Skip to contentSkip to conversation

Fix nohz_full vs cfs bandwidth

PATCH Applied Phil Auld 5 participants 22 messages Jul 12, 2023 – Aug 9, 2023 Scheduler

lore
Share & tools
lore

Phil Auld

author
#0 lore raw

v6 cover Fix nohz_full vs cfs bandwidth

This is v6 of patch 2/2 which is adding code to prevent the tick from being stopped when the single running task has bandwidth limits. Discussions had led to the idea of adding a bit to task_struct to help make this decision.

There was some complexity with doing it in the task which is avoided by using something in the cfs_rq. Looking into that lead me to the hierarchical_quota field in the cfs_bandwith struct. We spend a good deal of effort updating (or trying to, see patch 1/2) that value for the whole task_group tree when a quota is set/changed.

This new version first fixes that value to be meaningful for cgroupv2 and then leverages it to make the decisions about blocking the tick_stop.

Phil Auld (2):
sched, cgroup: Restore meaning to hierarchical_quota Sched/fair: Block nohz tick_stop when cfs bandwidth in use

 kernel/sched/core.c     | 23 ++++++++++++++---
 kernel/sched/fair.c     | 56 ++++++++++++++++++++++++++++++++++++++---
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  3 ++-
 4 files changed, 76 insertions(+), 8 deletions(-)
Signature · 1 line
2.31.1
author
#1 lore raw

PATCH v2 1/2 sched, cgroup: Restore meaning to hierarchical_quota

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

---

v2: Improve comment about how setting hierarchical_quota correctly

helps the scheduler. Remove extra parens.

 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +9 -4
index a68d1276bab0..f80697a79baf 100644
@@ -9904,7 +9904,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;
@@ -11038,11 +11038,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In cgroup2 this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */
 		if (cgroup_subsys_on_dfl(cpu_cgrp_subsys)) {
-			quota = min(quota, parent_quota);
+			if (quota == RUNTIME_INF)
+				quota = parent_quota;
+			else if (parent_quota != RUNTIME_INF)
+				quota = min(quota, parent_quota);
 		} else {
 			if (quota == RUNTIME_INF)
 				quota = parent_quota;
kernel/sched/fair.c +4 -3
index 373ff5f55884..d9b3d4617e16 100644
@@ -6005,13 +6005,14 @@ static enum hrtimer_restart sched_cfs_period_timer(struct hrtimer *timer)
 	return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent)
 {
 	raw_spin_lock_init(&cfs_b->lock);
 	cfs_b->runtime = 0;
 	cfs_b->quota = RUNTIME_INF;
 	cfs_b->period = ns_to_ktime(default_cfs_period());
 	cfs_b->burst = 0;
+	cfs_b->hierarchical_quota = parent ? parent->hierarchical_quota : RUNTIME_INF;
 
 	INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);
 	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED);
@@ -6168,7 +6169,7 @@ static inline int throttled_lb_pair(struct task_group *tg,
 	return 0;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent) {}
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) {}
@@ -12373,7 +12374,7 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
 
 	tg->shares = NICE_0_LOAD;
 
-	init_cfs_bandwidth(tg_cfs_bandwidth(tg));
+	init_cfs_bandwidth(tg_cfs_bandwidth(tg), tg_cfs_bandwidth(parent));
 
 	for_each_possible_cpu(i) {
 		cfs_rq = kzalloc_node(sizeof(struct cfs_rq),
kernel/sched/sched.h +1 -1
index ec7b3e0a2b20..63822c9238cc 100644
@@ -460,7 +460,7 @@ extern void unregister_fair_sched_group(struct task_group *tg);
 extern void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
 			struct sched_entity *se, int cpu,
 			struct sched_entity *parent);
-extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
+extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent);
 
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
Signature · 1 line
2.31.1
#2 lore raw

Re: [PATCH v2 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Quoted 49 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

---

v2: Improve comment about how setting hierarchical_quota correctly

helps the scheduler. Remove extra parens.

 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +1 -1
index a68d1276bab0..f80697a79baf 100644
@@ -9904,7 +9904,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11045

Quoted 10 lines
kernel/sched/core.c +4 -2
@@ -11038,11 +11038,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In cgroup2 this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */

It's still used for determining this on cgroup1 (and the cgroup1 code still works for that), right?

Quoted 60 lines
kernel/sched/core.c +4 -1
 		if (cgroup_subsys_on_dfl(cpu_cgrp_subsys)) {
-			quota = min(quota, parent_quota);
+			if (quota == RUNTIME_INF)
+				quota = parent_quota;
+			else if (parent_quota != RUNTIME_INF)
+				quota = min(quota, parent_quota);
 		} else {
 			if (quota == RUNTIME_INF)
 				quota = parent_quota;
kernel/sched/fair.c +4 -3
index 373ff5f55884..d9b3d4617e16 100644
@@ -6005,13 +6005,14 @@ static enum hrtimer_restart sched_cfs_period_timer(struct hrtimer *timer)
 	return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent)
 {
 	raw_spin_lock_init(&cfs_b->lock);
 	cfs_b->runtime = 0;
 	cfs_b->quota = RUNTIME_INF;
 	cfs_b->period = ns_to_ktime(default_cfs_period());
 	cfs_b->burst = 0;
+	cfs_b->hierarchical_quota = parent ? parent->hierarchical_quota : RUNTIME_INF;
 
 	INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);
 	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED);
@@ -6168,7 +6169,7 @@ static inline int throttled_lb_pair(struct task_group *tg,
 	return 0;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent) {}
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) {}
@@ -12373,7 +12374,7 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
 
 	tg->shares = NICE_0_LOAD;
 
-	init_cfs_bandwidth(tg_cfs_bandwidth(tg));
+	init_cfs_bandwidth(tg_cfs_bandwidth(tg), tg_cfs_bandwidth(parent));
 
 	for_each_possible_cpu(i) {
 		cfs_rq = kzalloc_node(sizeof(struct cfs_rq),
kernel/sched/sched.h +1 -1
index ec7b3e0a2b20..63822c9238cc 100644
@@ -460,7 +460,7 @@ extern void unregister_fair_sched_group(struct task_group *tg);
 extern void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
 			struct sched_entity *se, int cpu,
 			struct sched_entity *parent);
-extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
+extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent);
 
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
author
#3 lore raw

Re: [PATCH v2 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Quoted 52 lines · On Wed, Jul 12, 2023 at 03:09:31PM -0700 Benjamin Segall wrote:
Quoted 50 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

---

v2: Improve comment about how setting hierarchical_quota correctly

helps the scheduler. Remove extra parens.

