* [PATCH v2] sched/fair: Remove setting task's se->runnable_weight during PELT update
@ 2018-08-03 14:05 Dietmar Eggemann
2018-09-10 23:49 ` Dietmar Eggemann
2018-10-02 10:07 ` [tip:sched/core] " tip-bot for Dietmar Eggemann
0 siblings, 2 replies; 4+ messages in thread
From: Dietmar Eggemann @ 2018-08-03 14:05 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar
Cc: Vincent Guittot, Morten Rasmussen, Patrick Bellasi,
Quentin Perret, Joel Fernandes, linux-kernel
A CFS (SCHED_OTHER, SCHED_BATCH or SCHED_IDLE policy) task's
se->runnable_weight must always be in sync with its se->load.weight.
se->runnable_weight is set to se->load.weight when the task is
forked (init_entity_runnable_average()) or reniced (reweight_entity()).
There are two cases in set_load_weight() which since they currently only
set se->load.weight could lead to a situation in which se->load.weight
is different to se->runnable_weight for a CFS task:
(1) A task switches to SCHED_IDLE.
(2) A SCHED_FIFO, SCHED_RR or SCHED_DEADLINE task which has been reniced
(during which only its static priority gets set) switches to
SCHED_OTHER or SCHED_BATCH.
Set se->runnable_weight to se->load.weight in these two cases to prevent
this. This eliminates the need to explicitly set it to se->load.weight
during PELT updates in the CFS scheduler fastpath.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
---
Changes v1->v2:
- Rebased on latest tip/sched/core
This patch has been tested w/ appropriate BUG_ON()'s in
__update_load_avg_blocked_se() and __update_load_avg_se() on an Ubuntu
18.04 desktop.
kernel/sched/core.c | 2 ++
kernel/sched/pelt.c | 6 ------
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index deafa9fe602b..2a08418db3d1 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -701,6 +701,7 @@ static void set_load_weight(struct task_struct *p, bool update_load)
if (idle_policy(p->policy)) {
load->weight = scale_load(WEIGHT_IDLEPRIO);
load->inv_weight = WMULT_IDLEPRIO;
+ p->se.runnable_weight = load->weight;
return;
}
@@ -713,6 +714,7 @@ static void set_load_weight(struct task_struct *p, bool update_load)
} else {
load->weight = scale_load(sched_prio_to_weight[prio]);
load->inv_weight = sched_prio_to_wmult[prio];
+ p->se.runnable_weight = load->weight;
}
}
diff --git a/kernel/sched/pelt.c b/kernel/sched/pelt.c
index 35475c0c5419..d0016b16d23a 100644
--- a/kernel/sched/pelt.c
+++ b/kernel/sched/pelt.c
@@ -269,9 +269,6 @@ ___update_load_avg(struct sched_avg *sa, unsigned long load, unsigned long runna
int __update_load_avg_blocked_se(u64 now, int cpu, struct sched_entity *se)
{
- if (entity_is_task(se))
- se->runnable_weight = se->load.weight;
-
if (___update_load_sum(now, cpu, &se->avg, 0, 0, 0)) {
___update_load_avg(&se->avg, se_weight(se), se_runnable(se));
return 1;
@@ -282,9 +279,6 @@ int __update_load_avg_blocked_se(u64 now, int cpu, struct sched_entity *se)
int __update_load_avg_se(u64 now, int cpu, struct cfs_rq *cfs_rq, struct sched_entity *se)
{
- if (entity_is_task(se))
- se->runnable_weight = se->load.weight;
-
if (___update_load_sum(now, cpu, &se->avg, !!se->on_rq, !!se->on_rq,
cfs_rq->curr == se)) {
--
2.11.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH v2] sched/fair: Remove setting task's se->runnable_weight during PELT update
2018-08-03 14:05 [PATCH v2] sched/fair: Remove setting task's se->runnable_weight during PELT update Dietmar Eggemann
@ 2018-09-10 23:49 ` Dietmar Eggemann
2018-09-11 11:30 ` Peter Zijlstra
2018-10-02 10:07 ` [tip:sched/core] " tip-bot for Dietmar Eggemann
1 sibling, 1 reply; 4+ messages in thread
From: Dietmar Eggemann @ 2018-09-10 23:49 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar
Cc: Vincent Guittot, Morten Rasmussen, Patrick Bellasi,
Quentin Perret, Joel Fernandes, linux-kernel
Hi,
On 08/03/2018 07:05 AM, Dietmar Eggemann wrote:
> A CFS (SCHED_OTHER, SCHED_BATCH or SCHED_IDLE policy) task's
> se->runnable_weight must always be in sync with its se->load.weight.
