From: Sridhar Samudrala <sri@us.ibm.com>
Subject: Re: [PATCH] oops when unloading sunrpc module.
Date: Mon, 28 Jun 2004 12:25:41 -0700 (PDT)
Sender: nfs-admin@lists.sourceforge.net
Message-ID: <Pine.LNX.4.58.0406281209140.2385@localhost.localdomain>
References: <Pine.LNX.4.58.0406241542480.2427@localhost.localdomain>
 <200406281409.49253.vincent.roqueta@ext.bull.net>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=ISO-8859-1
Cc: nfs@lists.sourceforge.net
Return-path: <nfs-admin@lists.sourceforge.net>
Received: from sc8-sf-mx1-b.sourceforge.net ([10.3.1.11] helo=sc8-sf-mx1.sourceforge.net)
	by sc8-sf-list2.sourceforge.net with esmtp (Exim 4.30)
	id 1Bf1lp-0004ZU-Gv
	for nfs@lists.sourceforge.net; Mon, 28 Jun 2004 12:26:25 -0700
Received: from e5.ny.us.ibm.com ([32.97.182.105])
	by sc8-sf-mx1.sourceforge.net with esmtp (TLSv1:DES-CBC3-SHA:168)
	(Exim 4.34)
	id 1Bf1lo-0003qc-LK
	for nfs@lists.sourceforge.net; Mon, 28 Jun 2004 12:26:25 -0700
To: Vincent ROQUETA <vincent.roqueta@ext.bull.net>
In-Reply-To: <200406281409.49253.vincent.roqueta@ext.bull.net>
Errors-To: nfs-admin@lists.sourceforge.net
List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/nfs>,
	<mailto:nfs-request@lists.sourceforge.net?subject=unsubscribe>
List-Id: Discussion of NFS under Linux development,
	interoperability,
	and testing. <nfs.lists.sourceforge.net>
List-Post: <mailto:nfs@lists.sourceforge.net>
List-Help: <mailto:nfs-request@lists.sourceforge.net?subject=help>
List-Subscribe: <https://lists.sourceforge.net/lists/listinfo/nfs>,
	<mailto:nfs-request@lists.sourceforge.net?subject=subscribe>
List-Archive: <http://sourceforge.net/mailarchive/forum.php?forum=nfs>

On Mon, 28 Jun 2004, Vincent ROQUETA wrote:

> Le Vendredi 25 Juin 2004 00:49, Sridhar Samudrala a =E9crit :
> > I am seeing the following warning message followed by a oops when unloa=
ding
> > sunrpc module in linux 2.6.7. This happens only if lock operations are
> > performed on a file that is mounted on the client before unloading nfs
> > modules.
> >
> > Badness in remove_proc_entry at fs/proc/generic.c:685
> >  [<c018e799>] remove_proc_entry+0x109/0x150
> >  [<d091dfdc>] rpc_proc_exit+0x3c/0x50 [sunrpc]
> >  [<c0134386>] sys_delete_module+0x176/0x1b0
> >  [<c014c2d8>] do_munmap+0x178/0x1e0
> >  [<c010462b>] syscall_call+0x7/0xb
> >
> > Unable to handle kernel paging request at virtual address d0929c74
> >  printing eip:
> > c01258d0
> > *pde =3D 0fdc0067
> > *pte =3D 00000000
> > Oops: 0000 [#1]
> > PREEMPT SMP
> > Modules linked in: netconsole 3c59x e100
> > CPU:    0
> > EIP:    0060:[<c01258d0>]    Not tainted
> > EFLAGS: 00010006   (2.6.7)
> > EIP is at cascade+0x30/0x70
> > eax: cb5afeb8   ebx: d0929c58   ecx: 0009c000   edx: c1202e04
> > esi: c12030e4   edi: c12025a0   ebp: 00000027   esp: c0383f10
> > ds: 007b   es: 007b   ss: 0068
> > Process swapper (pid: 0, threadinfo=3Dc0382000 task=3Dc030b180)
> > Stack: c12025a0 cfdb01f0 c12025a0 00000000 0000000a c0382000 c0125fa7
> > c12025a0 c1202fac 00000027 c030dc8c c0382000 c0383f40 c0383f40 00000000
> > 00000001 c0381008 0000000a c0383f94 c0121527 c0381008 00000046 00000000
> > c03a70a4 Call Trace:
> >  [<c0125fa7>] run_timer_softirq+0x197/0x1e0
> >  [<c0121527>] __do_softirq+0xb7/0xc0
> >  [<c012155d>] do_softirq+0x2d/0x30
> >  [<c0111a47>] smp_apic_timer_interrupt+0xe7/0x160
> >  [<c0102310>] default_idle+0x0/0x40
> >  [<c010501a>] apic_timer_interrupt+0x1a/0x20
> >  [<c0102310>] default_idle+0x0/0x40
> >  [<c010233d>] default_idle+0x2d/0x40
> >  [<c01023d6>] cpu_idle+0x46/0x50
> >  [<c0384919>] start_kernel+0x179/0x1b0
> >  [<c0384380>] unknown_bootoption+0x0/0x140
> >
> > Code: 39 7b 1c 89 d8 75 21 8b 1b 89 3c 24 89 44 24 04 e8 9b f9 ff
> >  <0>Kernel panic: Fatal exception in interrupt
> > In interrupt handler - not syncing
> >
> > The following simple patch fixes the problem.
> >
> > diff -Nru a/fs/nfsd/lockd.c b/fs/nfsd/lockd.c
> > --- a/fs/nfsd/lockd.c   Thu Jun 24 15:38:29 2004
> > +++ b/fs/nfsd/lockd.c   Thu Jun 24 15:38:29 2004
> > @@ -40,7 +40,6 @@
> >                 mntget(filp->f_vfsmnt);
> >         }
> >         fh_put(&fh);
> > -       rqstp->rq_client =3D NULL;
> >         exp_readunlock();
> >         /* nlm and nfsd don't share error codes.
> >          * we invent: 0 =3D no error
> >
> > I am not sure why rqstp->rq_client is set to NULL in nlm_fopen. This
> > results in leaking of auth_domain cache entries.
> > I think we should do a auth_domain_put() before clearing rq_client. Thi=
s is
> > done in the release function(eg:svcauth_unix_release).
> >
> > Thanks
> > Sridhar
> >
>
> This bug appends only on the server or the client is also concerned ?

The OOPS occurs on the server. It can be easily reproduced on a server wher=
e
NFS Server Support and NFS File system support are enabled as modules and
following the steps below.

Server:
   modprobe nfsd
   service nfs start
Client:
   mount server:/export1 /mnt/mnt1
   Perform a locking operation on a file in the mounted filesystem.
   umount /mnt/mnt1
Server:
   service nfs stop
   rmmod nfsd
   rmmod lockd
   rmmod sunrpc

Thanks
Sridhar


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs