[PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Phil Auld]

The exising code uses housekeeping_any_cpu() to select a cpu for
a given housekeeping task. However, this often ends up calling
cpumask_any_and() which is defined as cpumask_first_and() which has
the effect of alyways using the first cpu among those available.

The same applies when multiple NUMA nodes are involved. In that
case the first cpu in the local node is chosen which does provide
a bit of spreading but with multiple HK cpus per node the same
issues arise.

We have numerous cases where a single HK cpu just cannot keep up
and the remote_tick warning fires. It also can lead to the other
things (orchastration sw, HA keepalives etc) on the HK cpus getting
starved which leads to other issues.  In these cases we recommend
increasing the number of HK cpus.  But... that only helps the
userspace tasks somewhat. It does not help the actual housekeeping
part.

Spread the HK work out by having housekeeping_any_cpu() and
sched_numa_find_closest() use cpumask_any_and_distribute()
instead of cpumask_any_and().

Signed-off-by: Phil Auld <pauld@redhat.com>
Reviewed-by: Waiman Long <longman@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Juri Lelli <juri.lelli@redhat.com>
Cc: Frederic Weisbecker <frederic@kernel.org>
Cc: Waiman Long <longman@redhat.com>
Cc: linux-kernel@vger.kernel.org
Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/

---

v2: Fix subject line. Update commit message. No code change. 

 kernel/sched/isolation.c | 2 +-
 kernel/sched/topology.c  | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/isolation.c b/kernel/sched/isolation.c
index 81bc8b329ef1..93b038d48900 100644
--- a/kernel/sched/isolation.c
+++ b/kernel/sched/isolation.c
@@ -40,7 +40,7 @@ int housekeeping_any_cpu(enum hk_type type)
 			if (cpu < nr_cpu_ids)
 				return cpu;
 
-			cpu = cpumask_any_and(housekeeping.cpumasks[type], cpu_online_mask);
+			cpu = cpumask_any_and_distribute(housekeeping.cpumasks[type], cpu_online_mask);
 			if (likely(cpu < nr_cpu_ids))
 				return cpu;
 			/*
diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index c49aea8c1025..94133f843485 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -2101,7 +2101,7 @@ int sched_numa_find_closest(const struct cpumask *cpus, int cpu)
 	for (i = 0; i < sched_domains_numa_levels; i++) {
 		if (!masks[i][j])
 			break;
-		cpu = cpumask_any_and(cpus, masks[i][j]);
+		cpu = cpumask_any_and_distribute(cpus, masks[i][j]);
 		if (cpu < nr_cpu_ids) {
 			found = cpu;
 			break;
-- 
2.47.1

Re: [PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Vishal Chourasia]

On Tue, Feb 18, 2025 at 06:46:18PM +0000, Phil Auld wrote:
> The exising code uses housekeeping_any_cpu() to select a cpu for
> a given housekeeping task. However, this often ends up calling
> cpumask_any_and() which is defined as cpumask_first_and() which has
> the effect of alyways using the first cpu among those available.
> 
> The same applies when multiple NUMA nodes are involved. In that
> case the first cpu in the local node is chosen which does provide
> a bit of spreading but with multiple HK cpus per node the same
> issues arise.
> 
> We have numerous cases where a single HK cpu just cannot keep up
> and the remote_tick warning fires. It also can lead to the other
> things (orchastration sw, HA keepalives etc) on the HK cpus getting
> starved which leads to other issues.  In these cases we recommend
> increasing the number of HK cpus.  But... that only helps the
> userspace tasks somewhat. It does not help the actual housekeeping
> part.
> 
> Spread the HK work out by having housekeeping_any_cpu() and
> sched_numa_find_closest() use cpumask_any_and_distribute()
> instead of cpumask_any_and().
> 
LGTM.

