diff mbox series

[v5,1/4] sched/core: Fixed missing rq clock update before calling set_rq_offline()

Message ID 20230613082012.49615-2-jiahao.os@bytedance.com
State New
Headers
Series Fix some warnings about rq clock |

Commit Message

Hao Jia June 13, 2023, 8:20 a.m. UTC 
  This is triggered during cpu offline when CONFIG_CPU_FREQ is enabled
and cpufreq is set to powersave:
------------[ cut here ]------------
rq->clock_update_flags < RQCF_ACT_SKIP
WARNING: CPU: 24 PID: 754 at kernel/sched/sched.h:1496
enqueue_top_rt_rq+0x139/0x160
Call Trace:
 <TASK>
 ? intel_pstate_update_util+0x3b0/0x3b0
 rq_offline_rt+0x1b7/0x250
 set_rq_offline.part.120+0x28/0x60
 rq_attach_root+0xc4/0xd0
 cpu_attach_domain+0x3dc/0x7f0
 ? __schedule+0x65e/0x1310
 partition_sched_domains_locked+0x2a5/0x3c0
 rebuild_sched_domains_locked+0x477/0x830
 ? percpu_rwsem_wait+0x140/0x140
 rebuild_sched_domains+0x1b/0x30
 cpuset_hotplug_workfn+0x2ca/0xc90
 ? balance_push+0x56/0x120
 ? _raw_spin_unlock+0x15/0x30
 ? finish_task_switch+0x98/0x2f0
 ? __switch_to+0x116/0x410
 ? __schedule+0x65e/0x1310 ? internal_add_timer+0x42/0x60
 ? _raw_spin_unlock_irqrestore+0x23/0x40
 ? add_timer_on+0xd5/0x130
 process_one_work+0x1bc/0x3d0
 worker_thread+0x4c/0x380
 ? preempt_count_add+0x56/0xa0
 ? rescuer_thread+0x310/0x310
 kthread+0xe6/0x110
 ? kthread_complete_and_exit+0x20/0x20
 ret_from_fork+0x1f/0x30

More detailed key function call graph:
rq_offline_rt()
  __disable_runtime()
    sched_rt_rq_enqueue()
      enqueue_top_rt_rq()
        cpufreq_update_util() <-- depends on CONFIG_CPU_FREQ
          data->func(data, *rq_clock(rq)*, flags)
            intel_pstate_update_util() <-- powersave policy callback function

Before calling rq_offline_rt() we need to update the rq clock to avoid
using the old rq clock, So we add update_rq_clock() to set_rq_offline()
to update rq clock. And we use rq_lock_irqsave()/rq_unlock_irqrestore()
to replace raw_spin_rq_lock_irqsave()/raw_spin_rq_unlock_irqrestore()
in rq_attach_root() to ensure that rq->clock_update_flags are cleared
before updating the rq clock.

Steps to reproduce:
1. Enable CONFIG_SMP and CONFIG_CPU_FREQ when compiling the kernel
2. echo 1 > /sys/kernel/debug/clear_warn_once
3. cpupower -c all frequency-set -g powersave
4. Run some rt tasks e.g. Create 5*n rt (100% running) tasks (on a
   system with n CPUs)
5. Offline cpu one by one until the warninng is triggered

Suggested-by: Ben Segall <bsegall@google.com>
Signed-off-by: Hao Jia <jiahao.os@bytedance.com>
---
 kernel/sched/core.c     | 2 +-
 kernel/sched/topology.c | 6 +++---
 2 files changed, 4 insertions(+), 4 deletions(-)

