Hi,
Whenever I start /etc/init.d/udftools. I get a warning from lockdep about
a possible recursive lockup... It looks like this:
---
Clocksource tsc unstable (delta = -80671344 ns)
pktcdvd: writer pktcdvd0 mapped to sr0
=============================================
[ INFO: possible recursive locking detected ]
2.6.24.4 #1
---------------------------------------------
scsi_id/4705 is trying to acquire lock:
(&bdev->bd_mutex){--..}, at: [<ffffffff802a71a8>] do_open+0x74/0x2a9
but task is already holding lock:
(&bdev->bd_mutex){--..}, at: [<ffffffff802a71a8>] do_open+0x74/0x2a9
other info that might help us debug this:
2 locks held by scsi_id/4705:
#0: (&bdev->bd_mutex){--..}, at: [<ffffffff802a71a8>] do_open+0x74/0x2a9
#1: (&ctl_mutex#2){--..}, at: [<ffffffff881ecf2a>] pkt_open+0x27/0xda1
[pktcdvd]
stack backtrace:
Pid: 4705, comm: scsi_id Not tainted 2.6.24.4 #1
Call Trace:
[<ffffffff80253b08>] __lock_acquire+0x1db/0xd3f
[<ffffffff802a71a8>] do_open+0x74/0x2a9
[<ffffffff802546f0>] lock_acquire+0x84/0xa8
[<ffffffff802a71a8>] do_open+0x74/0x2a9
[...]
(full dmesg + backtrace: http://www.pastebin.ca/955702 )
---
Well, as you can see the function that tries to get the lock... has it
already?! This sounds a bit bogus, what's going on here? Is this really
a locking bug or just lockdep noise?
(BTW: the pktcdvd device seems to work nonetheless...)
Regards,
Christian