[PATCH 0/3] sched: code refine/clean for cfs-bandwidth

[PATCH 0/3] sched: code refine/clean for cfs-bandwidth

[Michael Wang]

Code refine and clean patch set.

Michael Wang (3):
	[PATCH 1/3] sched: don't repeat the initialization in sched_init()
	[PATCH 2/3] sched: code refine in unthrottle_cfs_rq()
	[PATCH 3/3] sched: remove the useless declaration in kernel/sched/fair.c

--- 
 b/kernel/sched/core.c |   46 +++++++++++++++++++++++++---------------------
 b/kernel/sched/fair.c |    2 +-
 kernel/sched/fair.c   |    4 ----
 3 files changed, 26 insertions(+), 26 deletions(-)

[PATCH 1/3] sched: don't repeat the initialization in sched_init()

[Michael Wang]

In sched_init(), there is no need to initialize 'root_task_group.shares' and
'root_task_group.cfs_bandwidth' repeatedly.

CC: Ingo Molnar <mingo@kernel.org>
CC: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 kernel/sched/core.c |   46 +++++++++++++++++++++++++---------------------
 1 files changed, 25 insertions(+), 21 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 58453b8..c0c3716 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -6955,6 +6955,31 @@ void __init sched_init(void)
 
 #endif /* CONFIG_CGROUP_SCHED */
 
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	root_task_group.shares = ROOT_TASK_GROUP_LOAD;
+
+	/*
+	 * How much cpu bandwidth does root_task_group get?
+	 *
+	 * In case of task-groups formed thr' the cgroup filesystem, it
+	 * gets 100% of the cpu resources in the system. This overall
+	 * system cpu resource is divided among the tasks of
+	 * root_task_group and its child task-groups in a fair manner,
+	 * based on each entity's (task or task-group's) weight
+	 * (se->load.weight).
+	 *
+	 * In other words, if root_task_group has 10 tasks of weight
+	 * 1024) and two child groups A0 and A1 (of weight 1024 each),
+	 * then A0's share of the cpu resource is:
+	 *
+	 *	A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33%
+	 *
+	 * We achieve this by letting root_task_group's tasks sit
+	 * directly in rq->cfs (i.e root_task_group->se[] = NULL).
+	 */
+	init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+#endif
+
 	for_each_possible_cpu(i) {
 		struct rq *rq;
 
@@ -6966,28 +6991,7 @@ void __init sched_init(void)
 		init_cfs_rq(&rq->cfs);
 		init_rt_rq(&rq->rt, rq);
 #ifdef CONFIG_FAIR_GROUP_SCHED
-		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
 		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
-		/*
-		 * How much cpu bandwidth does root_task_group get?
-		 *
-		 * In case of task-groups formed thr' the cgroup filesystem, it
-		 * gets 100% of the cpu resources in the system. This overall
-		 * system cpu resource is divided among the tasks of
-		 * root_task_group and its child task-groups in a fair manner,
-		 * based on each entity's (task or task-group's) weight
-		 * (se->load.weight).
-		 *
-		 * In other words, if root_task_group has 10 tasks of weight
-		 * 1024) and two child groups A0 and A1 (of weight 1024 each),
-		 * then A0's share of the cpu resource is:
-		 *
-		 *	A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33%
-		 *
-		 * We achieve this by letting root_task_group's tasks sit
-		 * directly in rq->cfs (i.e root_task_group->se[] = NULL).
-		 */
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
 		init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 
-- 
1.7.4.1

[PATCH 2/3] sched: code refine in unthrottle_cfs_rq()

[Michael Wang]

Directly use rq to save some code.