Reviewed-by: Vishal Chourasia <vishalc@linux.ibm.com>

> Signed-off-by: Phil Auld <pauld@redhat.com>
> Reviewed-by: Waiman Long <longman@redhat.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Juri Lelli <juri.lelli@redhat.com>
> Cc: Frederic Weisbecker <frederic@kernel.org>
> Cc: Waiman Long <longman@redhat.com>
> Cc: linux-kernel@vger.kernel.org
> Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/
> 
> ---
> 
> v2: Fix subject line. Update commit message. No code change. 
> 
>  kernel/sched/isolation.c | 2 +-
>  kernel/sched/topology.c  | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/sched/isolation.c b/kernel/sched/isolation.c
> index 81bc8b329ef1..93b038d48900 100644
> --- a/kernel/sched/isolation.c
> +++ b/kernel/sched/isolation.c
> @@ -40,7 +40,7 @@ int housekeeping_any_cpu(enum hk_type type)
>  			if (cpu < nr_cpu_ids)
>  				return cpu;
>  
> -			cpu = cpumask_any_and(housekeeping.cpumasks[type], cpu_online_mask);
> +			cpu = cpumask_any_and_distribute(housekeeping.cpumasks[type], cpu_online_mask);
>  			if (likely(cpu < nr_cpu_ids))
>  				return cpu;
>  			/*
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> index c49aea8c1025..94133f843485 100644
> --- a/kernel/sched/topology.c
> +++ b/kernel/sched/topology.c
> @@ -2101,7 +2101,7 @@ int sched_numa_find_closest(const struct cpumask *cpus, int cpu)
>  	for (i = 0; i < sched_domains_numa_levels; i++) {
>  		if (!masks[i][j])
>  			break;
> -		cpu = cpumask_any_and(cpus, masks[i][j]);
> +		cpu = cpumask_any_and_distribute(cpus, masks[i][j]);
>  		if (cpu < nr_cpu_ids) {
>  			found = cpu;
>  			break;
> -- 
> 2.47.1
> 

Re: [PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Phil Auld]

Hi

On Tue, Feb 18, 2025 at 06:46:18PM +0000 Phil Auld wrote:
> The exising code uses housekeeping_any_cpu() to select a cpu for
> a given housekeeping task. However, this often ends up calling
> cpumask_any_and() which is defined as cpumask_first_and() which has
> the effect of alyways using the first cpu among those available.
> 
> The same applies when multiple NUMA nodes are involved. In that
> case the first cpu in the local node is chosen which does provide
> a bit of spreading but with multiple HK cpus per node the same
> issues arise.
> 
> We have numerous cases where a single HK cpu just cannot keep up
> and the remote_tick warning fires. It also can lead to the other
> things (orchastration sw, HA keepalives etc) on the HK cpus getting
> starved which leads to other issues.  In these cases we recommend
> increasing the number of HK cpus.  But... that only helps the
> userspace tasks somewhat. It does not help the actual housekeeping
> part.
> 
> Spread the HK work out by having housekeeping_any_cpu() and
> sched_numa_find_closest() use cpumask_any_and_distribute()
> instead of cpumask_any_and().
> 
> Signed-off-by: Phil Auld <pauld@redhat.com>
> Reviewed-by: Waiman Long <longman@redhat.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Juri Lelli <juri.lelli@redhat.com>
> Cc: Frederic Weisbecker <frederic@kernel.org>
> Cc: Waiman Long <longman@redhat.com>
> Cc: linux-kernel@vger.kernel.org
> Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/

Friendly ping...

Frederic, anyone else, thoughts?

Cheers,
Phil

> 
> ---
> 
> v2: Fix subject line. Update commit message. No code change. 
> 
>  kernel/sched/isolation.c | 2 +-
>  kernel/sched/topology.c  | 2 +-
>  2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/sched/isolation.c b/kernel/sched/isolation.c
> index 81bc8b329ef1..93b038d48900 100644
> --- a/kernel/sched/isolation.c
> +++ b/kernel/sched/isolation.c
> @@ -40,7 +40,7 @@ int housekeeping_any_cpu(enum hk_type type)
>  			if (cpu < nr_cpu_ids)
>  				return cpu;
>  
> -			cpu = cpumask_any_and(housekeeping.cpumasks[type], cpu_online_mask);
> +			cpu = cpumask_any_and_distribute(housekeeping.cpumasks[type], cpu_online_mask);
>  			if (likely(cpu < nr_cpu_ids))
>  				return cpu;
>  			/*
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> index c49aea8c1025..94133f843485 100644
> --- a/kernel/sched/topology.c
> +++ b/kernel/sched/topology.c
> @@ -2101,7 +2101,7 @@ int sched_numa_find_closest(const struct cpumask *cpus, int cpu)
>  	for (i = 0; i < sched_domains_numa_levels; i++) {
>  		if (!masks[i][j])
>  			break;
> -		cpu = cpumask_any_and(cpus, masks[i][j]);
> +		cpu = cpumask_any_and_distribute(cpus, masks[i][j]);
>  		if (cpu < nr_cpu_ids) {
>  			found = cpu;
>  			break;
> -- 
> 2.47.1
> 
> 

