Re: [PATCHv2 11/11] writeback: prevent unnecessary bdi threads wakeups

[Artem Bityutskiy <dedekind1@gmail.com>]

On Thu, 2010-07-22 at 05:00 -0400, Christoph Hellwig wrote:
> On Thu, Jul 22, 2010 at 10:41:55AM +1000, Dave Chinner wrote:
> > 	if (wakeup_bdi) {
> > 		trace_writeback_wakeup(bdi)
> > 		spin_lock(&bdi->wb_lock);
> > 		if (!bdi->wb.task) {{
> > 			trace_writeback_wakeup_nothread(bdi);
> > 			wake_up_process(default_backing_dev_info.wb.task);
> > 		} else
> > 			wake_up_process(bdi->wb.task);
> > 		spin_unlock(&bdi->wb_lock);
> 
> That gives us duplicate traces for the first case, what about:
> 
> 	 if (wakeup_bdi) {
> 	 	spin_lock(&bdi->wb_lock);
> 		if (bdi->wb.task) {
> 			trace_writeback_wake_thread(bdi);
> 			wake_up_process(bdi->wb.task);
> 		} else {
> 			trace_writeback_wake_forker_thread(bdi);
> 			wake_up_process(default_backing_dev_info.wb.task);
> 		}
> 		spin_unlock(&bdi->wb_lock);
> 	}

But Dave's version is what we have in 'bdi_queue_work()' as well. I it
is OK - first trace point is triggered every time the bdi thread is
_wanted_ to be woken up.

Then if it does not exist, we need to do something special to wake it up
(ask the forker thread which to create it). This is a different event
and we use a different trace point.

-- 
Best Regards,
Artem Bityutskiy (Артём Битюцкий)

--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html