Re: Linux 3.6-rc4

[Linus Torvalds <torvalds@linux-foundation.org>]

Al? Please look into this. I'm not entirely sure what's going on, but
lockdep complains about this:

 Possible interrupt unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(&(&p->alloc_lock)->rlock);
                               local_irq_disable();
                               lock(&(&new_timer->it_lock)->rlock);
                               lock(tasklist_lock);
  <Interrupt>
    lock(&(&new_timer->it_lock)->rlock);

 *** DEADLOCK ***

and it looks real. IOW, if I read that right, we have the task_lock ->
it_lock dependency through exit_itimers(), and then we have the
tasklist_lock -> task_lock dependency everywhere else. So now it_lock
-> tasklist_lock becomes a deadlock.

                  Linus

On Fri, Sep 7, 2012 at 10:55 AM, Sasha Levin <levinsasha928@gmail.com> wrote:
> On 09/04/2012 05:44 PM, Dave Jones wrote:
>> On Sat, Sep 01, 2012 at 03:10:58PM -0700, Linus Torvalds wrote:
>>  > The kernel summit is over, and most people have either returned or are
>>  > returning from San Diego.
>>
>> Still seeing this, that I started seeing just before leaving for San Diego..
>>
>>       Dave
>
> I've bisected this one down to
>
>
> d35abdb28824cf74f0a106a0f9c6f3ff700a35bf is the first bad commit
> commit d35abdb28824cf74f0a106a0f9c6f3ff700a35bf
> Author: Al Viro <viro@zeniv.linux.org.uk>
> Date:   Sat Jun 30 11:55:24 2012 +0400
>
>     hold task_lock around checks in keyctl
>
>     Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
>
>
> Just didn't get a response from anyone when I've mailed about it...