From owner-freebsd-afs@FreeBSD.ORG  Sat Oct  2 12:33:45 2004
Return-Path: <owner-freebsd-afs@FreeBSD.ORG>
Delivered-To: freebsd-afs@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 668E316A4CE
	for <freebsd-afs@freebsd.org>; Sat,  2 Oct 2004 12:33:45 +0000 (GMT)
Received: from citi.umich.edu (citi.umich.edu [141.211.133.111])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 43E4943D46
	for <freebsd-afs@freebsd.org>; Sat,  2 Oct 2004 12:33:45 +0000 (GMT)
	(envelope-from rees@citi.umich.edu)
Received: from citi.umich.edu (dsl093-001-248.det1.dsl.speakeasy.net
	[66.93.1.248])	by citi.umich.edu (Postfix) with ESMTP
	id 75F041BB05; Sat,  2 Oct 2004 08:33:44 -0400 (EDT)
To: port-freebsd@openafs.org, freebsd-afs@freebsd.org
From: Jim Rees <rees@umich.edu>
In-Reply-To: "David E. Cross", Sat, 02 Oct 2004 03:05:50 EDT
Date: Sat, 02 Oct 2004 08:33:43 -0400
Sender: rees@citi.umich.edu
Message-Id: <20041002123344.75F041BB05@citi.umich.edu>
Subject: Re: [OpenAFS-port-freebsd] Re: FreeBSD 5.2.1 Client Success! 
X-BeenThere: freebsd-afs@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: The Andrew File System and FreeBSD <freebsd-afs.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-afs>,
	<mailto:freebsd-afs-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-afs>
List-Post: <mailto:freebsd-afs@freebsd.org>
List-Help: <mailto:freebsd-afs-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-afs>,
	<mailto:freebsd-afs-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Sat, 02 Oct 2004 12:33:45 -0000

The FreeBSD client no longer has its own vnodes.  It uses the system pool
now.  But that has introduced a slew of new bugs.

Client bugs seem to fall into two broad classes, vnode ref count screwups,
and vnode locking screwups.  These usually manifest themselves as "unclean
vnode" panics, "locking against myself" panics, and hangs.  Turning on
WITNESS can help to track down some of them.  Part of this is due to the
code borrowed from OpenBSD, which has a different set of vnode locking
rules.

If anyone has patches, I urge you to send them here rather than fork the
source.