* [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list
@ 2022-04-27 16:05 Chengming Zhou
2022-05-11 12:05 ` Chengming Zhou
2022-05-23 16:23 ` Vincent Guittot
0 siblings, 2 replies; 6+ messages in thread
From: Chengming Zhou @ 2022-04-27 16:05 UTC (permalink / raw)
To: mingo, peterz, juri.lelli, vincent.guittot, dietmar.eggemann,
rostedt, bsegall, mgorman, bristot
Cc: linux-kernel, duanxiongchun, songmuchun, zhengqi.arch, Chengming Zhou
We notice the rq leaf_cfs_rq_list has two problems when do bugfix
backports and some test profiling.
1. cfs_rqs under throttled subtree could be added to the list, and
make their fully decayed ancestors on the list, even though not needed.
2. #1 also make the leaf_cfs_rq_list management complex and error prone,
this is the list of related bugfix so far:
commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
commit fe61468b2cbc ("sched/fair: Fix enqueue_task_fair warning")
commit b34cb07dde7c ("sched/fair: Fix enqueue_task_fair() warning some more")
commit 39f23ce07b93 ("sched/fair: Fix unthrottle_cfs_rq() for leaf_cfs_rq list")
commit 0258bdfaff5b ("sched/fair: Fix unfairness caused by missing load decay")
commit a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
commit fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling")
commit 2630cde26711 ("sched/fair: Add ancestors of unthrottled undecayed cfs_rq")
commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
delete every cfs_rq under throttled subtree from rq->leaf_cfs_rq_list,
and delete the throttled_hierarchy() test in update_blocked_averages(),
which optimized update_blocked_averages().
But those later bugfix add cfs_rqs under throttled subtree back to
rq->leaf_cfs_rq_list again, with their fully decayed ancestors, for
the integrity of rq->leaf_cfs_rq_list.
This patch takes another method, skip all cfs_rqs under throttled
hierarchy when list_add_leaf_cfs_rq(), to completely make cfs_rqs
under throttled subtree off the leaf_cfs_rq_list.
So we don't need to consider throttled related things in
enqueue_entity(), unthrottle_cfs_rq() and enqueue_task_fair(),
which simplify the code a lot. Also optimize update_blocked_averages()
since cfs_rqs under throttled hierarchy and their ancestors
won't be on the leaf_cfs_rq_list.
Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
---
kernel/sched/fair.c | 72 ++++++++++-----------------------------------
1 file changed, 16 insertions(+), 56 deletions(-)
diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 1ad18b5cc1b8..083c3d32c899 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -309,6 +309,8 @@ const struct sched_class fair_sched_class;
#ifdef CONFIG_FAIR_GROUP_SCHED
+static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
+
/* Walk up scheduling entities hierarchy */
#define for_each_sched_entity(se) \
for (; se; se = se->parent)
@@ -331,7 +333,7 @@ static inline bool list_add_leaf_cfs_rq(struct cfs_rq *cfs_rq)
struct rq *rq = rq_of(cfs_rq);
int cpu = cpu_of(rq);
- if (cfs_rq->on_list)
+ if (cfs_rq->on_list || throttled_hierarchy(cfs_rq))
return rq->tmp_alone_branch == &rq->leaf_cfs_rq_list;
cfs_rq->on_list = 1;
@@ -3242,8 +3244,6 @@ static long calc_group_shares(struct cfs_rq *cfs_rq)
}
#endif /* CONFIG_SMP */
-static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
-
/*
* Recomputes the group entity based on the current state of its group
* runqueue.
@@ -4356,16 +4356,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
__enqueue_entity(cfs_rq, se);
se->on_rq = 1;
- /*
- * When bandwidth control is enabled, cfs might have been removed
- * because of a parent been throttled but cfs->nr_running > 1. Try to
- * add it unconditionally.
- */
- if (cfs_rq->nr_running == 1 || cfs_bandwidth_used())
+ if (cfs_rq->nr_running == 1) {
list_add_leaf_cfs_rq(cfs_rq);
-
- if (cfs_rq->nr_running == 1)
check_enqueue_throttle(cfs_rq);
+ }
}
static void __clear_buddies_last(struct sched_entity *se)
@@ -4980,11 +4974,18 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
/* update hierarchical throttle state */
walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq);
- /* Nothing to run but something to decay (on_list)? Complete the branch */
if (!cfs_rq->load.weight) {
- if (cfs_rq->on_list)
- goto unthrottle_throttle;
- return;
+ if (!cfs_rq->on_list)
+ return;
+ /*
+ * Nothing to run but something to decay (on_list)?
+ * Complete the branch.
+ */
+ for_each_sched_entity(se) {
+ if (list_add_leaf_cfs_rq(cfs_rq_of(se)))
+ break;
+ }
+ goto unthrottle_throttle;
}
task_delta = cfs_rq->h_nr_running;
@@ -5022,31 +5023,12 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
/* end evaluation on encountering a throttled cfs_rq */
if (cfs_rq_throttled(qcfs_rq))
goto unthrottle_throttle;
-
- /*
- * One parent has been throttled and cfs_rq removed from the
- * list. Add it back to not break the leaf list.
- */
- if (throttled_hierarchy(qcfs_rq))
- list_add_leaf_cfs_rq(qcfs_rq);
}
/* At this point se is NULL and we are at root level*/
add_nr_running(rq, task_delta);
unthrottle_throttle:
- /*
- * The cfs_rq_throttled() breaks in the above iteration can result in
- * incomplete leaf list maintenance, resulting in triggering the
- * assertion below.
- */
- for_each_sched_entity(se) {
- struct cfs_rq *qcfs_rq = cfs_rq_of(se);
-
- if (list_add_leaf_cfs_rq(qcfs_rq))
- break;
- }
-
assert_list_leaf_cfs_rq(rq);
/* Determine whether we need to wake up potentially idle CPU: */
@@ -5701,13 +5683,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
/* end evaluation on encountering a throttled cfs_rq */
if (cfs_rq_throttled(cfs_rq))
goto enqueue_throttle;
-
- /*
- * One parent has been throttled and cfs_rq removed from the
- * list. Add it back to not break the leaf list.
