Re: [PATCH 4/7] sched: Add a preempt count base offset to __might_sleep()

[Peter Zijlstra <a.p.zijlstra@chello.nl>]

On Thu, 2009-07-16 at 02:28 -0400, Frederic Weisbecker wrote:
> +++ b/include/linux/hardirq.h
> @@ -103,6 +103,13 @@
>   */
>  #define in_atomic()    ((preempt_count() & ~PREEMPT_ACTIVE) != PREEMPT_INATOMIC_BASE)
>  
> +static inline int current_preempt_equals(int preempt_offset)
> +{
> +       int nested = preempt_count() & ~PREEMPT_ACTIVE;
> +
> +       return (nested == PREEMPT_INATOMIC_BASE + preempt_offset);
> +}

I'm not sure about it being in hardirq.h, I think we should keep this in
sched.c.

Also, I'm not sure about the name, but then I suck at naming too. How
about something like: preempt_count_equals() ?

Other than that the series looks nice and I've got it queued.