 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +1 -1
index a68d1276bab0..f80697a79baf 100644
@@ -9904,7 +9904,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;
@@ -11038,11 +11038,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
Quoted 13 lines

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11045

Quoted 9 lines
kernel/sched/core.c +4 -2
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In cgroup2 this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */

It's still used for determining this on cgroup1 (and the cgroup1 code still works for that), right?

It would, except that the enforcement of child quota <= parent quota means that cfs_rq->runtime_enabled will be set and we'll hit that first on cgroup1. So we don't really use it for this determination in cgroup1.

But I could generalize that comment if you want.

Thanks,
Phil

Quoted 61 lines
Quoted 60 lines
kernel/sched/core.c +4 -1
 		if (cgroup_subsys_on_dfl(cpu_cgrp_subsys)) {
-			quota = min(quota, parent_quota);
+			if (quota == RUNTIME_INF)
+				quota = parent_quota;
+			else if (parent_quota != RUNTIME_INF)
+				quota = min(quota, parent_quota);
 		} else {
 			if (quota == RUNTIME_INF)
 				quota = parent_quota;
kernel/sched/fair.c +4 -3
index 373ff5f55884..d9b3d4617e16 100644
@@ -6005,13 +6005,14 @@ static enum hrtimer_restart sched_cfs_period_timer(struct hrtimer *timer)
 	return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent)
 {
 	raw_spin_lock_init(&cfs_b->lock);
 	cfs_b->runtime = 0;
 	cfs_b->quota = RUNTIME_INF;
 	cfs_b->period = ns_to_ktime(default_cfs_period());
 	cfs_b->burst = 0;
+	cfs_b->hierarchical_quota = parent ? parent->hierarchical_quota : RUNTIME_INF;
 
 	INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);
 	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED);
@@ -6168,7 +6169,7 @@ static inline int throttled_lb_pair(struct task_group *tg,
 	return 0;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent) {}
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) {}
@@ -12373,7 +12374,7 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
 
 	tg->shares = NICE_0_LOAD;
 
-	init_cfs_bandwidth(tg_cfs_bandwidth(tg));
+	init_cfs_bandwidth(tg_cfs_bandwidth(tg), tg_cfs_bandwidth(parent));
 
 	for_each_possible_cpu(i) {
 		cfs_rq = kzalloc_node(sizeof(struct cfs_rq),
kernel/sched/sched.h +1 -1
index ec7b3e0a2b20..63822c9238cc 100644
@@ -460,7 +460,7 @@ extern void unregister_fair_sched_group(struct task_group *tg);
 extern void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
 			struct sched_entity *se, int cpu,
 			struct sched_entity *parent);
-extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
+extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent);
 
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
#4 lore raw

Re: [PATCH v2 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Quoted 67 lines · Phil Auld
Quoted 52 lines · On Wed, Jul 12, 2023 at 03:09:31PM -0700 Benjamin Segall wrote:
Quoted 50 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

---

v2: Improve comment about how setting hierarchical_quota correctly

helps the scheduler. Remove extra parens.

 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +1 -1
index a68d1276bab0..f80697a79baf 100644
@@ -9904,7 +9904,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;
@@ -11038,11 +11038,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
Quoted 13 lines

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11045

Quoted 9 lines
kernel/sched/core.c +4 -2
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In cgroup2 this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */

It's still used for determining this on cgroup1 (and the cgroup1 code still works for that), right?

Quoted 4 lines

It would, except that the enforcement of child quota <= parent quota means that cfs_rq->runtime_enabled will be set and we'll hit that first on cgroup1. So we don't really use it for this determination in cgroup1.

cgroup1 tg_cfs_schedulable_down only constricts child quota when it's set. You can set quota=RUNTIME_INF on any cgroup, no matter what its parent is. (The schedulable constraint is a little silly)

author
#5 lore raw

Re: [PATCH v2 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Quoted 69 lines · On Thu, Jul 13, 2023 at 01:12:04PM -0700 Benjamin Segall wrote:
Quoted 67 lines · Phil Auld
Quoted 65 lines · On Wed, Jul 12, 2023 at 03:09:31PM -0700 Benjamin Segall wrote:
Quoted 49 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

---

v2: Improve comment about how setting hierarchical_quota correctly

helps the scheduler. Remove extra parens.

 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +1 -1
index a68d1276bab0..f80697a79baf 100644
@@ -9904,7 +9904,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11045

Quoted 10 lines
kernel/sched/core.c +4 -2
@@ -11038,11 +11038,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In cgroup2 this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */

It's still used for determining this on cgroup1 (and the cgroup1 code still works for that), right?

Quoted 9 lines
Quoted 4 lines

It would, except that the enforcement of child quota <= parent quota means that cfs_rq->runtime_enabled will be set and we'll hit that first on cgroup1. So we don't really use it for this determination in cgroup1.

cgroup1 tg_cfs_schedulable_down only constricts child quota when it's set. You can set quota=RUNTIME_INF on any cgroup, no matter what its parent is. (The schedulable constraint is a little silly)

Aargh :) I had "In both cases this ..." there before I talked myself out
of it.

I'll update this one.

Cheers,
Phil

author
#6 lore raw

PATCH v3 1/2 sched, cgroup: Restore meaning to hierarchical_quota

In reply to Benjamin Segall

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

---
v3: Make comment apply to both cgroup1 and 2.

v2: Improve comment about how setting hierarchical_quota correctly helps the scheduler. Remove extra parens.

