Re: [RFC] Imprecise timers.

[Rene Herman <rene.herman@keyaccess.nl>]

On 22-07-08 14:54, Arjan van de Ven wrote:

> On Tue, 22 Jul 2008 09:19:02 +0200
> Rene Herman <rene.herman@keyaccess.nl> wrote:
> 
>> On 22-07-08 05:02, David Woodhouse wrote:
>>
>>> Many users of timers don't really care too much about exactly when
>>> their timer fires -- and waking a CPU to satisfy such a timer is a
>>> waste of power. This patch implements a 'range' timer which will
>>> fire at a 'convenient' moment within given constraints.
>>>
>>> It's implemented by a deferrable timer at the beginning of the
>>> range, which will run some time later when the CPU happens to be
>>> awake. And a non-deferrable timer at the hard deadline, to ensure
>>> it really does happen by then.
>> Are there actually users for this (not just in theory)? The
>> deferrable timer sort of sounds like all I'd ever want if I, as you
>> say, wouldn't really care...
> 
> there's a few; mostly around hardware timeout..For example Stephen want
> it for his drivers.

Hardware I've dealt with is (almost? can't remember anything else) 
exlusively minimal delays and as such this thing seemed like perhaps a 
bit over-apisized...

> EXT3 journal flushing is another one where we can easily say 
> "between 4 and 7 seconds" rather than "exactly at 5"

This a nice-ish example though. It might be considered necessary to make 
the current commit delay when set explicitly be the non-deferrable upper 
bound but almost none do I guess.

Rene.