From owner-freebsd-bugs Fri Nov 20 16:09:35 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA12009 for freebsd-bugs-outgoing; Fri, 20 Nov 1998 16:09:35 -0800 (PST) (envelope-from owner-freebsd-bugs@FreeBSD.ORG) Received: from troutmask.apl.washington.edu (troutmask.apl.washington.edu [128.95.76.54]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA12004 for ; Fri, 20 Nov 1998 16:09:34 -0800 (PST) (envelope-from kargl@troutmask.apl.washington.edu) Received: (from kargl@localhost) by troutmask.apl.washington.edu (8.9.1/8.9.1) id QAA00451; Fri, 20 Nov 1998 16:16:57 -0800 (PST) (envelope-from kargl) From: "Steven G. Kargl" Message-Id: <199811210016.QAA00451@troutmask.apl.washington.edu> Subject: Re: kern/8750 In-Reply-To: <199811190017.QAA13701@root.com> from David Greenman at "Nov 18, 1998 4:17:18 pm" To: dg@root.com Date: Fri, 20 Nov 1998 16:16:57 -0800 (PST) Cc: freebsd-bugs@FreeBSD.ORG 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-bugs@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org According to David Greenman: > >According to David Greenman: > >> Synopsis: SMP kernel panics and wipes out filesystems > >> > >> State-Changed-From-To: open-closed > >> State-Changed-By: dg > >> State-Changed-When: Wed Nov 18 16:04:22 PST 1998 > >> State-Changed-Why: > >> The bug was fixed by Kirk McKusick on the 17th. rev 1.74 of vfs_cluster.c. > >> > > > >Are you sure? (Yes, I know I shouldn't doubt you and Kirk). I > >saw Kirk's commit, and cvsup'd new sources. My system would > >still panic. However, my filesystems had been toasted, so perhaps, I > >had a corrupt /usr/src tree. > > > >Thanks for the quick response. I'm cvsup'ing new sources, now. > > The cause was fairly well understood and others confirmed that the bug > was fixed. 3 more crashes. I had rev 1.74 of vfs_cluster.c compiled into the kernel. The first dropped me into the debugger: dev=0x20404, bno = 6, bsize = 8192, size = 8192, fs = /var panic: ffs_blkfree: bad size mp_lock = 00000001; cpuid = 0; lapic.id = 01000000 Debugger("panic") Stopped at _Debugger+0x35: movb $0, _in_Debugger.98 db> trace _Debugger() at _Debugger+0x35 _panic() at _panic+0x9f _ffs_blkfree() at _ffs_blkfree+0xc2 _ffs_reallocblks() at _ffs_reallocblks+0x420 _cluster_write() at _cluster_write+0x15f _ffs_write() at _ffs_write+0x56f _vn_write() at _vn_write+0xef _write() at _write+0xba _syscall() at _syscall+0x187 _Xint0x80_syscall() at _Xint0x80_syscall+0x4c rm -rf /usr/src cvsup'd /usr/src on 19 Nov 98 at 0915. The next two crashed occurred while running X11R6, and the system froze. I could not break into the debugger. This is an SMP system with 2 processors. Is there a possible race condition due to SMP that Kirk's patch missed? troutmask:kargl[201] mount -v /dev/da0s1a on / (local, writes: sync 47 async 466) /dev/da1s1e on /ftp (local, writes: sync 2 async 0) /dev/da0s1g on /home (NFS exported, local, writes: sync 77 async 284) /dev/da0s1f on /usr (NFS exported, local, writes: sync 7 async 220) /dev/da0s1e on /var (local, writes: sync 209 async 496) procfs on /proc (local) -- Steve To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-bugs" in the body of the message