Re: [patch 05/15] Generic Mutex Subsystem, mutex-core.patch

[Ingo Molnar <mingo@elte.hu>]

* Oleg Nesterov <oleg@tv-sign.ru> wrote:

> > +start_mutex_timer(struct timer_list *timer, unsigned long time,
> > +		  unsigned long *expire)
> > +{
> > +	*expire = time + jiffies;
> > +	init_timer(timer);
> > +	timer->expires = *expire;
> > +	timer->data = (unsigned long)current;
> > +	timer->function = process_timeout;
> > +	add_timer(timer);
> > +}
> 
> How about
> 	setup_timer(&timer, process_timeout, (unsigned long)current);
> 	__mod_timer(&timer, *expire);
> ?

i've removed the timer code from the latest queue - because it's unused 
and we can add it back later. But you are right, it was racy (Arjan 
noticed the signal race too), because this codepath in -rt relied on 
guaranteed lock-passing. (while mutex.c does a trylock)

	Ingo