 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +9 -4
index a68d1276bab0..f7c1510fdac1 100644
@@ -9904,7 +9904,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;
@@ -11038,11 +11038,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In both cases this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */
 		if (cgroup_subsys_on_dfl(cpu_cgrp_subsys)) {
-			quota = min(quota, parent_quota);
+			if (quota == RUNTIME_INF)
+				quota = parent_quota;
+			else if (parent_quota != RUNTIME_INF)
+				quota = min(quota, parent_quota);
 		} else {
 			if (quota == RUNTIME_INF)
 				quota = parent_quota;
kernel/sched/fair.c +4 -3
index 373ff5f55884..d9b3d4617e16 100644
@@ -6005,13 +6005,14 @@ static enum hrtimer_restart sched_cfs_period_timer(struct hrtimer *timer)
 	return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent)
 {
 	raw_spin_lock_init(&cfs_b->lock);
 	cfs_b->runtime = 0;
 	cfs_b->quota = RUNTIME_INF;
 	cfs_b->period = ns_to_ktime(default_cfs_period());
 	cfs_b->burst = 0;
+	cfs_b->hierarchical_quota = parent ? parent->hierarchical_quota : RUNTIME_INF;
 
 	INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);
 	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED);
@@ -6168,7 +6169,7 @@ static inline int throttled_lb_pair(struct task_group *tg,
 	return 0;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent) {}
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) {}
@@ -12373,7 +12374,7 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
 
 	tg->shares = NICE_0_LOAD;
 
-	init_cfs_bandwidth(tg_cfs_bandwidth(tg));
+	init_cfs_bandwidth(tg_cfs_bandwidth(tg), tg_cfs_bandwidth(parent));
 
 	for_each_possible_cpu(i) {
 		cfs_rq = kzalloc_node(sizeof(struct cfs_rq),
kernel/sched/sched.h +1 -1
index ec7b3e0a2b20..63822c9238cc 100644
@@ -460,7 +460,7 @@ extern void unregister_fair_sched_group(struct task_group *tg);
 extern void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
 			struct sched_entity *se, int cpu,
 			struct sched_entity *parent);
-extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
+extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent);
 
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
Signature · 1 line
2.31.1
Acked-by on v3 1/2
#7 lore raw

Re: [PATCH v3 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Quoted 8 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

Quoted 6 lines

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")

Does this really fix anything observable? I wonder whether this is more misleading than helpful. In cgroup2, the value simply wasn't being used, right?

Quoted 11 lines
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11044

Quoted 4 lines
kernel/sched/core.c +4 -0
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In both cases this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.

The discussion on this comment is stretching too long and this is fine too but what's worth commenting for cgroup2 is that the limit value itself doesn't mean anything and we're just latching onto the value used by cgroup1 to track whether there's any limit active or not.

Thanks.

Signature · 1 line
tejun
author
#8 lore raw

Re: [PATCH v3 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Quoted 14 lines · On Mon, Jul 17, 2023 at 08:27:24AM -1000 Tejun Heo wrote:
Quoted 13 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Quoted 6 lines
Quoted 1 line
Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")

Does this really fix anything observable? I wonder whether this is more misleading than helpful. In cgroup2, the value simply wasn't being used, right?

It wasn't being used but was actively being set wrong. I mean if we are going to bother doing the __cfs_schedulable() tg tree walk we might as well have not been setting a bogus value. But that said, no it was not observable until I tried to use it.

I'm fine if that's dropped. I just wanted it set right going forward :)

Quoted 12 lines
Quoted 11 lines
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>
Quoted 2 lines

Thanks!

Quoted 9 lines

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11044

Quoted 4 lines
kernel/sched/core.c +4 -0
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In both cases this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.

The discussion on this comment is stretching too long and this is fine too but what's worth commenting for cgroup2 is that the limit value itself doesn't mean anything and we're just latching onto the value used by cgroup1 to track whether there's any limit active or not.

I thought that was implied by the wording. "If a given task has a bandwidth contraint" not "what a given task's bandwidth constraint is". In both cases that's how the other parts of the scheduler are using it. The actual non-RUNTIME_INF value only matters in this function (and only for cgroup1 indeed).

But... the value is just as accurate for cgroup2 and cgroup1. The task is still going to be limited by that bandwidth constraint even if its own bandwidth limit is nominally higher, right?

Cheers,
Phil

Quoted 6 lines

Thanks.

Signature · 1 line
tejun
author
#9 lore raw

Re: [PATCH v3 1/2] sched, cgroup: Restore meaning to hierarchical_quota

Hi Tejun,

Quoted 15 lines · On Tue, Jul 18, 2023 at 08:57:59AM -0400 Phil Auld wrote:
Quoted 14 lines · On Mon, Jul 17, 2023 at 08:27:24AM -1000 Tejun Heo wrote:
Quoted 13 lines · Phil Auld

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Quoted 6 lines
Quoted 6 lines
Quoted 1 line
Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")

Does this really fix anything observable? I wonder whether this is more misleading than helpful. In cgroup2, the value simply wasn't being used, right?

(Sorry, my editor bit me...I had added:)

I don't feel strongly about the fixes. What was there seemed broken to me so ... "Fixes". But it doesn't matter.

Quoted 6 lines

It wasn't being used but was actively being set wrong. I mean if we are going to bother doing the __cfs_schedulable() tg tree walk we might as well have not been setting a bogus value. But that said, no it was not observable until I tried to use it.

We have a field called hierarchical_quota, that was being unconditionally set to -1 for cgroup2. I figured it would be more correct to reflect the hieratchical quota. :)

Quoted 54 lines

I'm fine if that's dropped. I just wanted it set right going forward :)

Quoted 12 lines
Quoted 11 lines
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>
Cc Tejun Heo <[email protected]>
Quoted 2 lines

Thanks!

Quoted 9 lines

kernel/sched/core.c static int tg_cfs_schedulable_down(struct task_group *tg, void *data) · line 11044

Quoted 4 lines
kernel/sched/core.c +4 -0
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In both cases this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.

The discussion on this comment is stretching too long and this is fine too but what's worth commenting for cgroup2 is that the limit value itself doesn't mean anything and we're just latching onto the value used by cgroup1 to track whether there's any limit active or not.

I thought that was implied by the wording. "If a given task has a bandwidth contraint" not "what a given task's bandwidth constraint is". In both cases that's how the other parts of the scheduler are using it. The actual non-RUNTIME_INF value only matters in this function (and only for cgroup1 indeed).

But... the value is just as accurate for cgroup2 and cgroup1. The task is still going to be limited by that bandwidth constraint even if its own bandwidth limit is nominally higher, right?

Cheers,
Phil

Quoted 6 lines

Thanks.

Signature · 1 line
tejun
merge bot Applied
#10 lore raw

[tip: sched/core] sched, cgroup: Restore meaning to hierarchical_quota

In reply to Phil Auld
The following commit has been merged into the sched/core branch of tip:

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

Commit-ID:     c98c18270be115678f4295b10a5af5dcc9c4efa0
Gitweb:        https://git.kernel.org/tip/c98c18270be115678f4295b10a5af5dcc9c4efa0
Author:        Phil Auld <[email protected]>
AuthorDate:    Fri, 14 Jul 2023 08:57:46 -04:00
Committer:     Peter Zijlstra <[email protected]>
CommitterDate: Wed, 02 Aug 2023 16:19:26 +02:00

sched, cgroup: Restore meaning to hierarchical_quota

In cgroupv2 cfs_b->hierarchical_quota is set to -1 for all task groups due to the previous fix simply taking the min. It should reflect a limit imposed at that level or by an ancestor. Even though cgroupv2 does not require child quota to be less than or equal to that of its ancestors the task group will still be constrained by such a quota so this should be shown here. Cgroupv1 continues to set this correctly.