- */
- if (throttled_hierarchy(cfs_rq))
- list_add_leaf_cfs_rq(cfs_rq);
}
/* At this point se is NULL and we are at root level*/
@@ -5731,21 +5706,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
update_overutilized_status(rq);
enqueue_throttle:
- if (cfs_bandwidth_used()) {
- /*
- * When bandwidth control is enabled; the cfs_rq_throttled()
- * breaks in the above iteration can result in incomplete
- * leaf list maintenance, resulting in triggering the assertion
- * below.
- */
- for_each_sched_entity(se) {
- cfs_rq = cfs_rq_of(se);
-
- if (list_add_leaf_cfs_rq(cfs_rq))
- break;
- }
- }
-
assert_list_leaf_cfs_rq(rq);
hrtick_update(rq);
--
2.35.2
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list
2022-04-27 16:05 [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list Chengming Zhou
@ 2022-05-11 12:05 ` Chengming Zhou
2022-05-12 14:02 ` Vincent Guittot
2022-05-23 16:23 ` Vincent Guittot
1 sibling, 1 reply; 6+ messages in thread
From: Chengming Zhou @ 2022-05-11 12:05 UTC (permalink / raw)
To: mingo, peterz, juri.lelli, vincent.guittot, dietmar.eggemann,
rostedt, bsegall, mgorman, bristot
Cc: linux-kernel, duanxiongchun, songmuchun, zhengqi.arch
Hello, friendly ping...
On 2022/4/28 00:05, Chengming Zhou wrote:
> We notice the rq leaf_cfs_rq_list has two problems when do bugfix
> backports and some test profiling.
>
> 1. cfs_rqs under throttled subtree could be added to the list, and
> make their fully decayed ancestors on the list, even though not needed.
>
> 2. #1 also make the leaf_cfs_rq_list management complex and error prone,
> this is the list of related bugfix so far:
>
> commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
> commit fe61468b2cbc ("sched/fair: Fix enqueue_task_fair warning")
> commit b34cb07dde7c ("sched/fair: Fix enqueue_task_fair() warning some more")
> commit 39f23ce07b93 ("sched/fair: Fix unthrottle_cfs_rq() for leaf_cfs_rq list")
> commit 0258bdfaff5b ("sched/fair: Fix unfairness caused by missing load decay")
> commit a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
> commit fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling")
> commit 2630cde26711 ("sched/fair: Add ancestors of unthrottled undecayed cfs_rq")
>
> commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
> delete every cfs_rq under throttled subtree from rq->leaf_cfs_rq_list,
> and delete the throttled_hierarchy() test in update_blocked_averages(),
> which optimized update_blocked_averages().
>
> But those later bugfix add cfs_rqs under throttled subtree back to
> rq->leaf_cfs_rq_list again, with their fully decayed ancestors, for
> the integrity of rq->leaf_cfs_rq_list.
>
> This patch takes another method, skip all cfs_rqs under throttled
> hierarchy when list_add_leaf_cfs_rq(), to completely make cfs_rqs
> under throttled subtree off the leaf_cfs_rq_list.
>
> So we don't need to consider throttled related things in
> enqueue_entity(), unthrottle_cfs_rq() and enqueue_task_fair(),
> which simplify the code a lot. Also optimize update_blocked_averages()
> since cfs_rqs under throttled hierarchy and their ancestors
> won't be on the leaf_cfs_rq_list.
>
> Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
> ---
> kernel/sched/fair.c | 72 ++++++++++-----------------------------------
> 1 file changed, 16 insertions(+), 56 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 1ad18b5cc1b8..083c3d32c899 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -309,6 +309,8 @@ const struct sched_class fair_sched_class;
>
> #ifdef CONFIG_FAIR_GROUP_SCHED
>
> +static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
> +
> /* Walk up scheduling entities hierarchy */
> #define for_each_sched_entity(se) \
> for (; se; se = se->parent)
> @@ -331,7 +333,7 @@ static inline bool list_add_leaf_cfs_rq(struct cfs_rq *cfs_rq)
> struct rq *rq = rq_of(cfs_rq);
> int cpu = cpu_of(rq);
>
> - if (cfs_rq->on_list)
> + if (cfs_rq->on_list || throttled_hierarchy(cfs_rq))
> return rq->tmp_alone_branch == &rq->leaf_cfs_rq_list;
>
> cfs_rq->on_list = 1;
> @@ -3242,8 +3244,6 @@ static long calc_group_shares(struct cfs_rq *cfs_rq)
> }
> #endif /* CONFIG_SMP */
>
> -static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
> -
> /*
> * Recomputes the group entity based on the current state of its group
> * runqueue.
> @@ -4356,16 +4356,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
> __enqueue_entity(cfs_rq, se);
> se->on_rq = 1;
>
> - /*
> - * When bandwidth control is enabled, cfs might have been removed
> - * because of a parent been throttled but cfs->nr_running > 1. Try to
> - * add it unconditionally.
> - */
> - if (cfs_rq->nr_running == 1 || cfs_bandwidth_used())
> + if (cfs_rq->nr_running == 1) {
> list_add_leaf_cfs_rq(cfs_rq);
> -
> - if (cfs_rq->nr_running == 1)
> check_enqueue_throttle(cfs_rq);
> + }
> }
>
> static void __clear_buddies_last(struct sched_entity *se)
> @@ -4980,11 +4974,18 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> /* update hierarchical throttle state */
> walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq);
>
> - /* Nothing to run but something to decay (on_list)? Complete the branch */
> if (!cfs_rq->load.weight) {
> - if (cfs_rq->on_list)
> - goto unthrottle_throttle;
> - return;
> + if (!cfs_rq->on_list)
> + return;
> + /*
> + * Nothing to run but something to decay (on_list)?