-- 

Re: [PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Frederic Weisbecker]

Le Tue, Feb 18, 2025 at 06:46:18PM +0000, Phil Auld a écrit :
> The exising code uses housekeeping_any_cpu() to select a cpu for
> a given housekeeping task. However, this often ends up calling
> cpumask_any_and() which is defined as cpumask_first_and() which has
> the effect of alyways using the first cpu among those available.
> 
> The same applies when multiple NUMA nodes are involved. In that
> case the first cpu in the local node is chosen which does provide
> a bit of spreading but with multiple HK cpus per node the same
> issues arise.
> 
> We have numerous cases where a single HK cpu just cannot keep up
> and the remote_tick warning fires. It also can lead to the other
> things (orchastration sw, HA keepalives etc) on the HK cpus getting
> starved which leads to other issues.  In these cases we recommend
> increasing the number of HK cpus.  But... that only helps the
> userspace tasks somewhat. It does not help the actual housekeeping
> part.
> 
> Spread the HK work out by having housekeeping_any_cpu() and
> sched_numa_find_closest() use cpumask_any_and_distribute()
> instead of cpumask_any_and().
> 
> Signed-off-by: Phil Auld <pauld@redhat.com>
> Reviewed-by: Waiman Long <longman@redhat.com>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Juri Lelli <juri.lelli@redhat.com>
> Cc: Frederic Weisbecker <frederic@kernel.org>
> Cc: Waiman Long <longman@redhat.com>
> Cc: linux-kernel@vger.kernel.org
> Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/

Acked-by: Frederic Weisbecker <frederic@kernel.org>

Re: [PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Phil Auld]

On Tue, Mar 04, 2025 at 02:24:19PM +0100 Frederic Weisbecker wrote:
> Le Tue, Feb 18, 2025 at 06:46:18PM +0000, Phil Auld a écrit :
> > The exising code uses housekeeping_any_cpu() to select a cpu for
> > a given housekeeping task. However, this often ends up calling
> > cpumask_any_and() which is defined as cpumask_first_and() which has
> > the effect of alyways using the first cpu among those available.
> > 
> > The same applies when multiple NUMA nodes are involved. In that
> > case the first cpu in the local node is chosen which does provide
> > a bit of spreading but with multiple HK cpus per node the same
> > issues arise.
> > 
> > We have numerous cases where a single HK cpu just cannot keep up
> > and the remote_tick warning fires. It also can lead to the other
> > things (orchastration sw, HA keepalives etc) on the HK cpus getting
> > starved which leads to other issues.  In these cases we recommend
> > increasing the number of HK cpus.  But... that only helps the
> > userspace tasks somewhat. It does not help the actual housekeeping
> > part.
> > 
> > Spread the HK work out by having housekeeping_any_cpu() and
> > sched_numa_find_closest() use cpumask_any_and_distribute()
> > instead of cpumask_any_and().
> > 
> > Signed-off-by: Phil Auld <pauld@redhat.com>
> > Reviewed-by: Waiman Long <longman@redhat.com>
> > Cc: Peter Zijlstra <peterz@infradead.org>
> > Cc: Juri Lelli <juri.lelli@redhat.com>
> > Cc: Frederic Weisbecker <frederic@kernel.org>
> > Cc: Waiman Long <longman@redhat.com>
> > Cc: linux-kernel@vger.kernel.org
> > Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/
> 
> Acked-by: Frederic Weisbecker <frederic@kernel.org>
>

Thanks Frederic!

