[PATCH] kernel/sched: Update schedstats when migrating threads

[PATCH] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

The kernel manages per-task scheduler statistics or schedstats. Such
counters should be reinitialized when the thread is migrated to a
different core rq, except for the values recording number of migrations.

Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
---
 kernel/sched/core.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index fe53e510e711..d64c2a290176 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
 int migrate_task_to(struct task_struct *p, int target_cpu)
 {
 	struct migration_arg arg = { p, target_cpu };
+	uint64_t forced_migrations, migrations_cold;
 	int curr_cpu = task_cpu(p);
 
 	if (curr_cpu == target_cpu)
@@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
 	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
 		return -EINVAL;
 
-	/* TODO: This is not properly updating schedstats */
+	if (schedstat_enabled()) {
+		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
+		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
+		memset(&p->stats, 0, sizeof(p->stats));
+		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
+		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
+		schedstat_inc(p->stats.nr_migrations_cold);
+	}
 
 	trace_sched_move_numa(p, curr_cpu, target_cpu);
 	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
-- 
2.27.0

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

Hi,

On 1/26/2022 9:22 AM, Carlos Bilbao wrote:
> The kernel manages per-task scheduler statistics or schedstats. Such
> counters should be reinitialized when the thread is migrated to a
> different core rq, except for the values recording number of migrations.
> 
> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
> ---
>  kernel/sched/core.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index fe53e510e711..d64c2a290176 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
>  int migrate_task_to(struct task_struct *p, int target_cpu)
>  {
>  	struct migration_arg arg = { p, target_cpu };
> +	uint64_t forced_migrations, migrations_cold;
>  	int curr_cpu = task_cpu(p);
>  
>  	if (curr_cpu == target_cpu)
> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>  		return -EINVAL;
>  
> -	/* TODO: This is not properly updating schedstats */
> +	if (schedstat_enabled()) {
> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
> +		memset(&p->stats, 0, sizeof(p->stats));
> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
> +		schedstat_inc(p->stats.nr_migrations_cold);
> +	}
>  
>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);


I would love to hear some thoughts on this. 

Thanks,
Carlos

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Steven Rostedt]

On Tue, 22 Feb 2022 10:32:32 -0600
Carlos Bilbao <carlos.bilbao@amd.com> wrote:

> > @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
> >  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
> >  		return -EINVAL;
> >  
> > -	/* TODO: This is not properly updating schedstats */
> > +	if (schedstat_enabled()) {
> > +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
> > +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
> > +		memset(&p->stats, 0, sizeof(p->stats));
> > +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
> > +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
> > +		schedstat_inc(p->stats.nr_migrations_cold);
> > +	}
> >  
> >  	trace_sched_move_numa(p, curr_cpu, target_cpu);
> >  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);  
> 
> 
> I would love to hear some thoughts on this. 

I have no issues with this.

Peter?

-- Steve

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Peter Zijlstra]

On Wed, Jan 26, 2022 at 09:22:23AM -0600, Carlos Bilbao wrote:
> The kernel manages per-task scheduler statistics or schedstats. Such
> counters should be reinitialized when the thread is migrated to a
> different core rq, except for the values recording number of migrations.

I'm confused, why should we reset schedstats on migrate? I'm thinking
this breaks per-task, since tasks tend to bounce around quite a lot.

> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
> ---
>  kernel/sched/core.c | 10 +++++++++-
>  1 file changed, 9 insertions(+), 1 deletion(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index fe53e510e711..d64c2a290176 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
>  int migrate_task_to(struct task_struct *p, int target_cpu)
>  {
>  	struct migration_arg arg = { p, target_cpu };
> +	uint64_t forced_migrations, migrations_cold;
>  	int curr_cpu = task_cpu(p);
>  
>  	if (curr_cpu == target_cpu)
> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>  		return -EINVAL;
>  
> -	/* TODO: This is not properly updating schedstats */
> +	if (schedstat_enabled()) {
> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
> +		memset(&p->stats, 0, sizeof(p->stats));
> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
> +		schedstat_inc(p->stats.nr_migrations_cold);
> +	}
>  
>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
> -- 
> 2.27.0
> 