> + * Complete the branch.
> + */
> + for_each_sched_entity(se) {
> + if (list_add_leaf_cfs_rq(cfs_rq_of(se)))
> + break;
> + }
> + goto unthrottle_throttle;
> }
>
> task_delta = cfs_rq->h_nr_running;
> @@ -5022,31 +5023,12 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> /* end evaluation on encountering a throttled cfs_rq */
> if (cfs_rq_throttled(qcfs_rq))
> goto unthrottle_throttle;
> -
> - /*
> - * One parent has been throttled and cfs_rq removed from the
> - * list. Add it back to not break the leaf list.
> - */
> - if (throttled_hierarchy(qcfs_rq))
> - list_add_leaf_cfs_rq(qcfs_rq);
> }
>
> /* At this point se is NULL and we are at root level*/
> add_nr_running(rq, task_delta);
>
> unthrottle_throttle:
> - /*
> - * The cfs_rq_throttled() breaks in the above iteration can result in
> - * incomplete leaf list maintenance, resulting in triggering the
> - * assertion below.
> - */
> - for_each_sched_entity(se) {
> - struct cfs_rq *qcfs_rq = cfs_rq_of(se);
> -
> - if (list_add_leaf_cfs_rq(qcfs_rq))
> - break;
> - }
> -
> assert_list_leaf_cfs_rq(rq);
>
> /* Determine whether we need to wake up potentially idle CPU: */
> @@ -5701,13 +5683,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
> /* end evaluation on encountering a throttled cfs_rq */
> if (cfs_rq_throttled(cfs_rq))
> goto enqueue_throttle;
> -
> - /*
> - * One parent has been throttled and cfs_rq removed from the
> - * list. Add it back to not break the leaf list.
> - */
> - if (throttled_hierarchy(cfs_rq))
> - list_add_leaf_cfs_rq(cfs_rq);
> }
>
> /* At this point se is NULL and we are at root level*/
> @@ -5731,21 +5706,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
> update_overutilized_status(rq);
>
> enqueue_throttle:
> - if (cfs_bandwidth_used()) {
> - /*
> - * When bandwidth control is enabled; the cfs_rq_throttled()
> - * breaks in the above iteration can result in incomplete
> - * leaf list maintenance, resulting in triggering the assertion
> - * below.
> - */
> - for_each_sched_entity(se) {
> - cfs_rq = cfs_rq_of(se);
> -
> - if (list_add_leaf_cfs_rq(cfs_rq))
> - break;
> - }
> - }
> -
> assert_list_leaf_cfs_rq(rq);
>
> hrtick_update(rq);
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list
2022-05-11 12:05 ` Chengming Zhou
@ 2022-05-12 14:02 ` Vincent Guittot
2022-05-13 4:32 ` [External] " Chengming Zhou
0 siblings, 1 reply; 6+ messages in thread
From: Vincent Guittot @ 2022-05-12 14:02 UTC (permalink / raw)
To: Chengming Zhou
Cc: mingo, peterz, juri.lelli, dietmar.eggemann, rostedt, bsegall,
mgorman, bristot, linux-kernel, duanxiongchun, songmuchun,
zhengqi.arch
Hi Chengming,
Your patch is on my list but I have been busy on other stuff and we
had enough warnings and problems with this part that I want to
carefully review that all the cases are covered. I will review it soon
Vincent
On Wed, 11 May 2022 at 14:05, Chengming Zhou
<zhouchengming@bytedance.com> wrote:
>
> Hello, friendly ping...
>
>
> On 2022/4/28 00:05, Chengming Zhou wrote:
> > We notice the rq leaf_cfs_rq_list has two problems when do bugfix
> > backports and some test profiling.
> >
> > 1. cfs_rqs under throttled subtree could be added to the list, and
> > make their fully decayed ancestors on the list, even though not needed.
> >
> > 2. #1 also make the leaf_cfs_rq_list management complex and error prone,
> > this is the list of related bugfix so far:
> >
> > commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
> > commit fe61468b2cbc ("sched/fair: Fix enqueue_task_fair warning")
> > commit b34cb07dde7c ("sched/fair: Fix enqueue_task_fair() warning some more")
> > commit 39f23ce07b93 ("sched/fair: Fix unthrottle_cfs_rq() for leaf_cfs_rq list")
> > commit 0258bdfaff5b ("sched/fair: Fix unfairness caused by missing load decay")
> > commit a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
> > commit fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling")
> > commit 2630cde26711 ("sched/fair: Add ancestors of unthrottled undecayed cfs_rq")
> >
> > commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
> > delete every cfs_rq under throttled subtree from rq->leaf_cfs_rq_list,
> > and delete the throttled_hierarchy() test in update_blocked_averages(),
> > which optimized update_blocked_averages().
> >
> > But those later bugfix add cfs_rqs under throttled subtree back to
> > rq->leaf_cfs_rq_list again, with their fully decayed ancestors, for
> > the integrity of rq->leaf_cfs_rq_list.
> >
> > This patch takes another method, skip all cfs_rqs under throttled
> > hierarchy when list_add_leaf_cfs_rq(), to completely make cfs_rqs
> > under throttled subtree off the leaf_cfs_rq_list.
> >
> > So we don't need to consider throttled related things in
> > enqueue_entity(), unthrottle_cfs_rq() and enqueue_task_fair(),
> > which simplify the code a lot. Also optimize update_blocked_averages()
> > since cfs_rqs under throttled hierarchy and their ancestors
> > won't be on the leaf_cfs_rq_list.
