Re: [PATCH v3 8/9] KVM-GST: adjust scheduler cpu power

[Peter Zijlstra <peterz@xxxxxxxxxxxxx>]

On Wed, 2011-06-29 at 11:29 -0400, Glauber Costa wrote:
> +#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING
> +       if (static_branch((&paravirt_steal_rq_enabled))) {
> +               int is_idle;
> +               u64 st;
> +
> +               is_idle = ((rq->curr != rq->idle) ||
> +                               irq_count() != HARDIRQ_OFFSET);

Now that hurt my brain. If the vcpu is idle, why does it want to run?
How can an idle vcpu ever rack up steal time?

Also, what's that HARDIRQ_OFFSET bit about? sorely lacking in
explanation, and the Changelog to this patch is about as bad as the last
one.

> +               __touch_steal_time(is_idle, delta, &st);
> +
> +               steal = st * TICK_NSEC;
> +
> +               delta -= steal;
> +       }
> +#endif 
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html