In both cases, add initialization when a new task group is created based on the current parent's value (or RUNTIME_INF in the case of root_task_group). Otherwise, the field is wrong until a quota is changed after creation and __cfs_schedulable() is called.

Fixes c53593e5cb69 ("sched, cgroup: Don't reject lower cpu.max on ancestors")
---
 kernel/sched/core.c  | 13 +++++++++----
 kernel/sched/fair.c  |  7 ++++---
 kernel/sched/sched.h |  2 +-
 3 files changed, 14 insertions(+), 8 deletions(-)
kernel/sched/core.c +9 -4
index 83e3654..3af25ca 100644
@@ -9953,7 +9953,7 @@ void __init sched_init(void)
 		ptr += nr_cpu_ids * sizeof(void **);
 
 		root_task_group.shares = ROOT_TASK_GROUP_LOAD;
-		init_cfs_bandwidth(&root_task_group.cfs_bandwidth);
+		init_cfs_bandwidth(&root_task_group.cfs_bandwidth, NULL);
 #endif /* CONFIG_FAIR_GROUP_SCHED */
 #ifdef CONFIG_RT_GROUP_SCHED
 		root_task_group.rt_se = (struct sched_rt_entity **)ptr;
@@ -11087,11 +11087,16 @@ static int tg_cfs_schedulable_down(struct task_group *tg, void *data)
 
 		/*
 		 * Ensure max(child_quota) <= parent_quota.  On cgroup2,
-		 * always take the min.  On cgroup1, only inherit when no
-		 * limit is set:
+		 * always take the non-RUNTIME_INF min.  On cgroup1, only
+		 * inherit when no limit is set. In both cases this is used
+		 * by the scheduler to determine if a given CFS task has a
+		 * bandwidth constraint at some higher level.
 		 */
 		if (cgroup_subsys_on_dfl(cpu_cgrp_subsys)) {
-			quota = min(quota, parent_quota);
+			if (quota == RUNTIME_INF)
+				quota = parent_quota;
+			else if (parent_quota != RUNTIME_INF)
+				quota = min(quota, parent_quota);
 		} else {
 			if (quota == RUNTIME_INF)
 				quota = parent_quota;
kernel/sched/fair.c +4 -3
index f55b0a7..26bfbb6 100644
@@ -6045,13 +6045,14 @@ static enum hrtimer_restart sched_cfs_period_timer(struct hrtimer *timer)
 	return idle ? HRTIMER_NORESTART : HRTIMER_RESTART;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b)
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent)
 {
 	raw_spin_lock_init(&cfs_b->lock);
 	cfs_b->runtime = 0;
 	cfs_b->quota = RUNTIME_INF;
 	cfs_b->period = ns_to_ktime(default_cfs_period());
 	cfs_b->burst = 0;
+	cfs_b->hierarchical_quota = parent ? parent->hierarchical_quota : RUNTIME_INF;
 
 	INIT_LIST_HEAD(&cfs_b->throttled_cfs_rq);
 	hrtimer_init(&cfs_b->period_timer, CLOCK_MONOTONIC, HRTIMER_MODE_ABS_PINNED);
@@ -6217,7 +6218,7 @@ static inline int throttled_lb_pair(struct task_group *tg,
 	return 0;
 }
 
-void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
+void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent) {}
 
 #ifdef CONFIG_FAIR_GROUP_SCHED
 static void init_cfs_rq_runtime(struct cfs_rq *cfs_rq) {}
@@ -12599,7 +12600,7 @@ int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
 
 	tg->shares = NICE_0_LOAD;
 
-	init_cfs_bandwidth(tg_cfs_bandwidth(tg));
+	init_cfs_bandwidth(tg_cfs_bandwidth(tg), tg_cfs_bandwidth(parent));
 
 	for_each_possible_cpu(i) {
 		cfs_rq = kzalloc_node(sizeof(struct cfs_rq),
kernel/sched/sched.h +1 -1
index 9baeb1a..602de71 100644
@@ -454,7 +454,7 @@ extern void unregister_fair_sched_group(struct task_group *tg);
 extern void init_tg_cfs_entry(struct task_group *tg, struct cfs_rq *cfs_rq,
 			struct sched_entity *se, int cpu,
 			struct sched_entity *parent);
-extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
+extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth *parent);
 
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
author
#11 lore raw

PATCH v6 2/2 Sched/fair: Block nohz tick_stop when cfs bandwidth in use

In reply to Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>

---
v6: restore check for fair_sched_class

v5: Reworked checks to use newly-fixed cfs_b->hierarchical_quota to check for bw
constraints.

v4: Made checks for runtime_enabled hierarchical.

v3: Moved sched_cfs_bandwidth_active() prototype to sched.h outside of guards to
silence -Wmissing-prototypes.

v2: Ben pointed out that the bit could get cleared in the dequeue path if we migrate a newly enqueued task without preempting curr. Added a check for that edge case to sched_can_stop_tick. Removed the call to sched_can_stop_tick from sched_fair_update_stop_tick since it was redundant.

 kernel/sched/core.c     | 26 ++++++++++++++++++++++
 kernel/sched/fair.c     | 49 +++++++++++++++++++++++++++++++++++++++++
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  1 +
 4 files changed, 78 insertions(+)
kernel/sched/core.c +26 -0
index f80697a79baf..8a2ed4c0b709 100644
@@ -1194,6 +1194,20 @@ static void nohz_csd_func(void *info)
 #endif /* CONFIG_NO_HZ_COMMON */
 
 #ifdef CONFIG_NO_HZ_FULL
+static inline bool __need_bw_check(struct rq *rq, struct task_struct *p)
+{
+	if (rq->nr_running != 1)
+		return false;
+
+	if (p->sched_class != &fair_sched_class)
+		return false;
+
+	if (!task_on_rq_queued(p))
+		return false;
+
+	return true;
+}
+
 bool sched_can_stop_tick(struct rq *rq)
 {
 	int fifo_nr_running;
@@ -1229,6 +1243,18 @@ bool sched_can_stop_tick(struct rq *rq)
 	if (rq->nr_running > 1)
 		return false;
 
