[PATCH -next v2 0/3] cpuset: code cleanups

[PATCH -next v2 0/3] cpuset: code cleanups

[Chen Ridong]

From: Chen Ridong <chenridong@huawei.com>

Patch 1 simplifies the error handling path in cpuset_set_nodes() by
returning directly on failure, eliminating an unnecessary jump.

Patch 2 removes the global remote_children list and replaces it with
a boolean remote_partition flag, which provides a more direct way
to identify remote partitions.

Patch 3 removes need_rebuild_sched_domains.

---

v2: Patch 2 moves up 'remote_partition' and removes redundant
    initialization.

Chen Ridong (3):
  cpuset: simplify node setting on error
  cpuset: remove global remote_children list
  cpuset: remove need_rebuild_sched_domains

 kernel/cgroup/cpuset-internal.h | 10 ++++++---
 kernel/cgroup/cpuset.c          | 40 ++++++++++++---------------------
 2 files changed, 21 insertions(+), 29 deletions(-)

-- 
2.34.1

[PATCH -next v2 1/3] cpuset: simplify node setting on error

[Chen Ridong]

From: Chen Ridong <chenridong@huawei.com>

There is no need to jump to the 'done' label upon failure, as no cleanup
is required. Return the error code directly instead.

Signed-off-by: Chen Ridong <chenridong@huawei.com>
Reviewed-by: Waiman Long <longman@redhat.com>
Reviewed-by: Michal Koutný <mkoutny@suse.com>
---
 kernel/cgroup/cpuset.c | 21 +++++++++------------
 1 file changed, 9 insertions(+), 12 deletions(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 8238fd8c0c29..c90476d52f09 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -2897,21 +2897,19 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs,
 	 */
 	retval = nodelist_parse(buf, trialcs->mems_allowed);
 	if (retval < 0)
-		goto done;
+		return retval;
 
 	if (!nodes_subset(trialcs->mems_allowed,
-			  top_cpuset.mems_allowed)) {
-		retval = -EINVAL;
-		goto done;
-	}
+			  top_cpuset.mems_allowed))
+		return -EINVAL;
+
+	/* No change? nothing to do */
+	if (nodes_equal(cs->mems_allowed, trialcs->mems_allowed))
+		return 0;
 
-	if (nodes_equal(cs->mems_allowed, trialcs->mems_allowed)) {
-		retval = 0;		/* Too easy - nothing to do */
-		goto done;
-	}
 	retval = validate_change(cs, trialcs);
 	if (retval < 0)
-		goto done;
+		return retval;
 
 	check_insane_mems_config(&trialcs->mems_allowed);
 
@@ -2921,8 +2919,7 @@ static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs,
 
 	/* use trialcs->mems_allowed as a temp variable */
 	update_nodemasks_hier(cs, &trialcs->mems_allowed);
-done:
-	return retval;
+	return 0;
 }
 
 bool current_cpuset_is_being_rebound(void)
-- 
2.34.1

[PATCH -next v2 2/3] cpuset: remove global remote_children list

[Chen Ridong]

From: Chen Ridong <chenridong@huawei.com>

The remote_children list is used to track all remote partitions attached
to a cpuset. However, it serves no other purpose. Using a boolean flag to
indicate whether a cpuset is a remote partition is a more direct approach,
making remote_children unnecessary.

This patch replaces the list with a remote_partition flag in the cpuset
structure and removes remote_children entirely.

Signed-off-by: Chen Ridong <chenridong@huawei.com>
---
 kernel/cgroup/cpuset-internal.h | 10 +++++++---
 kernel/cgroup/cpuset.c          | 13 ++++---------
 2 files changed, 11 insertions(+), 12 deletions(-)

diff --git a/kernel/cgroup/cpuset-internal.h b/kernel/cgroup/cpuset-internal.h
index 5cac42c5fd97..01976c8e7d49 100644
--- a/kernel/cgroup/cpuset-internal.h
+++ b/kernel/cgroup/cpuset-internal.h
@@ -158,6 +158,13 @@ struct cpuset {
 	/* partition root state */
 	int partition_root_state;
 
+	/*
+	 * Whether cpuset is a remote partition.
+	 * It used to be a list anchoring all remote partitions — we can switch back
+	 * to a list if we need to iterate over the remote partitions.
+	 */
+	bool remote_partition;
+
 	/*
 	 * number of SCHED_DEADLINE tasks attached to this cpuset, so that we
 	 * know when to rebuild associated root domain bandwidth information.
@@ -172,9 +179,6 @@ struct cpuset {
 	/* Handle for cpuset.cpus.partition */
 	struct cgroup_file partition_file;
 
-	/* Remote partition silbling list anchored at remote_children */
-	struct list_head remote_sibling;
-
 	/* Used to merge intersecting subsets for generate_sched_domains */
 	struct uf_node node;
 };
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index c90476d52f09..aff3ddc67393 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -94,9 +94,6 @@ static bool isolated_cpus_updating;
 static cpumask_var_t	boot_hk_cpus;
 static bool		have_boot_isolcpus;
 