> >
> > Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
> > ---
> > kernel/sched/fair.c | 72 ++++++++++-----------------------------------
> > 1 file changed, 16 insertions(+), 56 deletions(-)
> >
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 1ad18b5cc1b8..083c3d32c899 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -309,6 +309,8 @@ const struct sched_class fair_sched_class;
> >
> > #ifdef CONFIG_FAIR_GROUP_SCHED
> >
> > +static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
> > +
> > /* Walk up scheduling entities hierarchy */
> > #define for_each_sched_entity(se) \
> > for (; se; se = se->parent)
> > @@ -331,7 +333,7 @@ static inline bool list_add_leaf_cfs_rq(struct cfs_rq *cfs_rq)
> > struct rq *rq = rq_of(cfs_rq);
> > int cpu = cpu_of(rq);
> >
> > - if (cfs_rq->on_list)
> > + if (cfs_rq->on_list || throttled_hierarchy(cfs_rq))
> > return rq->tmp_alone_branch == &rq->leaf_cfs_rq_list;
> >
> > cfs_rq->on_list = 1;
> > @@ -3242,8 +3244,6 @@ static long calc_group_shares(struct cfs_rq *cfs_rq)
> > }
> > #endif /* CONFIG_SMP */
> >
> > -static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
> > -
> > /*
> > * Recomputes the group entity based on the current state of its group
> > * runqueue.
> > @@ -4356,16 +4356,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
> > __enqueue_entity(cfs_rq, se);
> > se->on_rq = 1;
> >
> > - /*
> > - * When bandwidth control is enabled, cfs might have been removed
> > - * because of a parent been throttled but cfs->nr_running > 1. Try to
> > - * add it unconditionally.
> > - */
> > - if (cfs_rq->nr_running == 1 || cfs_bandwidth_used())
> > + if (cfs_rq->nr_running == 1) {
> > list_add_leaf_cfs_rq(cfs_rq);
> > -
> > - if (cfs_rq->nr_running == 1)
> > check_enqueue_throttle(cfs_rq);
> > + }
> > }
> >
> > static void __clear_buddies_last(struct sched_entity *se)
> > @@ -4980,11 +4974,18 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> > /* update hierarchical throttle state */
> > walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq);
> >
> > - /* Nothing to run but something to decay (on_list)? Complete the branch */
> > if (!cfs_rq->load.weight) {
> > - if (cfs_rq->on_list)
> > - goto unthrottle_throttle;
> > - return;
> > + if (!cfs_rq->on_list)
> > + return;
> > + /*
> > + * Nothing to run but something to decay (on_list)?
> > + * Complete the branch.
> > + */
> > + for_each_sched_entity(se) {
> > + if (list_add_leaf_cfs_rq(cfs_rq_of(se)))
> > + break;
> > + }
> > + goto unthrottle_throttle;
> > }
> >
> > task_delta = cfs_rq->h_nr_running;
> > @@ -5022,31 +5023,12 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> > /* end evaluation on encountering a throttled cfs_rq */
> > if (cfs_rq_throttled(qcfs_rq))
> > goto unthrottle_throttle;
> > -
> > - /*
> > - * One parent has been throttled and cfs_rq removed from the
> > - * list. Add it back to not break the leaf list.
> > - */
> > - if (throttled_hierarchy(qcfs_rq))
> > - list_add_leaf_cfs_rq(qcfs_rq);
> > }
> >
> > /* At this point se is NULL and we are at root level*/
> > add_nr_running(rq, task_delta);
> >
> > unthrottle_throttle:
> > - /*
> > - * The cfs_rq_throttled() breaks in the above iteration can result in
> > - * incomplete leaf list maintenance, resulting in triggering the
> > - * assertion below.
> > - */
> > - for_each_sched_entity(se) {
> > - struct cfs_rq *qcfs_rq = cfs_rq_of(se);
> > -
> > - if (list_add_leaf_cfs_rq(qcfs_rq))
> > - break;
> > - }
> > -
> > assert_list_leaf_cfs_rq(rq);
> >
> > /* Determine whether we need to wake up potentially idle CPU: */
> > @@ -5701,13 +5683,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
> > /* end evaluation on encountering a throttled cfs_rq */
> > if (cfs_rq_throttled(cfs_rq))
> > goto enqueue_throttle;
> > -
> > - /*
> > - * One parent has been throttled and cfs_rq removed from the
> > - * list. Add it back to not break the leaf list.
> > - */
> > - if (throttled_hierarchy(cfs_rq))
> > - list_add_leaf_cfs_rq(cfs_rq);
> > }
> >
> > /* At this point se is NULL and we are at root level*/
> > @@ -5731,21 +5706,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
> > update_overutilized_status(rq);
> >
> > enqueue_throttle:
> > - if (cfs_bandwidth_used()) {
> > - /*
> > - * When bandwidth control is enabled; the cfs_rq_throttled()
> > - * breaks in the above iteration can result in incomplete
> > - * leaf list maintenance, resulting in triggering the assertion
> > - * below.
> > - */
> > - for_each_sched_entity(se) {
> > - cfs_rq = cfs_rq_of(se);
> > -
> > - if (list_add_leaf_cfs_rq(cfs_rq))
> > - break;
> > - }
> > - }
> > -
> > assert_list_leaf_cfs_rq(rq);
> >
> > hrtick_update(rq);
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [External] Re: [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list
2022-05-12 14:02 ` Vincent Guittot
@ 2022-05-13 4:32 ` Chengming Zhou
0 siblings, 0 replies; 6+ messages in thread
From: Chengming Zhou @ 2022-05-13 4:32 UTC (permalink / raw)
To: Vincent Guittot
Cc: mingo, peterz, juri.lelli, dietmar.eggemann, rostedt, bsegall,
mgorman, bristot, linux-kernel, duanxiongchun, songmuchun,
zhengqi.arch
On 2022/5/12 22:02, Vincent Guittot wrote:
> Hi Chengming,
>
> Your patch is on my list but I have been busy on other stuff and we
> had enough warnings and problems with this part that I want to
> carefully review that all the cases are covered. I will review it soon
>
Hi Vincent,
Thanks for your reply. There's no rush, just take your time. I agree
that we should be very careful with this part due to previous problems.