+	/*
+	 * If there is one task and it has CFS runtime bandwidth constraints
+	 * and it's on the cpu now we don't want to stop the tick.
+	 * This check prevents clearing the bit if a newly enqueued task here is
+	 * dequeued by migrating while the constrained task continues to run.
+	 * E.g. going from 2->1 without going through pick_next_task().
+	 */
+	if (sched_feat(HZ_BW) && __need_bw_check(rq, rq->curr)) {
+		if (cfs_task_bw_constrained(rq->curr))
+			return false;
+	}
+
 	return true;
 }
 #endif /* CONFIG_NO_HZ_FULL */
kernel/sched/fair.c +49 -0

70 lines not on this page · raw message

kernel/sched/features.h +2 -0
index ee7f23c76bd3..6fdf1fdf6b17 100644
@@ -101,3 +101,5 @@ SCHED_FEAT(LATENCY_WARN, false)
 
 SCHED_FEAT(ALT_PERIOD, true)
 SCHED_FEAT(BASE_SLICE, true)
+
+SCHED_FEAT(HZ_BW, false)
kernel/sched/sched.h +1 -0
index 63822c9238cc..d6d346bc78aa 100644
@@ -465,6 +465,7 @@ extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
 extern void unthrottle_cfs_rq(struct cfs_rq *cfs_rq);
+extern bool cfs_task_bw_constrained(struct task_struct *p);
 
 extern void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
 		struct sched_rt_entity *rt_se, int cpu,
Signature · 1 line
2.31.1
reviewer Reviewed-by on v6 2/2
#12 lore raw

Re: [PATCH v6 2/2] Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 14 lines · Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Quoted 8 lines

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

I think this looks good now.

Quoted 178 lines
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>

---
v6: restore check for fair_sched_class

v5: Reworked checks to use newly-fixed cfs_b->hierarchical_quota to check for bw
constraints.

v4: Made checks for runtime_enabled hierarchical.

v3: Moved sched_cfs_bandwidth_active() prototype to sched.h outside of guards to
silence -Wmissing-prototypes.

v2: Ben pointed out that the bit could get cleared in the dequeue path if we migrate a newly enqueued task without preempting curr. Added a check for that edge case to sched_can_stop_tick. Removed the call to sched_can_stop_tick from sched_fair_update_stop_tick since it was redundant.

 kernel/sched/core.c     | 26 ++++++++++++++++++++++
 kernel/sched/fair.c     | 49 +++++++++++++++++++++++++++++++++++++++++
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  1 +
 4 files changed, 78 insertions(+)
kernel/sched/core.c +26 -0
index f80697a79baf..8a2ed4c0b709 100644
@@ -1194,6 +1194,20 @@ static void nohz_csd_func(void *info)
 #endif /* CONFIG_NO_HZ_COMMON */
 
 #ifdef CONFIG_NO_HZ_FULL
+static inline bool __need_bw_check(struct rq *rq, struct task_struct *p)
+{
+	if (rq->nr_running != 1)
+		return false;
+
+	if (p->sched_class != &fair_sched_class)
+		return false;
+
+	if (!task_on_rq_queued(p))
+		return false;
+
+	return true;
+}
+
 bool sched_can_stop_tick(struct rq *rq)
 {
 	int fifo_nr_running;
@@ -1229,6 +1243,18 @@ bool sched_can_stop_tick(struct rq *rq)
 	if (rq->nr_running > 1)
 		return false;
 
+	/*
+	 * If there is one task and it has CFS runtime bandwidth constraints
+	 * and it's on the cpu now we don't want to stop the tick.
+	 * This check prevents clearing the bit if a newly enqueued task here is
+	 * dequeued by migrating while the constrained task continues to run.
+	 * E.g. going from 2->1 without going through pick_next_task().
+	 */
+	if (sched_feat(HZ_BW) && __need_bw_check(rq, rq->curr)) {
+		if (cfs_task_bw_constrained(rq->curr))
+			return false;
+	}
+
 	return true;
 }
 #endif /* CONFIG_NO_HZ_FULL */
kernel/sched/fair.c +49 -0
index d9b3d4617e16..acd9f317aad1 100644
@@ -6140,6 +6140,46 @@ static void __maybe_unused unthrottle_offline_cfs_rqs(struct rq *rq)
 	rcu_read_unlock();
 }
 
+bool cfs_task_bw_constrained(struct task_struct *p)
+{
+	struct cfs_rq *cfs_rq = task_cfs_rq(p);
+
+	if (!cfs_bandwidth_used())
+		return false;
+
+	if (cfs_rq->runtime_enabled ||
+	    tg_cfs_bandwidth(cfs_rq->tg)->hierarchical_quota != RUNTIME_INF)
+		return true;
+
+	return false;
+}
+
+#ifdef CONFIG_NO_HZ_FULL
+/* called from pick_next_task_fair() */
+static void sched_fair_update_stop_tick(struct rq *rq, struct task_struct *p)
+{
+	int cpu = cpu_of(rq);
+
+	if (!sched_feat(HZ_BW) || !cfs_bandwidth_used())
+		return;
+
+	if (!tick_nohz_full_cpu(cpu))
+		return;
+
+	if (rq->nr_running != 1)
+		return;
+
+	/*
+	 *  We know there is only one task runnable and we've just picked it. The
+	 *  normal enqueue path will have cleared TICK_DEP_BIT_SCHED if we will
+	 *  be otherwise able to stop the tick. Just need to check if we are using
+	 *  bandwidth control.
+	 */
+	if (cfs_task_bw_constrained(p))
+		tick_nohz_dep_set_cpu(cpu, TICK_DEP_BIT_SCHED);
+}
+#endif
+
 #else /* CONFIG_CFS_BANDWIDTH */
 
 static inline bool cfs_bandwidth_used(void)
@@ -6182,9 +6222,17 @@ static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg)
 static inline void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
 static inline void update_runtime_enabled(struct rq *rq) {}
 static inline void unthrottle_offline_cfs_rqs(struct rq *rq) {}
+bool cfs_task_bw_constrained(struct task_struct *p)
+{
+	return false;
+}
 
 #endif /* CONFIG_CFS_BANDWIDTH */
 
+#if !defined(CONFIG_CFS_BANDWIDTH) || !defined(CONFIG_NO_HZ_FULL)
+static inline void sched_fair_update_stop_tick(struct rq *rq, struct task_struct *p) {}
+#endif
+
 /**************************************************
  * CFS operations on tasks:
  */
@@ -8098,6 +8146,7 @@ done: __maybe_unused;
 		hrtick_start_fair(rq, p);
 
 	update_misfit_status(p, rq);
+	sched_fair_update_stop_tick(rq, p);
 
 	return p;
 
kernel/sched/features.h +2 -0
index ee7f23c76bd3..6fdf1fdf6b17 100644
@@ -101,3 +101,5 @@ SCHED_FEAT(LATENCY_WARN, false)
 
