* [PATCH v2 0/2] Fix RT/DL utilization during policy change
@ 2021-06-21 10:37 Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 1/2] sched/rt: Fix RT utilization tracking " Vincent Donnefort
` (2 more replies)
0 siblings, 3 replies; 7+ messages in thread
From: Vincent Donnefort @ 2021-06-21 10:37 UTC (permalink / raw)
To: mingo, peterz, juri.lelli
Cc: vincent.guittot, dietmar.eggemann, valentin.schneider, rostedt,
linux-kernel, Vincent Donnefort
changelog since V1:
* Simplify and merge "if" conditions for RT's switched_to
* Collect Reviewed-by
Vincent Donnefort (2):
sched/rt: Fix RT utilization tracking during policy change
sched/rt: Fix Deadline utilization tracking during policy change
kernel/sched/deadline.c | 2 ++
kernel/sched/rt.c | 17 ++++++++++++-----
2 files changed, 14 insertions(+), 5 deletions(-)
--
2.7.4
^ permalink raw reply [flat|nested] 7+ messages in thread
* [PATCH v2 1/2] sched/rt: Fix RT utilization tracking during policy change
2021-06-21 10:37 [PATCH v2 0/2] Fix RT/DL utilization during policy change Vincent Donnefort
@ 2021-06-21 10:37 ` Vincent Donnefort
2021-06-21 12:03 ` Vincent Guittot
2021-06-23 8:19 ` [tip: sched/core] " tip-bot2 for Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 2/2] sched/rt: Fix Deadline " Vincent Donnefort
2021-06-22 13:16 ` [PATCH v2 0/2] Fix RT/DL utilization " Peter Zijlstra
2 siblings, 2 replies; 7+ messages in thread
From: Vincent Donnefort @ 2021-06-21 10:37 UTC (permalink / raw)
To: mingo, peterz, juri.lelli
Cc: vincent.guittot, dietmar.eggemann, valentin.schneider, rostedt,
linux-kernel, Vincent Donnefort
RT keeps track of the utilization on a per-rq basis with the structure
avg_rt. This utilization is updated during task_tick_rt(),
put_prev_task_rt() and set_next_task_rt(). However, when the current
running task changes its policy, set_next_task_rt() which would usually
take care of updating the utilization when the rq starts running RT tasks,
will not see a such change, leaving the avg_rt structure outdated. When
that very same task will be dequeued later, put_prev_task_rt() will then
update the utilization, based on a wrong last_update_time, leading to a
huge spike in the RT utilization signal.
The signal would eventually recover from this issue after few ms. Even if
no RT tasks are run, avg_rt is also updated in __update_blocked_others().
But as the CPU capacity depends partly on the avg_rt, this issue has
nonetheless a significant impact on the scheduler.
Fix this issue by ensuring a load update when a running task changes
its policy to RT.
Fixes: 371bf427 ("sched/rt: Add rt_rq utilization tracking")
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index a525447..3daf42a 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -2341,13 +2341,20 @@ void __init init_sched_rt_class(void)
static void switched_to_rt(struct rq *rq, struct task_struct *p)
{
/*
- * If we are already running, then there's nothing
- * that needs to be done. But if we are not running
- * we may need to preempt the current running task.
- * If that current running task is also an RT task
+ * If we are running, update the avg_rt tracking, as the running time
+ * will now on be accounted into the latter.
+ */
+ if (task_current(rq, p)) {
+ update_rt_rq_load_avg(rq_clock_pelt(rq), rq, 0);
+ return;
+ }
+
+ /*
+ * If we are not running we may need to preempt the current
+ * running task. If that current running task is also an RT task
* then see if we can move to another run queue.
