[PATCH v6 0/2] Make find_later_rq() choose a closer cpu in topology

[PATCH v6 0/2] Make find_later_rq() choose a closer cpu in topology

[Byungchul Park]

When cpudl_find() returns any among free_cpus, the cpu might not be
closer than others, considering sched domain. For example:

   this_cpu: 15
   free_cpus: 0, 1,..., 14 (== later_mask)
   best_cpu: 0

   topology:

   0 --+
       +--+
   1 --+  |
          +-- ... --+
   2 --+  |         |
       +--+         |
   3 --+            |

   ...             ...

   12 --+           |
        +--+        |
   13 --+  |        |
           +-- ... -+
   14 --+  |
        +--+
   15 --+

In this case, it would be best to select 14 since it's a free cpu and
closest to 15(this_cpu). However, currently the code select 0(best_cpu)
even though that's just any among free_cpus. Fix it.

Change from v5
   -. exclude two patches already picked up by peterz
      (sched/deadline: Make find_later_rq() choose a closer cpu in topology)
      (sched/deadline: Change return value of cpudl_find())
   -. apply what peterz fixed for 'prefer sibling', into deadline and rt

Change from v4
   -. remove a patch that might cause huge lock contention
      (by spin lock(&cpudl.lock) in a hot path of scheduler)

Change from v3
   -. rename closest_cpu to best_cpu so that it align with rt
   -. protect referring cpudl.elements with cpudl.lock
   -. change return value of cpudl_find() to bool

Change from v2
   -. add support for SD_PREFER_SIBLING

Change from v1
   -. clean up the patch

Byungchul Park (2):
  sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()
  sched/rt: Add support for SD_PREFER_SIBLING on find_lowest_rq()

 kernel/sched/deadline.c | 46 +++++++++++++++++++++++++++++++++++++++++++---
 kernel/sched/rt.c       | 47 ++++++++++++++++++++++++++++++++++++++++++++---
 2 files changed, 87 insertions(+), 6 deletions(-)

-- 
1.9.1

[PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Byungchul Park]

It would be better to avoid pushing tasks to other cpu within
a SD_PREFER_SIBLING domain, instead, get more chances to check other
siblings.

Signed-off-by: Byungchul Park <byungchul.park@lge.com>
---
 kernel/sched/deadline.c | 46 +++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 43 insertions(+), 3 deletions(-)

diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
index 0223694..2fd1591 100644
--- a/kernel/sched/deadline.c
+++ b/kernel/sched/deadline.c
@@ -1319,12 +1319,35 @@ static struct task_struct *pick_earliest_pushable_dl_task(struct rq *rq, int cpu
 
 static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask_dl);
 
+/*
+ * Find the first cpu in: mask & sd & ~prefer
+ */
+static int find_cpu(const struct cpumask *mask,
+		    const struct sched_domain *sd,
+		    const struct sched_domain *prefer)
+{
+	const struct cpumask *sds = sched_domain_span(sd);
+	const struct cpumask *ps  = prefer ? sched_domain_span(prefer) : NULL;
+	int cpu = -1;
+
+	while ((cpu = cpumask_next(cpu, mask)) < nr_cpu_ids) {
+		if (!cpumask_test_cpu(cpu, sds))
+			continue;
+		if (ps && cpumask_test_cpu(cpu, ps))
+			continue;
+		break;
+	}
+
+	return cpu;
+}
+
 static int find_later_rq(struct task_struct *task)
 {
-	struct sched_domain *sd;
+	struct sched_domain *sd, *prefer = NULL;
 	struct cpumask *later_mask = this_cpu_cpumask_var_ptr(local_cpu_mask_dl);
 	int this_cpu = smp_processor_id();
 	int cpu = task_cpu(task);
+	int fallback_cpu = -1;
 
 	/* Make sure the mask is initialized first */
 	if (unlikely(!later_mask))
@@ -1376,8 +1399,7 @@ static int find_later_rq(struct task_struct *task)
 				return this_cpu;
 			}
 
