Re: [PATCH v3 5/6] timerfd: Add support for deferrable timers

[Anton Vorontsov <anton@enomsg.org>]

On Thu, Feb 20, 2014 at 12:09:43PM +0100, Thomas Gleixner wrote:
> On Thu, 20 Feb 2014, Alexey Perevalov wrote:
> > From: Anton Vorontsov <anton@enomsg.org>
> > 
> > This patch implements a userland-side API for generic deferrable timers,
> > per linux/timer.h:
> > 
> >  * A deferrable timer will work normally when the system is busy, but
> >  * will not cause a CPU to come out of idle just to service it; instead,
> >  * the timer will be serviced when the CPU eventually wakes up with a
> >  * subsequent non-deferrable timer.
> > 
> > These timers are crucial for power saving, i.e. periodic tasks that want
> > to work in background when the system is under use, but don't want to
> > cause wakeups themselves.
> > 
> > The deferred timers are somewhat orthogonal to high-res external timers,
> > since the deferred timer is tied to the system load, not just to some
> > external decrementer source.
> 
> Again this changelog makes no sense. What's orthogonal to high-res
> timers and why are they external?

Not trying to defend the current series, just felt the need clarify this
one.

By orthogonal I meant that comparing to high resolution timers' use cases,
deferred timers can be super-low resolution, super inaccurate. We don't
know exactly when they will fire, all we know is something like "every 0.2
seconds, iff the system/user is doing something, otherwise don't bother."

As for external (my bad, shouldn't invent personal terminology): the
hrtimers are tied to some clock source (which is "external" to me), but
deferred timers are mostly tied to the system's activity.

Thanks,

Anton