Re: [PATCH] preempt abstraction

[Andrew Morton <akpm@zip.com.au>]

Christoph Hellwig wrote:
> 
> On Tue, Jan 08, 2002 at 01:57:28PM -0500, Robert Love wrote:
> > Why not use the more commonly named conditional_schedule instead of
> > preempt() ?  In addition to being more in-use (low-latency, lock-break,
> > and Andrea's aa patch all use it) I think it better conveys its meaning,
> > which is a schedule() but only conditionally.
> 
> I think the choice is very subjective, but I prefer preempt().
> It's nicely short to type (!) and similar in spirit to Ingo's yield()..
> 

naah.  preempt() means preempt.  But the implementation
is in fact maybe_preempt(), or preempt_if_needed().

I use (verbosely) (simplified):

#define conditional_schedule_needed()	unlikely(current->need_resched)
#define unconditional_schedule()	do {
						__set_current_state(TASK_RUNNING)
						schedule();
					} while(0);
#define conditional_schedule()		if (conditional_schedule_needed())
						unconditional_schedule();

...

foo()
{
	...
	conditional_schedule();
	...
}

bar()
{
	...
	if (conditional_schedule_needed()) {
		spin_unlock(&piggy_lock);
		unconditional_schedule();
		spin_lock(&piggy_lock);
		goto clean_up_mess;
	}
	...
}