From owner-freebsd-questions  Thu May 20  3:10:29 1999
Delivered-To: freebsd-questions@freebsd.org
Received: from serenity.mcc.ac.uk (serenity.mcc.ac.uk [130.88.200.93])
	by hub.freebsd.org (Postfix) with ESMTP id 26EC814F2B
	for <questions@freebsd.org>; Thu, 20 May 1999 03:10:24 -0700 (PDT)
	(envelope-from ip@albatross.mcc.ac.uk)
Received: from albatross.mcc.ac.uk ([130.88.202.16])
	by serenity.mcc.ac.uk with esmtp (Exim 1.92 #3)
	id 10kPmQ-000MFK-00; Thu, 20 May 1999 11:10:22 +0100
Received: (from ip@localhost)
	by albatross.mcc.ac.uk (8.9.3/8.9.1) id LAA14316;
	Thu, 20 May 1999 11:10:17 +0100 (BST)
	(envelope-from ip)
From: Ian Pallfreeman <ip@albatross.mcc.ac.uk>
Message-Id: <199905201010.LAA14316@albatross.mcc.ac.uk>
Subject: Re: panic: ffs_valloc: dup alloc & speaking klingon
In-Reply-To: <Pine.BSF.4.03.9905191600510.9085-100000@resnet.uoregon.edu> from Doug White at "May 19, 1999  4: 7:58 pm"
To: dwhite@resnet.uoregon.edu (Doug White)
Date: Thu, 20 May 1999 11:10:17 +0100 (BST)
Cc: questions@freebsd.org
Reply-To: ip@mcc.ac.uk
X-Mailer: ELM [version 2.4ME+ PL43 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-questions@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

On Wed 19 May 1999, Doug White wrote:
> On Wed, 19 May 1999, Ian Pallfreeman wrote:
>[...]
> > 	additional daemons: syslogd.
> > 	checking for core dump...savecore: warning: /kernel version mismatch:
> > 	        FdreeBSD 3.2-STABLsE #2: Tue May 18c 12:09:01 BST 19h99
> > 	csavecore: rebootk
> > 	May 19 11:18:5:9 curlew savecor e: reboot
> > 	savencore: system wenet down at Wed Magy 19 11:15:22 19a99
> > 	savecore: /vtar/crash/bounds:i No such file orv directory
> 
> This is fun.

Indeed. I've figured what it's trying to tell me. This is a combination of
the output from savecore(8), mixed up with a kernel message:

	dscheck: negative b_blkno -1047972

I've traced dscheck to sys/kern/subr_diskslice.c, but am none the wiser.
Clutching at straws, I do have a reasonably large swap partition on that
box:

	Device      1K-blocks     Used    Avail Capacity  Type
	/dev/da0s1b   3932462        0  3932334     0%    Interleaved

I'm able to dd it to /dev/null without provoking complaints from dscheck, so
I guess it isn't some legacy 32-bit code which should be off_t's. Another
theory down the pan. :-(

> The dup alloc is a bit suspicious.  Read the section on kernel debugging
> in the Handbook and try building a kernel with 'options DDB' so you can
> peg this in person.

Aww, no, not again. :-(

The dup alloc is _very_ suspicious. I've pulled a 3.1-STABLE kernel from late
April back off a tape, and that has stayed up all night. 3.2-STABLE will die
with the dup alloc within minutes. But that's another problem.

Cheers,

Ian.


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-questions" in the body of the message