linux-nfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jeff Layton <jlayton@redhat.com>
To: "J. Bruce Fields" <bfields@fieldses.org>
Cc: trond.myklebust@primarydata.com, linux-nfs@vger.kernel.org,
	raphoszap@laposte.net,
	Stanislav Kinsbursky <skinsbursky@parallels.com>
Subject: Re: [PATCH] lockd: ensure we tear down any live sockets when socket creation fails during lockd_up
Date: Thu, 27 Mar 2014 20:37:51 -0700	[thread overview]
Message-ID: <20140327203751.1fbaf01f@ipyr.poochiereds.net> (raw)
In-Reply-To: <20140328021234.GE27633@fieldses.org>

On Thu, 27 Mar 2014 22:12:34 -0400
"J. Bruce Fields" <bfields@fieldses.org> wrote:

> Thanks, applying.
> 
> (Can you tell if this has always been there, or if it was introduced
> recently?  I guess it should go to stable, anyway....)
> 
> --b.
> 

(cc'ing Stanislav so he's aware...)

Yeah, probably reasonable for stable...

Looks like this bug crept in with 786185b5f8abefa. Prior to that we
called svc_shutdown_net from svc_destroy, which cleaned up the sockets
before the BUG_ON.

FWIW, the Fedora bug report is here in case anyone is interested:

    https://bugzilla.redhat.com/show_bug.cgi?id=1079700

> On Tue, Mar 25, 2014 at 11:55:26AM -0700, Jeff Layton wrote:
> > We had a Fedora ABRT report with a stack trace like this:
> > 
> > kernel BUG at net/sunrpc/svc.c:550!
> > invalid opcode: 0000 [#1] SMP
> > [...]
> > CPU: 2 PID: 913 Comm: rpc.nfsd Not tainted 3.13.6-200.fc20.x86_64 #1
> > Hardware name: Hewlett-Packard HP ProBook 4740s/1846, BIOS 68IRR
> > Ver. F.40 01/29/2013 task: ffff880146b00000 ti: ffff88003f9b8000
> > task.ti: ffff88003f9b8000 RIP: 0010:[<ffffffffa0305fa8>]
> > [<ffffffffa0305fa8>] svc_destroy+0x128/0x130 [sunrpc] RSP:
> > 0018:ffff88003f9b9de0  EFLAGS: 00010206 RAX: ffff88003f829628 RBX:
> > ffff88003f829600 RCX: 00000000000041ee RDX: 0000000000000000 RSI:
> > 0000000000000286 RDI: 0000000000000286 RBP: ffff88003f9b9de8 R08:
> > 0000000000017360 R09: ffff88014fa97360 R10: ffffffff8114ce57 R11:
> > ffffea00051c9c00 R12: ffff88003f829600 R13: 00000000ffffff9e R14:
> > ffffffff81cc7cc0 R15: 0000000000000000 FS:  00007f4fde284840(0000)
> > GS:ffff88014fa80000(0000) knlGS:0000000000000000 CS:  0010 DS: 0000
> > ES: 0000 CR0: 0000000080050033 CR2: 00007f4fdf5192f8 CR3:
> > 00000000a569a000 CR4: 00000000001407e0 Stack:
> >  ffff88003f792300 ffff88003f9b9e18 ffffffffa02de02a 0000000000000000
> >  ffffffff81cc7cc0 ffff88003f9cb000 0000000000000008 ffff88003f9b9e60
> >  ffffffffa033bb35 ffffffff8131c86c ffff88003f9cb000 ffff8800a5715008
> > Call Trace:
> >  [<ffffffffa02de02a>] lockd_up+0xaa/0x330 [lockd]
> >  [<ffffffffa033bb35>] nfsd_svc+0x1b5/0x2f0 [nfsd]
> >  [<ffffffff8131c86c>] ? simple_strtoull+0x2c/0x50
> >  [<ffffffffa033c630>] ? write_pool_threads+0x280/0x280 [nfsd]
> >  [<ffffffffa033c6bb>] write_threads+0x8b/0xf0 [nfsd]
> >  [<ffffffff8114efa4>] ? __get_free_pages+0x14/0x50
> >  [<ffffffff8114eff6>] ? get_zeroed_page+0x16/0x20
> >  [<ffffffff811dec51>] ? simple_transaction_get+0xb1/0xd0
> >  [<ffffffffa033c098>] nfsctl_transaction_write+0x48/0x80 [nfsd]
> >  [<ffffffff811b8b34>] vfs_write+0xb4/0x1f0
> >  [<ffffffff811c3f99>] ? putname+0x29/0x40
> >  [<ffffffff811b9569>] SyS_write+0x49/0xa0
> >  [<ffffffff810fc2a6>] ? __audit_syscall_exit+0x1f6/0x2a0
> >  [<ffffffff816962e9>] system_call_fastpath+0x16/0x1b
> > Code: 31 c0 e8 82 db 37 e1 e9 2a ff ff ff 48 8b 07 8b 57 14 48 c7
> > c7 d5 c6 31 a0 48 8b 70 20 31 c0 e8 65 db 37 e1 e9 f4 fe ff ff 0f
> > 0b <0f> 0b 66 0f 1f 44 00 00 0f 1f 44 00 00 55 48 89 e5 41 56 41 55
> > RIP  [<ffffffffa0305fa8>] svc_destroy+0x128/0x130 [sunrpc] RSP
> > <ffff88003f9b9de0>
> > 
> > Evidently, we created some lockd sockets and then failed to create
> > others. make_socks then returned an error and we tried to tear down
> > the svc, but svc->sv_permsocks was not empty so we ended up
> > tripping over the BUG() in svc_destroy().
> > 
> > Fix this by ensuring that we tear down any live sockets we created
> > when socket creation is going to return an error.
> > 
> > Reported-by: Raphos <raphoszap@laposte.net>
> > Signed-off-by: Jeff Layton <jlayton@redhat.com>
> > ---
> >  fs/lockd/svc.c | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/fs/lockd/svc.c b/fs/lockd/svc.c
> > index 10d6c41aecad..6bf06a07f3e0 100644
> > --- a/fs/lockd/svc.c
> > +++ b/fs/lockd/svc.c
> > @@ -235,6 +235,7 @@ out_err:
> >  	if (warned++ == 0)
> >  		printk(KERN_WARNING
> >  			"lockd_up: makesock failed, error=%d\n",
> > err);
> > +	svc_shutdown_net(serv, net);
> >  	return err;
> >  }
> >  
> > -- 
> > 1.8.5.3
> > 



-- 
Jeff Layton <jlayton@redhat.com>

  reply	other threads:[~2014-03-28  3:38 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-03-25 18:55 [PATCH] lockd: ensure we tear down any live sockets when socket creation fails during lockd_up Jeff Layton
2014-03-28  2:12 ` J. Bruce Fields
2014-03-28  3:37   ` Jeff Layton [this message]
2014-03-28  6:32     ` Stanislav Kinsbursky
2014-03-28 14:44       ` J. Bruce Fields

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20140327203751.1fbaf01f@ipyr.poochiereds.net \
    --to=jlayton@redhat.com \
    --cc=bfields@fieldses.org \
    --cc=linux-nfs@vger.kernel.org \
    --cc=raphoszap@laposte.net \
    --cc=skinsbursky@parallels.com \
    --cc=trond.myklebust@primarydata.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).