-			best_cpu = cpumask_first_and(later_mask,
-							sched_domain_span(sd));
+			best_cpu = find_cpu(later_mask, sd, prefer);
 			/*
 			 * Last chance: if a cpu being in both later_mask
 			 * and current sd span is valid, that becomes our
@@ -1385,6 +1407,17 @@ static int find_later_rq(struct task_struct *task)
 			 * already under consideration through later_mask.
 			 */
 			if (best_cpu < nr_cpu_ids) {
+				/*
+				 * If current domain is SD_PREFER_SIBLING
+				 * flaged, we have to get more chances to
+				 * check other siblings.
+				 */
+				if (sd->flags & SD_PREFER_SIBLING) {
+					prefer = sd;
+					if (fallback_cpu == -1)
+						fallback_cpu = best_cpu;
+					continue;
+				}
 				rcu_read_unlock();
 				return best_cpu;
 			}
@@ -1393,6 +1426,13 @@ static int find_later_rq(struct task_struct *task)
 	rcu_read_unlock();
 
 	/*
+	 * If fallback_cpu is valid, all our guesses failed *except* for
+	 * SD_PREFER_SIBLING domain. Now, we can return the fallback cpu.
+	 */
+	if (fallback_cpu != -1)
+		return fallback_cpu;
+
+	/*
 	 * At this point, all our guesses failed, we just return
 	 * 'something', and let the caller sort the things out.
 	 */
-- 
1.9.1

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Steven Rostedt]

On Mon,  7 Aug 2017 12:50:33 +0900
Byungchul Park <byungchul.park@lge.com> wrote:

> It would be better to avoid pushing tasks to other cpu within
> a SD_PREFER_SIBLING domain, instead, get more chances to check other
> siblings.
> 
> Signed-off-by: Byungchul Park <byungchul.park@lge.com>
> ---
>  kernel/sched/deadline.c | 46 +++++++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 43 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/sched/deadline.c b/kernel/sched/deadline.c
> index 0223694..2fd1591 100644
> --- a/kernel/sched/deadline.c
> +++ b/kernel/sched/deadline.c
> @@ -1319,12 +1319,35 @@ static struct task_struct *pick_earliest_pushable_dl_task(struct rq *rq, int cpu
>  
>  static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask_dl);
>  
> +/*
> + * Find the first cpu in: mask & sd & ~prefer
> + */
> +static int find_cpu(const struct cpumask *mask,
> +		    const struct sched_domain *sd,
> +		    const struct sched_domain *prefer)
> +{
> +	const struct cpumask *sds = sched_domain_span(sd);
> +	const struct cpumask *ps  = prefer ? sched_domain_span(prefer) : NULL;
> +	int cpu = -1;
> +
> +	while ((cpu = cpumask_next(cpu, mask)) < nr_cpu_ids) {
> +		if (!cpumask_test_cpu(cpu, sds))
> +			continue;
> +		if (ps && cpumask_test_cpu(cpu, ps))
> +			continue;
> +		break;
> +	}
> +
> +	return cpu;
> +}
> +
>  static int find_later_rq(struct task_struct *task)
>  {
> -	struct sched_domain *sd;
> +	struct sched_domain *sd, *prefer = NULL;
>  	struct cpumask *later_mask = this_cpu_cpumask_var_ptr(local_cpu_mask_dl);
>  	int this_cpu = smp_processor_id();
>  	int cpu = task_cpu(task);
> +	int fallback_cpu = -1;
>  
>  	/* Make sure the mask is initialized first */
>  	if (unlikely(!later_mask))
> @@ -1376,8 +1399,7 @@ static int find_later_rq(struct task_struct *task)
>  				return this_cpu;
>  			}
>  
> -			best_cpu = cpumask_first_and(later_mask,
> -							sched_domain_span(sd));
> +			best_cpu = find_cpu(later_mask, sd, prefer);
>  			/*
>  			 * Last chance: if a cpu being in both later_mask
>  			 * and current sd span is valid, that becomes our
> @@ -1385,6 +1407,17 @@ static int find_later_rq(struct task_struct *task)
>  			 * already under consideration through later_mask.
>  			 */
>  			if (best_cpu < nr_cpu_ids) {
> +				/*
> +				 * If current domain is SD_PREFER_SIBLING
> +				 * flaged, we have to get more chances to
> +				 * check other siblings.
> +				 */
> +				if (sd->flags & SD_PREFER_SIBLING) {
> +					prefer = sd;

Is this how the SD_PREFER_SIBLING works? According to this, the
preferred sd is the next sd in for_each_domain(). Not to mention, the
prefer variable stays set if the next domain has no available CPUs. Is
that what we want?

