From: Stuart Anderson <anderson@ligo.caltech.edu>
Subject: Re: Kernel NFS nfs_update_inode Oops in 2.6.20.11
Date: Sat, 23 Jun 2007 14:51:31 -0700
Message-ID: <20070623215131.GC27551@ligo.caltech.edu>
References: <20070623153345.GA22117@ligo.caltech.edu>
	<1182615064.5859.2.camel@heimdal.trondhjem.org>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Cc: nfs@lists.sourceforge.net
To: Trond Myklebust <trond.myklebust@fys.uio.no>, MalteSch@gmx.de
Return-path: <nfs-bounces@lists.sourceforge.net>
Received: from sc8-sf-mx1-b.sourceforge.net ([10.3.1.91]
	helo=mail.sourceforge.net)
	by sc8-sf-list2-new.sourceforge.net with esmtp (Exim 4.43)
	id 1I2DWG-0002C5-1u
	for nfs@lists.sourceforge.net; Sat, 23 Jun 2007 14:51:48 -0700
Received: from acrux.ligo.caltech.edu ([131.215.115.14])
	by mail.sourceforge.net with esmtps (TLSv1:AES256-SHA:256)
	(Exim 4.44) id 1I2DWH-00042Q-Lp
	for nfs@lists.sourceforge.net; Sat, 23 Jun 2007 14:51:51 -0700
In-Reply-To: <1182615064.5859.2.camel@heimdal.trondhjem.org>
List-Id: "Discussion of NFS under Linux development, interoperability,
	and testing." <nfs.lists.sourceforge.net>
List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/nfs>,
	<mailto:nfs-request@lists.sourceforge.net?subject=unsubscribe>
List-Archive: <http://sourceforge.net/mailarchive/forum.php?forum_name=nfs>
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>
Sender: nfs-bounces@lists.sourceforge.net
Errors-To: nfs-bounces@lists.sourceforge.net

Trond,

This applied cleanly to 2.6.20.14 and so far so good with 2.5hr uptime
on each of two machines that both previously had this Oops on an interval
of minutes up to 24hr. The statistics are not conclusive yet, but it does
boot and I was able to successfully run "make -j 8 boostrap" on the gcc
source code using 8 out of 16 CPU-cores on an NFSV4 mount without crashing.

However, the gcc build did generate 19 of the following on this client machine:
NFS: v4 server returned a bad sequence-id error!
Are these serious? How to go about tracking these down?

The server is a Sun X4500 running ZFS on Solaris10 Update 3 and it did
not log any error messages during the gcc build.

Very many thanks for incredible turn around a kernel patch--less than 40
minutes from Oops posting to patch posting!


Malte,
	I suggest you give this patch a try as well to see if it solves your
similar/identical(?) Oops posted on lkml.