Thanks.
> Vincent
>
> On Wed, 11 May 2022 at 14:05, Chengming Zhou
> <zhouchengming@bytedance.com> wrote:
>>
>> Hello, friendly ping...
>>
>>
>> On 2022/4/28 00:05, Chengming Zhou wrote:
>>> We notice the rq leaf_cfs_rq_list has two problems when do bugfix
>>> backports and some test profiling.
>>>
>>> 1. cfs_rqs under throttled subtree could be added to the list, and
>>> make their fully decayed ancestors on the list, even though not needed.
>>>
>>> 2. #1 also make the leaf_cfs_rq_list management complex and error prone,
>>> this is the list of related bugfix so far:
>>>
>>> commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
>>> commit fe61468b2cbc ("sched/fair: Fix enqueue_task_fair warning")
>>> commit b34cb07dde7c ("sched/fair: Fix enqueue_task_fair() warning some more")
>>> commit 39f23ce07b93 ("sched/fair: Fix unthrottle_cfs_rq() for leaf_cfs_rq list")
>>> commit 0258bdfaff5b ("sched/fair: Fix unfairness caused by missing load decay")
>>> commit a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
>>> commit fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling")
>>> commit 2630cde26711 ("sched/fair: Add ancestors of unthrottled undecayed cfs_rq")
>>>
>>> commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
>>> delete every cfs_rq under throttled subtree from rq->leaf_cfs_rq_list,
>>> and delete the throttled_hierarchy() test in update_blocked_averages(),
>>> which optimized update_blocked_averages().
>>>
>>> But those later bugfix add cfs_rqs under throttled subtree back to
>>> rq->leaf_cfs_rq_list again, with their fully decayed ancestors, for
>>> the integrity of rq->leaf_cfs_rq_list.
>>>
>>> This patch takes another method, skip all cfs_rqs under throttled
>>> hierarchy when list_add_leaf_cfs_rq(), to completely make cfs_rqs
>>> under throttled subtree off the leaf_cfs_rq_list.
>>>
>>> So we don't need to consider throttled related things in
>>> enqueue_entity(), unthrottle_cfs_rq() and enqueue_task_fair(),
>>> which simplify the code a lot. Also optimize update_blocked_averages()
>>> since cfs_rqs under throttled hierarchy and their ancestors
>>> won't be on the leaf_cfs_rq_list.
>>>
>>> Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
>>> ---
>>> kernel/sched/fair.c | 72 ++++++++++-----------------------------------
>>> 1 file changed, 16 insertions(+), 56 deletions(-)
>>>
>>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>>> index 1ad18b5cc1b8..083c3d32c899 100644
>>> --- a/kernel/sched/fair.c
>>> +++ b/kernel/sched/fair.c
>>> @@ -309,6 +309,8 @@ const struct sched_class fair_sched_class;
>>>
>>> #ifdef CONFIG_FAIR_GROUP_SCHED
>>>
>>> +static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
>>> +
>>> /* Walk up scheduling entities hierarchy */
>>> #define for_each_sched_entity(se) \
>>> for (; se; se = se->parent)
>>> @@ -331,7 +333,7 @@ static inline bool list_add_leaf_cfs_rq(struct cfs_rq *cfs_rq)
>>> struct rq *rq = rq_of(cfs_rq);
>>> int cpu = cpu_of(rq);
>>>
>>> - if (cfs_rq->on_list)
>>> + if (cfs_rq->on_list || throttled_hierarchy(cfs_rq))
>>> return rq->tmp_alone_branch == &rq->leaf_cfs_rq_list;
>>>
>>> cfs_rq->on_list = 1;
>>> @@ -3242,8 +3244,6 @@ static long calc_group_shares(struct cfs_rq *cfs_rq)
>>> }
>>> #endif /* CONFIG_SMP */
>>>
>>> -static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
>>> -
>>> /*
>>> * Recomputes the group entity based on the current state of its group
>>> * runqueue.
>>> @@ -4356,16 +4356,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
>>> __enqueue_entity(cfs_rq, se);
>>> se->on_rq = 1;
>>>
>>> - /*
>>> - * When bandwidth control is enabled, cfs might have been removed
>>> - * because of a parent been throttled but cfs->nr_running > 1. Try to
>>> - * add it unconditionally.
>>> - */
>>> - if (cfs_rq->nr_running == 1 || cfs_bandwidth_used())
>>> + if (cfs_rq->nr_running == 1) {
>>> list_add_leaf_cfs_rq(cfs_rq);
>>> -
>>> - if (cfs_rq->nr_running == 1)
>>> check_enqueue_throttle(cfs_rq);
>>> + }
>>> }
>>>
>>> static void __clear_buddies_last(struct sched_entity *se)
>>> @@ -4980,11 +4974,18 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
>>> /* update hierarchical throttle state */
>>> walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq);
>>>
>>> - /* Nothing to run but something to decay (on_list)? Complete the branch */
>>> if (!cfs_rq->load.weight) {
>>> - if (cfs_rq->on_list)
>>> - goto unthrottle_throttle;
>>> - return;
>>> + if (!cfs_rq->on_list)
>>> + return;
>>> + /*
>>> + * Nothing to run but something to decay (on_list)?
>>> + * Complete the branch.
>>> + */
>>> + for_each_sched_entity(se) {
>>> + if (list_add_leaf_cfs_rq(cfs_rq_of(se)))
>>> + break;
>>> + }
>>> + goto unthrottle_throttle;
>>> }
>>>
>>> task_delta = cfs_rq->h_nr_running;
>>> @@ -5022,31 +5023,12 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
>>> /* end evaluation on encountering a throttled cfs_rq */
>>> if (cfs_rq_throttled(qcfs_rq))
>>> goto unthrottle_throttle;
>>> -
>>> - /*
>>> - * One parent has been throttled and cfs_rq removed from the
>>> - * list. Add it back to not break the leaf list.