-- Steve


> +					if (fallback_cpu == -1)
> +						fallback_cpu = best_cpu;
> +					continue;
> +				}
>  				rcu_read_unlock();
>  				return best_cpu;
>  			}
> @@ -1393,6 +1426,13 @@ static int find_later_rq(struct task_struct *task)
>  	rcu_read_unlock();
>  
>  	/*
> +	 * If fallback_cpu is valid, all our guesses failed *except* for
> +	 * SD_PREFER_SIBLING domain. Now, we can return the fallback cpu.
> +	 */
> +	if (fallback_cpu != -1)
> +		return fallback_cpu;
> +
> +	/*
>  	 * At this point, all our guesses failed, we just return
>  	 * 'something', and let the caller sort the things out.
>  	 */

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Byungchul Park]

On Tue, Aug 15, 2017 at 11:19:40AM -0400, Steven Rostedt wrote:
> > @@ -1385,6 +1407,17 @@ static int find_later_rq(struct task_struct *task)
> >  			 * already under consideration through later_mask.
> >  			 */
> >  			if (best_cpu < nr_cpu_ids) {
> > +				/*
> > +				 * If current domain is SD_PREFER_SIBLING
> > +				 * flaged, we have to get more chances to
> > +				 * check other siblings.
> > +				 */
> > +				if (sd->flags & SD_PREFER_SIBLING) {
> > +					prefer = sd;
> 
> Is this how the SD_PREFER_SIBLING works? According to this, the
> preferred sd is the next sd in for_each_domain(). Not to mention, the
> prefer variable stays set if the next domain has no available CPUs. Is
> that what we want?

Maybe I don't understand what you want to say. The variable, prefer, is
used to pick up the smallest sched domain among SD_PREFER_SIBLING
domains, if more than one SD_PREFER_SIBLING domain exist in the visit.

The prefer variable alway points to the previous SD_PREFER_SIBLING domain.
And that must stay set to be used as a fallback choise if the next domain
has no available CPUs.

Could you explain what I mis-understand?

Thanks,
Byungchul

> -- Steve
> 
> 
> > +					if (fallback_cpu == -1)
> > +						fallback_cpu = best_cpu;
> > +					continue;
> > +				}
> >  				rcu_read_unlock();
> >  				return best_cpu;
> >  			}
> > @@ -1393,6 +1426,13 @@ static int find_later_rq(struct task_struct *task)
> >  	rcu_read_unlock();
> >  
> >  	/*
> > +	 * If fallback_cpu is valid, all our guesses failed *except* for
> > +	 * SD_PREFER_SIBLING domain. Now, we can return the fallback cpu.
> > +	 */
> > +	if (fallback_cpu != -1)
> > +		return fallback_cpu;
> > +
> > +	/*
> >  	 * At this point, all our guesses failed, we just return
> >  	 * 'something', and let the caller sort the things out.
> >  	 */

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Steven Rostedt]

On Wed, 16 Aug 2017 09:38:11 +0900
Byungchul Park <byungchul.park@lge.com> wrote:

> On Tue, Aug 15, 2017 at 11:19:40AM -0400, Steven Rostedt wrote:
> > > @@ -1385,6 +1407,17 @@ static int find_later_rq(struct task_struct *task)
> > >  			 * already under consideration through later_mask.
> > >  			 */
> > >  			if (best_cpu < nr_cpu_ids) {
> > > +				/*
> > > +				 * If current domain is SD_PREFER_SIBLING
> > > +				 * flaged, we have to get more chances to
> > > +				 * check other siblings.

BTW, "we have to get more chances" doesn't really make sense. Do you
mean "we need to try other domains"?