*/
- if (task_on_rq_queued(p) && rq->curr != p) {
+ if (task_on_rq_queued(p)) {
#ifdef CONFIG_SMP
if (p->nr_cpus_allowed > 1 && rq->rt.overloaded)
rt_queue_push_tasks(rq);
--
2.7.4
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [PATCH v2 2/2] sched/rt: Fix Deadline utilization tracking during policy change
2021-06-21 10:37 [PATCH v2 0/2] Fix RT/DL utilization during policy change Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 1/2] sched/rt: Fix RT utilization tracking " Vincent Donnefort
@ 2021-06-21 10:37 ` Vincent Donnefort
2021-06-23 8:19 ` [tip: sched/core] " tip-bot2 for Vincent Donnefort
2021-06-22 13:16 ` [PATCH v2 0/2] Fix RT/DL utilization " Peter Zijlstra
2 siblings, 1 reply; 7+ messages in thread
From: Vincent Donnefort @ 2021-06-21 10:37 UTC (permalink / raw)
To: mingo, peterz, juri.lelli
Cc: vincent.guittot, dietmar.eggemann, valentin.schneider, rostedt,
linux-kernel, Vincent Donnefort
DL keeps track of the utilization on a per-rq basis with the structure
avg_dl. This utilization is updated during task_tick_dl(),
put_prev_task_dl() and set_next_task_dl(). However, when the current
running task changes its policy, set_next_task_dl() which would usually
take care of updating the utilization when the rq starts running DL
tasks, will not see a such change, leaving the avg_dl structure outdated.
When that very same task will be dequeued later, put_prev_task_dl() will
then update the utilization, based on a wrong last_update_time, leading to
a huge spike in the DL utilization signal.
The signal would eventually recover from this issue after few ms. Even
if no DL tasks are run, avg_dl is also updated in
__update_blocked_others(). But as the CPU capacity depends partly on the
avg_dl, this issue has nonetheless a significant impact on the scheduler.
Fix this issue by ensuring a load update when a running task changes
its policy to DL.
Fixes: 3727e0e ("sched/dl: Add dl_rq utilization tracking")
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 3829c5a..915227a 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -2497,6 +2497,8 @@ static void switched_to_dl(struct rq *rq, struct task_struct *p)
check_preempt_curr_dl(rq, p, 0);
else
resched_curr(rq);
+ } else {
+ update_dl_rq_load_avg(rq_clock_pelt(rq), rq, 0);
}
}
--
2.7.4
^ permalink raw reply related [flat|nested] 7+ messages in thread
* Re: [PATCH v2 1/2] sched/rt: Fix RT utilization tracking during policy change
2021-06-21 10:37 ` [PATCH v2 1/2] sched/rt: Fix RT utilization tracking " Vincent Donnefort
@ 2021-06-21 12:03 ` Vincent Guittot
2021-06-23 8:19 ` [tip: sched/core] " tip-bot2 for Vincent Donnefort
1 sibling, 0 replies; 7+ messages in thread
From: Vincent Guittot @ 2021-06-21 12:03 UTC (permalink / raw)
To: Vincent Donnefort
Cc: Ingo Molnar, Peter Zijlstra, Juri Lelli, Dietmar Eggemann,
Valentin Schneider, Steven Rostedt, linux-kernel
On Mon, 21 Jun 2021 at 12:39, Vincent Donnefort
<vincent.donnefort@arm.com> wrote:
>
> RT keeps track of the utilization on a per-rq basis with the structure
> avg_rt. This utilization is updated during task_tick_rt(),
> put_prev_task_rt() and set_next_task_rt(). However, when the current
> running task changes its policy, set_next_task_rt() which would usually
> take care of updating the utilization when the rq starts running RT tasks,
> will not see a such change, leaving the avg_rt structure outdated. When
> that very same task will be dequeued later, put_prev_task_rt() will then
> update the utilization, based on a wrong last_update_time, leading to a
> huge spike in the RT utilization signal.
>
> The signal would eventually recover from this issue after few ms. Even if
> no RT tasks are run, avg_rt is also updated in __update_blocked_others().
> But as the CPU capacity depends partly on the avg_rt, this issue has
> nonetheless a significant impact on the scheduler.
>
> Fix this issue by ensuring a load update when a running task changes
> its policy to RT.
>
> Fixes: 371bf427 ("sched/rt: Add rt_rq utilization tracking")
> Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
>
> diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
> index a525447..3daf42a 100644
> --- a/kernel/sched/rt.c
> +++ b/kernel/sched/rt.c
> @@ -2341,13 +2341,20 @@ void __init init_sched_rt_class(void)
> static void switched_to_rt(struct rq *rq, struct task_struct *p)
> {
> /*
> - * If we are already running, then there's nothing
> - * that needs to be done. But if we are not running
> - * we may need to preempt the current running task.
> - * If that current running task is also an RT task
> + * If we are running, update the avg_rt tracking, as the running time
> + * will now on be accounted into the latter.
> + */
> + if (task_current(rq, p)) {
> + update_rt_rq_load_avg(rq_clock_pelt(rq), rq, 0);
> + return;
> + }
> +
> + /*
> + * If we are not running we may need to preempt the current
> + * running task. If that current running task is also an RT task
> * then see if we can move to another run queue.