CC: Ingo Molnar <mingo@kernel.org>
CC: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 kernel/sched/fair.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index c61a614..1e10911 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2298,7 +2298,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
 	int enqueue = 1;
 	long task_delta;
 
-	se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))];
+	se = cfs_rq->tg->se[cpu_of(rq)];
 
 	cfs_rq->throttled = 0;
 	raw_spin_lock(&cfs_b->lock);
-- 
1.7.4.1

[PATCH 3/3] sched: remove the useless declaration in kernel/sched/fair.c

[Michael Wang]

default_cfs_period(), do_sched_cfs_period_timer(), do_sched_cfs_slack_timer()
already defined previously, no need to declare again.

CC: Ingo Molnar <mingo@kernel.org>
CC: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 kernel/sched/fair.c |    4 ----
 1 files changed, 0 insertions(+), 4 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index c61a614..73cad33 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2599,10 +2599,6 @@ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq)
 	throttle_cfs_rq(cfs_rq);
 }
 
-static inline u64 default_cfs_period(void);
-static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun);
-static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b);
-
 static enum hrtimer_restart sched_cfs_slack_timer(struct hrtimer *timer)
 {
 	struct cfs_bandwidth *cfs_b =
-- 
1.7.4.1

Re: [PATCH 1/3] sched: don't repeat the initialization in sched_init()

[Paul Turner]

On Mon, Jun 3, 2013 at 11:23 PM, Michael Wang
<wangyun@linux.vnet.ibm.com> wrote:
> In sched_init(), there is no need to initialize 'root_task_group.shares' and
> 'root_task_group.cfs_bandwidth' repeatedly.
>
> CC: Ingo Molnar <mingo@kernel.org>
> CC: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
> ---
>  kernel/sched/core.c |   46 +++++++++++++++++++++++++---------------------
>  1 files changed, 25 insertions(+), 21 deletions(-)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 58453b8..c0c3716 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -6955,6 +6955,31 @@ void __init sched_init(void)
>
>  #endif /* CONFIG_CGROUP_SCHED */
>
> +#ifdef CONFIG_FAIR_GROUP_SCHED
> +       root_task_group.shares = ROOT_TASK_GROUP_LOAD;
> +
> +       /*
> +        * How much cpu bandwidth does root_task_group get?
> +        *
> +        * In case of task-groups formed thr' the cgroup filesystem, it
> +        * gets 100% of the cpu resources in the system. This overall
> +        * system cpu resource is divided among the tasks of
> +        * root_task_group and its child task-groups in a fair manner,
> +        * based on each entity's (task or task-group's) weight
> +        * (se->load.weight).
> +        *
> +        * In other words, if root_task_group has 10 tasks of weight
> +        * 1024) and two child groups A0 and A1 (of weight 1024 each),
> +        * then A0's share of the cpu resource is:
> +        *
> +        *      A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33%
> +        *
> +        * We achieve this by letting root_task_group's tasks sit
> +        * directly in rq->cfs (i.e root_task_group->se[] = NULL).
> +        */

This comment has become unglued from what it's supposed to be attached
to (it's tied to root_task_group.shares & init_tg_cfs_entry, not
init_cfs_bandwidth).

> +       init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
> +#endif
> +
>         for_each_possible_cpu(i) {
>                 struct rq *rq;
>
> @@ -6966,28 +6991,7 @@ void __init sched_init(void)
>                 init_cfs_rq(&rq->cfs);
>                 init_rt_rq(&rq->rt, rq);
>  #ifdef CONFIG_FAIR_GROUP_SCHED
> -               root_task_group.shares = ROOT_TASK_GROUP_LOAD;
>                 INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
> -               /*
> -                * How much cpu bandwidth does root_task_group get?
> -                *
> -                * In case of task-groups formed thr' the cgroup filesystem, it
> -                * gets 100% of the cpu resources in the system. This overall
> -                * system cpu resource is divided among the tasks of
> -                * root_task_group and its child task-groups in a fair manner,
> -                * based on each entity's (task or task-group's) weight
> -                * (se->load.weight).
> -                *
> -                * In other words, if root_task_group has 10 tasks of weight
> -                * 1024) and two child groups A0 and A1 (of weight 1024 each),
> -                * then A0's share of the cpu resource is:
> -                *
> -                *      A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33%
> -                *
> -                * We achieve this by letting root_task_group's tasks sit
> -                * directly in rq->cfs (i.e root_task_group->se[] = NULL).
> -                */
> -               init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
>                 init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, NULL);
>  #endif /* CONFIG_FAIR_GROUP_SCHED */
>
> --
> 1.7.4.1
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: [PATCH 1/3] sched: don't repeat the initialization in sched_init()