> > > +				 */
> > > +				if (sd->flags & SD_PREFER_SIBLING) {
> > > +					prefer = sd;  
> > 
> > Is this how the SD_PREFER_SIBLING works? According to this, the
> > preferred sd is the next sd in for_each_domain(). Not to mention, the
> > prefer variable stays set if the next domain has no available CPUs. Is
> > that what we want?  
> 
> Maybe I don't understand what you want to say. The variable, prefer, is
> used to pick up the smallest sched domain among SD_PREFER_SIBLING
> domains, if more than one SD_PREFER_SIBLING domain exist in the visit.
> 
> The prefer variable alway points to the previous SD_PREFER_SIBLING domain.
> And that must stay set to be used as a fallback choise if the next domain
> has no available CPUs.
> 
> Could you explain what I mis-understand?
>

I may be the one confused here ;-)

I think I misread the patch. So, the SD_PREFER_SIBLING means to try to
find a CPU in another sd instead? Thus, we try to find a CPU in a sd
that does not have SD_PREFER_SIBLING set. And if there is none, we use
the preferred sd as a fallback. Is that correct?

I'm not familiar with the SD_PREFER_SIBLING flag, the only
documentation I can find about it is the comment that states:

  /* Prefer to place tasks in a sibling domain */

And the very informative git commit change log:

commit b5d978e0c7e79a7ff842e895c85a86b38c71f1cd
Date:   Tue Sep 1 10:34:33 2009 +0200

    sched: Add SD_PREFER_SIBLING
    
    Do the placement thing using SD flags.

 ;-)

-- Steve

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Byungchul Park]

On Tue, Aug 15, 2017 at 09:42:01PM -0400, Steven Rostedt wrote:
> > > > @@ -1385,6 +1407,17 @@ static int find_later_rq(struct task_struct *task)
> > > >  			 * already under consideration through later_mask.
> > > >  			 */
> > > >  			if (best_cpu < nr_cpu_ids) {
> > > > +				/*
> > > > +				 * If current domain is SD_PREFER_SIBLING
> > > > +				 * flaged, we have to get more chances to
> > > > +				 * check other siblings.
> 
> BTW, "we have to get more chances" doesn't really make sense. Do you
> mean "we need to try other domains"?

Yes, we need to try other domains first if current domain is
SD_PREFER_SIBLING flaged.

> > > > +				 */
> > > > +				if (sd->flags & SD_PREFER_SIBLING) {
> > > > +					prefer = sd;  
> > > 
> > > Is this how the SD_PREFER_SIBLING works? According to this, the
> > > preferred sd is the next sd in for_each_domain(). Not to mention, the
> > > prefer variable stays set if the next domain has no available CPUs. Is
> > > that what we want?  
> > 
> > Maybe I don't understand what you want to say. The variable, prefer, is
> > used to pick up the smallest sched domain among SD_PREFER_SIBLING
> > domains, if more than one SD_PREFER_SIBLING domain exist in the visit.
> > 
> > The prefer variable alway points to the previous SD_PREFER_SIBLING domain.
> > And that must stay set to be used as a fallback choise if the next domain
> > has no available CPUs.
> > 
> > Could you explain what I mis-understand?
> >
> 
> I may be the one confused here ;-)
> 
> I think I misread the patch. So, the SD_PREFER_SIBLING means to try to
> find a CPU in another sd instead? Thus, we try to find a CPU in a sd
> that does not have SD_PREFER_SIBLING set. And if there is none, we use
> the preferred sd as a fallback. Is that correct?

Yes, that's what I intended. IOW:

If (we found a proper sd, not having SD_PREFER_SIBLING?)
	use the sd;
else if (we found a proper sd, having SD_PREFER_SIBLING?)
	use the smallest sd among SD_PREFER_SIBLING sds;

Thanks,
Byungchul

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Steven Rostedt]

On Wed, 16 Aug 2017 11:17:36 +0900
Byungchul Park <byungchul.park@lge.com> wrote:


> Yes, that's what I intended. IOW:
> 
> If (we found a proper sd, not having SD_PREFER_SIBLING?)
> 	use the sd;
> else if (we found a proper sd, having SD_PREFER_SIBLING?)
> 	use the smallest sd among SD_PREFER_SIBLING sds;

BTW, what do you mean by "smallest sd"?

-- Steve

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Byungchul Park]

On Wed, Aug 16, 2017 at 09:32:44AM -0400, Steven Rostedt wrote:
> On Wed, 16 Aug 2017 11:17:36 +0900
> Byungchul Park <byungchul.park@lge.com> wrote:
> 
> 
> > Yes, that's what I intended. IOW:
> > 
> > If (we found a proper sd, not having SD_PREFER_SIBLING?)
> > 	use the sd;
> > else if (we found a proper sd, having SD_PREFER_SIBLING?)
> > 	use the smallest sd among SD_PREFER_SIBLING sds;
> 
> BTW, what do you mean by "smallest sd"?

