Re: INFO: possible recursive locking detected:autofs4_expire_indirect()

[Nick Bowler]

On 2011-10-25 14:48 +0200, Pawel Sikora wrote:
> the nfs client/server with fresh 3.0.8 + vserver + enabled debug options reports attached info.
> afaics, the vserver doesn't change autofs code, so it looks like a pure vanilla problem.

I reported this issue a long time ago, and a patch[1] was provided
quickly, but it seems that for some reason it never made it to mainline.

Adding Steven to CC.

[1] http://permalink.gmane.org/gmane.linux.kernel/1129741

> [ 3708.715749] =============================================
> [ 3708.715940] [ INFO: possible recursive locking detected ]
> [ 3708.716040] 3.0.8-vs2.3.1-dirty #6
> [ 3708.716131] ---------------------------------------------
> [ 3708.716230] automount/29215 is trying to acquire lock:
> [ 3708.716301]  (&(&dentry->d_lock)->rlock/1){+.+...}, at: [<ffffffffa0214fb0>] autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]
> [ 3708.716301] but task is already holding lock:
> [ 3708.716301]  (&(&dentry->d_lock)->rlock/1){+.+...}, at: [<ffffffffa0214fb0>] autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]
> [ 3708.716301] other info that might help us debug this:
> [ 3708.716301]  Possible unsafe locking scenario:
> [ 3708.716301]
> [ 3708.716301]        CPU0
> [ 3708.716301]        ----
> [ 3708.716301]   lock(&(&dentry->d_lock)->rlock);
> [ 3708.716301]   lock(&(&dentry->d_lock)->rlock);
> [ 3708.716301]
> [ 3708.716301]  *** DEADLOCK ***
> [ 3708.716301]
> [ 3708.716301]  May be due to missing lock nesting notation
> [ 3708.716301]
> [ 3708.716301] 2 locks held by automount/29215:
> [ 3708.716301]  #0:  (&(&sbi->lookup_lock)->rlock){+.+...}, at: [<ffffffffa0214f61>] autofs4_expire_indirect+0x91/0x4e0 [autofs4]
> [ 3708.716301]  #1:  (&(&dentry->d_lock)->rlock/1){+.+...}, at: [<ffffffffa0214fb0>] autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]
> [ 3708.716301] stack backtrace:
> [ 3708.716301] Pid: 29215, comm: automount Not tainted 3.0.8-vs2.3.1-dirty #6
> [ 3708.716301] Call Trace:
> [ 3708.716301]  [<ffffffff810925d6>] __lock_acquire+0x1606/0x1b50
> [ 3708.716301]  [<ffffffffa0214fb0>] ? autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffff81092c6a>] ? lock_release_non_nested+0x14a/0x310
> [ 3708.716301]  [<ffffffffa0214fb0>] ? autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffffa0214fb0>] ? autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffff810930c5>] lock_acquire+0x85/0x110
> [ 3708.716301]  [<ffffffffa0214fb0>] ? autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffffa02151bd>] ? autofs4_expire_indirect+0x2ed/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffff81453c1a>] _raw_spin_lock_nested+0x2a/0x40
> [ 3708.716301]  [<ffffffffa0214fb0>] ? autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffff81454386>] ? _raw_spin_unlock+0x26/0x30
> [ 3708.716301]  [<ffffffff8110f830>] ? might_fault+0x40/0x90
> [ 3708.716301]  [<ffffffffa0214fb0>] autofs4_expire_indirect+0xe0/0x4e0 [autofs4]
> [ 3708.716301]  [<ffffffffa021568d>] autofs4_do_expire_multi+0xed/0x130 [autofs4]
> [ 3708.716301]  [<ffffffffa0215a70>] ? autofs_dev_ioctl_askumount+0x30/0x30 [autofs4]
> [ 3708.716301]  [<ffffffffa0215a8a>] autofs_dev_ioctl_expire+0x1a/0x20 [autofs4]
> [ 3708.716301]  [<ffffffffa0216063>] _autofs_dev_ioctl+0x273/0x360 [autofs4]
> [ 3708.716301]  [<ffffffffa021615e>] autofs_dev_ioctl+0xe/0x20 [autofs4]
> [ 3708.716301]  [<ffffffff8115dc56>] do_vfs_ioctl+0x96/0x560
> [ 3708.716301]  [<ffffffff8114c289>] ? fget_light+0x99/0x130
> [ 3708.716301]  [<ffffffff8114c227>] ? fget_light+0x37/0x130
> [ 3708.716301]  [<ffffffff8115e1b1>] sys_ioctl+0x91/0xa0
> [ 3708.716301]  [<ffffffff8145bdbb>] system_call_fastpath+0x16/0x1b

Cheers,
-- 
Nick Bowler, Elliptic Technologies (http://www.elliptictech.com/)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/