[Michael Wang]

Hi, Paul

On 06/04/2013 02:52 PM, Paul Turner wrote:
> On Mon, Jun 3, 2013 at 11:23 PM, Michael Wang

[snip]

> 
> This comment has become unglued from what it's supposed to be attached
> to (it's tied to root_task_group.shares & init_tg_cfs_entry, not
> init_cfs_bandwidth).

Thanks for your review and notify :)

What about put the comment with init_tg_cfs_entry()?

'root_task_group.shares' may not needed to be covered under the comment,
after all, it won't have any peers to flaunt it's share...

Regards,
Michael Wang

> 
>> +       init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
>> +#endif
>> +
>>         for_each_possible_cpu(i) {
>>                 struct rq *rq;
>>
>> @@ -6966,28 +6991,7 @@ void __init sched_init(void)
>>                 init_cfs_rq(&rq->cfs);
>>                 init_rt_rq(&rq->rt, rq);
>>  #ifdef CONFIG_FAIR_GROUP_SCHED
>> -               root_task_group.shares = ROOT_TASK_GROUP_LOAD;
>>                 INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
>> -               /*
>> -                * How much cpu bandwidth does root_task_group get?
>> -                *
>> -                * In case of task-groups formed thr' the cgroup filesystem, it
>> -                * gets 100% of the cpu resources in the system. This overall
>> -                * system cpu resource is divided among the tasks of
>> -                * root_task_group and its child task-groups in a fair manner,
>> -                * based on each entity's (task or task-group's) weight
>> -                * (se->load.weight).
>> -                *
>> -                * In other words, if root_task_group has 10 tasks of weight
>> -                * 1024) and two child groups A0 and A1 (of weight 1024 each),
>> -                * then A0's share of the cpu resource is:
>> -                *
>> -                *      A0's bandwidth = 1024 / (10*1024 + 1024 + 1024) = 8.33%
>> -                *
>> -                * We achieve this by letting root_task_group's tasks sit
>> -                * directly in rq->cfs (i.e root_task_group->se[] = NULL).
>> -                */
>> -               init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
>>                 init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, NULL);
>>  #endif /* CONFIG_FAIR_GROUP_SCHED */
>>
>> --
>> 1.7.4.1
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at  http://www.tux.org/lkml/
> 

[PATCH v2 1/3] sched: don't repeat the initialization in sched_init()

[Michael Wang]

v2:
	Move comments back before init_tg_cfs_entry(). (Thanks for the notify from pjt)

In sched_init(), there is no need to initialize 'root_task_group.shares' and
'root_task_group.cfs_bandwidth' repeatedly.

CC: Paul Tuner <pjt@google.com>
CC: Ingo Molnar <mingo@kernel.org>
CC: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 kernel/sched/core.c |    7 +++++--
 1 files changed, 5 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 58453b8..96f69da 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -6955,6 +6955,11 @@ void __init sched_init(void)
 
 #endif /* CONFIG_CGROUP_SCHED */
 
+#ifdef CONFIG_FAIR_GROUP_SCHED
+	root_task_group.shares = ROOT_TASK_GROUP_LOAD;
+	init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+#endif
+
 	for_each_possible_cpu(i) {
 		struct rq *rq;
 