>
> se->runnable_weight is set to se->load.weight when the task is
> forked (init_entity_runnable_average()) or reniced (reweight_entity()).
>
> There are two cases in set_load_weight() which since they currently only
> set se->load.weight could lead to a situation in which se->load.weight
> is different to se->runnable_weight for a CFS task:
>
> (1) A task switches to SCHED_IDLE.
>
> (2) A SCHED_FIFO, SCHED_RR or SCHED_DEADLINE task which has been reniced
> (during which only its static priority gets set) switches to
> SCHED_OTHER or SCHED_BATCH.
>
> Set se->runnable_weight to se->load.weight in these two cases to prevent
> this. This eliminates the need to explicitly set it to se->load.weight
> during PELT updates in the CFS scheduler fastpath.
>
> Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
> ---
>
> Changes v1->v2:
> - Rebased on latest tip/sched/core
>
> This patch has been tested w/ appropriate BUG_ON()'s in
> __update_load_avg_blocked_se() and __update_load_avg_se() on an Ubuntu
> 18.04 desktop.
>
> kernel/sched/core.c | 2 ++
> kernel/sched/pelt.c | 6 ------
> 2 files changed, 2 insertions(+), 6 deletions(-)
>
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index deafa9fe602b..2a08418db3d1 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -701,6 +701,7 @@ static void set_load_weight(struct task_struct *p, bool update_load)
> if (idle_policy(p->policy)) {
> load->weight = scale_load(WEIGHT_IDLEPRIO);
> load->inv_weight = WMULT_IDLEPRIO;
> + p->se.runnable_weight = load->weight;
> return;
> }
>
> @@ -713,6 +714,7 @@ static void set_load_weight(struct task_struct *p, bool update_load)
> } else {
> load->weight = scale_load(sched_prio_to_weight[prio]);
> load->inv_weight = sched_prio_to_wmult[prio];
> + p->se.runnable_weight = load->weight;
> }
> }
>
> diff --git a/kernel/sched/pelt.c b/kernel/sched/pelt.c
> index 35475c0c5419..d0016b16d23a 100644
> --- a/kernel/sched/pelt.c
> +++ b/kernel/sched/pelt.c
> @@ -269,9 +269,6 @@ ___update_load_avg(struct sched_avg *sa, unsigned long load, unsigned long runna
>
> int __update_load_avg_blocked_se(u64 now, int cpu, struct sched_entity *se)
> {
> - if (entity_is_task(se))
> - se->runnable_weight = se->load.weight;
> -
> if (___update_load_sum(now, cpu, &se->avg, 0, 0, 0)) {
> ___update_load_avg(&se->avg, se_weight(se), se_runnable(se));
> return 1;
> @@ -282,9 +279,6 @@ int __update_load_avg_blocked_se(u64 now, int cpu, struct sched_entity *se)
>
> int __update_load_avg_se(u64 now, int cpu, struct cfs_rq *cfs_rq, struct sched_entity *se)
> {
> - if (entity_is_task(se))
> - se->runnable_weight = se->load.weight;
> -
> if (___update_load_sum(now, cpu, &se->avg, !!se->on_rq, !!se->on_rq,
> cfs_rq->curr == se)) {
>
>
Is there anything else I should do for this patch ?
Thanks,
-- Dietmar
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH v2] sched/fair: Remove setting task's se->runnable_weight during PELT update
2018-09-10 23:49 ` Dietmar Eggemann
@ 2018-09-11 11:30 ` Peter Zijlstra
0 siblings, 0 replies; 4+ messages in thread
From: Peter Zijlstra @ 2018-09-11 11:30 UTC (permalink / raw)
To: Dietmar Eggemann
Cc: Ingo Molnar, Vincent Guittot, Morten Rasmussen, Patrick Bellasi,
Quentin Perret, Joel Fernandes, linux-kernel
On Mon, Sep 10, 2018 at 04:49:05PM -0700, Dietmar Eggemann wrote:
>
> Is there anything else I should do for this patch ?
This. Got it now, thanks!
