Date: Wed, 19 Mar 1997 00:18:21 -0600 From: Jonathan Lemon <jlemon@americantv.com> To: freebsd-current@freebsd.org Subject: Panics on reboot/halt/shutdown with MFS Message-ID: <19970319001821.63996@right.PCS> In-Reply-To: <19970318212528.28756@klemm.gtn.com>; from Andreas Klemm on Mar 03, 1997 at 09:25:28PM %2B0100 References: <199703131421.GAA28883@freefall.freebsd.org> <m0w6jY4-000SF6C@olivaw.olivetti.nl> <19970318081322.59650@klemm.gtn.com> <m0w735q-000SKuC@olivaw.olivetti.nl> <19970318212528.28756@klemm.gtn.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mar 03, 1997 at 09:25:28PM +0100, Andreas Klemm wrote:
> > I noticed that if I used shutdown I got a message that some process
> > refused to die. The only process which might cause it was mount_mfs.
> > I just disabled mounting an mfs on /tmp and now I can halt cleanly.
> > Are you using mfs ??? If so try disabling it.
>
> Well, yes, I use mfs ... but ... who or what broke it in 2.2.0 and 3.0 ?
> I'm using mfs for months and had no trouble.
I spent some time digging into this. The symptoms are that the system
panics/locks up when executing reboot() when there is a MFS mounted
filesystem. (freebsd bugs #3012, #3026)
When MFS is not mounted, the system shuts down without a problem.
The culprit appears to be this line in spec_strategy():
(*bdevsw[major(ap->a_bp->b_dev)]->d_strategy)(ap->a_bp);
major(ap->a_bp->b_dev) == 0xff. d_strategy for this bdevsv[] == 0.
#0 spec_strategy (ap=0xefbffe64) at ../../miscfs/specfs/spec_vnops.c:509
509 {
#1 0xf012e2a8 in bwrite (bp=0xf3a23024) at vnode_if.h:1116
1116 return (VCALL((bp)->b_vp, VOFFSET(vop_strategy), &a));
#2 0xf019599e in ffs_sbupdate (mp=0xf1a20c00, waitfor=1)
at ../../ufs/ffs/ffs_vfsops.c:1057
1057 error = bwrite(bp);
#3 0xf019524c in ffs_unmount (mp=0xf1a0d200, mntflags=524288, p=0xf1a20800)
at ../../ufs/ffs/ffs_vfsops.c:672
672 ffs_sbupdate(ump, MNT_WAIT);
#4 0xf0134f1c in dounmount (mp=0xf1a0d200, flags=524288, p=0xf1a20800)
at ../../kern/vfs_syscalls.c:309
309 error = VFS_UNMOUNT(mp, flags, p);
#5 0xf0132f79 in vfs_unmountall () at ../../kern/vfs_subr.c:245
245 error = dounmount(mp, MNT_FORCE, initproc);
#6 0xf0114e1f in boot (howto=8) at ../../kern/kern_shutdown.c:217
217 vfs_unmountall();
Perhaps someone with a CVS tree available could look at the recent changes
that were made to 2_2_RELENG? This only just started happening.
--
Jonathan
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970319001821.63996>