Anyone with commit powers willing to pick this up?

Thanks,
Phil

-- 

Re: [PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Waiman Long]

On 3/5/25 12:14 PM, Phil Auld wrote:
> On Tue, Mar 04, 2025 at 02:24:19PM +0100 Frederic Weisbecker wrote:
>> Le Tue, Feb 18, 2025 at 06:46:18PM +0000, Phil Auld a écrit :
>>> The exising code uses housekeeping_any_cpu() to select a cpu for
>>> a given housekeeping task. However, this often ends up calling
>>> cpumask_any_and() which is defined as cpumask_first_and() which has
>>> the effect of alyways using the first cpu among those available.
>>>
>>> The same applies when multiple NUMA nodes are involved. In that
>>> case the first cpu in the local node is chosen which does provide
>>> a bit of spreading but with multiple HK cpus per node the same
>>> issues arise.
>>>
>>> We have numerous cases where a single HK cpu just cannot keep up
>>> and the remote_tick warning fires. It also can lead to the other
>>> things (orchastration sw, HA keepalives etc) on the HK cpus getting
>>> starved which leads to other issues.  In these cases we recommend
>>> increasing the number of HK cpus.  But... that only helps the
>>> userspace tasks somewhat. It does not help the actual housekeeping
>>> part.
>>>
>>> Spread the HK work out by having housekeeping_any_cpu() and
>>> sched_numa_find_closest() use cpumask_any_and_distribute()
>>> instead of cpumask_any_and().
>>>
>>> Signed-off-by: Phil Auld <pauld@redhat.com>
>>> Reviewed-by: Waiman Long <longman@redhat.com>
>>> Cc: Peter Zijlstra <peterz@infradead.org>
>>> Cc: Juri Lelli <juri.lelli@redhat.com>
>>> Cc: Frederic Weisbecker <frederic@kernel.org>
>>> Cc: Waiman Long <longman@redhat.com>
>>> Cc: linux-kernel@vger.kernel.org
>>> Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/
>> Acked-by: Frederic Weisbecker <frederic@kernel.org>
>>
> Thanks Frederic!
>
> Anyone with commit powers willing to pick this up?

Is this patch eligible to be merged into sched/core for the current 
cycle or will have to wait until the next one?

Thanks,
Longman

Re: [PATCH v2] sched/isolation: Make use of more than one housekeeping cpu

[Phil Auld]

Hi Peter,

On Fri, Mar 14, 2025 at 07:38:22PM -0400 Waiman Long wrote:
> On 3/5/25 12:14 PM, Phil Auld wrote:
> > On Tue, Mar 04, 2025 at 02:24:19PM +0100 Frederic Weisbecker wrote:
> > > Le Tue, Feb 18, 2025 at 06:46:18PM +0000, Phil Auld a écrit :
> > > > The exising code uses housekeeping_any_cpu() to select a cpu for
> > > > a given housekeeping task. However, this often ends up calling
> > > > cpumask_any_and() which is defined as cpumask_first_and() which has
> > > > the effect of alyways using the first cpu among those available.
> > > > 
> > > > The same applies when multiple NUMA nodes are involved. In that
> > > > case the first cpu in the local node is chosen which does provide
> > > > a bit of spreading but with multiple HK cpus per node the same
> > > > issues arise.
> > > > 
> > > > We have numerous cases where a single HK cpu just cannot keep up
> > > > and the remote_tick warning fires. It also can lead to the other
> > > > things (orchastration sw, HA keepalives etc) on the HK cpus getting
> > > > starved which leads to other issues.  In these cases we recommend
> > > > increasing the number of HK cpus.  But... that only helps the
> > > > userspace tasks somewhat. It does not help the actual housekeeping
> > > > part.
> > > > 
> > > > Spread the HK work out by having housekeeping_any_cpu() and
> > > > sched_numa_find_closest() use cpumask_any_and_distribute()
> > > > instead of cpumask_any_and().
> > > > 
> > > > Signed-off-by: Phil Auld <pauld@redhat.com>
> > > > Reviewed-by: Waiman Long <longman@redhat.com>
> > > > Cc: Peter Zijlstra <peterz@infradead.org>
> > > > Cc: Juri Lelli <juri.lelli@redhat.com>
> > > > Cc: Frederic Weisbecker <frederic@kernel.org>
> > > > Cc: Waiman Long <longman@redhat.com>
> > > > Cc: linux-kernel@vger.kernel.org
> > > > Link: https://lore.kernel.org/lkml/20250211141437.GA349314@pauld.westford.csb/
> > > Acked-by: Frederic Weisbecker <frederic@kernel.org>
> > > 
> > Thanks Frederic!
> > 
> > Anyone with commit powers willing to pick this up?
> 
> Is this patch eligible to be merged into sched/core for the current cycle or
> will have to wait until the next one?
> 
> Thanks,
> Longman
> 

