2014-10-27 03:13:37

by Yao Dongdong

Subject: [PATCH] sched:skip loop non-idle cpus after find an idle cpu while find_idlest_cpu

Idle cpu is idler than non-idle cpu, so we needn't loop non-idle cpus after find an idle cpu.

Signed-off-by:[email protected]

---
kernel/sched/fair.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 0b069bf..2445a23 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -4446,7 +4446,7 @@ find_idlest_cpu(struct sched_group *group, struct task_struct *p, int this_cpu)
latest_idle_timestamp = rq->idle_stamp;
shallowest_idle_cpu = i;
}
- } else {
+ } else if (shallowest_idle_cpu == -1) {
load = weighted_cpuload(i);
if (load < min_load || (load == min_load && i == this_cpu)) {
min_load = load;
--
1.8.0.1

2014-10-27 08:05:31

by Srikar Dronamraju

Subject: Re: [PATCH] sched:skip loop non-idle cpus after find an idle cpu while find_idlest_cpu

* Yao Dongdong <[email protected]> [2014-10-27 11:13:17]:

> Idle cpu is idler than non-idle cpu, so we needn't loop non-idle cpus after find an idle cpu.
>

While this check looks good,
I dont see how we are avoiding a loop?

Are you suggesting that we break of the for loop once we set
shallowest_idle_cpu?

--
Thanks and Regards
Srikar


> Signed-off-by:[email protected]
>
> ---
> kernel/sched/fair.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 0b069bf..2445a23 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -4446,7 +4446,7 @@ find_idlest_cpu(struct sched_group *group, struct task_struct *p, int this_cpu)
> latest_idle_timestamp = rq->idle_stamp;
> shallowest_idle_cpu = i;
> }
> - } else {
> + } else if (shallowest_idle_cpu == -1) {
> load = weighted_cpuload(i);
> if (load < min_load || (load == min_load && i == this_cpu)) {
> min_load = load;
> --
> 1.8.0.1
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/

--
Thanks and Regards
Srikar Dronamraju

2014-10-27 08:58:14

by Yao Dongdong

Subject: Re: [PATCH] sched:skip loop non-idle cpus after find an idle cpu while find_idlest_cpu

On 2014/10/27 16:04, Srikar Dronamraju wrote:
> * Yao Dongdong <[email protected]> [2014-10-27 11:13:17]:
>
>> Idle cpu is idler than non-idle cpu, so we needn't loop non-idle cpus after find an idle cpu.
>>
> While this check looks good,
> I dont see how we are avoiding a loop?
>
> Are you suggesting that we break of the for loop once we set
> shallowest_idle_cpu?
>
Thanks for your review.

I agree that the submit message is a bit misleading, what this patch do is just avoiding search for least_loaded_cpu
in non-idle cpus.

sched:add check of shallowest_idle_cpu before search for least_loaded_cpu

Idle cpu is idler than non-idle cpu, so we needn't search for least_loaded_cpu
after we have found an idle cpu.

what about this?

Thanks,
Yao Dongdong

2014-10-27 09:23:29

by Srikar Dronamraju

Subject: Re: [PATCH] sched:skip loop non-idle cpus after find an idle cpu while find_idlest_cpu

* Yao Dongdong <[email protected]> [2014-10-27 16:58:05]:

> On 2014/10/27 16:04, Srikar Dronamraju wrote:
> > * Yao Dongdong <[email protected]> [2014-10-27 11:13:17]:
> >
> >> Idle cpu is idler than non-idle cpu, so we needn't loop non-idle cpus after find an idle cpu.
> >>
> > While this check looks good,
> > I dont see how we are avoiding a loop?
> >
> > Are you suggesting that we break of the for loop once we set
> > shallowest_idle_cpu?
> >
> Thanks for your review.
>
> I agree that the submit message is a bit misleading, what this patch do is just avoiding search for least_loaded_cpu
> in non-idle cpus.
>
> sched:add check of shallowest_idle_cpu before search for least_loaded_cpu
>
> Idle cpu is idler than non-idle cpu, so we needn't search for least_loaded_cpu
> after we have found an idle cpu.
>

Looks good for me.

Reviewed-by: Srikar Dronamraju <[email protected]>


--
Thanks and Regards
Srikar Dronamraju