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

On 2/23/2022 4:19 AM, Peter Zijlstra wrote:
> On Wed, Jan 26, 2022 at 09:22:23AM -0600, Carlos Bilbao wrote:
>> The kernel manages per-task scheduler statistics or schedstats. Such
>> counters should be reinitialized when the thread is migrated to a
>> different core rq, except for the values recording number of migrations.
> 
> I'm confused, why should we reset schedstats on migrate? I'm thinking
> this breaks per-task, since tasks tend to bounce around quite a lot.
> 

Thanks for your comments, Peter. 

Looking at the documentation of schedstats I see that most values are 
actually linked to the particular CPU: time spent on the cpu, timeslices 
run on this cpu, number of times _something_ was called when the cpu was 
idle, and so forth. Those values lose their meaning after migration and we 
should reinitialize their counters. However, reviewing sched_statistics I 
identify two fields that we should definitely keep increasing even after 
migration (nr_migrations_cold, nr_forced_migrations).

So this patch will have to be upgraded if there's some other value(s) in
schedstats that we do not want to reinitialize either.

>> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
>> ---
>>  kernel/sched/core.c | 10 +++++++++-
>>  1 file changed, 9 insertions(+), 1 deletion(-)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index fe53e510e711..d64c2a290176 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
>>  int migrate_task_to(struct task_struct *p, int target_cpu)
>>  {
>>  	struct migration_arg arg = { p, target_cpu };
>> +	uint64_t forced_migrations, migrations_cold;
>>  	int curr_cpu = task_cpu(p);
>>  
>>  	if (curr_cpu == target_cpu)
>> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>>  		return -EINVAL;
>>  
>> -	/* TODO: This is not properly updating schedstats */
>> +	if (schedstat_enabled()) {
>> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
>> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
>> +		memset(&p->stats, 0, sizeof(p->stats));
>> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
>> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
>> +		schedstat_inc(p->stats.nr_migrations_cold);
>> +	}
>>  
>>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
>> -- 
>> 2.27.0
>>

Thanks,
Carlos

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Phil Auld]

On Wed, Feb 23, 2022 at 09:14:45AM -0600 Carlos Bilbao wrote:
> On 2/23/2022 4:19 AM, Peter Zijlstra wrote:
> > On Wed, Jan 26, 2022 at 09:22:23AM -0600, Carlos Bilbao wrote:
> >> The kernel manages per-task scheduler statistics or schedstats. Such
> >> counters should be reinitialized when the thread is migrated to a
> >> different core rq, except for the values recording number of migrations.
> > 
> > I'm confused, why should we reset schedstats on migrate? I'm thinking
> > this breaks per-task, since tasks tend to bounce around quite a lot.
> > 
> 
> Thanks for your comments, Peter. 
> 
> Looking at the documentation of schedstats I see that most values are 
> actually linked to the particular CPU: time spent on the cpu, timeslices 
> run on this cpu, number of times _something_ was called when the cpu was 
> idle, and so forth. Those values lose their meaning after migration and we 
> should reinitialize their counters. However, reviewing sched_statistics I 
> identify two fields that we should definitely keep increasing even after 
> migration (nr_migrations_cold, nr_forced_migrations).
>

The documentation is a little off. I think it should say "any cpu" instead
of "this cpu".  If you reset these per task counters (time on cpu, number
of timeslices etc) on every migration then they become meaningless (and
useless).


