From owner-freebsd-bugs Wed Jul 19 6:40:10 2000 Delivered-To: freebsd-bugs@freebsd.org Received: from freefall.freebsd.org (freefall.FreeBSD.ORG [204.216.27.21]) by hub.freebsd.org (Postfix) with ESMTP id C830637B6A7 for ; Wed, 19 Jul 2000 06:40:05 -0700 (PDT) (envelope-from gnats@FreeBSD.org) Received: (from gnats@localhost) by freefall.freebsd.org (8.9.3/8.9.2) id GAA79352; Wed, 19 Jul 2000 06:40:05 -0700 (PDT) (envelope-from gnats@FreeBSD.org) Date: Wed, 19 Jul 2000 06:40:05 -0700 (PDT) Message-Id: <200007191340.GAA79352@freefall.freebsd.org> To: freebsd-bugs@FreeBSD.org Cc: From: Sheldon Hearn Subject: Re: kern/16740: panic: ffs_clusteralloc: map mismatch Reply-To: Sheldon Hearn Sender: owner-freebsd-bugs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org The following reply was made to PR kern/16740; it has been noted by GNATS. From: Sheldon Hearn To: freebsd-gnats-submit@FreeBSD.org Cc: Subject: Re: kern/16740: panic: ffs_clusteralloc: map mismatch Date: Wed, 19 Jul 2000 15:34:22 +0200 ------- Forwarded Message Date: Wed, 19 Jul 2000 15:16:12 +0200 From: Joakim Henriksson To: Sheldon Hearn Subject: Re: misc/20031: kernel randomly panics with ffs_clusteralloc: map mismatch Message-Id: <200007191316.PAA08680@rmstar.campus.luth.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii > > kern/16740. I can make available a crashdump from today if you > > wish. Just tell if ftp or http or scp is most convenient. I will need > > the hostname you wish to access from also. My host access file is > > pretty restrictive. > > I won't be able to solve this myself, but I might know the right person > to approach for help. This would be appreciated, i have had my machine panic twice(!) today due to this bug. And all i've got are these lousy t-sh^D^D^D^Dcrashdumps. > The Environment section of PR 16740 indicates that the panics you > received were on a 4.0-CURRENT box in February. I think you said you're > now using 4.1-RC, which means you're tracking the RELENG_4 stable > branch, right? You bet! > If that's the case, one thing I can think of is that you may have > _copied_ the softupdates source files instead of creating symbolic links > for them. Of course, now we don't use symlinks, because the soft > updates license has changed. Neg on that. I did have the links and got bitten by the old "cvsup don't know what to do with symlinks trick" > Could you please verify that the following files on your filesystem are > regular files? Also, please check the versions that you have, as per > the version numbers in brackets. You can use the ident(1) utility to > find the RCS/CVS version of a file. > > /usr/src/sys/ufs/ffs/ffs_softdep.c [1.57.2.1] > /usr/src/sys/ufs/ffs/ffs_softdep_stub.c [1.7] > /usr/src/sys/ufs/ffs/softdep.h [1.7.2.1] murduth@rmstar /usr/src/sys/ufs/ffs >ident ffs_softdep.c ffs_softdep.c: $FreeBSD: src/sys/ufs/ffs/ffs_softdep.c,v 1.57.2.1 2000/06/22 19:27:42 peter Exp $ murduth@rmstar /usr/src/sys/ufs/ffs >ident ffs_softdep_stub.c ffs_softdep_stub.c: $FreeBSD: src/sys/ufs/ffs/ffs_softdep_stub.c,v 1.7 2000/01/10 00:24:22 mckusick Exp $ murduth@rmstar /usr/src/sys/ufs/ffs >ident softdep.h softdep.h: $FreeBSD: src/sys/ufs/ffs/softdep.h,v 1.7.2.1 2000/06/22 19:27:42 peter Exp $ murduth@rmstar /usr/src/sys/ufs/ffs > So nothing deviant about this... > I've made the assumption that you are actually using soft updates. Are > you? :-) Yep, but i seem to remember turning them of to see if that helped. And it didn't if i remember correctly. It's been a couple of months since this test. > I'm not too interested in PR 20031, since it relates to an old release > on the RELENG_3 branch. I'm much more interested in your PR, especially > if you can confirm that the panics occur on a recent RELENG_4 system. I would think that it occurs in 5.0-CURRENT too, but i'm too much of a wuss to try it out. What with all the controversia about the interrupt threads ;) Othermail question > Oops, I forgot another thing. Some mail I saw in the archives suggested > that this might be cause by weird compiler optimizations. Are you using > COPTFLAGS other than "-O -pipe"? murduth@rmstar /usr/src/sys/compile/RMSTAR #cat /etc/make.conf RSAREF=NO USA_RESIDENT=NO MASTER_SITE_OVERRIDE=ftp:/ftp.se.freebsd.org/pub/FreeBSD/ports/distfiles/${DIST _SUBDIR}/ murduth@rmstar /usr/src/sys/compile/RMSTAR # I always build my kernel with "make depend all" so nothing strange here either. John Baldwin mailed freebsd-fs about this last week or so and he had the same chipset in the mobo. My guess is that there is a race, bug or something in the ata driver. But i can't substansiate it. We'll see what the originator of 20031 says. I mailed him about his chipset. - -- regards/ Joakim ------- End of Forwarded Message To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message