sched/psi: avoid unnecessary resetting min update period when destroy trigger

[PATCH linux-next] sched/psi: avoid unnecessary resetting min update period when destroy trigger

Posted by yang.yang29@zte.com.cn 2 years, 8 months ago

From: Yang Yang <yang.yang19@zte.com.cn>

Psi_group's poll_min_period is determined by the min window size of
psi_trigger when creating new triggers. While destroying a psi_trigger,
there is no need to reset poll_min_period if the destroying psi_trigger
not had the min windows size, since in this condition poll_min_period
will keep the same as before.

Signed-off-by: Yang Yang <yang.yang29@zte.com.cn>
---
 kernel/sched/psi.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
index 02e011cabe91..12869585cf89 100644
--- a/kernel/sched/psi.c
+++ b/kernel/sched/psi.c
@@ -1359,11 +1359,16 @@ void psi_trigger_destroy(struct psi_trigger *t)
 		group->nr_triggers[t->state]--;
 		if (!group->nr_triggers[t->state])
 			group->poll_states &= ~(1 << t->state);
-		/* reset min update period for the remaining triggers */
-		list_for_each_entry(tmp, &group->triggers, node)
-			period = min(period, div_u64(tmp->win.size,
-					UPDATES_PER_WINDOW));
-		group->poll_min_period = period;
+		/*
+		 * Reset min update period for the remaining triggers iff the destroying
+		 * trigger had the min window size.
+		 */
+		if (group->poll_min_period == div_u64(t->win.size, UPDATES_PER_WINDOW)) {
+			list_for_each_entry(tmp, &group->triggers, node)
+				period = min(period, div_u64(tmp->win.size,
+						UPDATES_PER_WINDOW));
+			group->poll_min_period = period;
+		}
 		/* Destroy poll_task when the last trigger is destroyed */
 		if (group->poll_states == 0) {
 			group->polling_until = 0;
-- 
2.25.1

Re: [PATCH linux-next] sched/psi: avoid unnecessary resetting min update period when destroy trigger

Posted by Suren Baghdasaryan 2 years, 8 months ago

On Mon, Apr 17, 2023 at 1:12 AM <yang.yang29@zte.com.cn> wrote:
>
> From: Yang Yang <yang.yang19@zte.com.cn>
>
> Psi_group's poll_min_period is determined by the min window size of
> psi_trigger when creating new triggers. While destroying a psi_trigger,
> there is no need to reset poll_min_period if the destroying psi_trigger
> not had the min windows size, since in this condition poll_min_period
> will keep the same as before.

Nice optimization.

>
> Signed-off-by: Yang Yang <yang.yang29@zte.com.cn>

Acked-by: Suren Baghdasaryan <surenb@google.com>

> ---
>  kernel/sched/psi.c | 15 ++++++++++-----
>  1 file changed, 10 insertions(+), 5 deletions(-)
>
> diff --git a/kernel/sched/psi.c b/kernel/sched/psi.c
> index 02e011cabe91..12869585cf89 100644
> --- a/kernel/sched/psi.c
> +++ b/kernel/sched/psi.c
> @@ -1359,11 +1359,16 @@ void psi_trigger_destroy(struct psi_trigger *t)
>                 group->nr_triggers[t->state]--;
>                 if (!group->nr_triggers[t->state])
>                         group->poll_states &= ~(1 << t->state);
> -               /* reset min update period for the remaining triggers */
> -               list_for_each_entry(tmp, &group->triggers, node)
> -                       period = min(period, div_u64(tmp->win.size,
> -                                       UPDATES_PER_WINDOW));
> -               group->poll_min_period = period;
> +               /*
> +                * Reset min update period for the remaining triggers iff the destroying
> +                * trigger had the min window size.
> +                */
> +               if (group->poll_min_period == div_u64(t->win.size, UPDATES_PER_WINDOW)) {
> +                       list_for_each_entry(tmp, &group->triggers, node)
> +                               period = min(period, div_u64(tmp->win.size,
> +                                               UPDATES_PER_WINDOW));
> +                       group->poll_min_period = period;
> +               }
>                 /* Destroy poll_task when the last trigger is destroyed */
>                 if (group->poll_states == 0) {
>                         group->polling_until = 0;
> --
> 2.25.1

Re: [PATCH linux-next] sched/psi: avoid unnecessary resetting min update period when destroy trigger

Posted by Yang Yang 2 years, 7 months ago

>On Mon, Apr 17, 2023 at 1:12 AM <yang.yang29@zte.com.cn> wrote:
>>
>> From: Yang Yang <yang.yang19@zte.com.cn>
>>
>> Psi_group's poll_min_period is determined by the min window size of
>> psi_trigger when creating new triggers. While destroying a psi_trigger,
>> there is no need to reset poll_min_period if the destroying psi_trigger
>> not had the min windows size, since in this condition poll_min_period
>> will keep the same as before.
>
> Nice optimization.
>
>> Signed-off-by: Yang Yang <yang.yang29@zte.com.cn>
> Acked-by: Suren Baghdasaryan <surenb@google.com>

Hi, would you please take a review?
Thanks!

Re: [PATCH linux-next] sched/psi: avoid unnecessary resetting min update period when destroy trigger

Posted by Yang Yang 2 years, 7 months ago

>On Mon, Apr 17, 2023 at 1:12 AM <yang.yang29@zte.com.cn> wrote:
>>
>> From: Yang Yang <yang.yang19@zte.com.cn>
>>
>> Psi_group's poll_min_period is determined by the min window size of
>> psi_trigger when creating new triggers. While destroying a psi_trigger,
>> there is no need to reset poll_min_period if the destroying psi_trigger
>> not had the min windows size, since in this condition poll_min_period
>> will keep the same as before.
>
> Nice optimization.
>
>> Signed-off-by: Yang Yang <yang.yang29@zte.com.cn>
> Acked-by: Suren Baghdasaryan <surenb@google.com>

Hi, would you please take a review?
Thanks!