2022-04-06 15:48:53

by zgpeng

[permalink] [raw]
Subject: [PATCH] sched/fair: Move calculate of avg_load to a better location

In calculate_imbalance function, when the value of local->avg_load is
greater than or equal to busiest->avg_load, the calculated sds->avg_load is
not used. So this calculation can be placed in a more appropriate position.

Signed-off-by: zgpeng <[email protected]>
Reviewed-by: Samuel Liao <[email protected]>
---
kernel/sched/fair.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index d4bd299..601f8bd 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -9460,8 +9460,6 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
local->avg_load = (local->group_load * SCHED_CAPACITY_SCALE) /
local->group_capacity;

- sds->avg_load = (sds->total_load * SCHED_CAPACITY_SCALE) /
- sds->total_capacity;
/*
* If the local group is more loaded than the selected
* busiest group don't try to pull any tasks.
@@ -9470,6 +9468,9 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
env->imbalance = 0;
return;
}
+
+ sds->avg_load = (sds->total_load * SCHED_CAPACITY_SCALE) /
+ sds->total_capacity;
}

/*
--
2.9.5


2022-04-06 17:22:48

by Vincent Guittot

[permalink] [raw]
Subject: Re: [PATCH] sched/fair: Move calculate of avg_load to a better location

On Wed, 6 Apr 2022 at 11:57, zgpeng <[email protected]> wrote:
>
> In calculate_imbalance function, when the value of local->avg_load is
> greater than or equal to busiest->avg_load, the calculated sds->avg_load is
> not used. So this calculation can be placed in a more appropriate position.
>
> Signed-off-by: zgpeng <[email protected]>
> Reviewed-by: Samuel Liao <[email protected]>

Reviewed-by: Vincent Guittot <[email protected]>

> ---
> kernel/sched/fair.c | 5 +++--
> 1 file changed, 3 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index d4bd299..601f8bd 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -9460,8 +9460,6 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
> local->avg_load = (local->group_load * SCHED_CAPACITY_SCALE) /
> local->group_capacity;
>
> - sds->avg_load = (sds->total_load * SCHED_CAPACITY_SCALE) /
> - sds->total_capacity;
> /*
> * If the local group is more loaded than the selected
> * busiest group don't try to pull any tasks.
> @@ -9470,6 +9468,9 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
> env->imbalance = 0;
> return;
> }
> +
> + sds->avg_load = (sds->total_load * SCHED_CAPACITY_SCALE) /
> + sds->total_capacity;
> }
>
> /*
> --
> 2.9.5
>

Subject: [tip: sched/core] sched/fair: Move calculate of avg_load to a better location

The following commit has been merged into the sched/core branch of tip:

Commit-ID: 06354900787f25bf5be3c07a68e3cdbc5bf0fa69
Gitweb: https://git.kernel.org/tip/06354900787f25bf5be3c07a68e3cdbc5bf0fa69
Author: zgpeng <[email protected]>
AuthorDate: Wed, 06 Apr 2022 17:57:05 +08:00
Committer: Peter Zijlstra <[email protected]>
CommitterDate: Fri, 22 Apr 2022 12:14:07 +02:00

sched/fair: Move calculate of avg_load to a better location

In calculate_imbalance function, when the value of local->avg_load is
greater than or equal to busiest->avg_load, the calculated sds->avg_load is
not used. So this calculation can be placed in a more appropriate position.

Signed-off-by: zgpeng <[email protected]>
Signed-off-by: Peter Zijlstra (Intel) <[email protected]>
Reviewed-by: Samuel Liao <[email protected]>
Reviewed-by: Vincent Guittot <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
kernel/sched/fair.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index d4bd299..601f8bd 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -9460,8 +9460,6 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
local->avg_load = (local->group_load * SCHED_CAPACITY_SCALE) /
local->group_capacity;

- sds->avg_load = (sds->total_load * SCHED_CAPACITY_SCALE) /
- sds->total_capacity;
/*
* If the local group is more loaded than the selected
* busiest group don't try to pull any tasks.
@@ -9470,6 +9468,9 @@ static inline void calculate_imbalance(struct lb_env *env, struct sd_lb_stats *s
env->imbalance = 0;
return;
}
+
+ sds->avg_load = (sds->total_load * SCHED_CAPACITY_SCALE) /
+ sds->total_capacity;
}

/*