There might be more than one SD_PREFER_SIBLING domain in its hierachy.
In that case, we have to choose one of them. Imagine the following
example, in case that the source cpu is cpu 0:

[Domain hierachy for cpu 0]

cpu 0 -+ domain 1                 -+
       | SD_PREFER_SIBLING flaged  |
cpu 1 -+                           +- domain 2
                                   |  SD_PREFER_SIBLING flaged
cpu 2 -+---------------------------+
       |
cpu 3 -+

In this case, we have to choose domain 1 than 2, because cpus in domain 1
are closer to the source cpu, cpu 0. That's what I meant.

Re: [PATCH v6 1/2] sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()

[Steven Rostedt]

On Wed, 16 Aug 2017 23:04:14 +0900
Byungchul Park <byungchul.park@lge.com> wrote:

> On Wed, Aug 16, 2017 at 09:32:44AM -0400, Steven Rostedt wrote:
> > On Wed, 16 Aug 2017 11:17:36 +0900
> > Byungchul Park <byungchul.park@lge.com> wrote:
> > 
> >   
> > > Yes, that's what I intended. IOW:
> > > 
> > > If (we found a proper sd, not having SD_PREFER_SIBLING?)
> > > 	use the sd;
> > > else if (we found a proper sd, having SD_PREFER_SIBLING?)
> > > 	use the smallest sd among SD_PREFER_SIBLING sds;  
> > 
> > BTW, what do you mean by "smallest sd"?  
> 
> There might be more than one SD_PREFER_SIBLING domain in its hierachy.
> In that case, we have to choose one of them. Imagine the following
> example, in case that the source cpu is cpu 0:
> 
> [Domain hierachy for cpu 0]
> 
> cpu 0 -+ domain 1                 -+
>        | SD_PREFER_SIBLING flaged  |
> cpu 1 -+                           +- domain 2
>                                    |  SD_PREFER_SIBLING flaged
> cpu 2 -+---------------------------+
>        |
> cpu 3 -+
> 
> In this case, we have to choose domain 1 than 2, because cpus in domain 1
> are closer to the source cpu, cpu 0. That's what I meant.

Then you mean "closest sd", at least that makes more sense in the
context.

-- Steve

[PATCH v6 2/2] sched/rt: Add support for SD_PREFER_SIBLING on find_lowest_rq()

[Byungchul Park]

It would be better to avoid pushing tasks to other cpu within
a SD_PREFER_SIBLING domain, instead, get more chances to check other
siblings.

Signed-off-by: Byungchul Park <byungchul.park@lge.com>
---
 kernel/sched/rt.c | 47 ++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 44 insertions(+), 3 deletions(-)

diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
index 979b734..50639e5 100644
--- a/kernel/sched/rt.c
+++ b/kernel/sched/rt.c
@@ -1618,12 +1618,35 @@ static struct task_struct *pick_highest_pushable_task(struct rq *rq, int cpu)
 
 static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask);
 
+/*
+ * Find the first cpu in: mask & sd & ~prefer
+ */
+static int find_cpu(const struct cpumask *mask,
+		    const struct sched_domain *sd,
+		    const struct sched_domain *prefer)
+{
+	const struct cpumask *sds = sched_domain_span(sd);
+	const struct cpumask *ps  = prefer ? sched_domain_span(prefer) : NULL;
+	int cpu = -1;
+
+	while ((cpu = cpumask_next(cpu, mask)) < nr_cpu_ids) {
+		if (!cpumask_test_cpu(cpu, sds))
+			continue;
+		if (ps && cpumask_test_cpu(cpu, ps))
+			continue;
+		break;
+	}
+
+	return cpu;
+}
+
 static int find_lowest_rq(struct task_struct *task)
 {
-	struct sched_domain *sd;
+	struct sched_domain *sd, *prefer = NULL;
 	struct cpumask *lowest_mask = this_cpu_cpumask_var_ptr(local_cpu_mask);
 	int this_cpu = smp_processor_id();
 	int cpu      = task_cpu(task);
+	int fallback_cpu = -1;
 