> */
> - if (task_on_rq_queued(p) && rq->curr != p) {
> + if (task_on_rq_queued(p)) {
> #ifdef CONFIG_SMP
> if (p->nr_cpus_allowed > 1 && rq->rt.overloaded)
> rt_queue_push_tasks(rq);
> --
> 2.7.4
>
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v2 0/2] Fix RT/DL utilization during policy change
2021-06-21 10:37 [PATCH v2 0/2] Fix RT/DL utilization during policy change Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 1/2] sched/rt: Fix RT utilization tracking " Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 2/2] sched/rt: Fix Deadline " Vincent Donnefort
@ 2021-06-22 13:16 ` Peter Zijlstra
2 siblings, 0 replies; 7+ messages in thread
From: Peter Zijlstra @ 2021-06-22 13:16 UTC (permalink / raw)
To: Vincent Donnefort
Cc: mingo, juri.lelli, vincent.guittot, dietmar.eggemann,
valentin.schneider, rostedt, linux-kernel
On Mon, Jun 21, 2021 at 11:37:50AM +0100, Vincent Donnefort wrote:
> changelog since V1:
> * Simplify and merge "if" conditions for RT's switched_to
> * Collect Reviewed-by
>
> Vincent Donnefort (2):
> sched/rt: Fix RT utilization tracking during policy change
> sched/rt: Fix Deadline utilization tracking during policy change
>
> kernel/sched/deadline.c | 2 ++
> kernel/sched/rt.c | 17 ++++++++++++-----
> 2 files changed, 14 insertions(+), 5 deletions(-)
Thanks!
^ permalink raw reply [flat|nested] 7+ messages in thread
* [tip: sched/core] sched/rt: Fix Deadline utilization tracking during policy change
2021-06-21 10:37 ` [PATCH v2 2/2] sched/rt: Fix Deadline " Vincent Donnefort
@ 2021-06-23 8:19 ` tip-bot2 for Vincent Donnefort
0 siblings, 0 replies; 7+ messages in thread
From: tip-bot2 for Vincent Donnefort @ 2021-06-23 8:19 UTC (permalink / raw)
To: linux-tip-commits
Cc: Vincent Donnefort, Peter Zijlstra (Intel),
Vincent Guittot, x86, linux-kernel
The following commit has been merged into the sched/core branch of tip:
Commit-ID: d7d607096ae6d378b4e92d49946d22739c047d4c
Gitweb: https://git.kernel.org/tip/d7d607096ae6d378b4e92d49946d22739c047d4c
Author: Vincent Donnefort <vincent.donnefort@arm.com>
AuthorDate: Mon, 21 Jun 2021 11:37:52 +01:00
Committer: Peter Zijlstra <peterz@infradead.org>
CommitterDate: Tue, 22 Jun 2021 16:41:59 +02:00
sched/rt: Fix Deadline utilization tracking during policy change
DL keeps track of the utilization on a per-rq basis with the structure
avg_dl. This utilization is updated during task_tick_dl(),
put_prev_task_dl() and set_next_task_dl(). However, when the current
running task changes its policy, set_next_task_dl() which would usually
take care of updating the utilization when the rq starts running DL
tasks, will not see a such change, leaving the avg_dl structure outdated.
When that very same task will be dequeued later, put_prev_task_dl() will
then update the utilization, based on a wrong last_update_time, leading to
a huge spike in the DL utilization signal.
The signal would eventually recover from this issue after few ms. Even
if no DL tasks are run, avg_dl is also updated in
__update_blocked_others(). But as the CPU capacity depends partly on the
avg_dl, this issue has nonetheless a significant impact on the scheduler.
Fix this issue by ensuring a load update when a running task changes
its policy to DL.
