b_vp
being NULL when they request it. It was meant as a reference to the
suggestion to put this statement in the BUF_INIT() macro.
Cheers,
Reinoud