Comments

Hao Jia June 13, 2023, 8:31 a.m. UTC | #1 
On 2023/6/13 Hao Jia wrote:
> This is triggered during cpu offline when CONFIG_CPU_FREQ is enabled
> and cpufreq is set to powersave:
> ------------[ cut here ]------------
> rq->clock_update_flags < RQCF_ACT_SKIP
> WARNING: CPU: 24 PID: 754 at kernel/sched/sched.h:1496
> enqueue_top_rt_rq+0x139/0x160
> Call Trace:
>   <TASK>
>   ? intel_pstate_update_util+0x3b0/0x3b0
>   rq_offline_rt+0x1b7/0x250
>   set_rq_offline.part.120+0x28/0x60
>   rq_attach_root+0xc4/0xd0
>   cpu_attach_domain+0x3dc/0x7f0
>   ? __schedule+0x65e/0x1310
>   partition_sched_domains_locked+0x2a5/0x3c0
>   rebuild_sched_domains_locked+0x477/0x830
>   ? percpu_rwsem_wait+0x140/0x140
>   rebuild_sched_domains+0x1b/0x30
>   cpuset_hotplug_workfn+0x2ca/0xc90
>   ? balance_push+0x56/0x120
>   ? _raw_spin_unlock+0x15/0x30
>   ? finish_task_switch+0x98/0x2f0
>   ? __switch_to+0x116/0x410
>   ? __schedule+0x65e/0x1310 ? internal_add_timer+0x42/0x60
>   ? _raw_spin_unlock_irqrestore+0x23/0x40
>   ? add_timer_on+0xd5/0x130
>   process_one_work+0x1bc/0x3d0
>   worker_thread+0x4c/0x380
>   ? preempt_count_add+0x56/0xa0
>   ? rescuer_thread+0x310/0x310
>   kthread+0xe6/0x110
>   ? kthread_complete_and_exit+0x20/0x20
>   ret_from_fork+0x1f/0x30
> 
> More detailed key function call graph:
> rq_offline_rt()
>    __disable_runtime()
>      sched_rt_rq_enqueue()
>        enqueue_top_rt_rq()
>          cpufreq_update_util() <-- depends on CONFIG_CPU_FREQ
>            data->func(data, *rq_clock(rq)*, flags)
>              intel_pstate_update_util() <-- powersave policy callback function
> 
> Before calling rq_offline_rt() we need to update the rq clock to avoid
> using the old rq clock, So we add update_rq_clock() to set_rq_offline()
> to update rq clock. And we use rq_lock_irqsave()/rq_unlock_irqrestore()
> to replace raw_spin_rq_lock_irqsave()/raw_spin_rq_unlock_irqrestore()
> in rq_attach_root() to ensure that rq->clock_update_flags are cleared
> before updating the rq clock.
> 
> Steps to reproduce:
> 1. Enable CONFIG_SMP and CONFIG_CPU_FREQ when compiling the kernel
> 2. echo 1 > /sys/kernel/debug/clear_warn_once
> 3. cpupower -c all frequency-set -g powersave
> 4. Run some rt tasks e.g. Create 5*n rt (100% running) tasks (on a
>     system with n CPUs)
> 5. Offline cpu one by one until the warninng is triggered
> 
> Suggested-by: Ben Segall <bsegall@google.com>
> Signed-off-by: Hao Jia <jiahao.os@bytedance.com>

Hi Vincent,

   Benjamin Segall suggested adding update_rq_clock() to set_rq_offline()
instead of calling update_rq_clock() before set_rq_offline() runs.
The code of patch1 has changed, so I removed your tag "Reviewed-by". 
Please review again.