@@ -6966,7 +6971,6 @@ void __init sched_init(void)
 		init_cfs_rq(&rq->cfs);
 		init_rt_rq(&rq->rt, rq);
 #ifdef CONFIG_FAIR_GROUP_SCHED
-		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
 		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
 		/*
 		 * How much cpu bandwidth does root_task_group get?
@@ -6987,7 +6991,6 @@ void __init sched_init(void)
 		 * We achieve this by letting root_task_group's tasks sit
 		 * directly in rq->cfs (i.e root_task_group->se[] = NULL).
 		 */
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
 		init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 
-- 
1.7.4.1

Re: [PATCH v2 1/3] sched: don't repeat the initialization in sched_init()

[Peter Zijlstra]

On Wed, Jun 05, 2013 at 10:24:18AM +0800, Michael Wang wrote:
> v2:
> 	Move comments back before init_tg_cfs_entry(). (Thanks for the notify from pjt)
> 
> In sched_init(), there is no need to initialize 'root_task_group.shares' and
> 'root_task_group.cfs_bandwidth' repeatedly.
> 
> CC: Paul Tuner <pjt@google.com>
> CC: Ingo Molnar <mingo@kernel.org>
> CC: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
> ---
>  kernel/sched/core.c |    7 +++++--
>  1 files changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index 58453b8..96f69da 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -6955,6 +6955,11 @@ void __init sched_init(void)
>  
>  #endif /* CONFIG_CGROUP_SCHED */
>  
> +#ifdef CONFIG_FAIR_GROUP_SCHED
> +	root_task_group.shares = ROOT_TASK_GROUP_LOAD;
> +	init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
> +#endif
> +
>  	for_each_possible_cpu(i) {
>  		struct rq *rq;
>  
> @@ -6966,7 +6971,6 @@ void __init sched_init(void)
>  		init_cfs_rq(&rq->cfs);
>  		init_rt_rq(&rq->rt, rq);
>  #ifdef CONFIG_FAIR_GROUP_SCHED
> -		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
>  		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
>  		/*
>  		 * How much cpu bandwidth does root_task_group get?
> @@ -6987,7 +6991,6 @@ void __init sched_init(void)
>  		 * We achieve this by letting root_task_group's tasks sit
>  		 * directly in rq->cfs (i.e root_task_group->se[] = NULL).
>  		 */
> -		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
>  		init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, NULL);
>  #endif /* CONFIG_FAIR_GROUP_SCHED */

I would actually like a patch reducing the #ifdef forest there, not
adding to it.

There's no actual harm in doing the initialization mutliple times,
right?

Re: [PATCH 2/3] sched: code refine in unthrottle_cfs_rq()

[Peter Zijlstra]

On Tue, Jun 04, 2013 at 02:23:39PM +0800, Michael Wang wrote:
> Directly use rq to save some code.
> 
> CC: Ingo Molnar <mingo@kernel.org>
> CC: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>

Please send patches against tip/master; the below didn't apply cleanly.
It was a trivial conflict so I applied force and made it fit.

Thanks!

> ---
>  kernel/sched/fair.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index c61a614..1e10911 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -2298,7 +2298,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
>  	int enqueue = 1;
>  	long task_delta;
>  
> -	se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))];
> +	se = cfs_rq->tg->se[cpu_of(rq)];
>  
>  	cfs_rq->throttled = 0;
>  	raw_spin_lock(&cfs_b->lock);
> -- 
> 1.7.4.1
> 

Re: [PATCH 3/3] sched: remove the useless declaration in kernel/sched/fair.c

[Peter Zijlstra]

On Tue, Jun 04, 2013 at 02:24:08PM +0800, Michael Wang wrote:
> default_cfs_period(), do_sched_cfs_period_timer(), do_sched_cfs_slack_timer()
> already defined previously, no need to declare again.
> 
> CC: Ingo Molnar <mingo@kernel.org>
> CC: Peter Zijlstra <peterz@infradead.org>
> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>

Thanks!

