On Tue, Oct 14, 2008 at 11:40:16PM +0900, [EMAIL PROTECTED] wrote: > Louis Rilling: > > > The message about the possible recursive locking is due to my recent > > > reducing lockdep_off/on work. I guess it will be fixed by reverting them > > > (but narrowing down its region). > > > > Surely a good thing to do. > > Here is a patch for the lockdep warning. > Will you test it and let me know the result? > Of course, you may meet the deadlock again because this patch is > unrelated to the deadlock.
I obtained another lockdep alert, which I guess was hidden by the first one:
[ 54.188183] =============================================
[ 54.188316] [ INFO: possible recursive locking detected ]
[ 54.188395] 2.6.25-elric #10
[ 54.188461] ---------------------------------------------
[ 54.188535] nfsd/2598 is trying to acquire lock:
[ 54.188601] (&rw->rwsem){----}, at: [<ffffffff883f5abd>]
au_do_open+0xf6/0x551 [aufs]
[ 54.188861]
[ 54.188862] but task is already holding lock:
[ 54.188982] (&rw->rwsem){----}, at: [<ffffffff8840db9a>]
aufs_fh_to_dentry+0x106/0x48c [aufs]
[ 54.189233]
[ 54.189234] other info that might help us debug this:
[ 54.189356] 2 locks held by nfsd/2598:
[ 54.189420] #0: (hash_sem){..--}, at: [<ffffffff882e9801>]
nfsd+0x189/0x2b4 [nfsd]
[ 54.189701] #1: (&rw->rwsem){----}, at: [<ffffffff8840db9a>]
aufs_fh_to_dentry+0x106/0x48c [aufs]
[ 54.189992]
[ 54.189993] stack backtrace:
[ 54.190110] Pid: 2598, comm: nfsd Not tainted 2.6.25-elric #10
[ 54.190178]
[ 54.190179] Call Trace:
[ 54.190297] [<ffffffff80253b7b>] __lock_acquire+0x8e0/0xc7a
[ 54.190375] [<ffffffff80254362>] lock_acquire+0x88/0xb2
[ 54.190458] [<ffffffff883f5abd>] :aufs:au_do_open+0xf6/0x551
[ 54.190544] [<ffffffff883f83f5>] :aufs:aufs_open_dir+0x0/0x75
[ 54.190618] [<ffffffff8042f7a0>] down_read+0x32/0x3b
[ 54.190700] [<ffffffff883f5abd>] :aufs:au_do_open+0xf6/0x551
[ 54.190785] [<ffffffff883f8ff0>] :aufs:do_open_dir+0x0/0x543
[ 54.190858] [<ffffffff802a2fad>] file_move+0x1d/0x51
[ 54.190942] [<ffffffff883f83f5>] :aufs:aufs_open_dir+0x0/0x75
[ 54.191089] [<ffffffff802a0a41>] __dentry_open+0xe9/0x1c0
[ 54.191180] [<ffffffff8840c2d3>] :aufs:au_lkup_by_ino+0x7a/0x211
[ 54.191254] [<ffffffff802abc20>] do_path_lookup+0x165/0x1de
[ 54.191339] [<ffffffff8840d591>] :aufs:decode_by_path+0x4f8/0x589
[ 54.191415] [<ffffffff802b43e9>] ifind_fast+0x1a/0x92
[ 54.191478] [<ffffffff80430685>] _spin_unlock+0x17/0x20
[ 54.191478] [<ffffffff802b4458>] ifind_fast+0x89/0x92
[ 54.191478] [<ffffffff8840dd9b>] :aufs:aufs_fh_to_dentry+0x307/0x48c
[ 54.191478] [<ffffffff8829d46e>] :sunrpc:sunrpc_cache_lookup+0x30/0x132
[ 54.191478] [<ffffffff8043063b>] _read_unlock+0x17/0x20
[ 54.191478] [<ffffffff8829c3d7>] :sunrpc:cache_check+0x66/0x435
[ 54.191478] [<ffffffff88289593>] :exportfs:exportfs_decode_fh+0x30/0x1e1
[ 54.191478] [<ffffffff882ebc32>] :nfsd:nfsd_acceptable+0x0/0xd3
[ 54.191478] [<ffffffff8029dc3d>] kfree+0x20d/0x224
[ 54.191478] [<ffffffff80252e07>] trace_hardirqs_on+0xf5/0x118
[ 54.191478] [<ffffffff80242023>] set_current_groups+0x15e/0x16a
[ 54.191478] [<ffffffff882f1e96>] :nfsd:nfsd_setuser+0x11f/0x20d
[ 54.191478] [<ffffffff882eb7bb>] :nfsd:nfsd_setuser_and_check_port+0x52/0x57
[ 54.191478] [<ffffffff882ebf52>] :nfsd:fh_verify+0x24d/0x48f
[ 54.191478] [<ffffffff8029d142>] __kmalloc+0x142/0x176
[ 54.191478] [<ffffffff882f3ea3>] :nfsd:nfsd3_proc_getattr+0x7d/0xb5
[ 54.191478] [<ffffffff882e9249>] :nfsd:nfsd_dispatch+0xde/0x1b6
[ 54.191478] [<ffffffff88296854>] :sunrpc:svc_process+0x3eb/0x707
[ 54.191478] [<ffffffff882e9678>] :nfsd:nfsd+0x0/0x2b4
[ 54.191478] [<ffffffff882e9818>] :nfsd:nfsd+0x1a0/0x2b4
[ 54.191478] [<ffffffff8020d028>] child_rip+0xa/0x12
[ 54.191478] [<ffffffff8020c5bc>] restore_args+0x0/0x30
[ 54.191478] [<ffffffff882e9678>] :nfsd:nfsd+0x0/0x2b4
[ 54.191478] [<ffffffff8020d01e>] child_rip+0x0/0x12
[ 54.191478]
>
> To investigate your deadlock, I'd like to ask you to enable
> both of CONFIG_AUFS_DEBUG_LOCK and CONFIG_MAGIC_SYSRQ, and press
> MagicSysRq + A after the deadlock happend.
Tests running, I'll let you know in later reply.
Thanks!
Louis
>
> I will try LTP by myself in a few days.
>
>
> Junjiro R. Okajima
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
--
Dr Louis Rilling Kerlabs
Skype: louis.rilling Batiment Germanium
Phone: (+33|0) 6 80 89 08 23 80 avenue des Buttes de Coesmes
http://www.kerlabs.com/ 35700 Rennes
signature.asc
Description: Digital signature
------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/
