Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1405249 > unrolled thread

[PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up Morten Rasmussen <morten.rasmussen@arm.com> - 2016-05-23 13:00 +0200
    Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Mike Galbraith <mgalbraith@suse.de> - 2016-05-24 09:10 +0200
      Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Mike Galbraith <umgwanakikbuti@gmail.com> - 2016-05-24 09:20 +0200
    Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Yuyang Du <yuyang.du@intel.com> - 2016-05-24 10:10 +0200
      Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Morten Rasmussen <morten.rasmussen@arm.com> - 2016-05-24 10:20 +0200
    Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Wanpeng Li <kernellwp@gmail.com> - 2016-05-25 09:00 +0200
      Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Morten Rasmussen <morten.rasmussen@arm.com> - 2016-05-25 11:50 +0200
        Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Wanpeng Li <kernellwp@gmail.com> - 2016-05-25 12:30 +0200
          Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Morten Rasmussen <morten.rasmussen@arm.com> - 2016-05-25 13:00 +0200
            Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations  balance at wake-up Wanpeng Li <kernellwp@gmail.com> - 2016-05-25 13:20 +0200

#1405249 — [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
configurations SD_WAKE_AFFINE is only desirable if the waking task's
compute demand (utilization) is suitable for the cpu capacities
available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
balancing take over (find_idlest_{group, cpu}()).

The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
containing cpus with different capacities. This is enforced by a
previous patch based on the SD_ASYM_CPUCAPACITY flag.

Ideally, we shouldn't set 'want_affine' in the first place, but we don't
know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
traversing them.

cc: Ingo Molnar <mingo@redhat.com>
cc: Peter Zijlstra <peterz@infradead.org>

Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
---
 kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
 1 file changed, 27 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 564215d..ce44fa7 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
 unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
 #endif
 
+/*
+ * The margin used when comparing utilization with cpu capacity:
+ * util * 1024 < capacity * margin
+ */
+unsigned int capacity_margin = 1280; /* ~20% */
+
 static inline void update_load_add(struct load_weight *lw, unsigned long inc)
 {
 	lw->weight += inc;
@@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
 	return (util >= capacity) ? capacity : util;
 }
 
+static inline int task_util(struct task_struct *p)
+{
+	return p->se.avg.util_avg;
+}
+
+static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
+{
+	long delta;
+	long prev_cap = capacity_of(prev_cpu);
+
+	delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
+
+	/* prev_cpu is fairly close to max, no need to abort wake_affine */
+	if (delta < prev_cap >> 3)
+		return 0;
+
+	return prev_cap * 1024 < task_util(p) * capacity_margin;
+}
+
 /*
  * select_task_rq_fair: Select target runqueue for the waking task in domains
  * that have the 'sd_flag' flag set. In practice, this is SD_BALANCE_WAKE,
@@ -5316,7 +5341,8 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
 
 	if (sd_flag & SD_BALANCE_WAKE) {
 		record_wakee(p);
-		want_affine = !wake_wide(p) && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
+		want_affine = !wake_wide(p) && !wake_cap(p, cpu, prev_cpu)
+			      && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
 	}
 
 	rcu_read_lock();
-- 
1.9.1

[toc] | [next] | [standalone]

#1405897 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

On Mon, 2016-05-23 at 11:58 +0100, Morten Rasmussen wrote:
> Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
> SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
> configurations SD_WAKE_AFFINE is only desirable if the waking task's
> compute demand (utilization) is suitable for the cpu capacities
> available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
> balancing take over (find_idlest_{group, cpu}()).
> 
> The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
> containing cpus with different capacities. This is enforced by a
> previous patch based on the SD_ASYM_CPUCAPACITY flag.
> 
> Ideally, we shouldn't set 'want_affine' in the first place, but we don't
> know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
> traversing them.

This doesn't look like it's restricted to big/little setups, so could
overrule wake_wide() wanting to NAK a x-node pull.