Thanks,
Hao

> ---
>   kernel/sched/core.c     | 2 +-
>   kernel/sched/topology.c | 6 +++---
>   2 files changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index a68d1276bab0..a8be5415daba 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -9548,6 +9548,7 @@ void set_rq_offline(struct rq *rq)
>   	if (rq->online) {
>   		const struct sched_class *class;
>   
> +		update_rq_clock(rq);
>   		for_each_class(class) {
>   			if (class->rq_offline)
>   				class->rq_offline(rq);
> @@ -9689,7 +9690,6 @@ int sched_cpu_deactivate(unsigned int cpu)
>   
>   	rq_lock_irqsave(rq, &rf);
>   	if (rq->rd) {
> -		update_rq_clock(rq);
>   		BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
>   		set_rq_offline(rq);
>   	}
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> index 6682535e37c8..52976eadfee9 100644
> --- a/kernel/sched/topology.c
> +++ b/kernel/sched/topology.c
> @@ -487,9 +487,9 @@ static void free_rootdomain(struct rcu_head *rcu)
>   void rq_attach_root(struct rq *rq, struct root_domain *rd)
>   {
>   	struct root_domain *old_rd = NULL;
> -	unsigned long flags;
> +	struct rq_flags rf;
>   
> -	raw_spin_rq_lock_irqsave(rq, flags);
> +	rq_lock_irqsave(rq, &rf);
>   
>   	if (rq->rd) {
>   		old_rd = rq->rd;
> @@ -515,7 +515,7 @@ void rq_attach_root(struct rq *rq, struct root_domain *rd)
>   	if (cpumask_test_cpu(rq->cpu, cpu_active_mask))
>   		set_rq_online(rq);
>   
> -	raw_spin_rq_unlock_irqrestore(rq, flags);
> +	rq_unlock_irqrestore(rq, &rf);
>   
>   	if (old_rd)
>   		call_rcu(&old_rd->rcu, free_rootdomain);
Vincent Guittot June 13, 2023, 9:13 a.m. UTC | #2 
On Tue, 13 Jun 2023 at 10:31, Hao Jia <jiahao.os@bytedance.com> wrote:
>
>
>
> On 2023/6/13 Hao Jia wrote:
> > This is triggered during cpu offline when CONFIG_CPU_FREQ is enabled
> > and cpufreq is set to powersave:
> > ------------[ cut here ]------------
> > rq->clock_update_flags < RQCF_ACT_SKIP
> > WARNING: CPU: 24 PID: 754 at kernel/sched/sched.h:1496
> > enqueue_top_rt_rq+0x139/0x160
> > Call Trace:
> >   <TASK>
> >   ? intel_pstate_update_util+0x3b0/0x3b0
> >   rq_offline_rt+0x1b7/0x250
> >   set_rq_offline.part.120+0x28/0x60
> >   rq_attach_root+0xc4/0xd0
> >   cpu_attach_domain+0x3dc/0x7f0
> >   ? __schedule+0x65e/0x1310
> >   partition_sched_domains_locked+0x2a5/0x3c0
> >   rebuild_sched_domains_locked+0x477/0x830
> >   ? percpu_rwsem_wait+0x140/0x140
> >   rebuild_sched_domains+0x1b/0x30
> >   cpuset_hotplug_workfn+0x2ca/0xc90
> >   ? balance_push+0x56/0x120
> >   ? _raw_spin_unlock+0x15/0x30
> >   ? finish_task_switch+0x98/0x2f0
> >   ? __switch_to+0x116/0x410
> >   ? __schedule+0x65e/0x1310 ? internal_add_timer+0x42/0x60
> >   ? _raw_spin_unlock_irqrestore+0x23/0x40
> >   ? add_timer_on+0xd5/0x130
> >   process_one_work+0x1bc/0x3d0
> >   worker_thread+0x4c/0x380
> >   ? preempt_count_add+0x56/0xa0
> >   ? rescuer_thread+0x310/0x310
> >   kthread+0xe6/0x110
> >   ? kthread_complete_and_exit+0x20/0x20
> >   ret_from_fork+0x1f/0x30
> >
> > More detailed key function call graph:
> > rq_offline_rt()
> >    __disable_runtime()
> >      sched_rt_rq_enqueue()
> >        enqueue_top_rt_rq()
> >          cpufreq_update_util() <-- depends on CONFIG_CPU_FREQ
> >            data->func(data, *rq_clock(rq)*, flags)
> >              intel_pstate_update_util() <-- powersave policy callback function
> >
> > Before calling rq_offline_rt() we need to update the rq clock to avoid
> > using the old rq clock, So we add update_rq_clock() to set_rq_offline()
> > to update rq clock. And we use rq_lock_irqsave()/rq_unlock_irqrestore()
> > to replace raw_spin_rq_lock_irqsave()/raw_spin_rq_unlock_irqrestore()
> > in rq_attach_root() to ensure that rq->clock_update_flags are cleared
> > before updating the rq clock.
> >
> > Steps to reproduce:
> > 1. Enable CONFIG_SMP and CONFIG_CPU_FREQ when compiling the kernel
> > 2. echo 1 > /sys/kernel/debug/clear_warn_once
> > 3. cpupower -c all frequency-set -g powersave
> > 4. Run some rt tasks e.g. Create 5*n rt (100% running) tasks (on a
> >     system with n CPUs)
> > 5. Offline cpu one by one until the warninng is triggered
> >
> > Suggested-by: Ben Segall <bsegall@google.com>
> > Signed-off-by: Hao Jia <jiahao.os@bytedance.com>
>
> Hi Vincent,
>
>    Benjamin Segall suggested adding update_rq_clock() to set_rq_offline()
> instead of calling update_rq_clock() before set_rq_offline() runs.
> The code of patch1 has changed, so I removed your tag "Reviewed-by".
> Please review again.