Cheers,
Phil

> So this patch will have to be upgraded if there's some other value(s) in
> schedstats that we do not want to reinitialize either.
> 
> >> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
> >> ---
> >>  kernel/sched/core.c | 10 +++++++++-
> >>  1 file changed, 9 insertions(+), 1 deletion(-)
> >>
> >> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >> index fe53e510e711..d64c2a290176 100644
> >> --- a/kernel/sched/core.c
> >> +++ b/kernel/sched/core.c
> >> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
> >>  int migrate_task_to(struct task_struct *p, int target_cpu)
> >>  {
> >>  	struct migration_arg arg = { p, target_cpu };
> >> +	uint64_t forced_migrations, migrations_cold;
> >>  	int curr_cpu = task_cpu(p);
> >>  
> >>  	if (curr_cpu == target_cpu)
> >> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
> >>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
> >>  		return -EINVAL;
> >>  
> >> -	/* TODO: This is not properly updating schedstats */
> >> +	if (schedstat_enabled()) {
> >> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
> >> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
> >> +		memset(&p->stats, 0, sizeof(p->stats));
> >> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
> >> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
> >> +		schedstat_inc(p->stats.nr_migrations_cold);
> >> +	}
> >>  
> >>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
> >>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
> >> -- 
> >> 2.27.0
> >>
> 
> Thanks,
> Carlos
> 

-- 

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

On 2/23/2022 9:28 AM, Phil Auld wrote:
> On Wed, Feb 23, 2022 at 09:14:45AM -0600 Carlos Bilbao wrote:
>> On 2/23/2022 4:19 AM, Peter Zijlstra wrote:
>>> On Wed, Jan 26, 2022 at 09:22:23AM -0600, Carlos Bilbao wrote:
>>>> The kernel manages per-task scheduler statistics or schedstats. Such
>>>> counters should be reinitialized when the thread is migrated to a
>>>> different core rq, except for the values recording number of migrations.
>>>
>>> I'm confused, why should we reset schedstats on migrate? I'm thinking
>>> this breaks per-task, since tasks tend to bounce around quite a lot.
>>>
>>
>> Thanks for your comments, Peter. 
>>
>> Looking at the documentation of schedstats I see that most values are 
>> actually linked to the particular CPU: time spent on the cpu, timeslices 
>> run on this cpu, number of times _something_ was called when the cpu was 
>> idle, and so forth. Those values lose their meaning after migration and we 
>> should reinitialize their counters. However, reviewing sched_statistics I 
>> identify two fields that we should definitely keep increasing even after 
>> migration (nr_migrations_cold, nr_forced_migrations).
>>
> 
> The documentation is a little off. I think it should say "any cpu" instead
> of "this cpu".  If you reset these per task counters (time on cpu, number
> of timeslices etc) on every migration then they become meaningless (and
> useless).
> 
> 
> Cheers,
> Phil
> 

Well that clarifies it! Then, let me ask the opposite question... What
fields of schedstats should we clear when migrating? If there isn't any,
I will just increase the number of migrations.

>> So this patch will have to be upgraded if there's some other value(s) in
>> schedstats that we do not want to reinitialize either.
>>
>>>> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
>>>> ---
>>>>  kernel/sched/core.c | 10 +++++++++-
>>>>  1 file changed, 9 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>>> index fe53e510e711..d64c2a290176 100644
>>>> --- a/kernel/sched/core.c
>>>> +++ b/kernel/sched/core.c
>>>> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
>>>>  int migrate_task_to(struct task_struct *p, int target_cpu)
>>>>  {
>>>>  	struct migration_arg arg = { p, target_cpu };
>>>> +	uint64_t forced_migrations, migrations_cold;
>>>>  	int curr_cpu = task_cpu(p);
>>>>  
>>>>  	if (curr_cpu == target_cpu)
>>>> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>>>>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>>>>  		return -EINVAL;
>>>>  
>>>> -	/* TODO: This is not properly updating schedstats */
>>>> +	if (schedstat_enabled()) {
>>>> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
>>>> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
>>>> +		memset(&p->stats, 0, sizeof(p->stats));
>>>> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
>>>> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
>>>> +		schedstat_inc(p->stats.nr_migrations_cold);
>>>> +	}
>>>>  
>>>>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>>>>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
>>>> -- 
>>>> 2.27.0
>>>>
>>
>> Thanks,
>> Carlos
>>
> 