> > 
> > cc: Ingo Molnar <mingo@redhat.com>
> > cc: Peter Zijlstra <peterz@infradead.org>
> > 
> > Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
> > ---
> >  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
> >  1 file changed, 27 insertions(+), 1 deletion(-)
> > 
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 564215d..ce44fa7 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
> >  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
> >  #endif
> >  
> > +/*
> > + * The margin used when comparing utilization with cpu capacity:
> > + * util * 1024 < capacity * margin
> > + */
> > +unsigned int capacity_margin = 1280; /* ~20% */
> > +
> >  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
> >  {
> >  > > 	> > lw->weight += inc;
> > @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
> >  > > 	> > return (util >= capacity) ? capacity : util;
> >  }
> >  
> > +static inline int task_util(struct task_struct *p)
> > +{
> > +> > 	> > return p->se.avg.util_avg;
> > +}
> > +
> > +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
> > +{
> > +> > 	> > long delta;
> > +> > 	> > long prev_cap = capacity_of(prev_cpu);
> > +
> > +> > 	> > delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
> > +
> > +> > 	> > /* prev_cpu is fairly close to max, no need to abort wake_affine */
> > +> > 	> > if (delta < prev_cap >> 3)
> > +> > 	> > 	> > return 0;
> > +
> > +> > 	> > return prev_cap * 1024 < task_util(p) * capacity_margin;
> > +}
> > +
> >  /*
> >   * select_task_rq_fair: Select target runqueue for the waking task in domains
> >   * that have the 'sd_flag' flag set. In practice, this is SD_BALANCE_WAKE,
> > @@ -5316,7 +5341,8 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
> >  
> >  > > 	> > if (sd_flag & SD_BALANCE_WAKE) {
> >  > > 	> > 	> > record_wakee(p);
> > -> > 	> > 	> > want_affine = !wake_wide(p) && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
> > +> > 	> > 	> > want_affine = !wake_wide(p) && !wake_cap(p, cpu, prev_cpu)
> > +> > 	> > 	> > 	> >       && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
> >  > > 	> > }
> >  
> >  > > 	> > rcu_read_lock();

[toc] | [prev] | [next] | [standalone]

#1405903 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

On Tue, 2016-05-24 at 09:03 +0200, Mike Galbraith wrote:

> This doesn't look like it's restricted to big/little setups, so could
> overrule wake_wide() wanting to NAK a x-node pull.

Bah, nevermind.

[toc] | [prev] | [next] | [standalone]

#1405923 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

On Mon, May 23, 2016 at 11:58:51AM +0100, Morten Rasmussen wrote:
> Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
> SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
> configurations SD_WAKE_AFFINE is only desirable if the waking task's
> compute demand (utilization) is suitable for the cpu capacities
> available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
> balancing take over (find_idlest_{group, cpu}()).
> 
> The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
> containing cpus with different capacities. This is enforced by a
> previous patch based on the SD_ASYM_CPUCAPACITY flag.
> 
> Ideally, we shouldn't set 'want_affine' in the first place, but we don't
> know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
> traversing them.
> 
> cc: Ingo Molnar <mingo@redhat.com>
> cc: Peter Zijlstra <peterz@infradead.org>
> 
> Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
> ---
>  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
>  1 file changed, 27 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 564215d..ce44fa7 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
>  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
>  #endif
>  
> +/*
> + * The margin used when comparing utilization with cpu capacity:
> + * util * 1024 < capacity * margin
> + */
> +unsigned int capacity_margin = 1280; /* ~20% */
> +
>  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
>  {
>  	lw->weight += inc;
> @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
>  	return (util >= capacity) ? capacity : util;
>  }
>  
> +static inline int task_util(struct task_struct *p)
> +{
> +	return p->se.avg.util_avg;
> +}
> +
> +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
> +{
> +	long delta;
> +	long prev_cap = capacity_of(prev_cpu);
> +
> +	delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
> +
> +	/* prev_cpu is fairly close to max, no need to abort wake_affine */
> +	if (delta < prev_cap >> 3)
> +		return 0;

delta can be negative? still return 0?