Re: [PATCH v2 1/3] sched: don't repeat the initialization in sched_init()

[Michael Wang]

On 06/05/2013 07:06 PM, Peter Zijlstra wrote:
> On Wed, Jun 05, 2013 at 10:24:18AM +0800, Michael Wang wrote:
>> v2:
>> 	Move comments back before init_tg_cfs_entry(). (Thanks for the notify from pjt)
>>
>> In sched_init(), there is no need to initialize 'root_task_group.shares' and
>> 'root_task_group.cfs_bandwidth' repeatedly.
>>
>> CC: Paul Tuner <pjt@google.com>
>> CC: Ingo Molnar <mingo@kernel.org>
>> CC: Peter Zijlstra <peterz@infradead.org>
>> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
>> ---
>>  kernel/sched/core.c |    7 +++++--
>>  1 files changed, 5 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index 58453b8..96f69da 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -6955,6 +6955,11 @@ void __init sched_init(void)
>>  
>>  #endif /* CONFIG_CGROUP_SCHED */
>>  
>> +#ifdef CONFIG_FAIR_GROUP_SCHED
>> +	root_task_group.shares = ROOT_TASK_GROUP_LOAD;
>> +	init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
>> +#endif
>> +
>>  	for_each_possible_cpu(i) {
>>  		struct rq *rq;
>>  
>> @@ -6966,7 +6971,6 @@ void __init sched_init(void)
>>  		init_cfs_rq(&rq->cfs);
>>  		init_rt_rq(&rq->rt, rq);
>>  #ifdef CONFIG_FAIR_GROUP_SCHED
>> -		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
>>  		INIT_LIST_HEAD(&rq->leaf_cfs_rq_list);
>>  		/*
>>  		 * How much cpu bandwidth does root_task_group get?
>> @@ -6987,7 +6991,6 @@ void __init sched_init(void)
>>  		 * We achieve this by letting root_task_group's tasks sit
>>  		 * directly in rq->cfs (i.e root_task_group->se[] = NULL).
>>  		 */
>> -		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
>>  		init_tg_cfs_entry(&root_task_group, &rq->cfs, NULL, i, NULL);
>>  #endif /* CONFIG_FAIR_GROUP_SCHED */
> 
> I would actually like a patch reducing the #ifdef forest there, not
> adding to it.

I see :)

> 
> There's no actual harm in doing the initialization mutliple times,
> right?

Yeah, it's safe to redo the init, cost some cycles but not so expensive.

Regards,
Michael Wang

> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

Re: [PATCH 2/3] sched: code refine in unthrottle_cfs_rq()

[Michael Wang]

On 06/05/2013 07:15 PM, Peter Zijlstra wrote:
> On Tue, Jun 04, 2013 at 02:23:39PM +0800, Michael Wang wrote:
>> Directly use rq to save some code.
>>
>> CC: Ingo Molnar <mingo@kernel.org>
>> CC: Peter Zijlstra <peterz@infradead.org>
>> Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
> 
> Please send patches against tip/master; the below didn't apply cleanly.
> It was a trivial conflict so I applied force and made it fit.

My sincere apologies on that, please allow me to resend the accepted
patches based on latest tip/master, forgive me for create extra work
like that...

Regards,
Michael Wang

> 
> Thanks!
> 
>> ---
>>  kernel/sched/fair.c |    2 +-
>>  1 files changed, 1 insertions(+), 1 deletions(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index c61a614..1e10911 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -2298,7 +2298,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
>>  	int enqueue = 1;
>>  	long task_delta;
>>  
>> -	se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))];
>> +	se = cfs_rq->tg->se[cpu_of(rq)];
>>  
>>  	cfs_rq->throttled = 0;
>>  	raw_spin_lock(&cfs_b->lock);
>> -- 
>> 1.7.4.1
>>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
> 

[PATCH v2 2/3] sched: code refine in unthrottle_cfs_rq()

[Michael Wang]

v2:
	re-based on latest tip/master

Directly use rq to save some code.