On Sat, Jun 23, 2007 at 12:11:04PM -0400, Trond Myklebust wrote:
> Does the attached patch (against 2.6.22) fix it?
> 
> Trond
> 
> 
> On Sat, 2007-06-23 at 08:33 -0700, Stuart Anderson wrote:
> > We started receiving frequent kernel (2.6.20.11) Oops messages in
> > nfs:nfs_update_inode on a pair of Sun X4600M2 machines once we started
> > mounting an NFS V4 filesystem from a Solaris x86 ZFS server.
> > 
> > Any help in tracking this down would be greatly appreciated.
> > 
> > Thanks.
> > 
> > Jun 22 18:15:49 ldas-grid kernel: Unable to handle kernel NULL pointer dereference at 0000000000000020 RIP: 
> > Jun 22 18:15:49 ldas-grid kernel:  [<ffffffff88170989>] :nfs:nfs_update_inode+0x99/0x5a0
> > Jun 22 18:15:49 ldas-grid kernel: PGD 3edd83067 PUD 3edd8b067 PMD 0 
> > Jun 22 18:15:49 ldas-grid kernel: Oops: 0000 [1] SMP 
> > Jun 22 18:15:49 ldas-grid kernel: CPU 7 
> > Jun 22 18:15:49 ldas-grid kernel: Modules linked in: nfsd exportfs autofs4 eeprom adm1026 hwmon_vid hwmon i2c_isa i2c_amd756
> >  i2c_amd8111 nfs lockd nfs_acl sunrpc ipt_REJECT xt_state usb_storage ip_conntrack nfnetlink xt_tcpudp iptable_filter ip_tab
> > les x_tables usbhid dm_mod ohci_hcd ehci_hcd i2c_nforce2 i2c_core e1000 usbcore mptsas scsi_transport_sas mptscsih mptbase s
> > d_mod scsi_mod
> > Jun 22 18:15:49 ldas-grid kernel: Pid: 33, comm: events/7 Not tainted 2.6.20.11-CIT #1
> > Jun 22 18:15:49 ldas-grid kernel: RIP: 0010:[<ffffffff88170989>]  [<ffffffff88170989>] :nfs:nfs_update_inode+0x99/0x5a0
> > Jun 22 18:15:49 ldas-grid kernel: RSP: 0018:ffff8103fbc3dc10  EFLAGS: 00010246
> > Jun 22 18:15:49 ldas-grid kernel: RAX: 0000000000000000 RBX: ffff81035d704ce8 RCX: 0000000000008180
> > Jun 22 18:15:49 ldas-grid kernel: RDX: ffff8101003cf8c0 RSI: ffff8103ace9c8d0 RDI: ffff81035d704ce8
> > Jun 22 18:15:49 ldas-grid kernel: RBP: ffff8103ace9c8d0 R08: 0000000000008180 R09: ffff8103eddf0030
> > Jun 22 18:15:49 ldas-grid kernel: R10: 0000000000000026 R11: 0000000000000003 R12: ffff81035d704b10
> > Jun 22 18:15:49 ldas-grid kernel: R13: ffff81035d704ce8 R14: ffff8101fbc3e6c0 R15: ffff8103ace9c8d0
> > Jun 22 18:15:49 ldas-grid kernel: FS:  00002b526bf797a0(0000) GS:ffff810300141d40(0000) knlGS:00000000f7dbb6c0
> > Jun 22 18:15:49 ldas-grid kernel: CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
> > Jun 22 18:15:49 ldas-grid kernel: CR2: 0000000000000020 CR3: 00000003edd89000 CR4: 00000000000006e0
> > Jun 22 18:15:49 ldas-grid kernel: Process events/7 (pid: 33, threadinfo ffff8103fbc3c000, task ffff8102fbcb90c0)
> > Jun 22 18:15:49 ldas-grid kernel: Stack:  ffff81035d704ce8 ffff8103ace9c8d0 ffff81035d704da0 ffff81035d704ce8
> > Jun 22 18:15:49 ldas-grid kernel:  ffff8101fbc3e6c0 ffffffff88170edb 0000000000000000 ffff8103ace9c800
> > Jun 22 18:15:49 ldas-grid kernel:  ffff810322664c00 ffffffff881817f0 ffff810376445de0 ffff8101003cf8c0
> > Jun 22 18:15:49 ldas-grid kernel: Call Trace:
> > Jun 22 18:15:49 ldas-grid kernel:  [<ffffffff88170edb>] :nfs:nfs_post_op_update_inode+0x4b/0x70
> > Jun 22 18:15:49 ldas-grid kernel:  [<ffffffff881817f0>] :nfs:nfs4_proc_delegreturn+0x160/0x1e0
> > Jun 22 18:15:49 ldas-grid kernel:  [<ffffffff8818e2ae>] :nfs:nfs_do_return_delegation+0x1e/0x40
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8816d0e3>] :nfs:nfs_dentry_iput+0x23/0x70
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff802999a2>] shrink_dcache_for_umount_subtree+0x212/0x270
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80299a53>] shrink_dcache_for_umount+0x53/0x70
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80288ce9>] generic_shutdown_super+0x19/0x100
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80288e29>] kill_anon_super+0x9/0x40
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff881724ad>] :nfs:nfs_kill_super+0xd/0x20
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80289046>] deactivate_super+0x76/0xb0
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff802a1573>] expire_mount_list+0x133/0x180
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8817a7a0>] :nfs:nfs_expire_automounts+0x0/0x40
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff802a176b>] mark_mounts_for_expiry+0xab/0xc0
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8817a7b0>] :nfs:nfs_expire_automounts+0x10/0x40
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8023fd6e>] run_workqueue+0xae/0x160
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8023fee0>] worker_thread+0x0/0x190
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80240031>] worker_thread+0x151/0x190
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80229490>] default_wake_function+0x0/0x10
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8023fee0>] worker_thread+0x0/0x190
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80243b79>] kthread+0xd9/0x120
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff802298cc>] schedule_tail+0x4c/0xb0
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8020a998>] child_rip+0xa/0x12
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80219250>] flat_send_IPI_mask+0x0/0x60
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff80243aa0>] kthread+0x0/0x120
> > Jun 22 18:15:50 ldas-grid kernel:  [<ffffffff8020a98e>] child_rip+0x0/0x12
> > Jun 22 18:15:50 ldas-grid kernel: 
> > Jun 22 18:15:50 ldas-grid kernel: 
> > Jun 22 18:15:50 ldas-grid kernel: Code: 48 3b 58 20 75 31 48 8b 45 60 48 39 82 b0 00 00 00 48 8d 75 
> > Jun 22 18:15:50 ldas-grid kernel: RIP  [<ffffffff88170989>] :nfs:nfs_update_inode+0x99/0x5a0
> > Jun 22 18:15:50 ldas-grid kernel:  RSP <ffff8103fbc3dc10>
> > Jun 22 18:15:50 ldas-grid kernel: CR2: 0000000000000020
> > 