> +
> +	return prev_cap * 1024 < task_util(p) * capacity_margin;
> +}

[toc] | [prev] | [next] | [standalone]

#1405929 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

On Tue, May 24, 2016 at 08:04:24AM +0800, Yuyang Du wrote:
> On Mon, May 23, 2016 at 11:58:51AM +0100, Morten Rasmussen wrote:
> > Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
> > SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
> > configurations SD_WAKE_AFFINE is only desirable if the waking task's
> > compute demand (utilization) is suitable for the cpu capacities
> > available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
> > balancing take over (find_idlest_{group, cpu}()).
> > 
> > The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
> > containing cpus with different capacities. This is enforced by a
> > previous patch based on the SD_ASYM_CPUCAPACITY flag.
> > 
> > Ideally, we shouldn't set 'want_affine' in the first place, but we don't
> > know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
> > traversing them.
> > 
> > cc: Ingo Molnar <mingo@redhat.com>
> > cc: Peter Zijlstra <peterz@infradead.org>
> > 
> > Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
> > ---
> >  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
> >  1 file changed, 27 insertions(+), 1 deletion(-)
> > 
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 564215d..ce44fa7 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
> >  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
> >  #endif
> >  
> > +/*
> > + * The margin used when comparing utilization with cpu capacity:
> > + * util * 1024 < capacity * margin
> > + */
> > +unsigned int capacity_margin = 1280; /* ~20% */
> > +
> >  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
> >  {
> >  	lw->weight += inc;
> > @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
> >  	return (util >= capacity) ? capacity : util;
> >  }
> >  
> > +static inline int task_util(struct task_struct *p)
> > +{
> > +	return p->se.avg.util_avg;
> > +}
> > +
> > +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
> > +{
> > +	long delta;
> > +	long prev_cap = capacity_of(prev_cpu);
> > +
> > +	delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
> > +
> > +	/* prev_cpu is fairly close to max, no need to abort wake_affine */
> > +	if (delta < prev_cap >> 3)
> > +		return 0;
> 
> delta can be negative? still return 0?

I could add an abs() around delta.

Do you have a specific scenario in mind? Under normal circumstances, I
don't think it can be negative?

[toc] | [prev] | [next] | [standalone]

#1406685 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

2016-05-23 18:58 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
> Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
> SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
> configurations SD_WAKE_AFFINE is only desirable if the waking task's
> compute demand (utilization) is suitable for the cpu capacities
> available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
> balancing take over (find_idlest_{group, cpu}()).
>
> The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
> containing cpus with different capacities. This is enforced by a
> previous patch based on the SD_ASYM_CPUCAPACITY flag.
>
> Ideally, we shouldn't set 'want_affine' in the first place, but we don't
> know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
> traversing them.
>
> cc: Ingo Molnar <mingo@redhat.com>
> cc: Peter Zijlstra <peterz@infradead.org>
>
> Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
> ---
>  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
>  1 file changed, 27 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 564215d..ce44fa7 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
>  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
>  #endif
>
> +/*
> + * The margin used when comparing utilization with cpu capacity:
> + * util * 1024 < capacity * margin
> + */
> +unsigned int capacity_margin = 1280; /* ~20% */
> +
>  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
>  {
>         lw->weight += inc;
> @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
>         return (util >= capacity) ? capacity : util;
>  }
>
> +static inline int task_util(struct task_struct *p)
> +{
> +       return p->se.avg.util_avg;
> +}
> +
> +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
> +{
> +       long delta;
> +       long prev_cap = capacity_of(prev_cpu);
> +
> +       delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
> +
> +       /* prev_cpu is fairly close to max, no need to abort wake_affine */
> +       if (delta < prev_cap >> 3)
> +               return 0;
> +
> +       return prev_cap * 1024 < task_util(p) * capacity_margin;
> +}

If one task util_avg is SCHED_CAPACITY_SCALE and running on x86 box w/
SMT enabled, then each HT has capacity 589, wake_cap() will result in
always not wake affine, right?

Regards,
Wanpeng Li

> +
>  /*
>   * select_task_rq_fair: Select target runqueue for the waking task in domains
>   * that have the 'sd_flag' flag set. In practice, this is SD_BALANCE_WAKE,
> @@ -5316,7 +5341,8 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
>
>         if (sd_flag & SD_BALANCE_WAKE) {
>                 record_wakee(p);
> -               want_affine = !wake_wide(p) && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
> +               want_affine = !wake_wide(p) && !wake_cap(p, cpu, prev_cpu)
> +                             && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
>         }
>
>         rcu_read_lock();
> --
> 1.9.1
>



-- 
Regards,
Wanpeng Li

[toc] | [prev] | [next] | [standalone]