 	/* Make sure the mask is initialized first */
 	if (unlikely(!lowest_mask))
@@ -1668,9 +1691,20 @@ static int find_lowest_rq(struct task_struct *task)
 				return this_cpu;
 			}
 
-			best_cpu = cpumask_first_and(lowest_mask,
-						     sched_domain_span(sd));
+			best_cpu = find_cpu(lowest_mask, sd, prefer);
+
 			if (best_cpu < nr_cpu_ids) {
+				/*
+				 * If current domain is SD_PREFER_SIBLING
+				 * flaged, we have to get more chances to
+				 * check other siblings.
+				 */
+				if (sd->flags & SD_PREFER_SIBLING) {
+					prefer = sd;
+					if (fallback_cpu == -1)
+						fallback_cpu = best_cpu;
+					continue;
+				}
 				rcu_read_unlock();
 				return best_cpu;
 			}
@@ -1679,6 +1713,13 @@ static int find_lowest_rq(struct task_struct *task)
 	rcu_read_unlock();
 
 	/*
+	 * If fallback_cpu is valid, all our quesses failed *except* for
+	 * SD_PREFER_SIBLING domain. Now, we can return the fallback cpu.
+	 */
+	if (fallback_cpu != -1)
+		return fallback_cpu;
+
+	/*
 	 * And finally, if there were no matches within the domains
 	 * just give the caller *something* to work with from the compatible
 	 * locations.
-- 
1.9.1

Re: [PATCH v6 2/2] sched/rt: Add support for SD_PREFER_SIBLING on find_lowest_rq()

[Byungchul Park]

On Mon, Aug 07, 2017 at 12:50:34PM +0900, Byungchul Park wrote:
> It would be better to avoid pushing tasks to other cpu within
> a SD_PREFER_SIBLING domain, instead, get more chances to check other
> siblings.

I applied your suggestion. Could you let me know your opinions about
this?

> 
> Signed-off-by: Byungchul Park <byungchul.park@lge.com>
> ---
>  kernel/sched/rt.c | 47 ++++++++++++++++++++++++++++++++++++++++++++---
>  1 file changed, 44 insertions(+), 3 deletions(-)
> 
> diff --git a/kernel/sched/rt.c b/kernel/sched/rt.c
> index 979b734..50639e5 100644
> --- a/kernel/sched/rt.c
> +++ b/kernel/sched/rt.c
> @@ -1618,12 +1618,35 @@ static struct task_struct *pick_highest_pushable_task(struct rq *rq, int cpu)
>  
>  static DEFINE_PER_CPU(cpumask_var_t, local_cpu_mask);
>  
> +/*
> + * Find the first cpu in: mask & sd & ~prefer
> + */
> +static int find_cpu(const struct cpumask *mask,
> +		    const struct sched_domain *sd,
> +		    const struct sched_domain *prefer)
> +{
> +	const struct cpumask *sds = sched_domain_span(sd);
> +	const struct cpumask *ps  = prefer ? sched_domain_span(prefer) : NULL;
> +	int cpu = -1;
> +
> +	while ((cpu = cpumask_next(cpu, mask)) < nr_cpu_ids) {
> +		if (!cpumask_test_cpu(cpu, sds))
> +			continue;
> +		if (ps && cpumask_test_cpu(cpu, ps))
> +			continue;
> +		break;
> +	}
> +
> +	return cpu;
> +}
> +
>  static int find_lowest_rq(struct task_struct *task)
>  {
> -	struct sched_domain *sd;
> +	struct sched_domain *sd, *prefer = NULL;
>  	struct cpumask *lowest_mask = this_cpu_cpumask_var_ptr(local_cpu_mask);
>  	int this_cpu = smp_processor_id();
>  	int cpu      = task_cpu(task);
> +	int fallback_cpu = -1;
>  
>  	/* Make sure the mask is initialized first */
>  	if (unlikely(!lowest_mask))
> @@ -1668,9 +1691,20 @@ static int find_lowest_rq(struct task_struct *task)
>  				return this_cpu;
>  			}
>  
> -			best_cpu = cpumask_first_and(lowest_mask,
> -						     sched_domain_span(sd));
> +			best_cpu = find_cpu(lowest_mask, sd, prefer);
> +
>  			if (best_cpu < nr_cpu_ids) {
> +				/*
> +				 * If current domain is SD_PREFER_SIBLING
> +				 * flaged, we have to get more chances to
> +				 * check other siblings.
> +				 */
> +				if (sd->flags & SD_PREFER_SIBLING) {
> +					prefer = sd;
> +					if (fallback_cpu == -1)
> +						fallback_cpu = best_cpu;
> +					continue;
> +				}
>  				rcu_read_unlock();
>  				return best_cpu;
>  			}
> @@ -1679,6 +1713,13 @@ static int find_lowest_rq(struct task_struct *task)
>  	rcu_read_unlock();
>  
>  	/*
> +	 * If fallback_cpu is valid, all our quesses failed *except* for
> +	 * SD_PREFER_SIBLING domain. Now, we can return the fallback cpu.
> +	 */
> +	if (fallback_cpu != -1)
> +		return fallback_cpu;
> +
> +	/*
>  	 * And finally, if there were no matches within the domains
>  	 * just give the caller *something* to work with from the compatible
>  	 * locations.
> -- 
> 1.9.1