Thanks,
Carlos

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Phil Auld]

On Wed, Feb 23, 2022 at 09:33:59AM -0600 Carlos Bilbao wrote:
> On 2/23/2022 9:28 AM, Phil Auld wrote:
> > On Wed, Feb 23, 2022 at 09:14:45AM -0600 Carlos Bilbao wrote:
> >> On 2/23/2022 4:19 AM, Peter Zijlstra wrote:
> >>> On Wed, Jan 26, 2022 at 09:22:23AM -0600, Carlos Bilbao wrote:
> >>>> The kernel manages per-task scheduler statistics or schedstats. Such
> >>>> counters should be reinitialized when the thread is migrated to a
> >>>> different core rq, except for the values recording number of migrations.
> >>>
> >>> I'm confused, why should we reset schedstats on migrate? I'm thinking
> >>> this breaks per-task, since tasks tend to bounce around quite a lot.
> >>>
> >>
> >> Thanks for your comments, Peter. 
> >>
> >> Looking at the documentation of schedstats I see that most values are 
> >> actually linked to the particular CPU: time spent on the cpu, timeslices 
> >> run on this cpu, number of times _something_ was called when the cpu was 
> >> idle, and so forth. Those values lose their meaning after migration and we 
> >> should reinitialize their counters. However, reviewing sched_statistics I 
> >> identify two fields that we should definitely keep increasing even after 
> >> migration (nr_migrations_cold, nr_forced_migrations).
> >>
> > 
> > The documentation is a little off. I think it should say "any cpu" instead
> > of "this cpu".  If you reset these per task counters (time on cpu, number
> > of timeslices etc) on every migration then they become meaningless (and
> > useless).
> > 
> > 
> > Cheers,
> > Phil
> > 
> 
> Well that clarifies it! Then, let me ask the opposite question... What
> fields of schedstats should we clear when migrating? If there isn't any,
> I will just increase the number of migrations.
>

I don't think any should be cleared on migration. They're per task and
should be monotically increasing. If they ever reset it becomes hard to
know what they mean when you read them.


Cheers,
Phil


> >> So this patch will have to be upgraded if there's some other value(s) in
> >> schedstats that we do not want to reinitialize either.
> >>
> >>>> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
> >>>> ---
> >>>>  kernel/sched/core.c | 10 +++++++++-
> >>>>  1 file changed, 9 insertions(+), 1 deletion(-)
> >>>>
> >>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> >>>> index fe53e510e711..d64c2a290176 100644
> >>>> --- a/kernel/sched/core.c
> >>>> +++ b/kernel/sched/core.c
> >>>> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
> >>>>  int migrate_task_to(struct task_struct *p, int target_cpu)
> >>>>  {
> >>>>  	struct migration_arg arg = { p, target_cpu };
> >>>> +	uint64_t forced_migrations, migrations_cold;
> >>>>  	int curr_cpu = task_cpu(p);
> >>>>  
> >>>>  	if (curr_cpu == target_cpu)
> >>>> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
> >>>>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
> >>>>  		return -EINVAL;
> >>>>  
> >>>> -	/* TODO: This is not properly updating schedstats */
> >>>> +	if (schedstat_enabled()) {
> >>>> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
> >>>> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
> >>>> +		memset(&p->stats, 0, sizeof(p->stats));
> >>>> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
> >>>> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
> >>>> +		schedstat_inc(p->stats.nr_migrations_cold);
> >>>> +	}
> >>>>  
> >>>>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
> >>>>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
> >>>> -- 
> >>>> 2.27.0
> >>>>
> >>
> >> Thanks,
> >> Carlos
> >>
> > 
> 
> Thanks,
> Carlos
> 