#1406780 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

On Wed, May 25, 2016 at 02:57:00PM +0800, Wanpeng Li wrote:
> 2016-05-23 18:58 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
> > Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
> > SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
> > configurations SD_WAKE_AFFINE is only desirable if the waking task's
> > compute demand (utilization) is suitable for the cpu capacities
> > available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
> > balancing take over (find_idlest_{group, cpu}()).
> >
> > The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
> > containing cpus with different capacities. This is enforced by a
> > previous patch based on the SD_ASYM_CPUCAPACITY flag.
> >
> > Ideally, we shouldn't set 'want_affine' in the first place, but we don't
> > know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
> > traversing them.
> >
> > cc: Ingo Molnar <mingo@redhat.com>
> > cc: Peter Zijlstra <peterz@infradead.org>
> >
> > Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
> > ---
> >  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
> >  1 file changed, 27 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> > index 564215d..ce44fa7 100644
> > --- a/kernel/sched/fair.c
> > +++ b/kernel/sched/fair.c
> > @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
> >  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
> >  #endif
> >
> > +/*
> > + * The margin used when comparing utilization with cpu capacity:
> > + * util * 1024 < capacity * margin
> > + */
> > +unsigned int capacity_margin = 1280; /* ~20% */
> > +
> >  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
> >  {
> >         lw->weight += inc;
> > @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
> >         return (util >= capacity) ? capacity : util;
> >  }
> >
> > +static inline int task_util(struct task_struct *p)
> > +{
> > +       return p->se.avg.util_avg;
> > +}
> > +
> > +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
> > +{
> > +       long delta;
> > +       long prev_cap = capacity_of(prev_cpu);
> > +
> > +       delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
> > +
> > +       /* prev_cpu is fairly close to max, no need to abort wake_affine */
> > +       if (delta < prev_cap >> 3)
> > +               return 0;
> > +
> > +       return prev_cap * 1024 < task_util(p) * capacity_margin;
> > +}
> 
> If one task util_avg is SCHED_CAPACITY_SCALE and running on x86 box w/
> SMT enabled, then each HT has capacity 589, wake_cap() will result in
> always not wake affine, right?

The idea is that SMT systems would bail out already at the previous
condition. We should have max_cpu_capacity == prev_cap == 589, delta
should then be zero and make the first condition true and make
wake_cap() always return 0 for any system with symmetric capacities
regardless of their actual capacity values.

Note that this isn't entirely true as I used capacity_of() for prev_cap,
if I change that to capacity_orig_of() it should be true.

By making the !wake_cap() condition always true for want_affine, we
should preserve existing behaviour for SMT/SMP. The only overhead is the
capacity delta computation and comparison, which should be cheap.

Does that make sense?

Btw, task util_avg == SCHED_CAPACITY_SCALE should only be possible
temporarily, it should decay to util_avg <=
capacity_orig_of(task_cpu(p)) over time. That doesn't affect your
question though as the second condition would still evaluate true if
util_avg == capacity_orig_of(task_cpu(p)), but as said above the first
condition should bail out before we get here.

Morten

> > +
> >  /*
> >   * select_task_rq_fair: Select target runqueue for the waking task in domains
> >   * that have the 'sd_flag' flag set. In practice, this is SD_BALANCE_WAKE,
> > @@ -5316,7 +5341,8 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f
> >
> >         if (sd_flag & SD_BALANCE_WAKE) {
> >                 record_wakee(p);
> > -               want_affine = !wake_wide(p) && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
> > +               want_affine = !wake_wide(p) && !wake_cap(p, cpu, prev_cpu)
> > +                             && cpumask_test_cpu(cpu, tsk_cpus_allowed(p));
> >         }
> >
> >         rcu_read_lock();
> > --
> > 1.9.1
> >

[toc] | [prev] | [next] | [standalone]