> From: Trond Myklebust <Trond.Myklebust@netapp.com>
> Date: Tue, 5 Jun 2007 13:26:15 -0400
> NFS: Fix nfs_reval_fsid()
> Subject: No Subject
> 
> We don't need to revalidate the fsid on the root directory. It suffices to
> revalidate it on the current directory.
> 
> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
> ---
> 
>  fs/nfs/dir.c   |    9 ++++-----
>  fs/nfs/inode.c |    4 ++--
>  2 files changed, 6 insertions(+), 7 deletions(-)
> 
> diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
> index 4948ec1..c02a796 100644
> --- a/fs/nfs/dir.c
> +++ b/fs/nfs/dir.c
> @@ -897,14 +897,13 @@ int nfs_is_exclusive_create(struct inode *dir, struct nameidata *nd)
>  	return (nd->intent.open.flags & O_EXCL) != 0;
>  }
>  
> -static inline int nfs_reval_fsid(struct vfsmount *mnt, struct inode *dir,
> -				 struct nfs_fh *fh, struct nfs_fattr *fattr)
> +static inline int nfs_reval_fsid(struct inode *dir, const struct nfs_fattr *fattr)
>  {
>  	struct nfs_server *server = NFS_SERVER(dir);
>  
>  	if (!nfs_fsid_equal(&server->fsid, &fattr->fsid))
> -		/* Revalidate fsid on root dir */
> -		return __nfs_revalidate_inode(server, mnt->mnt_root->d_inode);
> +		/* Revalidate fsid using the parent directory */
> +		return __nfs_revalidate_inode(server, dir);
>  	return 0;
>  }
>  
> @@ -946,7 +945,7 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru
>  		res = ERR_PTR(error);
>  		goto out_unlock;
>  	}
> -	error = nfs_reval_fsid(nd->mnt, dir, &fhandle, &fattr);
> +	error = nfs_reval_fsid(dir, &fattr);
>  	if (error < 0) {
>  		res = ERR_PTR(error);
>  		goto out_unlock;
> diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
> index 23ecf03..7bcb3df 100644
> --- a/fs/nfs/inode.c
> +++ b/fs/nfs/inode.c
> @@ -961,8 +961,8 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
>  		goto out_changed;
>  
>  	server = NFS_SERVER(inode);
> -	/* Update the fsid if and only if this is the root directory */
> -	if (inode == inode->i_sb->s_root->d_inode
> +	/* Update the fsid? */
> +	if (S_ISDIR(inode->i_mode)
>  			&& !nfs_fsid_equal(&server->fsid, &fattr->fsid))
>  		server->fsid = fattr->fsid;
>  


-- 
Stuart Anderson  anderson@ligo.caltech.edu
http://www.ligo.caltech.edu/~anderson

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs