Re: [PATCH tip/core/rcu 01/15] rcu: Add PROVE_RCU_DELAY to provokedifficult races

[Peter Zijlstra]

On Thu, 2012-09-06 at 13:51 -0700, Paul E. McKenney wrote:
> On Thu, Sep 06, 2012 at 04:38:32PM +0200, Peter Zijlstra wrote:
> > On Thu, 2012-08-30 at 11:56 -0700, Paul E. McKenney wrote:
> > > +#ifdef CONFIG_PROVE_RCU_DELAY
> > > +               udelay(10); /* Make preemption more probable. */
> > 		cond_resched(); /* for extra fun? */
> 
> The additional fun could include "scheduling while atomic", so I will
> pass.  ;-)
> 
> (The problem is that __rcu_read_unlock() can be called with interrupts
> disabled, among other things.)

Hmm, too bad. Without a preemption point here you're relying on forced
preemption, which of course can only happen on PREEMPT=y kernels.

> > > +#endif /* #ifdef CONFIG_PROVE_RCU_DELAY */ 
> > 
> > 
> 

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/