-- 

Re: [PATCH] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

On 2/23/2022 9:47 AM, Phil Auld wrote:
> On Wed, Feb 23, 2022 at 09:33:59AM -0600 Carlos Bilbao wrote:
>> On 2/23/2022 9:28 AM, Phil Auld wrote:
>>> On Wed, Feb 23, 2022 at 09:14:45AM -0600 Carlos Bilbao wrote:
>>>> On 2/23/2022 4:19 AM, Peter Zijlstra wrote:
>>>>> On Wed, Jan 26, 2022 at 09:22:23AM -0600, Carlos Bilbao wrote:
>>>>>> The kernel manages per-task scheduler statistics or schedstats. Such
>>>>>> counters should be reinitialized when the thread is migrated to a
>>>>>> different core rq, except for the values recording number of migrations.
>>>>>
>>>>> I'm confused, why should we reset schedstats on migrate? I'm thinking
>>>>> this breaks per-task, since tasks tend to bounce around quite a lot.
>>>>>
>>>>
>>>> Thanks for your comments, Peter. 
>>>>
>>>> Looking at the documentation of schedstats I see that most values are 
>>>> actually linked to the particular CPU: time spent on the cpu, timeslices 
>>>> run on this cpu, number of times _something_ was called when the cpu was 
>>>> idle, and so forth. Those values lose their meaning after migration and we 
>>>> should reinitialize their counters. However, reviewing sched_statistics I 
>>>> identify two fields that we should definitely keep increasing even after 
>>>> migration (nr_migrations_cold, nr_forced_migrations).
>>>>
>>>
>>> The documentation is a little off. I think it should say "any cpu" instead
>>> of "this cpu".  If you reset these per task counters (time on cpu, number
>>> of timeslices etc) on every migration then they become meaningless (and
>>> useless).
>>>
>>>
>>> Cheers,
>>> Phil
>>>
>>
>> Well that clarifies it! Then, let me ask the opposite question... What
>> fields of schedstats should we clear when migrating? If there isn't any,
>> I will just increase the number of migrations.
>>
> 
> I don't think any should be cleared on migration. They're per task and
> should be monotically increasing. If they ever reset it becomes hard to
> know what they mean when you read them.
> 
> 
> Cheers,
> Phil
> 
> 

Agreed, I'll send a PATCH v2 only increasing the migration counter.  

>>>> So this patch will have to be upgraded if there's some other value(s) in
>>>> schedstats that we do not want to reinitialize either.
>>>>
>>>>>> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
>>>>>> ---
>>>>>>  kernel/sched/core.c | 10 +++++++++-
>>>>>>  1 file changed, 9 insertions(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>>>>>> index fe53e510e711..d64c2a290176 100644
>>>>>> --- a/kernel/sched/core.c
>>>>>> +++ b/kernel/sched/core.c
>>>>>> @@ -8757,6 +8757,7 @@ bool sched_smp_initialized __read_mostly;
>>>>>>  int migrate_task_to(struct task_struct *p, int target_cpu)
>>>>>>  {
>>>>>>  	struct migration_arg arg = { p, target_cpu };
>>>>>> +	uint64_t forced_migrations, migrations_cold;
>>>>>>  	int curr_cpu = task_cpu(p);
>>>>>>  
>>>>>>  	if (curr_cpu == target_cpu)
>>>>>> @@ -8765,7 +8766,14 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>>>>>>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>>>>>>  		return -EINVAL;
>>>>>>  
>>>>>> -	/* TODO: This is not properly updating schedstats */
>>>>>> +	if (schedstat_enabled()) {
>>>>>> +		forced_migrations = schedstat_val(p->stats.nr_forced_migrations);
>>>>>> +		migrations_cold = schedstat_val(p->stats.nr_migrations_cold);
>>>>>> +		memset(&p->stats, 0, sizeof(p->stats));
>>>>>> +		schedstat_set(p->stats.nr_forced_migrations, forced_migrations);
>>>>>> +		schedstat_set(p->stats.nr_migrations_cold, migrations_cold);
>>>>>> +		schedstat_inc(p->stats.nr_migrations_cold);
>>>>>> +	}
>>>>>>  
>>>>>>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>>>>>>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
>>>>>> -- 
>>>>>> 2.27.0
>>>>>>
>>>>
>>>> Thanks,
>>>> Carlos
>>>>
>>>
>>
>> Thanks,
>> Carlos
>>
> 

[PATCH v2] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

The kernel manages per-task scheduler statistics or schedstats. Update
function migrate_task_to() to increase the counter for migrations.

Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
---
Changelog:
	v2: Update commit message, don't reinitialize sched fields.
---
 kernel/sched/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index fcf0c180617c..1360e501c737 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -8751,7 +8751,7 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
 	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
 		return -EINVAL;
 
-	/* TODO: This is not properly updating schedstats */
+	schedstat_inc(p->stats.nr_migrations_cold);
 
 	trace_sched_move_numa(p, curr_cpu, target_cpu);
 	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
-- 
2.27.0

Re: [PATCH v2] kernel/sched: Update schedstats when migrating threads

[Phil Auld]

On Wed, Feb 23, 2022 at 10:13:50AM -0600 Carlos Bilbao wrote:
> The kernel manages per-task scheduler statistics or schedstats. Update
> function migrate_task_to() to increase the counter for migrations.
> 
> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
> ---
> Changelog:
> 	v2: Update commit message, don't reinitialize sched fields.
> ---
>  kernel/sched/core.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
> index fcf0c180617c..1360e501c737 100644
> --- a/kernel/sched/core.c
> +++ b/kernel/sched/core.c
> @@ -8751,7 +8751,7 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>  		return -EINVAL;
>  
> -	/* TODO: This is not properly updating schedstats */
> +	schedstat_inc(p->stats.nr_migrations_cold);
>

I was going to give a reviewed by since I was looking at this,
but I can't convince myself that nr_migrations_cold is right.
This looks more like a "hot" migration (using stop_cpu to force
it). But nr_migrations_cold is not incremented anywhere else so
maybe it's a terminology thing.

Can you tell me why this is the right counter?


Thanks,
Phil


>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
> -- 
> 2.27.0
> 

-- 

Re: [PATCH v2] kernel/sched: Update schedstats when migrating threads

[Carlos Bilbao]

On 2/24/2022 8:03 AM, Phil Auld wrote:
> On Wed, Feb 23, 2022 at 10:13:50AM -0600 Carlos Bilbao wrote:
>> The kernel manages per-task scheduler statistics or schedstats. Update
>> function migrate_task_to() to increase the counter for migrations.
>>
>> Signed-off-by: Carlos Bilbao <carlos.bilbao@amd.com>
>> ---
>> Changelog:
>> 	v2: Update commit message, don't reinitialize sched fields.
>> ---
>>  kernel/sched/core.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index fcf0c180617c..1360e501c737 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -8751,7 +8751,7 @@ int migrate_task_to(struct task_struct *p, int target_cpu)
>>  	if (!cpumask_test_cpu(target_cpu, p->cpus_ptr))
>>  		return -EINVAL;
>>  
>> -	/* TODO: This is not properly updating schedstats */
>> +	schedstat_inc(p->stats.nr_migrations_cold);
>>
> 
> I was going to give a reviewed by since I was looking at this,
> but I can't convince myself that nr_migrations_cold is right.
> This looks more like a "hot" migration (using stop_cpu to force
> it). But nr_migrations_cold is not incremented anywhere else so
> maybe it's a terminology thing.
> 
> Can you tell me why this is the right counter?
> 
>
The alternative to nr_migrations_cold is nr_forced_migrations, and since
we call this function in the context of NUMA balancing it didn't seem like
an appropriate fit at first glance. But documentation in this is scarce and 
I'm not very confident with the terminology either. Maybe Ingo Molar -the
original author of the structure- can clarify the specifics of these two 
counters so we can be sure about this? Maybe we need a new counter?

> Thanks,
> Phil
> 
> 
>>  	trace_sched_move_numa(p, curr_cpu, target_cpu);
>>  	return stop_one_cpu(curr_cpu, migration_cpu_stop, &arg);
>> -- 
>> 2.27.0
>>
> 

Thanks,
Carlos