Fixes: 3727e0e ("sched/dl: Add dl_rq utilization tracking")
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/1624271872-211872-3-git-send-email-vincent.donnefort@arm.com
---
kernel/sched/deadline.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 22878cd..aaacd6c 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -2497,6 +2497,8 @@ static void switched_to_dl(struct rq *rq, struct task_struct *p)
check_preempt_curr_dl(rq, p, 0);
else
resched_curr(rq);
+ } else {
+ update_dl_rq_load_avg(rq_clock_pelt(rq), rq, 0);
}
}
^ permalink raw reply related [flat|nested] 7+ messages in thread
* [tip: sched/core] sched/rt: Fix RT utilization tracking during policy change
2021-06-21 10:37 ` [PATCH v2 1/2] sched/rt: Fix RT utilization tracking " Vincent Donnefort
2021-06-21 12:03 ` Vincent Guittot
@ 2021-06-23 8:19 ` tip-bot2 for Vincent Donnefort
1 sibling, 0 replies; 7+ messages in thread
From: tip-bot2 for Vincent Donnefort @ 2021-06-23 8:19 UTC (permalink / raw)
To: linux-tip-commits
Cc: Vincent Donnefort, Peter Zijlstra (Intel),
Vincent Guittot, x86, linux-kernel
The following commit has been merged into the sched/core branch of tip:
Commit-ID: fecfcbc288e9f4923f40fd23ca78a6acdc7fdf6c
Gitweb: https://git.kernel.org/tip/fecfcbc288e9f4923f40fd23ca78a6acdc7fdf6c
Author: Vincent Donnefort <vincent.donnefort@arm.com>
AuthorDate: Mon, 21 Jun 2021 11:37:51 +01:00
Committer: Peter Zijlstra <peterz@infradead.org>
CommitterDate: Tue, 22 Jun 2021 16:41:59 +02:00
sched/rt: Fix RT utilization tracking during policy change
RT keeps track of the utilization on a per-rq basis with the structure
avg_rt. This utilization is updated during task_tick_rt(),
put_prev_task_rt() and set_next_task_rt(). However, when the current
running task changes its policy, set_next_task_rt() which would usually
take care of updating the utilization when the rq starts running RT tasks,
will not see a such change, leaving the avg_rt structure outdated. When
that very same task will be dequeued later, put_prev_task_rt() will then
update the utilization, based on a wrong last_update_time, leading to a
huge spike in the RT utilization signal.
The signal would eventually recover from this issue after few ms. Even if
no RT tasks are run, avg_rt is also updated in __update_blocked_others().
But as the CPU capacity depends partly on the avg_rt, this issue has
nonetheless a significant impact on the scheduler.
Fix this issue by ensuring a load update when a running task changes
its policy to RT.
Fixes: 371bf427 ("sched/rt: Add rt_rq utilization tracking")
Signed-off-by: Vincent Donnefort <vincent.donnefort@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lore.kernel.org/r/1624271872-211872-2-git-send-email-vincent.donnefort@arm.com
---
kernel/sched/rt.c | 17 ++++++++++++-----
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index a525447..3daf42a 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -2341,13 +2341,20 @@ void __init init_sched_rt_class(void)
static void switched_to_rt(struct rq *rq, struct task_struct *p)
{
/*
- * If we are already running, then there's nothing
- * that needs to be done. But if we are not running
- * we may need to preempt the current running task.
- * If that current running task is also an RT task
+ * If we are running, update the avg_rt tracking, as the running time
+ * will now on be accounted into the latter.
+ */
+ if (task_current(rq, p)) {
+ update_rt_rq_load_avg(rq_clock_pelt(rq), rq, 0);
+ return;
+ }
+
+ /*
+ * If we are not running we may need to preempt the current
+ * running task. If that current running task is also an RT task
* then see if we can move to another run queue.
*/
- if (task_on_rq_queued(p) && rq->curr != p) {
+ if (task_on_rq_queued(p)) {
#ifdef CONFIG_SMP
if (p->nr_cpus_allowed > 1 && rq->rt.overloaded)
rt_queue_push_tasks(rq);
^ permalink raw reply related [flat|nested] 7+ messages in thread
end of thread, other threads:[~2021-06-23 8:19 UTC | newest]
Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-06-21 10:37 [PATCH v2 0/2] Fix RT/DL utilization during policy change Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 1/2] sched/rt: Fix RT utilization tracking " Vincent Donnefort
2021-06-21 12:03 ` Vincent Guittot
2021-06-23 8:19 ` [tip: sched/core] " tip-bot2 for Vincent Donnefort
2021-06-21 10:37 ` [PATCH v2 2/2] sched/rt: Fix Deadline " Vincent Donnefort
2021-06-23 8:19 ` [tip: sched/core] " tip-bot2 for Vincent Donnefort
2021-06-22 13:16 ` [PATCH v2 0/2] Fix RT/DL utilization " Peter Zijlstra
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.