Re: [PATCH v6 0/2] Make find_later_rq() choose a closer cpu in topology

[Byungchul Park]

On Mon, Aug 07, 2017 at 12:50:32PM +0900, Byungchul Park wrote:
> When cpudl_find() returns any among free_cpus, the cpu might not be
> closer than others, considering sched domain. For example:
> 
>    this_cpu: 15
>    free_cpus: 0, 1,..., 14 (== later_mask)
>    best_cpu: 0
> 
>    topology:
> 
>    0 --+
>        +--+
>    1 --+  |
>           +-- ... --+
>    2 --+  |         |
>        +--+         |
>    3 --+            |
> 
>    ...             ...
> 
>    12 --+           |
>         +--+        |
>    13 --+  |        |
>            +-- ... -+
>    14 --+  |
>         +--+
>    15 --+
> 
> In this case, it would be best to select 14 since it's a free cpu and
> closest to 15(this_cpu). However, currently the code select 0(best_cpu)
> even though that's just any among free_cpus. Fix it.

Could you let me know your opinions about this?

> Change from v5
>    -. exclude two patches already picked up by peterz
>       (sched/deadline: Make find_later_rq() choose a closer cpu in topology)
>       (sched/deadline: Change return value of cpudl_find())
>    -. apply what peterz fixed for 'prefer sibling', into deadline and rt
> 
> Change from v4
>    -. remove a patch that might cause huge lock contention
>       (by spin lock(&cpudl.lock) in a hot path of scheduler)
> 
> Change from v3
>    -. rename closest_cpu to best_cpu so that it align with rt
>    -. protect referring cpudl.elements with cpudl.lock
>    -. change return value of cpudl_find() to bool
> 
> Change from v2
>    -. add support for SD_PREFER_SIBLING
> 
> Change from v1
>    -. clean up the patch
> 
> Byungchul Park (2):
>   sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()
>   sched/rt: Add support for SD_PREFER_SIBLING on find_lowest_rq()
> 
>  kernel/sched/deadline.c | 46 +++++++++++++++++++++++++++++++++++++++++++---
>  kernel/sched/rt.c       | 47 ++++++++++++++++++++++++++++++++++++++++++++---
>  2 files changed, 87 insertions(+), 6 deletions(-)
> 
> -- 
> 1.9.1

Re: [PATCH v6 0/2] Make find_later_rq() choose a closer cpu in topology

[Joel Fernandes (Google)]

On Thu, Aug 17, 2017 at 6:25 PM, Byungchul Park <byungchul.park@lge.com> wrote:
> On Mon, Aug 07, 2017 at 12:50:32PM +0900, Byungchul Park wrote:
>> When cpudl_find() returns any among free_cpus, the cpu might not be
>> closer than others, considering sched domain. For example:
>>
>>    this_cpu: 15
>>    free_cpus: 0, 1,..., 14 (== later_mask)
>>    best_cpu: 0
>>
>>    topology:
>>
>>    0 --+
>>        +--+
>>    1 --+  |
>>           +-- ... --+
>>    2 --+  |         |
>>        +--+         |
>>    3 --+            |
>>
>>    ...             ...
>>
>>    12 --+           |
>>         +--+        |
>>    13 --+  |        |
>>            +-- ... -+
>>    14 --+  |
>>         +--+
>>    15 --+
>>
>> In this case, it would be best to select 14 since it's a free cpu and
>> closest to 15(this_cpu). However, currently the code select 0(best_cpu)
>> even though that's just any among free_cpus. Fix it.
>
> Could you let me know your opinions about this?

