Re: [PATCH] O16int for interactivity

[Nick Piggin]

Timothy Miller wrote:

> Con Kolivas wrote:
>
> > > A hardware timer interrupt happens at timeslice granularity.  If the
> > > interrupt occurs, but the timeslice is not expired, then NORMALLY, the
> > > ISR would just return right back to the running task, but sometimes, it
> > > might decided to end the timeslice early and run some other task.
> > >
> > > Right?
> >
> >
> >
> > No, the timeslice granularity is a hard cut off where a task gets 
> > rescheduled and put at the back of the queue again. If there is no 
> > other task of equal or better priority it will just start again.
>
>
>
> Hmmm... I'm still having trouble making sense of this.
>
> So, it seems that you're saying that all tasks, regardless of 
> timeslice length, are interrupted every 10ms (at 100hz). 


Interrupted - in that scheduler_tick is run, yes. They don't get switched.

>   If another task exists at a higher priority, then it gets run at 
> that point.


Well loosely, yes. Actually, it happens if the task exists and is "running",
and has timeslice left. That only happens in scheduler_tick when the 
current
task has finished its timeslice and the priority arrays are about to be
switched. The required conditions for preemption can also occur when a task
is being woken up, (after sleeping or newly forked).

> However, if there is more than one task at a given priority level, 
> then they will not round-robin until the current task has used up all 
> of its timeslice (some integer multiple of 10ms). 


Thats right. Although I think Con or Ingo recently changed this in mm

> Am I finally correct, or do I still have it wrong?  :)

Sounds right to me.

-
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/