tty_lock lockdep warning

tty_lock lockdep warning

[John Stultz]

Hey Alan,
	I've been seeing the following lockdep warning w/ Linus' 3.5-rc HEAD 
from today. This is running x86_64 in a kvm environment.

Let me know if there are any other details you'd like from me.

thanks
-john


=============================================
[ INFO: possible recursive locking detected ]
3.4.0john+ #347 Not tainted
---------------------------------------------
plymouthd/1345 is trying to acquire lock:
  (&tty->legacy_mutex){+.+.+.}, at: [<ffffffff81b10091>] tty_lock+0x61/0xb0

but task is already holding lock:
  (&tty->legacy_mutex){+.+.+.}, at: [<ffffffff81b10091>] tty_lock+0x61/0xb0

other info that might help us debug this:
  Possible unsafe locking scenario:

        CPU0
        ----
   lock(&tty->legacy_mutex);
   lock(&tty->legacy_mutex);

  *** DEADLOCK ***

  May be due to missing lock nesting notation

2 locks held by plymouthd/1345:
  #0:  (tty_mutex){+.+.+.}, at: [<ffffffff81417122>] tty_release+0x1e2/0x600
  #1:  (&tty->legacy_mutex){+.+.+.}, at: [<ffffffff81b10091>] 
tty_lock+0x61/0xb0

stack backtrace:
Pid: 1345, comm: plymouthd Not tainted 3.4.0john+ #347
Call Trace:
  [<ffffffff81044727>] ? vprintk_emit+0x287/0x5d0
  [<ffffffff810b63b6>] validate_chain+0xf36/0x1380
  [<ffffffff8108b433>] ? sched_clock_local+0x43/0xb0
  [<ffffffff8108b6df>] ? local_clock+0x6f/0x80
  [<ffffffff810b0ee9>] ? trace_hardirqs_off_caller+0x29/0x140
  [<ffffffff810b6c3e>] __lock_acquire+0x43e/0xb80
  [<ffffffff810b7431>] lock_acquire+0xb1/0x210
  [<ffffffff81b10091>] ? tty_lock+0x61/0xb0
  [<ffffffff81b0c273>] mutex_lock_nested+0x73/0x4b0
  [<ffffffff81b10091>] ? tty_lock+0x61/0xb0
  [<ffffffff81b10091>] ? tty_lock+0x61/0xb0
  [<ffffffff81b10091>] tty_lock+0x61/0xb0
  [<ffffffff81b1013d>] tty_lock_pair+0x5d/0x64
  [<ffffffff8141712d>] tty_release+0x1ed/0x600
  [<ffffffff811851e8>] fput+0x138/0x2e0
  [<ffffffff81416d18>] tty_ioctl+0xb68/0xd40
  [<ffffffff81b0fee1>] ? _raw_spin_unlock_irqrestore+0x71/0xa0
  [<ffffffff81198d94>] do_vfs_ioctl+0xa4/0x5c0
  [<ffffffff8141d12c>] ? put_ldisc+0x11c/0x190
  [<ffffffff8141d1de>] ? tty_ldisc_deref+0xe/0x10
  [<ffffffff81413887>] ? tty_compat_ioctl+0xc7/0xf0
  [<ffffffff811ded87>] compat_sys_ioctl+0x257/0x1220
  [<ffffffff81b12138>] ? sysenter_dispatch+0x19/0x26
  [<ffffffff81392ffd>] ? trace_hardirqs_off_thunk+0x3a/0x3c
  [<ffffffff81b12126>] sysenter_dispatch+0x7/0x26
  [<ffffffff81392fbe>] ? trace_hardirqs_on_thunk+0x3a/0x3f

Re: tty_lock lockdep warning

[Greg KH]

On Thu, May 24, 2012 at 09:43:37PM -0700, John Stultz wrote:
> Hey Alan,
> 	I've been seeing the following lockdep warning w/ Linus' 3.5-rc
> HEAD from today. This is running x86_64 in a kvm environment.
> 
> Let me know if there are any other details you'd like from me.

Known issue, see the thread with patches for it on lkml.  Any help with
those (in testing), is appreciated.

thanks,

greg k-h