From owner-freebsd-hackers Sun Jun 13 10: 6:25 1999 Delivered-To: freebsd-hackers@freebsd.org Received: from apollo.backplane.com (apollo.backplane.com [209.157.86.2]) by hub.freebsd.org (Postfix) with ESMTP id D4A5C150B5 for ; Sun, 13 Jun 1999 10:06:23 -0700 (PDT) (envelope-from dillon@apollo.backplane.com) Received: (from dillon@localhost) by apollo.backplane.com (8.9.3/8.9.1) id KAA04284; Sun, 13 Jun 1999 10:06:20 -0700 (PDT) (envelope-from dillon) Date: Sun, 13 Jun 1999 10:06:20 -0700 (PDT) From: Matthew Dillon Message-Id: <199906131706.KAA04284@apollo.backplane.com> To: David Scheidt Cc: Bernd Walter , freebsd-hackers@FreeBSD.ORG Subject: Re: softupdates problem? References: Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG :On Sun, 13 Jun 1999, Bernd Walter wrote: : :> On Sat, Jun 12, 1999 at 12:13:43AM -0500, David Scheidt wrote: :> > I had a 3.2 stable (from 30 May 1999)machine panic tonight, trying :> > to load the oss driver, which is not too shocking. What was shocking :> > was the damage done to my filesystem. The automatic fsck failed, :> > with an UNEXPECTED SOFT UPDATES INCONSISTNCY. PARTIALLY ALLOCATED :> > INODE I=39684. :> > :> What kernel-config are you using? :> I have had several fs-crashes because of a to high configured maxusers. : :The only diffs between my config file and GENERIC are: :1. softupdates :2. no support for hardware I haven't got :3. pty is 64 : :I am not actually complaining about the panic; I think the oss :driver is a piece of crap. I just happened to have a soundcard :that they support that the FreeBSD drivers don't, and a .wav file :I wanted to listen to. What does disturb me is the damage done to :a pretty quiet filesystem. I wouldn't have thought that toasting :a whole bunch of files that were not in use would be the normal :behavior. I have seen lots of panics, but have never seen filesystem :damage as bad as this. : :David Scheidt This smells like a memory corruption problem rather then a softupdates problem. i.e. that something in the kernel overwrote random kernel memory which happened to correspond to a filesystem buffer. There was a buffer corruption bug found in VFS/BIO recently, but it was fixed prior to the 3.2 release ( it was an getnewbuf() race, fixed in 1.193.2.4 of kern/vfs_bio.c ). -Matt Matthew Dillon To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message