 SCHED_FEAT(ALT_PERIOD, true)
 SCHED_FEAT(BASE_SLICE, true)
+
+SCHED_FEAT(HZ_BW, false)
kernel/sched/sched.h +1 -0
index 63822c9238cc..d6d346bc78aa 100644
@@ -465,6 +465,7 @@ extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
 extern void unthrottle_cfs_rq(struct cfs_rq *cfs_rq);
+extern bool cfs_task_bw_constrained(struct task_struct *p);
 
 extern void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
 		struct sched_rt_entity *rt_se, int cpu,
author
#13 lore raw

Re: [PATCH v6 2/2] Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 19 lines · On Wed, Jul 12, 2023 at 03:11:32PM -0700 Benjamin Segall wrote:
Quoted 17 lines · Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Quoted 9 lines
Quoted 5 lines

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

I think this looks good now.

Thanks for your help and patience!

Cheers,
Phil

Quoted 179 lines
Quoted 178 lines
Cc Ingo Molnar <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Vincent Guittot <[email protected]>
Cc Juri Lelli <[email protected]>
Cc Dietmar Eggemann <[email protected]>
Cc Valentin Schneider <[email protected]>
Cc Ben Segall <[email protected]>
Cc Frederic Weisbecker <[email protected]>

---
v6: restore check for fair_sched_class

v5: Reworked checks to use newly-fixed cfs_b->hierarchical_quota to check for bw
constraints.

v4: Made checks for runtime_enabled hierarchical.

v3: Moved sched_cfs_bandwidth_active() prototype to sched.h outside of guards to
silence -Wmissing-prototypes.

v2: Ben pointed out that the bit could get cleared in the dequeue path if we migrate a newly enqueued task without preempting curr. Added a check for that edge case to sched_can_stop_tick. Removed the call to sched_can_stop_tick from sched_fair_update_stop_tick since it was redundant.

 kernel/sched/core.c     | 26 ++++++++++++++++++++++
 kernel/sched/fair.c     | 49 +++++++++++++++++++++++++++++++++++++++++
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  1 +
 4 files changed, 78 insertions(+)
kernel/sched/core.c +26 -0
index f80697a79baf..8a2ed4c0b709 100644
@@ -1194,6 +1194,20 @@ static void nohz_csd_func(void *info)
 #endif /* CONFIG_NO_HZ_COMMON */
 
 #ifdef CONFIG_NO_HZ_FULL
+static inline bool __need_bw_check(struct rq *rq, struct task_struct *p)
+{
+	if (rq->nr_running != 1)
+		return false;
+
+	if (p->sched_class != &fair_sched_class)
+		return false;
+
+	if (!task_on_rq_queued(p))
+		return false;
+
+	return true;
+}
+
 bool sched_can_stop_tick(struct rq *rq)
 {
 	int fifo_nr_running;
@@ -1229,6 +1243,18 @@ bool sched_can_stop_tick(struct rq *rq)
 	if (rq->nr_running > 1)
 		return false;
 
+	/*
+	 * If there is one task and it has CFS runtime bandwidth constraints
+	 * and it's on the cpu now we don't want to stop the tick.
+	 * This check prevents clearing the bit if a newly enqueued task here is
+	 * dequeued by migrating while the constrained task continues to run.
+	 * E.g. going from 2->1 without going through pick_next_task().
+	 */
+	if (sched_feat(HZ_BW) && __need_bw_check(rq, rq->curr)) {
+		if (cfs_task_bw_constrained(rq->curr))
+			return false;
+	}
+
 	return true;
 }
 #endif /* CONFIG_NO_HZ_FULL */
kernel/sched/fair.c +49 -0
index d9b3d4617e16..acd9f317aad1 100644
@@ -6140,6 +6140,46 @@ static void __maybe_unused unthrottle_offline_cfs_rqs(struct rq *rq)
 	rcu_read_unlock();
 }
 
+bool cfs_task_bw_constrained(struct task_struct *p)
+{
+	struct cfs_rq *cfs_rq = task_cfs_rq(p);
+
+	if (!cfs_bandwidth_used())
+		return false;
+
+	if (cfs_rq->runtime_enabled ||
+	    tg_cfs_bandwidth(cfs_rq->tg)->hierarchical_quota != RUNTIME_INF)
+		return true;
+
+	return false;
+}
+
+#ifdef CONFIG_NO_HZ_FULL
+/* called from pick_next_task_fair() */
+static void sched_fair_update_stop_tick(struct rq *rq, struct task_struct *p)
+{
+	int cpu = cpu_of(rq);
+
+	if (!sched_feat(HZ_BW) || !cfs_bandwidth_used())
+		return;
+
+	if (!tick_nohz_full_cpu(cpu))
+		return;
+
+	if (rq->nr_running != 1)
+		return;
+
+	/*
+	 *  We know there is only one task runnable and we've just picked it. The
+	 *  normal enqueue path will have cleared TICK_DEP_BIT_SCHED if we will
+	 *  be otherwise able to stop the tick. Just need to check if we are using
+	 *  bandwidth control.
+	 */
+	if (cfs_task_bw_constrained(p))
+		tick_nohz_dep_set_cpu(cpu, TICK_DEP_BIT_SCHED);
+}
+#endif
+
 #else /* CONFIG_CFS_BANDWIDTH */
 
 static inline bool cfs_bandwidth_used(void)
@@ -6182,9 +6222,17 @@ static inline struct cfs_bandwidth *tg_cfs_bandwidth(struct task_group *tg)
 static inline void destroy_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
 static inline void update_runtime_enabled(struct rq *rq) {}
 static inline void unthrottle_offline_cfs_rqs(struct rq *rq) {}
+bool cfs_task_bw_constrained(struct task_struct *p)
+{
+	return false;
+}
 