>>> - */
>>> - if (throttled_hierarchy(qcfs_rq))
>>> - list_add_leaf_cfs_rq(qcfs_rq);
>>> }
>>>
>>> /* At this point se is NULL and we are at root level*/
>>> add_nr_running(rq, task_delta);
>>>
>>> unthrottle_throttle:
>>> - /*
>>> - * The cfs_rq_throttled() breaks in the above iteration can result in
>>> - * incomplete leaf list maintenance, resulting in triggering the
>>> - * assertion below.
>>> - */
>>> - for_each_sched_entity(se) {
>>> - struct cfs_rq *qcfs_rq = cfs_rq_of(se);
>>> -
>>> - if (list_add_leaf_cfs_rq(qcfs_rq))
>>> - break;
>>> - }
>>> -
>>> assert_list_leaf_cfs_rq(rq);
>>>
>>> /* Determine whether we need to wake up potentially idle CPU: */
>>> @@ -5701,13 +5683,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
>>> /* end evaluation on encountering a throttled cfs_rq */
>>> if (cfs_rq_throttled(cfs_rq))
>>> goto enqueue_throttle;
>>> -
>>> - /*
>>> - * One parent has been throttled and cfs_rq removed from the
>>> - * list. Add it back to not break the leaf list.
>>> - */
>>> - if (throttled_hierarchy(cfs_rq))
>>> - list_add_leaf_cfs_rq(cfs_rq);
>>> }
>>>
>>> /* At this point se is NULL and we are at root level*/
>>> @@ -5731,21 +5706,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
>>> update_overutilized_status(rq);
>>>
>>> enqueue_throttle:
>>> - if (cfs_bandwidth_used()) {
>>> - /*
>>> - * When bandwidth control is enabled; the cfs_rq_throttled()
>>> - * breaks in the above iteration can result in incomplete
>>> - * leaf list maintenance, resulting in triggering the assertion
>>> - * below.
>>> - */
>>> - for_each_sched_entity(se) {
>>> - cfs_rq = cfs_rq_of(se);
>>> -
>>> - if (list_add_leaf_cfs_rq(cfs_rq))
>>> - break;
>>> - }
>>> - }
>>> -
>>> assert_list_leaf_cfs_rq(rq);
>>>
>>> hrtick_update(rq);
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list
2022-04-27 16:05 [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list Chengming Zhou
2022-05-11 12:05 ` Chengming Zhou
@ 2022-05-23 16:23 ` Vincent Guittot
2022-05-24 9:35 ` [External] " Chengming Zhou
1 sibling, 1 reply; 6+ messages in thread
From: Vincent Guittot @ 2022-05-23 16:23 UTC (permalink / raw)
To: Chengming Zhou
Cc: mingo, peterz, juri.lelli, dietmar.eggemann, rostedt, bsegall,
mgorman, bristot, linux-kernel, duanxiongchun, songmuchun,
zhengqi.arch
On Wed, 27 Apr 2022 at 18:07, Chengming Zhou
<zhouchengming@bytedance.com> wrote:
>
> We notice the rq leaf_cfs_rq_list has two problems when do bugfix
> backports and some test profiling.
>
> 1. cfs_rqs under throttled subtree could be added to the list, and
> make their fully decayed ancestors on the list, even though not needed.
>
> 2. #1 also make the leaf_cfs_rq_list management complex and error prone,
> this is the list of related bugfix so far:
>
> commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
> commit fe61468b2cbc ("sched/fair: Fix enqueue_task_fair warning")
> commit b34cb07dde7c ("sched/fair: Fix enqueue_task_fair() warning some more")
> commit 39f23ce07b93 ("sched/fair: Fix unthrottle_cfs_rq() for leaf_cfs_rq list")
> commit 0258bdfaff5b ("sched/fair: Fix unfairness caused by missing load decay")
> commit a7b359fc6a37 ("sched/fair: Correctly insert cfs_rq's to list on unthrottle")
> commit fdaba61ef8a2 ("sched/fair: Ensure that the CFS parent is added after unthrottling")
> commit 2630cde26711 ("sched/fair: Add ancestors of unthrottled undecayed cfs_rq")
>
> commit 31bc6aeaab1d ("sched/fair: Optimize update_blocked_averages()")
> delete every cfs_rq under throttled subtree from rq->leaf_cfs_rq_list,
> and delete the throttled_hierarchy() test in update_blocked_averages(),
> which optimized update_blocked_averages().
>
> But those later bugfix add cfs_rqs under throttled subtree back to
> rq->leaf_cfs_rq_list again, with their fully decayed ancestors, for
> the integrity of rq->leaf_cfs_rq_list.
>
> This patch takes another method, skip all cfs_rqs under throttled
> hierarchy when list_add_leaf_cfs_rq(), to completely make cfs_rqs
> under throttled subtree off the leaf_cfs_rq_list.
>
> So we don't need to consider throttled related things in
> enqueue_entity(), unthrottle_cfs_rq() and enqueue_task_fair(),
> which simplify the code a lot. Also optimize update_blocked_averages()
> since cfs_rqs under throttled hierarchy and their ancestors
> won't be on the leaf_cfs_rq_list.
>
> Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
> ---
> kernel/sched/fair.c | 72 ++++++++++-----------------------------------
> 1 file changed, 16 insertions(+), 56 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 1ad18b5cc1b8..083c3d32c899 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -309,6 +309,8 @@ const struct sched_class fair_sched_class;
>
> #ifdef CONFIG_FAIR_GROUP_SCHED
>
> +static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
> +
> /* Walk up scheduling entities hierarchy */
> #define for_each_sched_entity(se) \
> for (; se; se = se->parent)
> @@ -331,7 +333,7 @@ static inline bool list_add_leaf_cfs_rq(struct cfs_rq *cfs_rq)
> struct rq *rq = rq_of(cfs_rq);
> int cpu = cpu_of(rq);
>
> - if (cfs_rq->on_list)
> + if (cfs_rq->on_list || throttled_hierarchy(cfs_rq))
Please move throttled_hierarchy() outside list_add_leaf_cfs_rq()
because the task will not be added in this case which is quite
misleading
I will continue to check the various corner cases but I haven't seen
problem so far with your method
> return rq->tmp_alone_branch == &rq->leaf_cfs_rq_list;
>
> cfs_rq->on_list = 1;
> @@ -3242,8 +3244,6 @@ static long calc_group_shares(struct cfs_rq *cfs_rq)
> }
> #endif /* CONFIG_SMP */
>
> -static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
> -
> /*
> * Recomputes the group entity based on the current state of its group
> * runqueue.