^ permalink raw reply [flat|nested] 4+ messages in thread
* [tip:sched/core] sched/fair: Remove setting task's se->runnable_weight during PELT update
2018-08-03 14:05 [PATCH v2] sched/fair: Remove setting task's se->runnable_weight during PELT update Dietmar Eggemann
2018-09-10 23:49 ` Dietmar Eggemann
@ 2018-10-02 10:07 ` tip-bot for Dietmar Eggemann
1 sibling, 0 replies; 4+ messages in thread
From: tip-bot for Dietmar Eggemann @ 2018-10-02 10:07 UTC (permalink / raw)
To: linux-tip-commits
Cc: joelaf, patrick.bellasi, dietmar.eggemann, mingo, hpa,
quentin.perret, vincent.guittot, linux-kernel, tglx, peterz,
morten.rasmussen, torvalds
Commit-ID: 4a465e3ebbc8004ce4f7f08f6022ee8315a94edf
Gitweb: https://git.kernel.org/tip/4a465e3ebbc8004ce4f7f08f6022ee8315a94edf
Author: Dietmar Eggemann <dietmar.eggemann@arm.com>
AuthorDate: Fri, 3 Aug 2018 15:05:38 +0100
Committer: Ingo Molnar <mingo@kernel.org>
CommitDate: Tue, 2 Oct 2018 09:45:03 +0200
sched/fair: Remove setting task's se->runnable_weight during PELT update
A CFS (SCHED_OTHER, SCHED_BATCH or SCHED_IDLE policy) task's
se->runnable_weight must always be in sync with its se->load.weight.
se->runnable_weight is set to se->load.weight when the task is
forked (init_entity_runnable_average()) or reniced (reweight_entity()).
There are two cases in set_load_weight() which since they currently only
set se->load.weight could lead to a situation in which se->load.weight
is different to se->runnable_weight for a CFS task:
(1) A task switches to SCHED_IDLE.
(2) A SCHED_FIFO, SCHED_RR or SCHED_DEADLINE task which has been reniced
(during which only its static priority gets set) switches to
SCHED_OTHER or SCHED_BATCH.
Set se->runnable_weight to se->load.weight in these two cases to prevent
this. This eliminates the need to explicitly set it to se->load.weight
during PELT updates in the CFS scheduler fastpath.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Joel Fernandes <joelaf@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Morten Rasmussen <morten.rasmussen@arm.com>
Cc: Patrick Bellasi <patrick.bellasi@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Quentin Perret <quentin.perret@arm.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Link: http://lkml.kernel.org/r/20180803140538.1178-1-dietmar.eggemann@arm.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
kernel/sched/core.c | 2 ++
kernel/sched/pelt.c | 6 ------
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index f2caf1bae4a3..56b3c1781276 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -700,6 +700,7 @@ static void set_load_weight(struct task_struct *p, bool update_load)
if (idle_policy(p->policy)) {
load->weight = scale_load(WEIGHT_IDLEPRIO);
load->inv_weight = WMULT_IDLEPRIO;
+ p->se.runnable_weight = load->weight;
return;
}
@@ -712,6 +713,7 @@ static void set_load_weight(struct task_struct *p, bool update_load)
} else {
load->weight = scale_load(sched_prio_to_weight[prio]);
load->inv_weight = sched_prio_to_wmult[prio];
+ p->se.runnable_weight = load->weight;
}
}
diff --git a/kernel/sched/pelt.c b/kernel/sched/pelt.c
index 48a126486435..90fb5bc12ad4 100644
--- a/kernel/sched/pelt.c
+++ b/kernel/sched/pelt.c
@@ -269,9 +269,6 @@ ___update_load_avg(struct sched_avg *sa, unsigned long load, unsigned long runna
int __update_load_avg_blocked_se(u64 now, int cpu, struct sched_entity *se)
{
- if (entity_is_task(se))
- se->runnable_weight = se->load.weight;
-
if (___update_load_sum(now, cpu, &se->avg, 0, 0, 0)) {
___update_load_avg(&se->avg, se_weight(se), se_runnable(se));
return 1;
@@ -282,9 +279,6 @@ int __update_load_avg_blocked_se(u64 now, int cpu, struct sched_entity *se)
int __update_load_avg_se(u64 now, int cpu, struct cfs_rq *cfs_rq, struct sched_entity *se)
{
- if (entity_is_task(se))
- se->runnable_weight = se->load.weight;
-
if (___update_load_sum(now, cpu, &se->avg, !!se->on_rq, !!se->on_rq,
cfs_rq->curr == se)) {
^ permalink raw reply related [flat|nested] 4+ messages in thread
end of thread, other threads:[~2018-10-02 10:07 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2018-08-03 14:05 [PATCH v2] sched/fair: Remove setting task's se->runnable_weight during PELT update Dietmar Eggemann
2018-09-10 23:49 ` Dietmar Eggemann
2018-09-11 11:30 ` Peter Zijlstra
2018-10-02 10:07 ` [tip:sched/core] " tip-bot for Dietmar Eggemann
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).