#1406814 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

2016-05-25 17:49 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
> On Wed, May 25, 2016 at 02:57:00PM +0800, Wanpeng Li wrote:
>> 2016-05-23 18:58 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
>> > Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
>> > SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
>> > configurations SD_WAKE_AFFINE is only desirable if the waking task's
>> > compute demand (utilization) is suitable for the cpu capacities
>> > available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
>> > balancing take over (find_idlest_{group, cpu}()).
>> >
>> > The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
>> > containing cpus with different capacities. This is enforced by a
>> > previous patch based on the SD_ASYM_CPUCAPACITY flag.
>> >
>> > Ideally, we shouldn't set 'want_affine' in the first place, but we don't
>> > know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
>> > traversing them.
>> >
>> > cc: Ingo Molnar <mingo@redhat.com>
>> > cc: Peter Zijlstra <peterz@infradead.org>
>> >
>> > Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
>> > ---
>> >  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
>> >  1 file changed, 27 insertions(+), 1 deletion(-)
>> >
>> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> > index 564215d..ce44fa7 100644
>> > --- a/kernel/sched/fair.c
>> > +++ b/kernel/sched/fair.c
>> > @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
>> >  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
>> >  #endif
>> >
>> > +/*
>> > + * The margin used when comparing utilization with cpu capacity:
>> > + * util * 1024 < capacity * margin
>> > + */
>> > +unsigned int capacity_margin = 1280; /* ~20% */
>> > +
>> >  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
>> >  {
>> >         lw->weight += inc;
>> > @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
>> >         return (util >= capacity) ? capacity : util;
>> >  }
>> >
>> > +static inline int task_util(struct task_struct *p)
>> > +{
>> > +       return p->se.avg.util_avg;
>> > +}
>> > +
>> > +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
>> > +{
>> > +       long delta;
>> > +       long prev_cap = capacity_of(prev_cpu);
>> > +
>> > +       delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
>> > +
>> > +       /* prev_cpu is fairly close to max, no need to abort wake_affine */
>> > +       if (delta < prev_cap >> 3)
>> > +               return 0;
>> > +
>> > +       return prev_cap * 1024 < task_util(p) * capacity_margin;
>> > +}
>>
>> If one task util_avg is SCHED_CAPACITY_SCALE and running on x86 box w/
>> SMT enabled, then each HT has capacity 589, wake_cap() will result in
>> always not wake affine, right?
>
> The idea is that SMT systems would bail out already at the previous
> condition. We should have max_cpu_capacity == prev_cap == 589, delta
> should then be zero and make the first condition true and make
> wake_cap() always return 0 for any system with symmetric capacities
> regardless of their actual capacity values.
>
> Note that this isn't entirely true as I used capacity_of() for prev_cap,
> if I change that to capacity_orig_of() it should be true.
>
> By making the !wake_cap() condition always true for want_affine, we
> should preserve existing behaviour for SMT/SMP. The only overhead is the
> capacity delta computation and comparison, which should be cheap.
>
> Does that make sense?

Fair enough, thanks for your explanation.

>
> Btw, task util_avg == SCHED_CAPACITY_SCALE should only be possible
> temporarily, it should decay to util_avg <=
> capacity_orig_of(task_cpu(p)) over time. That doesn't affect your

Sorry, I didn't find it will decay to capacity_orig in
__update_load_avg(), could you elaborate?

Regards,
Wanpeng Li

[toc] | [prev] | [next] | [standalone]