> @@ -4356,16 +4356,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
> __enqueue_entity(cfs_rq, se);
> se->on_rq = 1;
>
> - /*
> - * When bandwidth control is enabled, cfs might have been removed
> - * because of a parent been throttled but cfs->nr_running > 1. Try to
> - * add it unconditionally.
> - */
> - if (cfs_rq->nr_running == 1 || cfs_bandwidth_used())
> + if (cfs_rq->nr_running == 1) {
> list_add_leaf_cfs_rq(cfs_rq);
> -
> - if (cfs_rq->nr_running == 1)
> check_enqueue_throttle(cfs_rq);
> + }
> }
>
> static void __clear_buddies_last(struct sched_entity *se)
> @@ -4980,11 +4974,18 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> /* update hierarchical throttle state */
> walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq);
>
> - /* Nothing to run but something to decay (on_list)? Complete the branch */
> if (!cfs_rq->load.weight) {
> - if (cfs_rq->on_list)
> - goto unthrottle_throttle;
> - return;
> + if (!cfs_rq->on_list)
> + return;
> + /*
> + * Nothing to run but something to decay (on_list)?
> + * Complete the branch.
> + */
> + for_each_sched_entity(se) {
> + if (list_add_leaf_cfs_rq(cfs_rq_of(se)))
> + break;
> + }
> + goto unthrottle_throttle;
> }
>
> task_delta = cfs_rq->h_nr_running;
> @@ -5022,31 +5023,12 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
> /* end evaluation on encountering a throttled cfs_rq */
> if (cfs_rq_throttled(qcfs_rq))
> goto unthrottle_throttle;
> -
> - /*
> - * One parent has been throttled and cfs_rq removed from the
> - * list. Add it back to not break the leaf list.
> - */
> - if (throttled_hierarchy(qcfs_rq))
> - list_add_leaf_cfs_rq(qcfs_rq);
> }
>
> /* At this point se is NULL and we are at root level*/
> add_nr_running(rq, task_delta);
>
> unthrottle_throttle:
> - /*
> - * The cfs_rq_throttled() breaks in the above iteration can result in
> - * incomplete leaf list maintenance, resulting in triggering the
> - * assertion below.
> - */
> - for_each_sched_entity(se) {
> - struct cfs_rq *qcfs_rq = cfs_rq_of(se);
> -
> - if (list_add_leaf_cfs_rq(qcfs_rq))
> - break;
> - }
> -
> assert_list_leaf_cfs_rq(rq);
>
> /* Determine whether we need to wake up potentially idle CPU: */
> @@ -5701,13 +5683,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
> /* end evaluation on encountering a throttled cfs_rq */
> if (cfs_rq_throttled(cfs_rq))
> goto enqueue_throttle;
> -
> - /*
> - * One parent has been throttled and cfs_rq removed from the
> - * list. Add it back to not break the leaf list.
> - */
> - if (throttled_hierarchy(cfs_rq))
> - list_add_leaf_cfs_rq(cfs_rq);
> }
>
> /* At this point se is NULL and we are at root level*/
> @@ -5731,21 +5706,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
> update_overutilized_status(rq);
>
> enqueue_throttle:
> - if (cfs_bandwidth_used()) {
> - /*
> - * When bandwidth control is enabled; the cfs_rq_throttled()
> - * breaks in the above iteration can result in incomplete
> - * leaf list maintenance, resulting in triggering the assertion
> - * below.
> - */
> - for_each_sched_entity(se) {
> - cfs_rq = cfs_rq_of(se);
> -
> - if (list_add_leaf_cfs_rq(cfs_rq))
> - break;
> - }
> - }
> -
> assert_list_leaf_cfs_rq(rq);
>
> hrtick_update(rq);
> --
> 2.35.2
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [External] Re: [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list
2022-05-23 16:23 ` Vincent Guittot
@ 2022-05-24 9:35 ` Chengming Zhou
0 siblings, 0 replies; 6+ messages in thread
From: Chengming Zhou @ 2022-05-24 9:35 UTC (permalink / raw)
To: Vincent Guittot
Cc: mingo, peterz, juri.lelli, dietmar.eggemann, rostedt, bsegall,
mgorman, bristot, linux-kernel, duanxiongchun, songmuchun,
zhengqi.arch
On 2022/5/24 00:23, Vincent Guittot wrote:
> On Wed, 27 Apr 2022 at 18:07, Chengming Zhou
> <zhouchengming@bytedance.com> wrote:
[...]
>> kernel/sched/fair.c | 72 ++++++++++-----------------------------------
>> 1 file changed, 16 insertions(+), 56 deletions(-)
>>
>> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> index 1ad18b5cc1b8..083c3d32c899 100644
>> --- a/kernel/sched/fair.c
>> +++ b/kernel/sched/fair.c
>> @@ -309,6 +309,8 @@ const struct sched_class fair_sched_class;
>>
>> #ifdef CONFIG_FAIR_GROUP_SCHED
>>
>> +static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
>> +
>> /* Walk up scheduling entities hierarchy */
>> #define for_each_sched_entity(se) \
>> for (; se; se = se->parent)
>> @@ -331,7 +333,7 @@ static inline bool list_add_leaf_cfs_rq(struct cfs_rq *cfs_rq)
>> struct rq *rq = rq_of(cfs_rq);
>> int cpu = cpu_of(rq);
>>
>> - if (cfs_rq->on_list)
>> + if (cfs_rq->on_list || throttled_hierarchy(cfs_rq))
>
> Please move throttled_hierarchy() outside list_add_leaf_cfs_rq()
> because the task will not be added in this case which is quite
> misleading
Ok, will do. I will move throttled_hierarchy() outside.
