[PATCH] sched/numa: initialize numa statistics when forking new task

[PATCH] sched/numa: initialize numa statistics when forking new task

[Honglei Wang]

The child processes will inherit numa_pages_migrated and
total_numa_faults from the parent. It means even if there is no numa
fault happen on the child, the statistics in /proc/$pid of the child
process might show huge amount. This is a bit weird. Let's initialize
them when do fork.

Signed-off-by: Honglei Wang <wanghonglei@didichuxing.com>
---
 kernel/sched/fair.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 6e476f6d9435..1aa0ec123a4b 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2826,6 +2826,8 @@ void init_numa_balancing(unsigned long clone_flags, struct task_struct *p)
 	/* Protect against double add, see task_tick_numa and task_numa_work */
 	p->numa_work.next		= &p->numa_work;
 	p->numa_faults			= NULL;
+	p->numa_pages_migrated		= 0;
+	p->total_numa_faults		= 0;
 	RCU_INIT_POINTER(p->numa_group, NULL);
 	p->last_task_numa_placement	= 0;
 	p->last_sum_exec_runtime	= 0;
-- 
2.14.1

Re:[PATCH] sched/numa: initialize numa statistics when forking new task

[Honglei Wang]

Hello, friendly ping... any comments about this?

At 2022-01-13 21:39:20, "Honglei Wang" <wanghonglei@didichuxing.com> wrote:
>The child processes will inherit numa_pages_migrated and
>total_numa_faults from the parent. It means even if there is no numa
>fault happen on the child, the statistics in /proc/$pid of the child
>process might show huge amount. This is a bit weird. Let's initialize
>them when do fork.
>
>Signed-off-by: Honglei Wang <wanghonglei@didichuxing.com>
>---
> kernel/sched/fair.c | 2 ++
> 1 file changed, 2 insertions(+)
>
>diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>index 6e476f6d9435..1aa0ec123a4b 100644
>--- a/kernel/sched/fair.c
>+++ b/kernel/sched/fair.c
>@@ -2826,6 +2826,8 @@ void init_numa_balancing(unsigned long clone_flags, struct task_struct *p)
> 	/* Protect against double add, see task_tick_numa and task_numa_work */
> 	p->numa_work.next		= &p->numa_work;
> 	p->numa_faults			= NULL;
>+	p->numa_pages_migrated		= 0;
>+	p->total_numa_faults		= 0;
> 	RCU_INIT_POINTER(p->numa_group, NULL);
> 	p->last_task_numa_placement	= 0;
> 	p->last_sum_exec_runtime	= 0;
>-- 
>2.14.1

Re: [PATCH] sched/numa: initialize numa statistics when forking new task

[Honglei Wang]

Hello, friendly ping... any comments about this?

At 2022-01-13 21:39:20, "Honglei Wang" <wanghonglei@didichuxing.com> wrote:
> The child processes will inherit numa_pages_migrated and
> total_numa_faults from the parent. It means even if there is no numa
> fault happen on the child, the statistics in /proc/$pid of the child
> process might show huge amount. This is a bit weird. Let's initialize
> them when do fork.
>
> Signed-off-by: Honglei Wang <wanghonglei@didichuxing.com>
> ---
>   kernel/sched/fair.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 6e476f6d9435..1aa0ec123a4b 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -2826,6 +2826,8 @@ void init_numa_balancing(unsigned long clone_flags, struct task_struct *p)
> 	/* Protect against double add, see task_tick_numa and task_numa_work */
> 	p->numa_work.next		= &p->numa_work;
> 	p->numa_faults			= NULL;
> +	p->numa_pages_migrated		= 0;
> +	p->total_numa_faults		= 0;
> 	RCU_INIT_POINTER(p->numa_group, NULL);
> 	p->last_task_numa_placement	= 0;
> 	p->last_sum_exec_runtime	= 0;
> -- 
> 2.14.1

Re: [PATCH] sched/numa: initialize numa statistics when forking new task

[Mel Gorman]

On Thu, Jan 13, 2022 at 09:39:20PM +0800, Honglei Wang wrote:
> The child processes will inherit numa_pages_migrated and
> total_numa_faults from the parent. It means even if there is no numa
> fault happen on the child, the statistics in /proc/$pid of the child
> process might show huge amount. This is a bit weird. Let's initialize
> them when do fork.
> 
> Signed-off-by: Honglei Wang <wanghonglei@didichuxing.com>

Acked-by: Mel Gorman <mgorman@suse.de>

-- 
Mel Gorman
SUSE Labs

[tip: sched/core] sched/numa: initialize numa statistics when forking new task

[tip-bot2 for Honglei Wang]

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

Commit-ID:     12bf8a7eb84e4d3547ebfd89bb0a9255a0f2acc7
Gitweb:        https://git.kernel.org/tip/12bf8a7eb84e4d3547ebfd89bb0a9255a0f2acc7
Author:        Honglei Wang <wanghonglei@didichuxing.com>
AuthorDate:    Thu, 13 Jan 2022 21:39:20 +08:00
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Thu, 27 Jan 2022 12:57:18 +01:00

sched/numa: initialize numa statistics when forking new task

The child processes will inherit numa_pages_migrated and
total_numa_faults from the parent. It means even if there is no numa
fault happen on the child, the statistics in /proc/$pid of the child
process might show huge amount. This is a bit weird. Let's initialize
them when do fork.

Signed-off-by: Honglei Wang <wanghonglei@didichuxing.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Mel Gorman <mgorman@suse.de>
Link: https://lore.kernel.org/r/20220113133920.49900-1-wanghonglei@didichuxing.com
---
 kernel/sched/fair.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 095b0aa..5dca13f 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2825,6 +2825,8 @@ void init_numa_balancing(unsigned long clone_flags, struct task_struct *p)
 	/* Protect against double add, see task_tick_numa and task_numa_work */
 	p->numa_work.next		= &p->numa_work;
 	p->numa_faults			= NULL;
+	p->numa_pages_migrated		= 0;
+	p->total_numa_faults		= 0;
 	RCU_INIT_POINTER(p->numa_group, NULL);
 	p->last_task_numa_placement	= 0;
 	p->last_sum_exec_runtime	= 0;