Skip to contentSkip to conversation

3.3-rc2 snd_pcm lockdep backtrace

DISCUSSION Josh Boyer 2 participants 4 messages Feb 6, 2012 – Feb 8, 2012

lore
Share & tools
lore

Josh Boyer

author
#0 lore raw

Hi All,

We've had a report[1] of a lockdep backtrace from the snd_pcm driver. I was wondering if anyone had hit this already or had some decent ideas on what the issue might be.

josh

[1] https://bugzilla.redhat.com/show_bug.cgi?id=787319

backtrace:
:[ INFO: possible recursive locking detected ] :3.3.0-0.rc2.git2.1.fc17.x86_64 #1 Not tainted :---------------------------------------------
:pulseaudio/954 is trying to acquire lock:
: (&(&substream->self_group.lock)->rlock/1){......}, at: [<ffffffffa00d9093>] snd_pcm_action_group+0xa3/0x240 [snd_pcm]
:but task is already holding lock:
: (&(&substream->self_group.lock)->rlock/1){......}, at: [<ffffffffa00d9093>] snd_pcm_action_group+0xa3/0x240 [snd_pcm]
:other info that might help us debug this:
: Possible unsafe locking scenario:
: CPU0
: ----
: lock(&(&substream->self_group.lock)->rlock/1);
: lock(&(&substream->self_group.lock)->rlock/1);
: *** DEADLOCK ***
: May be due to missing lock nesting notation
:4 locks held by pulseaudio/954:
: #0: (snd_pcm_link_rwlock){......}, at: [<ffffffffa00d9e62>] snd_pcm_drop+0x62/0x110 [snd_pcm]
: #1: (&(&substream->self_group.lock)->rlock){......}, at:
[<ffffffffa00d9e6a>] snd_pcm_drop+0x6a/0x110 [snd_pcm]
: #2: (&(&substream->group->lock)->rlock){......}, at: [<ffffffffa00d93ce>] snd_pcm_action+0x3e/0xb0 [snd_pcm]
: #3: (&(&substream->self_group.lock)->rlock/1){......}, at:
[<ffffffffa00d9093>] snd_pcm_action_group+0xa3/0x240 [snd_pcm] :stack backtrace:
:Pid: 954, comm: pulseaudio Not tainted 3.3.0-0.rc2.git2.1.fc17.x86_64 #1 :Call Trace:
: [<ffffffff810cb7ec>] __lock_acquire+0x160c/0x1ad0
: [<ffffffff810ca4f6>] ? __lock_acquire+0x316/0x1ad0
: [<ffffffff81020f99>] ? sched_clock+0x9/0x10
: [<ffffffff810a24a5>] ? sched_clock_local+0x25/0xa0
: [<ffffffff810cc381>] lock_acquire+0xa1/0x1e0
: [<ffffffffa00d9093>] ? snd_pcm_action_group+0xa3/0x240 [snd_pcm] : [<ffffffff8169cca4>] _raw_spin_lock_nested+0x44/0x80
: [<ffffffffa00d9093>] ? snd_pcm_action_group+0xa3/0x240 [snd_pcm] : [<ffffffffa00d9093>] snd_pcm_action_group+0xa3/0x240 [snd_pcm] : [<ffffffffa00d9401>] snd_pcm_action+0x71/0xb0 [snd_pcm]
: [<ffffffffa00d945a>] snd_pcm_stop+0x1a/0x20 [snd_pcm]
: [<ffffffffa00d9e84>] snd_pcm_drop+0x84/0x110 [snd_pcm]
: [<ffffffffa00dbba8>] snd_pcm_common_ioctl1+0x4a8/0xbe0 [snd_pcm] : [<ffffffffa00dc650>] snd_pcm_playback_ioctl1+0x60/0x2d0 [snd_pcm] : [<ffffffff812c1481>] ? file_has_perm+0xe1/0xf0
: [<ffffffffa00dc8f4>] snd_pcm_playback_ioctl+0x34/0x40 [snd_pcm] : [<ffffffff811cf8d9>] do_vfs_ioctl+0x99/0x5a0
: [<ffffffff811cfe79>] sys_ioctl+0x99/0xa0
: [<ffffffff816a63e9>] system_call_fastpath+0x16/0x1b

#1 lore raw
Quoted 69 lines · On poniedziałek, 6 lutego 2012 o 15:56:22 Josh Boyer wrote:

Hi All,

We've had a report[1] of a lockdep backtrace from the snd_pcm driver. I was wondering if anyone had hit this already or had some decent ideas on what the issue might be.

josh

[1] https://bugzilla.redhat.com/show_bug.cgi?id=787319

backtrace:
:[ INFO: possible recursive locking detected ] :3.3.0-0.rc2.git2.1.fc17.x86_64 #1 Not tainted :--------------------------------------------- :
:pulseaudio/954 is trying to acquire lock:
: (&(&substream->self_group.lock)->rlock/1){......}, at:
: [<ffffffffa00d9093>]

snd_pcm_action_group+0xa3/0x240 [snd_pcm]

:but task is already holding lock:
: (&(&substream->self_group.lock)->rlock/1){......}, at:
: [<ffffffffa00d9093>]

snd_pcm_action_group+0xa3/0x240 [snd_pcm]