Patch looks good to me, I would also add a comment ontop of
fallback_cpu (I think Steve mentioned similar thing at [1])

/*
 * fallback is the closest CPU in the closest SD incase
 * all domains are PREFER_SIBLING
 */
 if (fallback_cpu == -1)
     fallback_cpu = best_cpu;

And clarify this in the commit message.


thanks,

-Joel

[1] https://patchwork.kernel.org/patch/9884383/


>
>> Change from v5
>>    -. exclude two patches already picked up by peterz
>>       (sched/deadline: Make find_later_rq() choose a closer cpu in topology)
>>       (sched/deadline: Change return value of cpudl_find())
>>    -. apply what peterz fixed for 'prefer sibling', into deadline and rt
>>
>> Change from v4
>>    -. remove a patch that might cause huge lock contention
>>       (by spin lock(&cpudl.lock) in a hot path of scheduler)
>>
>> Change from v3
>>    -. rename closest_cpu to best_cpu so that it align with rt
>>    -. protect referring cpudl.elements with cpudl.lock
>>    -. change return value of cpudl_find() to bool
>>
>> Change from v2
>>    -. add support for SD_PREFER_SIBLING
>>
>> Change from v1
>>    -. clean up the patch
>>
>> Byungchul Park (2):
>>   sched/deadline: Add support for SD_PREFER_SIBLING on find_later_rq()
>>   sched/rt: Add support for SD_PREFER_SIBLING on find_lowest_rq()
>>
>>  kernel/sched/deadline.c | 46 +++++++++++++++++++++++++++++++++++++++++++---
>>  kernel/sched/rt.c       | 47 ++++++++++++++++++++++++++++++++++++++++++++---
>>  2 files changed, 87 insertions(+), 6 deletions(-)
>>
>> --
>> 1.9.1

Re: [PATCH v6 0/2] Make find_later_rq() choose a closer cpu in topology

[Byungchul Park]

On Thu, Aug 17, 2017 at 09:51:34PM -0700, Joel Fernandes (Google) wrote:
> On Thu, Aug 17, 2017 at 6:25 PM, Byungchul Park <byungchul.park@lge.com> wrote:
> > On Mon, Aug 07, 2017 at 12:50:32PM +0900, Byungchul Park wrote:
> >> When cpudl_find() returns any among free_cpus, the cpu might not be
> >> closer than others, considering sched domain. For example:
> >>
> >>    this_cpu: 15
> >>    free_cpus: 0, 1,..., 14 (== later_mask)
> >>    best_cpu: 0
> >>
> >>    topology:
> >>
> >>    0 --+
> >>        +--+
> >>    1 --+  |
> >>           +-- ... --+
> >>    2 --+  |         |
> >>        +--+         |
> >>    3 --+            |
> >>
> >>    ...             ...
> >>
> >>    12 --+           |
> >>         +--+        |
> >>    13 --+  |        |
> >>            +-- ... -+
> >>    14 --+  |
> >>         +--+
> >>    15 --+
> >>
> >> In this case, it would be best to select 14 since it's a free cpu and
> >> closest to 15(this_cpu). However, currently the code select 0(best_cpu)
> >> even though that's just any among free_cpus. Fix it.
> >
> > Could you let me know your opinions about this?
> 
> Patch looks good to me, I would also add a comment ontop of
> fallback_cpu (I think Steve mentioned similar thing at [1])
> 
> /*
>  * fallback is the closest CPU in the closest SD incase
>  * all domains are PREFER_SIBLING
>  */
>  if (fallback_cpu == -1)
>      fallback_cpu = best_cpu;
> 
> And clarify this in the commit message.

Right. I will add it.

Thank you very much,
Byungchul