Re: [PATCH] O12.2int for interactivity

[Timothy Miller]

Con Kolivas wrote:
> On Fri, 15 Aug 2003 06:00, Timothy Miller wrote:
>
> > If my guess from my previous email was correct (that is pri 5 gets
> > shorter timeslide than pri 6), then that means that tasks of higher
> > static priority have are penalized more than lower pri tasks for expiring.
> >
> > Say a task has to run for 15ms.  If it's at a priority that gives it a
> > 10ms timeslice, then it'll expire and get demoted.  If it's at a
> > priority that gives it a 20ms timeslice, then it'll not expire and
> > therefore get promoted.
> >
> > Is that fair?
>
>
> Yes, it's a simple cutoff at the end of the timeslice. If you use up the 
> timeslice allocated to you, then you have to pass a test to see if you can go 
> onto the active array or get expired. Since higher static priority (lower 
> nice) tasks get longer timeslices, they are less likely to expire unless they 
> are purely cpu bound and never sleep.


Ok, I'm just a little confused, because of this inversion of "high 
priority" with "low numbers".

First, am I correct in understanding that a lower number means a higher 
priority?

And for a higher priority, in addition to begin run before all tasks of 
lower priority, they also get a longer timeslice?


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/