v5 looks good to me

Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>

>
> Thanks,
> Hao
>
> > ---
> >   kernel/sched/core.c     | 2 +-
> >   kernel/sched/topology.c | 6 +++---
> >   2 files changed, 4 insertions(+), 4 deletions(-)
> >
> > diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> > index a68d1276bab0..a8be5415daba 100644
> > --- a/kernel/sched/core.c
> > +++ b/kernel/sched/core.c
> > @@ -9548,6 +9548,7 @@ void set_rq_offline(struct rq *rq)
> >       if (rq->online) {
> >               const struct sched_class *class;
> >
> > +             update_rq_clock(rq);
> >               for_each_class(class) {
> >                       if (class->rq_offline)
> >                               class->rq_offline(rq);
> > @@ -9689,7 +9690,6 @@ int sched_cpu_deactivate(unsigned int cpu)
> >
> >       rq_lock_irqsave(rq, &rf);
> >       if (rq->rd) {
> > -             update_rq_clock(rq);
> >               BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
> >               set_rq_offline(rq);
> >       }
> > diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> > index 6682535e37c8..52976eadfee9 100644
> > --- a/kernel/sched/topology.c
> > +++ b/kernel/sched/topology.c
> > @@ -487,9 +487,9 @@ static void free_rootdomain(struct rcu_head *rcu)
> >   void rq_attach_root(struct rq *rq, struct root_domain *rd)
> >   {
> >       struct root_domain *old_rd = NULL;
> > -     unsigned long flags;
> > +     struct rq_flags rf;
> >
> > -     raw_spin_rq_lock_irqsave(rq, flags);
> > +     rq_lock_irqsave(rq, &rf);
> >
> >       if (rq->rd) {
> >               old_rd = rq->rd;
> > @@ -515,7 +515,7 @@ void rq_attach_root(struct rq *rq, struct root_domain *rd)
> >       if (cpumask_test_cpu(rq->cpu, cpu_active_mask))
> >               set_rq_online(rq);
> >
> > -     raw_spin_rq_unlock_irqrestore(rq, flags);
> > +     rq_unlock_irqrestore(rq, &rf);
> >
> >       if (old_rd)
> >               call_rcu(&old_rd->rcu, free_rootdomain);
diff mbox series

Patch

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index a68d1276bab0..a8be5415daba 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -9548,6 +9548,7 @@  void set_rq_offline(struct rq *rq)
 	if (rq->online) {
 		const struct sched_class *class;
 
+		update_rq_clock(rq);
 		for_each_class(class) {
 			if (class->rq_offline)
 				class->rq_offline(rq);
@@ -9689,7 +9690,6 @@  int sched_cpu_deactivate(unsigned int cpu)
 
 	rq_lock_irqsave(rq, &rf);
 	if (rq->rd) {
-		update_rq_clock(rq);
 		BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span));
 		set_rq_offline(rq);
 	}
diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index 6682535e37c8..52976eadfee9 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -487,9 +487,9 @@  static void free_rootdomain(struct rcu_head *rcu)
 void rq_attach_root(struct rq *rq, struct root_domain *rd)
 {
 	struct root_domain *old_rd = NULL;
-	unsigned long flags;
+	struct rq_flags rf;
 
-	raw_spin_rq_lock_irqsave(rq, flags);
+	rq_lock_irqsave(rq, &rf);
 
 	if (rq->rd) {
 		old_rd = rq->rd;
@@ -515,7 +515,7 @@  void rq_attach_root(struct rq *rq, struct root_domain *rd)
 	if (cpumask_test_cpu(rq->cpu, cpu_active_mask))
 		set_rq_online(rq);
 
-	raw_spin_rq_unlock_irqrestore(rq, flags);
+	rq_unlock_irqrestore(rq, &rf);
 
 	if (old_rd)
 		call_rcu(&old_rd->rcu, free_rootdomain);