lkml.org 
[lkml]   [2008]   [Jan]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] printk deadlocks if called with runqueue lock held

On Fri, 18 Jan 2008, Jan Kiszka wrote:

> Steven Rostedt wrote:
> ....
> > @@ -978,7 +980,13 @@ void release_console_sem(void)
> > console_locked = 0;
> > up(&console_sem);
>
> Hmm, just looking at this fragment: Doesn't up() include the risk of
> running onto the runqueue lock as well?

Very little risk (if any). If printk fails to get the console_sem it
doesn't block. So there would be no waiters on the semaphore, and thus
try_to_wake_up would not be called. The only place I see the down
actually being called is in suspend code, and even then, we would need to
lock the rq of the task that is trying to grab the console_sem and the
deadlock would only occur if that was on the same CPU. And honestly, I'm
not sure that's even possible.

-- Steve


>
> > spin_unlock_irqrestore(&logbuf_lock, flags);
> > - if (wake_klogd)
> > + /*
> > + * If we try to wake up klogd while printing with the runqueue lock
> > + * held, this will deadlock. We don't have access to the runqueue
> > + * lock from here, but just checking for interrupts disabled
> > + * should be enough.
> > + */
> > + if (!irqs_disabled() && wake_klogd)
> > wake_up_klogd();
> > }
> > EXPORT_SYMBOL(release_console_sem);
>
> Jan
>


\
 
 \ /
  Last update: 2008-01-18 17:31    [W:0.105 / U:0.144 seconds]
©2003-2011 Jasper Spaans. Advertise on this site