Fwd: lockdep pops with -rc5 kernel

[Steve French]

---------- Forwarded message ----------
From: Steve French <smfrench at gmail.com>
Date: Tue, May 1, 2012 at 9:22 PM
Subject: Re: lockdep pops with -rc5 kernel
To: Jeff Layton <jlayton at samba.org>
Cc: Pavel Shilovsky <piastryyy at gmail.com>, linux-cifs at vger.kernel.org




On Tue, May 1, 2012 at 4:40 PM, Jeff Layton <jlayton at samba.org> wrote:
>
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On Tue, 1 May 2012 17:22:20 -0400
> Jeff Layton <jlayton at samba.org> wrote:
>
> > I'm still seeing this lockdep pop with -rc5 kernel. I thought Pavel had
> > fixed this, but maybe his patch hasn't been merged yet? Or am I
> > mistaken? In any case, it looks like it just needs some lockdep
> > annotation and we probably want this fixed before 3.4 ships...
> >
> > [   55.218753] INFO: trying to register non-static key.
> > [   55.219396] the code is fine but needs lockdep annotation.
> > [   55.219396] turning off the locking correctness validator.
> > [   55.219396] Pid: 855, comm: mount.cifs Not tainted 3.4.0-rc5+ #15
> > [   55.219396] Call Trace:
> > [   55.219396]  [<ffffffff810cd7b6>] __lock_acquire+0x1306/0x1ad0
> > [   55.219396]  [<ffffffff810ccbe3>] ? __lock_acquire+0x733/0x1ad0
> > [   55.219396]  [<ffffffff810cf2ad>] ? trace_hardirqs_on+0xd/0x10
> > [   55.219396]  [<ffffffff8132ff8d>] ? pointer.isra.11+0x1fd/0x2d0
> > [   55.219396]  [<ffffffff810ce661>] lock_acquire+0xa1/0x1f0
> > [   55.219396]  [<ffffffffa00f96fe>] ? cifs_set_credits+0x2e/0x70 [cifs]
> > [   55.219396]  [<ffffffffa00e8ca0>] ? cifs_get_smb_ses+0x160/0x4c0
> > [cifs]
> > [   55.219396]  [<ffffffff816ad646>] _raw_spin_lock+0x46/0x80
> > [   55.219396]  [<ffffffffa00f96fe>] ? cifs_set_credits+0x2e/0x70 [cifs]
> > [   55.219396]  [<ffffffff816aa323>] ? mutex_lock_nested+0x2c3/0x390
> > [   55.219396]  [<ffffffffa00f96fe>] cifs_set_credits+0x2e/0x70 [cifs]
> > [   55.219396]  [<ffffffffa00e88ad>] cifs_negotiate_protocol+0x4d/0xe0
> > [cifs]
> > [   55.219396]  [<ffffffffa00e8cab>] cifs_get_smb_ses+0x16b/0x4c0 [cifs]
> > [   55.219396]  [<ffffffffa00e9099>] cifs_mount+0x99/0x6f0 [cifs]
> > [   55.219396]  [<ffffffffa00d6efc>] cifs_do_mount+0xac/0x4d0 [cifs]
> > [   55.219396]  [<ffffffff811c28b3>] mount_fs+0x43/0x1b0
> > [   55.219396]  [<ffffffff811df38f>] vfs_kern_mount+0x6f/0x100
> > [   55.219396]  [<ffffffff811dff14>] do_kern_mount+0x54/0x110
> > [   55.219396]  [<ffffffff811e177a>] do_mount+0x26a/0x840
> > [   55.219396]  [<ffffffff81155324>] ? __get_free_pages+0x14/0x50
> > [   55.219396]  [<ffffffff811e137a>] ? copy_mount_options+0x3a/0x180
> > [   55.219396]  [<ffffffff811e1e8d>] sys_mount+0x8d/0xe0
> > [   55.219396]  [<ffffffff816b6f29>] system_call_fastpath+0x16/0x1b
> >
>
> Ok, I think I've found the problem. The req_lock wasn't being
> initialized (ouch). I've just sent a patch that I think is pretty
> obvious.
>
> Steve, that should be pushed to Linus ASAP...
>
agreed



--
Thanks,

Steve



--
Thanks,

Steve