Re: check_usage_backwards() && forwards? (Was: [2.6.33-rc5] starting emacs makes lockdep warning)

["Américo Wang" <xiyou.wangcong@gmail.com>]

On Wed, Jan 27, 2010 at 2:16 AM, Oleg Nesterov <oleg@redhat.com> wrote:
> (add  lockdep gurus)
>
> Lockdep has found the real bug, but the output doesn't look right to me
>
> On 01/26, KOSAKI Motohiro wrote:
>>
>> =========================================================
>> [ INFO: possible irq lock inversion dependency detected ]
>> 2.6.33-rc5 #77
>> ---------------------------------------------------------
>> emacs/1609 just changed the state of lock:
>>  (&(&tty->ctrl_lock)->rlock){+.....}, at: [<ffffffff8127c648>] tty_fasync+0xe8/0x190
>> but this lock took another, HARDIRQ-unsafe lock in the past:
>>  (&(&sighand->siglock)->rlock){-.....}
>
> "HARDIRQ-unsafe" and "this lock took another" looks wrong, afaics.
>
>>   ... key      at: [<ffffffff81c054a4>] __key.46539+0x0/0x8
>>   ... acquired at:
>>    [<ffffffff81089af6>] __lock_acquire+0x1056/0x15a0
>>    [<ffffffff8108a0df>] lock_acquire+0x9f/0x120
>>    [<ffffffff81423012>] _raw_spin_lock_irqsave+0x52/0x90
>>    [<ffffffff8127c1be>] __proc_set_tty+0x3e/0x150
>>    [<ffffffff8127e01d>] tty_open+0x51d/0x5e0
>
> The stack-trace shows that this lock (ctrl_lock) was taken under
> ->siglock (which is hopefully irq-safe).
>
> Typo in check_usage_backwards() ?
>
> Oleg.
>
> --- a/kernel/lockdep.c
> +++ b/kernel/lockdep.c
> @@ -2147,7 +2147,7 @@ check_usage_backwards(struct task_struct
>                return ret;
>
>        return print_irq_inversion_bug(curr, &root, target_entry,
> -                                       this, 1, irqclass);
> +                                       this, 0, irqclass);
>  }
>
>  void print_irqtrace_events(struct task_struct *curr)
>
>

Yes!! Almost definitely... You are so careful!
ACK, please submit it as a normal patch.

Thanks.