This is the patch (in this thread somewhere...) I mentioned the other
day at OSPM. 

Thanks!


Cheers,
Phil

-- 

[tip: sched/core] sched/isolation: Make use of more than one housekeeping cpu

[tip-bot2 for Phil Auld]

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

Commit-ID:     6432e163ba1b7d80b5876792ce53e511f041ab91
Gitweb:        https://git.kernel.org/tip/6432e163ba1b7d80b5876792ce53e511f041ab91
Author:        Phil Auld <pauld@redhat.com>
AuthorDate:    Tue, 18 Feb 2025 18:46:18 
Committer:     Peter Zijlstra <peterz@infradead.org>
CommitterDate: Tue, 08 Apr 2025 20:55:55 +02:00

sched/isolation: Make use of more than one housekeeping cpu

The exising code uses housekeeping_any_cpu() to select a cpu for
a given housekeeping task. However, this often ends up calling
cpumask_any_and() which is defined as cpumask_first_and() which has
the effect of alyways using the first cpu among those available.

The same applies when multiple NUMA nodes are involved. In that
case the first cpu in the local node is chosen which does provide
a bit of spreading but with multiple HK cpus per node the same
issues arise.

We have numerous cases where a single HK cpu just cannot keep up
and the remote_tick warning fires. It also can lead to the other
things (orchastration sw, HA keepalives etc) on the HK cpus getting
starved which leads to other issues.  In these cases we recommend
increasing the number of HK cpus.  But... that only helps the
userspace tasks somewhat. It does not help the actual housekeeping
part.

Spread the HK work out by having housekeeping_any_cpu() and
sched_numa_find_closest() use cpumask_any_and_distribute()
instead of cpumask_any_and().

Signed-off-by: Phil Auld <pauld@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Waiman Long <longman@redhat.com>
Reviewed-by: Vishal Chourasia <vishalc@linux.ibm.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20250218184618.1331715-1-pauld@redhat.com
---
 kernel/sched/isolation.c | 2 +-
 kernel/sched/topology.c  | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/sched/isolation.c b/kernel/sched/isolation.c
index 81bc8b3..93b038d 100644
--- a/kernel/sched/isolation.c
+++ b/kernel/sched/isolation.c
@@ -40,7 +40,7 @@ int housekeeping_any_cpu(enum hk_type type)
 			if (cpu < nr_cpu_ids)
 				return cpu;
 
-			cpu = cpumask_any_and(housekeeping.cpumasks[type], cpu_online_mask);
+			cpu = cpumask_any_and_distribute(housekeeping.cpumasks[type], cpu_online_mask);
 			if (likely(cpu < nr_cpu_ids))
 				return cpu;
 			/*
diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
index b334f25..bbc2fc2 100644
--- a/kernel/sched/topology.c
+++ b/kernel/sched/topology.c
@@ -2098,7 +2098,7 @@ int sched_numa_find_closest(const struct cpumask *cpus, int cpu)
 	for (i = 0; i < sched_domains_numa_levels; i++) {
 		if (!masks[i][j])
 			break;
-		cpu = cpumask_any_and(cpus, masks[i][j]);
+		cpu = cpumask_any_and_distribute(cpus, masks[i][j]);
 		if (cpu < nr_cpu_ids) {
 			found = cpu;
 			break;