#1406838 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

On Wed, May 25, 2016 at 06:29:33PM +0800, Wanpeng Li wrote:
> 2016-05-25 17:49 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
> > On Wed, May 25, 2016 at 02:57:00PM +0800, Wanpeng Li wrote:
> >> 2016-05-23 18:58 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
> >> > Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
> >> > SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
> >> > configurations SD_WAKE_AFFINE is only desirable if the waking task's
> >> > compute demand (utilization) is suitable for the cpu capacities
> >> > available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
> >> > balancing take over (find_idlest_{group, cpu}()).
> >> >
> >> > The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
> >> > containing cpus with different capacities. This is enforced by a
> >> > previous patch based on the SD_ASYM_CPUCAPACITY flag.
> >> >
> >> > Ideally, we shouldn't set 'want_affine' in the first place, but we don't
> >> > know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
> >> > traversing them.
> >> >
> >> > cc: Ingo Molnar <mingo@redhat.com>
> >> > cc: Peter Zijlstra <peterz@infradead.org>
> >> >
> >> > Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
> >> > ---
> >> >  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
> >> >  1 file changed, 27 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> >> > index 564215d..ce44fa7 100644
> >> > --- a/kernel/sched/fair.c
> >> > +++ b/kernel/sched/fair.c
> >> > @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
> >> >  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
> >> >  #endif
> >> >
> >> > +/*
> >> > + * The margin used when comparing utilization with cpu capacity:
> >> > + * util * 1024 < capacity * margin
> >> > + */
> >> > +unsigned int capacity_margin = 1280; /* ~20% */
> >> > +
> >> >  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
> >> >  {
> >> >         lw->weight += inc;
> >> > @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
> >> >         return (util >= capacity) ? capacity : util;
> >> >  }
> >> >
> >> > +static inline int task_util(struct task_struct *p)
> >> > +{
> >> > +       return p->se.avg.util_avg;
> >> > +}
> >> > +
> >> > +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
> >> > +{
> >> > +       long delta;
> >> > +       long prev_cap = capacity_of(prev_cpu);
> >> > +
> >> > +       delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
> >> > +
> >> > +       /* prev_cpu is fairly close to max, no need to abort wake_affine */
> >> > +       if (delta < prev_cap >> 3)
> >> > +               return 0;
> >> > +
> >> > +       return prev_cap * 1024 < task_util(p) * capacity_margin;
> >> > +}
> >>
> >> If one task util_avg is SCHED_CAPACITY_SCALE and running on x86 box w/
> >> SMT enabled, then each HT has capacity 589, wake_cap() will result in
> >> always not wake affine, right?
> >
> > The idea is that SMT systems would bail out already at the previous
> > condition. We should have max_cpu_capacity == prev_cap == 589, delta
> > should then be zero and make the first condition true and make
> > wake_cap() always return 0 for any system with symmetric capacities
> > regardless of their actual capacity values.
> >
> > Note that this isn't entirely true as I used capacity_of() for prev_cap,
> > if I change that to capacity_orig_of() it should be true.
> >
> > By making the !wake_cap() condition always true for want_affine, we
> > should preserve existing behaviour for SMT/SMP. The only overhead is the
> > capacity delta computation and comparison, which should be cheap.
> >
> > Does that make sense?
> 
> Fair enough, thanks for your explanation.
> 
> >
> > Btw, task util_avg == SCHED_CAPACITY_SCALE should only be possible
> > temporarily, it should decay to util_avg <=
> > capacity_orig_of(task_cpu(p)) over time. That doesn't affect your
> 
> Sorry, I didn't find it will decay to capacity_orig in
> __update_load_avg(), could you elaborate?

I should have checked the code before writing that :-( I thought the
scaling by arch_scale_cpu_capacity() in __update_load_avg() would do
that, but it turns out that the default implementation of
arch_scale_cpu_capacity() doesn't do that when we pass a NULL pointer
for the sched_domain, it would have returned smt_gain/span_weight ==
capacity_orig_of(cpu) otherwise.

Sorry for the confusion, though I'm not sure if it is right to return
SCHED_CAPACITY_SCALE for SMT systems.

[toc] | [prev] | [next] | [standalone]

#1406846 — Re: [PATCH 09/16] sched/fair: Let asymmetric cpu configurations balance at wake-up