 #endif /* CONFIG_CFS_BANDWIDTH */
 
+#if !defined(CONFIG_CFS_BANDWIDTH) || !defined(CONFIG_NO_HZ_FULL)
+static inline void sched_fair_update_stop_tick(struct rq *rq, struct task_struct *p) {}
+#endif
+
 /**************************************************
  * CFS operations on tasks:
  */
@@ -8098,6 +8146,7 @@ done: __maybe_unused;
 		hrtick_start_fair(rq, p);
 
 	update_misfit_status(p, rq);
+	sched_fair_update_stop_tick(rq, p);
 
 	return p;
 
kernel/sched/features.h +2 -0
index ee7f23c76bd3..6fdf1fdf6b17 100644
@@ -101,3 +101,5 @@ SCHED_FEAT(LATENCY_WARN, false)
 
 SCHED_FEAT(ALT_PERIOD, true)
 SCHED_FEAT(BASE_SLICE, true)
+
+SCHED_FEAT(HZ_BW, false)
kernel/sched/sched.h +1 -0
index 63822c9238cc..d6d346bc78aa 100644
@@ -465,6 +465,7 @@ extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
 extern void unthrottle_cfs_rq(struct cfs_rq *cfs_rq);
+extern bool cfs_task_bw_constrained(struct task_struct *p);
 
 extern void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
 		struct sched_rt_entity *rt_se, int cpu,
maintainer
#14 lore raw

Re: [PATCH v6 2/2] Sched/fair: Block nohz tick_stop when cfs bandwidth in use

In reply to Phil Auld
Quoted 14 lines · Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Quoted 5 lines

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

These appear fine to me, except:

Quoted 2 lines

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

What was the thinking here? This means nobody will be using this -- why would you want this default disabled?

author
#15 lore raw

Re: [PATCH v6 2/2] Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 21 lines · On Tue, Aug 01, 2023 at 12:49:34AM +0200 Peter Zijlstra wrote:
Quoted 14 lines · Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Quoted 5 lines

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

Quoted 8 lines

These appear fine to me, except:

Quoted 2 lines

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

What was the thinking here? This means nobody will be using this -- why would you want this default disabled?

That was just a hedge in case it caused issues. I'd probably have had to enable it in RHEL anyway. Using a feature was to make it inocuous when disabled. Would you prefer me to enable it or remove the sched_feat entirely? (or do you want to just switch that to true when you apply it?)

Thanks,
Phil

maintainer
#16 lore raw

Re: [PATCH v6 2/2] Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 27 lines · Phil Auld
Quoted 26 lines · On Tue, Aug 01, 2023 at 12:49:34AM +0200 Peter Zijlstra wrote:
Quoted 14 lines · Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Quoted 5 lines

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

These appear fine to me, except:

Quoted 2 lines

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

Quoted 8 lines
Quoted 3 lines

What was the thinking here? This means nobody will be using this -- why would you want this default disabled?

That was just a hedge in case it caused issues. I'd probably have had to enable it in RHEL anyway. Using a feature was to make it inocuous when disabled. Would you prefer me to enable it or remove the sched_feat entirely? (or do you want to just switch that to true when you apply it?)

I've edited it to default enabled -- we can pull the feature flag eventually I suppose.

Things didn't readily apply, so I've kicked at it a little. Should be in queue/sched/core for the robots to chew on.

author
#17 lore raw

Re: [PATCH v6 2/2] Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 37 lines · On Tue, Aug 01, 2023 at 05:37:31PM +0200 Peter Zijlstra wrote:
Quoted 27 lines · Phil Auld
Quoted 26 lines · On Tue, Aug 01, 2023 at 12:49:34AM +0200 Peter Zijlstra wrote:
Quoted 14 lines · Phil Auld

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Quoted 5 lines

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

These appear fine to me, except:

Quoted 2 lines

Add sched_feat HZ_BW (off by default) to control the tick_stop behavior.

Quoted 8 lines
Quoted 3 lines

What was the thinking here? This means nobody will be using this -- why would you want this default disabled?

That was just a hedge in case it caused issues. I'd probably have had to enable it in RHEL anyway. Using a feature was to make it inocuous when disabled. Would you prefer me to enable it or remove the sched_feat entirely? (or do you want to just switch that to true when you apply it?)

Quoted 6 lines

I've edited it to default enabled -- we can pull the feature flag eventually I suppose.

Things didn't readily apply, so I've kicked at it a little. Should be in queue/sched/core for the robots to chew on.

Or "choke on" as the case may be :)
I sent you something that will hopefully clean that up.

Thanks!

Cheers,
Phil

merge bot Applied
#18 lore raw

[tip: sched/core] sched/fair: Block nohz tick_stop when cfs bandwidth in use

In reply to Phil Auld
The following commit has been merged into the sched/core branch of tip:

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

Commit-ID:     88c56cfeaec4642aee8aac58b38d5708c6aae0d3
Gitweb:        https://git.kernel.org/tip/88c56cfeaec4642aee8aac58b38d5708c6aae0d3
Author:        Phil Auld <[email protected]>
AuthorDate:    Wed, 12 Jul 2023 09:33:57 -04:00
Committer:     Peter Zijlstra <[email protected]>
CommitterDate: Wed, 02 Aug 2023 16:19:26 +02:00

sched/fair: Block nohz tick_stop when cfs bandwidth in use

CFS bandwidth limits and NOHZ full don't play well together. Tasks can easily run well past their quotas before a remote tick does accounting. This leads to long, multi-period stalls before such tasks can run again. Currently, when presented with these conflicting requirements the scheduler is favoring nohz_full and letting the tick be stopped. However, nohz tick stopping is already best-effort, there are a number of conditions that can prevent it, whereas cfs runtime bandwidth is expected to be enforced.

Make the scheduler favor bandwidth over stopping the tick by setting TICK_DEP_BIT_SCHED when the only running task is a cfs task with runtime limit enabled. We use cfs_b->hierarchical_quota to determine if the task requires the tick.

Add check in pick_next_task_fair() as well since that is where we have a handle on the task that is actually going to be running.

Add check in sched_can_stop_tick() to cover some edge cases such as nr_running going from 2->1 and the 1 remains the running task.