CC: Ingo Molnar <mingo@kernel.org>
CC: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 kernel/sched/fair.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 143dcdb..0cea941 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2275,7 +2275,7 @@ static void throttle_cfs_rq(struct cfs_rq *cfs_rq)
 	struct sched_entity *se;
 	long task_delta, dequeue = 1;
 
-	se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))];
+	se = cfs_rq->tg->se[cpu_of(rq)];
 
 	/* freeze hierarchy runnable averages while throttled */
 	rcu_read_lock();
-- 
1.7.4.1

[PATCH v2 3/3] sched: remove the useless declaration in kernel/sched/fair.c

[Michael Wang]

v2:
	re-based on latest tip/master

default_cfs_period(), do_sched_cfs_period_timer(), do_sched_cfs_slack_timer()
already defined previously, no need to declare again.

CC: Ingo Molnar <mingo@kernel.org>
CC: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
---
 kernel/sched/fair.c |    4 ----
 1 files changed, 0 insertions(+), 4 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 0cea941..9efc50f 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2618,10 +2618,6 @@ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq)
 	throttle_cfs_rq(cfs_rq);
 }
 
-static inline u64 default_cfs_period(void);
-static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun);
-static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b);
-
 static enum hrtimer_restart sched_cfs_slack_timer(struct hrtimer *timer)
 {
 	struct cfs_bandwidth *cfs_b =
-- 
1.7.4.1

[tip:sched/core] sched: Refine the code in unthrottle_cfs_rq()

[tip-bot for Michael Wang]

Commit-ID:  22b958d8cc5127d22d2ad2141277d312d93fad6c
Gitweb:     http://git.kernel.org/tip/22b958d8cc5127d22d2ad2141277d312d93fad6c
Author:     Michael Wang <wangyun@linux.vnet.ibm.com>
AuthorDate: Tue, 4 Jun 2013 14:23:39 +0800
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 19 Jun 2013 12:58:41 +0200

sched: Refine the code in unthrottle_cfs_rq()

Directly use rq to save some code.

Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/51AD87EB.1070605@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 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 143dcdb..47a30be 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2315,7 +2315,7 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
 	int enqueue = 1;
 	long task_delta;
 
-	se = cfs_rq->tg->se[cpu_of(rq_of(cfs_rq))];
+	se = cfs_rq->tg->se[cpu_of(rq)];
 
 	cfs_rq->throttled = 0;
 

[tip:sched/core] sched: Femove the useless declaration in kernel/ sched/fair.c

[tip-bot for Michael Wang]

Commit-ID:  8404c90d050733b3404dc36c500f63ccb0c972ce
Gitweb:     http://git.kernel.org/tip/8404c90d050733b3404dc36c500f63ccb0c972ce
Author:     Michael Wang <wangyun@linux.vnet.ibm.com>
AuthorDate: Tue, 4 Jun 2013 14:24:08 +0800
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Wed, 19 Jun 2013 12:58:41 +0200

sched: Femove the useless declaration in kernel/sched/fair.c

default_cfs_period(), do_sched_cfs_period_timer(), do_sched_cfs_slack_timer()
already defined previously, no need to declare again.

Signed-off-by: Michael Wang <wangyun@linux.vnet.ibm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/51AD8808.7020608@linux.vnet.ibm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/fair.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 47a30be..c0ac2c3 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -2618,10 +2618,6 @@ static void check_cfs_rq_runtime(struct cfs_rq *cfs_rq)
 	throttle_cfs_rq(cfs_rq);
 }
 
-static inline u64 default_cfs_period(void);
-static int do_sched_cfs_period_timer(struct cfs_bandwidth *cfs_b, int overrun);
-static void do_sched_cfs_slack_timer(struct cfs_bandwidth *cfs_b);
-
 static enum hrtimer_restart sched_cfs_slack_timer(struct hrtimer *timer)
 {
 	struct cfs_bandwidth *cfs_b =