>
> I will continue to check the various corner cases but I haven't seen
> problem so far with your method
Thanks!
>
>> return rq->tmp_alone_branch == &rq->leaf_cfs_rq_list;
>>
>> cfs_rq->on_list = 1;
>> @@ -3242,8 +3244,6 @@ static long calc_group_shares(struct cfs_rq *cfs_rq)
>> }
>> #endif /* CONFIG_SMP */
>>
>> -static inline int throttled_hierarchy(struct cfs_rq *cfs_rq);
>> -
>> /*
>> * Recomputes the group entity based on the current state of its group
>> * runqueue.
>> @@ -4356,16 +4356,10 @@ enqueue_entity(struct cfs_rq *cfs_rq, struct sched_entity *se, int flags)
>> __enqueue_entity(cfs_rq, se);
>> se->on_rq = 1;
>>
>> - /*
>> - * When bandwidth control is enabled, cfs might have been removed
>> - * because of a parent been throttled but cfs->nr_running > 1. Try to
>> - * add it unconditionally.
>> - */
>> - if (cfs_rq->nr_running == 1 || cfs_bandwidth_used())
>> + if (cfs_rq->nr_running == 1) {
>> list_add_leaf_cfs_rq(cfs_rq);
>> -
>> - if (cfs_rq->nr_running == 1)
>> check_enqueue_throttle(cfs_rq);
>> + }
>> }
>>
>> static void __clear_buddies_last(struct sched_entity *se)
>> @@ -4980,11 +4974,18 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
>> /* update hierarchical throttle state */
>> walk_tg_tree_from(cfs_rq->tg, tg_nop, tg_unthrottle_up, (void *)rq);
>>
>> - /* Nothing to run but something to decay (on_list)? Complete the branch */
>> if (!cfs_rq->load.weight) {
>> - if (cfs_rq->on_list)
>> - goto unthrottle_throttle;
>> - return;
>> + if (!cfs_rq->on_list)
>> + return;
>> + /*
>> + * Nothing to run but something to decay (on_list)?
>> + * Complete the branch.
>> + */
>> + for_each_sched_entity(se) {
>> + if (list_add_leaf_cfs_rq(cfs_rq_of(se)))
>> + break;
>> + }
>> + goto unthrottle_throttle;
>> }
>>
>> task_delta = cfs_rq->h_nr_running;
>> @@ -5022,31 +5023,12 @@ void unthrottle_cfs_rq(struct cfs_rq *cfs_rq)
>> /* end evaluation on encountering a throttled cfs_rq */
>> if (cfs_rq_throttled(qcfs_rq))
>> goto unthrottle_throttle;
>> -
>> - /*
>> - * One parent has been throttled and cfs_rq removed from the
>> - * list. Add it back to not break the leaf list.
>> - */
>> - if (throttled_hierarchy(qcfs_rq))
>> - list_add_leaf_cfs_rq(qcfs_rq);
>> }
>>
>> /* At this point se is NULL and we are at root level*/
>> add_nr_running(rq, task_delta);
>>
>> unthrottle_throttle:
>> - /*
>> - * The cfs_rq_throttled() breaks in the above iteration can result in
>> - * incomplete leaf list maintenance, resulting in triggering the
>> - * assertion below.
>> - */
>> - for_each_sched_entity(se) {
>> - struct cfs_rq *qcfs_rq = cfs_rq_of(se);
>> -
>> - if (list_add_leaf_cfs_rq(qcfs_rq))
>> - break;
>> - }
>> -
>> assert_list_leaf_cfs_rq(rq);
>>
>> /* Determine whether we need to wake up potentially idle CPU: */
>> @@ -5701,13 +5683,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
>> /* end evaluation on encountering a throttled cfs_rq */
>> if (cfs_rq_throttled(cfs_rq))
>> goto enqueue_throttle;
>> -
>> - /*
>> - * One parent has been throttled and cfs_rq removed from the
>> - * list. Add it back to not break the leaf list.
>> - */
>> - if (throttled_hierarchy(cfs_rq))
>> - list_add_leaf_cfs_rq(cfs_rq);
>> }
>>
>> /* At this point se is NULL and we are at root level*/
>> @@ -5731,21 +5706,6 @@ enqueue_task_fair(struct rq *rq, struct task_struct *p, int flags)
>> update_overutilized_status(rq);
>>
>> enqueue_throttle:
>> - if (cfs_bandwidth_used()) {
>> - /*
>> - * When bandwidth control is enabled; the cfs_rq_throttled()
>> - * breaks in the above iteration can result in incomplete
>> - * leaf list maintenance, resulting in triggering the assertion
>> - * below.
>> - */
>> - for_each_sched_entity(se) {
>> - cfs_rq = cfs_rq_of(se);
>> -
>> - if (list_add_leaf_cfs_rq(cfs_rq))
>> - break;
>> - }
>> - }
>> -
>> assert_list_leaf_cfs_rq(rq);
>>
>> hrtick_update(rq);
>> --
>> 2.35.2
>>
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-05-24 9:35 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-04-27 16:05 [PATCH] sched/fair: optimize and simplify rq leaf_cfs_rq_list Chengming Zhou
2022-05-11 12:05 ` Chengming Zhou
2022-05-12 14:02 ` Vincent Guittot
2022-05-13 4:32 ` [External] " Chengming Zhou
2022-05-23 16:23 ` Vincent Guittot
2022-05-24 9:35 ` [External] " Chengming Zhou
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).