---
 kernel/sched/core.c     | 26 ++++++++++++++++++++-
 kernel/sched/fair.c     | 52 +++++++++++++++++++++++++++++++++++++++-
 kernel/sched/features.h |  2 ++-
 kernel/sched/sched.h    |  2 ++-
 4 files changed, 81 insertions(+), 1 deletion(-)
kernel/sched/core.c +26 -0
index 3af25ca..614271a 100644
@@ -1194,6 +1194,20 @@ static void nohz_csd_func(void *info)
 #endif /* CONFIG_NO_HZ_COMMON */
 
 #ifdef CONFIG_NO_HZ_FULL
+static inline bool __need_bw_check(struct rq *rq, struct task_struct *p)
+{
+	if (rq->nr_running != 1)
+		return false;
+
+	if (p->sched_class != &fair_sched_class)
+		return false;
+
+	if (!task_on_rq_queued(p))
+		return false;
+
+	return true;
+}
+
 bool sched_can_stop_tick(struct rq *rq)
 {
 	int fifo_nr_running;
@@ -1229,6 +1243,18 @@ bool sched_can_stop_tick(struct rq *rq)
 	if (rq->nr_running > 1)
 		return false;
 
+	/*
+	 * If there is one task and it has CFS runtime bandwidth constraints
+	 * and it's on the cpu now we don't want to stop the tick.
+	 * This check prevents clearing the bit if a newly enqueued task here is
+	 * dequeued by migrating while the constrained task continues to run.
+	 * E.g. going from 2->1 without going through pick_next_task().
+	 */
+	if (sched_feat(HZ_BW) && __need_bw_check(rq, rq->curr)) {
+		if (cfs_task_bw_constrained(rq->curr))
+			return false;
+	}
+
 	return true;
 }
 #endif /* CONFIG_NO_HZ_FULL */
kernel/sched/fair.c +51 -1

72 lines not on this page · raw message

kernel/sched/features.h +2 -0
index ee7f23c..e10074c 100644
@@ -101,3 +101,5 @@ SCHED_FEAT(LATENCY_WARN, false)
 
 SCHED_FEAT(ALT_PERIOD, true)
 SCHED_FEAT(BASE_SLICE, true)
+
+SCHED_FEAT(HZ_BW, true)
kernel/sched/sched.h +2 -0
index 602de71..19af176 100644
@@ -459,6 +459,7 @@ extern void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b, struct cfs_bandwidth
 extern void __refill_cfs_bandwidth_runtime(struct cfs_bandwidth *cfs_b);
 extern void start_cfs_bandwidth(struct cfs_bandwidth *cfs_b);
 extern void unthrottle_cfs_rq(struct cfs_rq *cfs_rq);
+extern bool cfs_task_bw_constrained(struct task_struct *p);
 
 extern void init_tg_rt_entry(struct task_group *tg, struct rt_rq *rt_rq,
 		struct sched_rt_entity *rt_se, int cpu,
@@ -494,6 +495,7 @@ static inline void set_task_rq_fair(struct sched_entity *se,
 #else /* CONFIG_CGROUP_SCHED */
 
 struct cfs_bandwidth { };
+static inline bool cfs_task_bw_constrained(struct task_struct *p) { return false; }
 
 #endif	/* CONFIG_CGROUP_SCHED */
 
author
#19 lore raw
In reply to Phil Auld

Hi Peter,

Quoted 20 lines · On Wed, Jul 12, 2023 at 09:33:55AM -0400 Phil Auld wrote:

This is v6 of patch 2/2 which is adding code to prevent the tick from being stopped when the single running task has bandwidth limits. Discussions had led to the idea of adding a bit to task_struct to help make this decision.

There was some complexity with doing it in the task which is avoided by using something in the cfs_rq. Looking into that lead me to the hierarchical_quota field in the cfs_bandwith struct. We spend a good deal of effort updating (or trying to, see patch 1/2) that value for the whole task_group tree when a quota is set/changed.

This new version first fixes that value to be meaningful for cgroupv2 and then leverages it to make the decisions about blocking the tick_stop.

Phil Auld (2):
sched, cgroup: Restore meaning to hierarchical_quota Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 9 lines
 kernel/sched/core.c     | 23 ++++++++++++++---
 kernel/sched/fair.c     | 56 ++++++++++++++++++++++++++++++++++++++---
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  3 ++-
 4 files changed, 76 insertions(+), 8 deletions(-)
Signature · 1 line
2.31.1

Ping :)

Any thoughts on these now?

Cheers,
Phil

author
#20 lore raw
In reply to Phil Auld

Hi Peter,

Quoted 20 lines · On Wed, Jul 12, 2023 at 09:33:55AM -0400 Phil Auld wrote:

This is v6 of patch 2/2 which is adding code to prevent the tick from being stopped when the single running task has bandwidth limits. Discussions had led to the idea of adding a bit to task_struct to help make this decision.

There was some complexity with doing it in the task which is avoided by using something in the cfs_rq. Looking into that lead me to the hierarchical_quota field in the cfs_bandwith struct. We spend a good deal of effort updating (or trying to, see patch 1/2) that value for the whole task_group tree when a quota is set/changed.

This new version first fixes that value to be meaningful for cgroupv2 and then leverages it to make the decisions about blocking the tick_stop.

Phil Auld (2):
sched, cgroup: Restore meaning to hierarchical_quota Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 9 lines
 kernel/sched/core.c     | 23 ++++++++++++++---
 kernel/sched/fair.c     | 56 ++++++++++++++++++++++++++++++++++++++---
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  3 ++-
 4 files changed, 76 insertions(+), 8 deletions(-)
Signature · 1 line
2.31.1

Ping :)

Any thoughts on these now?

Cheers,
Phil

author
#21 lore raw

Sorry, ignore duplicate, please. Apparently I forgot email while on PTO last week :)

Quoted 44 lines · On Mon, Jul 31, 2023 at 12:38:37PM -0400 Phil Auld wrote:

Hi Peter,

Quoted 20 lines · On Wed, Jul 12, 2023 at 09:33:55AM -0400 Phil Auld wrote:

This is v6 of patch 2/2 which is adding code to prevent the tick from being stopped when the single running task has bandwidth limits. Discussions had led to the idea of adding a bit to task_struct to help make this decision.

There was some complexity with doing it in the task which is avoided by using something in the cfs_rq. Looking into that lead me to the hierarchical_quota field in the cfs_bandwith struct. We spend a good deal of effort updating (or trying to, see patch 1/2) that value for the whole task_group tree when a quota is set/changed.

This new version first fixes that value to be meaningful for cgroupv2 and then leverages it to make the decisions about blocking the tick_stop.

Phil Auld (2):
sched, cgroup: Restore meaning to hierarchical_quota Sched/fair: Block nohz tick_stop when cfs bandwidth in use

Quoted 9 lines
 kernel/sched/core.c     | 23 ++++++++++++++---
 kernel/sched/fair.c     | 56 ++++++++++++++++++++++++++++++++++++++---
 kernel/sched/features.h |  2 ++
 kernel/sched/sched.h    |  3 ++-
 4 files changed, 76 insertions(+), 8 deletions(-)
Signature · 1 line
2.31.1

Ping :)

Any thoughts on these now?

Cheers,
Phil