Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 2 Jul 2001 23:04:22 -0700
From:      "Robert L Sowders" <rsowders@usgs.gov>
To:        Tor.Egge@fast.no
Cc:        freebsd-stable@FreeBSD.ORG, ohartman@klima.physik.uni-mainz.de, owner-freebsd-stable@FreeBSD.ORG
Subject:   Re: HELP! Server crashes since last cvsupdate!
Message-ID:  <OF3C932387.AE1F7911-ON88256A7E.0020AD8E@wr.usgs.gov>

next in thread | raw e-mail | index | archive | help
Ok, Now I really confused,

Checking back thru cvsweb it seems that there is some confusion going on, 
or maybe it's me, but the mfs_vfsops.c version 1.81.2.1 is dated 
2001/06/17 at 5:05 and version 1.82, is dated 2000/04/15 03:34:21.  Seems 
to be a slight mixup here, It's marked as being a year earlier that the 
latter version.  Hope this URL doesn't wrap.
http://www.FreeBSD.org/cgi/cvsweb.cgi/src/sys/ufs/mfs/Attic/mfs_vfsops.c.diff?r1=1.82&r2=1.81.2.1&f=h





Tor.Egge@fast.no
Sent by: owner-freebsd-stable@FreeBSD.ORG
07/02/2001 04:55 PM

 
        To:     ohartman@klima.physik.uni-mainz.de
        cc:     freebsd-stable@FreeBSD.ORG
        Subject:        Re: HELP! Server crashes since last cvsupdate!

> Hello.
> 
> Since our last update Friday, 29th June, both SMP machines run
> into a "stuck" condition after a while. This happened now two times
> and I do not know what happens.

I updated one machine from a 22th June -stable kernel and immediately
got a 'hang' at boot when starting sendmail.  Breaking out to ddb
showed the following backtrace.

db> trace
SMP_breakpoint(c44d4c00,c46d3200,0,c0273d5b,f0f60) at SMP_breakpoint+0x1e
siointr1(c44d4c00,c03c3008,0,e6442f3c,c027258a) at siointr1+0x13d
siointr(c44d4c00) at siointr+0x17
Xfastintr4(c46d3200,0,e6442f58,c01ac825,e6442f90) at Xfastintr4+0x2a
vop_noislocked(e6442f90,e6442f9c,c01af24b,e6442f90,0) at 
vop_noislocked+0x19
vop_defaultop(e6442f90) at vop_defaultop+0x15
sched_sync(0) at sched_sync+0xaf
fork_trampoline() at fork_trampoline+0x30


The machine hung in lockstatus (called from vop_noislocked) trying to
obtain the interlock of what lockstatus believed to be a struct lock
but wasn't.  A suggested fix is enclosed.

- Tor Egge

Index: sys/ufs/mfs/mfs_vfsops.c
===================================================================
RCS file: /home/ncvs/src/sys/ufs/mfs/Attic/mfs_vfsops.c,v
retrieving revision 1.81.2.1
diff -u -2 -r1.81.2.1 mfs_vfsops.c
--- sys/ufs/mfs/mfs_vfsops.c             2001/06/17 05:05:00 1.81.2.1
+++ sys/ufs/mfs/mfs_vfsops.c             2001/07/02 23:28:25
@@ -249,4 +249,5 @@
                 addaliasu(devvp, makeudev(253, mfs_minor++));
                 devvp->v_data = mfsp;
+                lockinit(&mfsp->mfs_lock, PVFS, "mfslck", 0, LK_NOPAUSE);
                 mfsp->mfs_baseoff = args.base;
                 mfsp->mfs_size = args.size;
Index: sys/ufs/mfs/mfsnode.h
===================================================================
RCS file: /home/ncvs/src/sys/ufs/mfs/Attic/mfsnode.h,v
retrieving revision 1.12
diff -u -2 -r1.12 mfsnode.h
--- sys/ufs/mfs/mfsnode.h                1999/08/28 00:52:26 1.12
+++ sys/ufs/mfs/mfsnode.h                2001/07/02 23:28:25
@@ -43,4 +43,5 @@
 
 struct mfsnode {
+                struct lock mfs_lock;
                 struct          vnode *mfs_vnode;               /* vnode 
associated with this mfsnode */
                 caddr_t                 mfs_baseoff;  /* base of file 
system in memory */




To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?OF3C932387.AE1F7911-ON88256A7E.0020AD8E>