2016-05-25 18:54 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
> On Wed, May 25, 2016 at 06:29:33PM +0800, Wanpeng Li wrote:
>> 2016-05-25 17:49 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
>> > On Wed, May 25, 2016 at 02:57:00PM +0800, Wanpeng Li wrote:
>> >> 2016-05-23 18:58 GMT+08:00 Morten Rasmussen <morten.rasmussen@arm.com>:
>> >> > Currently, SD_WAKE_AFFINE always takes priority over wakeup balancing if
>> >> > SD_BALANCE_WAKE is set on the sched_domains. For asymmetric
>> >> > configurations SD_WAKE_AFFINE is only desirable if the waking task's
>> >> > compute demand (utilization) is suitable for the cpu capacities
>> >> > available within the SD_WAKE_AFFINE sched_domain. If not, let wakeup
>> >> > balancing take over (find_idlest_{group, cpu}()).
>> >> >
>> >> > The assumption is that SD_WAKE_AFFINE is never set for a sched_domain
>> >> > containing cpus with different capacities. This is enforced by a
>> >> > previous patch based on the SD_ASYM_CPUCAPACITY flag.
>> >> >
>> >> > Ideally, we shouldn't set 'want_affine' in the first place, but we don't
>> >> > know if SD_BALANCE_WAKE is enabled on the sched_domain(s) until we start
>> >> > traversing them.
>> >> >
>> >> > cc: Ingo Molnar <mingo@redhat.com>
>> >> > cc: Peter Zijlstra <peterz@infradead.org>
>> >> >
>> >> > Signed-off-by: Morten Rasmussen <morten.rasmussen@arm.com>
>> >> > ---
>> >> >  kernel/sched/fair.c | 28 +++++++++++++++++++++++++++-
>> >> >  1 file changed, 27 insertions(+), 1 deletion(-)
>> >> >
>> >> > diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
>> >> > index 564215d..ce44fa7 100644
>> >> > --- a/kernel/sched/fair.c
>> >> > +++ b/kernel/sched/fair.c
>> >> > @@ -114,6 +114,12 @@ unsigned int __read_mostly sysctl_sched_shares_window = 10000000UL;
>> >> >  unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL;
>> >> >  #endif
>> >> >
>> >> > +/*
>> >> > + * The margin used when comparing utilization with cpu capacity:
>> >> > + * util * 1024 < capacity * margin
>> >> > + */
>> >> > +unsigned int capacity_margin = 1280; /* ~20% */
>> >> > +
>> >> >  static inline void update_load_add(struct load_weight *lw, unsigned long inc)
>> >> >  {
>> >> >         lw->weight += inc;
>> >> > @@ -5293,6 +5299,25 @@ static int cpu_util(int cpu)
>> >> >         return (util >= capacity) ? capacity : util;
>> >> >  }
>> >> >
>> >> > +static inline int task_util(struct task_struct *p)
>> >> > +{
>> >> > +       return p->se.avg.util_avg;
>> >> > +}
>> >> > +
>> >> > +static int wake_cap(struct task_struct *p, int cpu, int prev_cpu)
>> >> > +{
>> >> > +       long delta;
>> >> > +       long prev_cap = capacity_of(prev_cpu);
>> >> > +
>> >> > +       delta = cpu_rq(cpu)->rd->max_cpu_capacity - prev_cap;
>> >> > +
>> >> > +       /* prev_cpu is fairly close to max, no need to abort wake_affine */
>> >> > +       if (delta < prev_cap >> 3)
>> >> > +               return 0;
>> >> > +
>> >> > +       return prev_cap * 1024 < task_util(p) * capacity_margin;
>> >> > +}
>> >>
>> >> If one task util_avg is SCHED_CAPACITY_SCALE and running on x86 box w/
>> >> SMT enabled, then each HT has capacity 589, wake_cap() will result in
>> >> always not wake affine, right?
>> >
>> > The idea is that SMT systems would bail out already at the previous
>> > condition. We should have max_cpu_capacity == prev_cap == 589, delta
>> > should then be zero and make the first condition true and make
>> > wake_cap() always return 0 for any system with symmetric capacities
>> > regardless of their actual capacity values.
>> >
>> > Note that this isn't entirely true as I used capacity_of() for prev_cap,
>> > if I change that to capacity_orig_of() it should be true.
>> >
>> > By making the !wake_cap() condition always true for want_affine, we
>> > should preserve existing behaviour for SMT/SMP. The only overhead is the
>> > capacity delta computation and comparison, which should be cheap.
>> >
>> > Does that make sense?
>>
>> Fair enough, thanks for your explanation.
>>
>> >
>> > Btw, task util_avg == SCHED_CAPACITY_SCALE should only be possible
>> > temporarily, it should decay to util_avg <=
>> > capacity_orig_of(task_cpu(p)) over time. That doesn't affect your
>>
>> Sorry, I didn't find it will decay to capacity_orig in
>> __update_load_avg(), could you elaborate?
>
> I should have checked the code before writing that :-( I thought the
> scaling by arch_scale_cpu_capacity() in __update_load_avg() would do
> that, but it turns out that the default implementation of
> arch_scale_cpu_capacity() doesn't do that when we pass a NULL pointer
> for the sched_domain, it would have returned smt_gain/span_weight ==
> capacity_orig_of(cpu) otherwise.

Thanks for the explanation. :)

Regards,
Wanpeng Li

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web