2010-11-07 18:47:15

by Paul E. McKenney

[permalink] [raw]
Subject: Re: 2.6.36-git11 -- kernel/pid.c:419 invoked rcu_dereference_check() without protection!

On Fri, Oct 29, 2010 at 01:22:28AM -0400, Miles Lane wrote:
> [ INFO: suspicious rcu_dereference_check() usage. ]
> ---------------------------------------------------
> kernel/pid.c:419 invoked rcu_dereference_check() without protection!
>
> other info that might help us debug this:
>
>
> rcu_scheduler_active = 1, debug_locks = 1
> 1 lock held by ureadahead/943:
> #0: (tasklist_lock){.+.+..}, at: [<ffffffff810ec497>]
> sys_ioprio_set+0x8e/0x2aa
>
> stack backtrace:
> Pid: 943, comm: ureadahead Not tainted 2.6.36-git11 #8
> Call Trace:
> [<ffffffff8105d335>] lockdep_rcu_dereference+0x9d/0xa5
> [<ffffffff8104c825>] find_task_by_pid_ns+0x3f/0x58
> [<ffffffff8104c85b>] find_task_by_vpid+0x1d/0x1f
> [<ffffffff810ec4c6>] sys_ioprio_set+0xbd/0x2aa
> [<ffffffff813547a3>] ? trace_hardirqs_on_thunk+0x3a/0x3f
> [<ffffffff81002aab>] system_call_fastpath+0x16/0x1b

Could you please try the patch from Sergey Senozhatsky, currently
in -mm, that may be found at http://lkml.org/lkml/2010/10/29/168?

Thanx, Paul


2010-11-08 21:00:53

by Miles Lane

[permalink] [raw]
Subject: Re: 2.6.36-git11 -- kernel/pid.c:419 invoked rcu_dereference_check() without protection!

Great! Works here. :-)

On Sun, Nov 7, 2010 at 1:47 PM, Paul E. McKenney
<[email protected]> wrote:
> On Fri, Oct 29, 2010 at 01:22:28AM -0400, Miles Lane wrote:
>> [ INFO: suspicious rcu_dereference_check() usage. ]
>> ---------------------------------------------------
>> kernel/pid.c:419 invoked rcu_dereference_check() without protection!
>>
>> other info that might help us debug this:
>>
>>
>> rcu_scheduler_active = 1, debug_locks = 1
>> 1 lock held by ureadahead/943:
>> ?#0: ?(tasklist_lock){.+.+..}, at: [<ffffffff810ec497>]
>> sys_ioprio_set+0x8e/0x2aa
>>
>> stack backtrace:
>> Pid: 943, comm: ureadahead Not tainted 2.6.36-git11 #8
>> Call Trace:
>> ?[<ffffffff8105d335>] lockdep_rcu_dereference+0x9d/0xa5
>> ?[<ffffffff8104c825>] find_task_by_pid_ns+0x3f/0x58
>> ?[<ffffffff8104c85b>] find_task_by_vpid+0x1d/0x1f
>> ?[<ffffffff810ec4c6>] sys_ioprio_set+0xbd/0x2aa
>> ?[<ffffffff813547a3>] ? trace_hardirqs_on_thunk+0x3a/0x3f
>> ?[<ffffffff81002aab>] system_call_fastpath+0x16/0x1b
>
> Could you please try the patch from Sergey Senozhatsky, currently
> in -mm, that may be found at http://lkml.org/lkml/2010/10/29/168?
>
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?Thanx, Paul
>

2010-11-08 21:08:48

by Paul E. McKenney

[permalink] [raw]
Subject: Re: 2.6.36-git11 -- kernel/pid.c:419 invoked rcu_dereference_check() without protection!

On Mon, Nov 08, 2010 at 04:00:51PM -0500, Miles Lane wrote:
> Great! Works here. :-)

Thank you for testing it, Miles!

Thanx, Paul

> On Sun, Nov 7, 2010 at 1:47 PM, Paul E. McKenney
> <[email protected]> wrote:
> > On Fri, Oct 29, 2010 at 01:22:28AM -0400, Miles Lane wrote:
> >> [ INFO: suspicious rcu_dereference_check() usage. ]
> >> ---------------------------------------------------
> >> kernel/pid.c:419 invoked rcu_dereference_check() without protection!
> >>
> >> other info that might help us debug this:
> >>
> >>
> >> rcu_scheduler_active = 1, debug_locks = 1
> >> 1 lock held by ureadahead/943:
> >> ?#0: ?(tasklist_lock){.+.+..}, at: [<ffffffff810ec497>]
> >> sys_ioprio_set+0x8e/0x2aa
> >>
> >> stack backtrace:
> >> Pid: 943, comm: ureadahead Not tainted 2.6.36-git11 #8
> >> Call Trace:
> >> ?[<ffffffff8105d335>] lockdep_rcu_dereference+0x9d/0xa5
> >> ?[<ffffffff8104c825>] find_task_by_pid_ns+0x3f/0x58
> >> ?[<ffffffff8104c85b>] find_task_by_vpid+0x1d/0x1f
> >> ?[<ffffffff810ec4c6>] sys_ioprio_set+0xbd/0x2aa
> >> ?[<ffffffff813547a3>] ? trace_hardirqs_on_thunk+0x3a/0x3f
> >> ?[<ffffffff81002aab>] system_call_fastpath+0x16/0x1b
> >
> > Could you please try the patch from Sergey Senozhatsky, currently
> > in -mm, that may be found at http://lkml.org/lkml/2010/10/29/168?
> >
> > ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?Thanx, Paul
> >