:other info that might help us debug this:
: Possible unsafe locking scenario:
: CPU0
: ----
:
: lock(&(&substream->self_group.lock)->rlock/1);
: lock(&(&substream->self_group.lock)->rlock/1);
:
: *** DEADLOCK ***
: May be due to missing lock nesting notation
:
:4 locks held by pulseaudio/954:
: #0: (snd_pcm_link_rwlock){......}, at: [<ffffffffa00d9e62>]

snd_pcm_drop+0x62/0x110 [snd_pcm]

: #1: (&(&substream->self_group.lock)->rlock){......}, at:
[<ffffffffa00d9e6a>] snd_pcm_drop+0x6a/0x110 [snd_pcm]

: #2: (&(&substream->group->lock)->rlock){......}, at:
: [<ffffffffa00d93ce>]

snd_pcm_action+0x3e/0xb0 [snd_pcm]

: #3: (&(&substream->self_group.lock)->rlock/1){......}, at:
[<ffffffffa00d9093>] snd_pcm_action_group+0xa3/0x240 [snd_pcm]

:stack backtrace:
:Pid: 954, comm: pulseaudio Not tainted 3.3.0-0.rc2.git2.1.fc17.x86_64 #1 :
:Call Trace:
: [<ffffffff810cb7ec>] __lock_acquire+0x160c/0x1ad0
: [<ffffffff810ca4f6>] ? __lock_acquire+0x316/0x1ad0
: [<ffffffff81020f99>] ? sched_clock+0x9/0x10
: [<ffffffff810a24a5>] ? sched_clock_local+0x25/0xa0
: [<ffffffff810cc381>] lock_acquire+0xa1/0x1e0
: [<ffffffffa00d9093>] ? snd_pcm_action_group+0xa3/0x240 [snd_pcm] : [<ffffffff8169cca4>] _raw_spin_lock_nested+0x44/0x80
: [<ffffffffa00d9093>] ? snd_pcm_action_group+0xa3/0x240 [snd_pcm] : [<ffffffffa00d9093>] snd_pcm_action_group+0xa3/0x240 [snd_pcm] : [<ffffffffa00d9401>] snd_pcm_action+0x71/0xb0 [snd_pcm]
: [<ffffffffa00d945a>] snd_pcm_stop+0x1a/0x20 [snd_pcm]

Quoted 9 lines

: [<ffffffffa00d9e84>] snd_pcm_drop+0x84/0x110 [snd_pcm] : [<ffffffffa00dbba8>] snd_pcm_common_ioctl1+0x4a8/0xbe0 [snd_pcm] : [<ffffffffa00dc650>] snd_pcm_playback_ioctl1+0x60/0x2d0 [snd_pcm] : [<ffffffff812c1481>] ? file_has_perm+0xe1/0xf0
: [<ffffffffa00dc8f4>] snd_pcm_playback_ioctl+0x34/0x40 [snd_pcm] : [<ffffffff811cf8d9>] do_vfs_ioctl+0x99/0x5a0
: [<ffffffff811cfe79>] sys_ioctl+0x99/0xa0
: [<ffffffff816a63e9>] system_call_fastpath+0x16/0x1b

I created a Bugzilla entry at
https://bugzilla.kernel.org/show_bug.cgi?id=42746
for your bug report, please add your address to the CC list in there, thanks!

Signature · 2 lines
Maciej Rutecki
http://www.mrutecki.pl
author
#2 lore raw
Quoted 5 lines · Maciej Rutecki

On poniedziałek, 6 lutego 2012 o 15:56:22 Josh Boyer wrote:

I created a Bugzilla entry at
https://bugzilla.kernel.org/show_bug.cgi?id=42746
for your bug report, please add your address to the CC list in there, thanks!

Er... why exactly?

I mean, great I don't really mind at all that a bugzilla.kernel.org bug was opened, but

1) there's already a bug opened in the RH bugzilla for this
2) I've already emailed the upstream developers and list about it
3) ALSA actually has it's own bug tool

so unless the ALSA devs are going to do something with that particular bug I don't see a huge amount of value.

As an aside, I'm basically an intermediary here. The original reporter in the RHT bug should probably add their email to CC.

josh

#3 lore raw
Quoted 17 lines · On środa, 8 lutego 2012 o 19:10:15 Josh Boyer wrote:
Quoted 6 lines · Maciej Rutecki

On poniedziałek, 6 lutego 2012 o 15:56:22 Josh Boyer wrote:

I created a Bugzilla entry at
https://bugzilla.kernel.org/show_bug.cgi?id=42746
for your bug report, please add your address to the CC list in there, thanks!

Er... why exactly?

I mean, great I don't really mind at all that a bugzilla.kernel.org bug was opened, but

1) there's already a bug opened in the RH bugzilla for this
2) I've already emailed the upstream developers and list about it
3) ALSA actually has it's own bug tool

Quoted 7 lines

so unless the ALSA devs are going to do something with that particular bug I don't see a huge amount of value.

As an aside, I'm basically an intermediary here. The original reporter in the RHT bug should probably add their email to CC.

josh

We use bugzilla.kernel.org to tracking regressions in one place (as database). I search LKML for it. It does not matter that you use another case for resolve bug. But it will be useful, when you send information about status of this bug (solved, invalid etc.).

Regards

Signature · 2 lines
Maciej Rutecki
http://www.mrutecki.pl