-/* List of remote partition root children */
-static struct list_head remote_children;
-
 /*
  * A flag to force sched domain rebuild at the end of an operation.
  * It can be set in
@@ -219,7 +216,7 @@ static struct cpuset top_cpuset = {
 		 BIT(CS_MEM_EXCLUSIVE) | BIT(CS_SCHED_LOAD_BALANCE),
 	.partition_root_state = PRS_ROOT,
 	.relax_domain_level = -1,
-	.remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
+	.remote_partition = false,
 };
 
 /*
@@ -1572,7 +1569,7 @@ static int compute_trialcs_excpus(struct cpuset *trialcs, struct cpuset *cs)
 
 static inline bool is_remote_partition(struct cpuset *cs)
 {
-	return !list_empty(&cs->remote_sibling);
+	return cs->remote_partition;
 }
 
 static inline bool is_local_partition(struct cpuset *cs)
@@ -1621,7 +1618,7 @@ static int remote_partition_enable(struct cpuset *cs, int new_prs,
 
 	spin_lock_irq(&callback_lock);
 	partition_xcpus_add(new_prs, NULL, tmp->new_cpus);
-	list_add(&cs->remote_sibling, &remote_children);
+	cs->remote_partition = true;
 	cpumask_copy(cs->effective_xcpus, tmp->new_cpus);
 	spin_unlock_irq(&callback_lock);
 	update_isolation_cpumasks();
@@ -1651,7 +1648,7 @@ static void remote_partition_disable(struct cpuset *cs, struct tmpmasks *tmp)
 	WARN_ON_ONCE(!cpumask_subset(cs->effective_xcpus, subpartitions_cpus));
 
 	spin_lock_irq(&callback_lock);
-	list_del_init(&cs->remote_sibling);
+	cs->remote_partition = false;
 	partition_xcpus_del(cs->partition_root_state, NULL, cs->effective_xcpus);
 	if (cs->prs_err)
 		cs->partition_root_state = -cs->partition_root_state;
@@ -3603,7 +3600,6 @@ cpuset_css_alloc(struct cgroup_subsys_state *parent_css)
 	__set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
 	fmeter_init(&cs->fmeter);
 	cs->relax_domain_level = -1;
-	INIT_LIST_HEAD(&cs->remote_sibling);
 
 	/* Set CS_MEMORY_MIGRATE for default hierarchy */
 	if (cpuset_v2())
@@ -3874,7 +3870,6 @@ int __init cpuset_init(void)
 	nodes_setall(top_cpuset.effective_mems);
 
 	fmeter_init(&top_cpuset.fmeter);
-	INIT_LIST_HEAD(&remote_children);
 
 	BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
 
-- 
2.34.1

Re: [PATCH -next v2 2/3] cpuset: remove global remote_children list

[Waiman Long]

On 11/11/25 8:24 AM, Chen Ridong wrote:
> From: Chen Ridong <chenridong@huawei.com>
>
> The remote_children list is used to track all remote partitions attached
> to a cpuset. However, it serves no other purpose. Using a boolean flag to
> indicate whether a cpuset is a remote partition is a more direct approach,
> making remote_children unnecessary.
>
> This patch replaces the list with a remote_partition flag in the cpuset
> structure and removes remote_children entirely.
>
> Signed-off-by: Chen Ridong <chenridong@huawei.com>
> ---
>   kernel/cgroup/cpuset-internal.h | 10 +++++++---
>   kernel/cgroup/cpuset.c          | 13 ++++---------
>   2 files changed, 11 insertions(+), 12 deletions(-)
>
> diff --git a/kernel/cgroup/cpuset-internal.h b/kernel/cgroup/cpuset-internal.h
> index 5cac42c5fd97..01976c8e7d49 100644
> --- a/kernel/cgroup/cpuset-internal.h
> +++ b/kernel/cgroup/cpuset-internal.h
> @@ -158,6 +158,13 @@ struct cpuset {
>   	/* partition root state */
>   	int partition_root_state;
>   
> +	/*
> +	 * Whether cpuset is a remote partition.
> +	 * It used to be a list anchoring all remote partitions — we can switch back
> +	 * to a list if we need to iterate over the remote partitions.
> +	 */
> +	bool remote_partition;
> +
>   	/*
>   	 * number of SCHED_DEADLINE tasks attached to this cpuset, so that we
>   	 * know when to rebuild associated root domain bandwidth information.
> @@ -172,9 +179,6 @@ struct cpuset {
>   	/* Handle for cpuset.cpus.partition */
>   	struct cgroup_file partition_file;
>   
> -	/* Remote partition silbling list anchored at remote_children */
> -	struct list_head remote_sibling;
> -
>   	/* Used to merge intersecting subsets for generate_sched_domains */
>   	struct uf_node node;
>   };
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index c90476d52f09..aff3ddc67393 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -94,9 +94,6 @@ static bool isolated_cpus_updating;
>   static cpumask_var_t	boot_hk_cpus;
>   static bool		have_boot_isolcpus;
>   
> -/* List of remote partition root children */
> -static struct list_head remote_children;
> -
>   /*
>    * A flag to force sched domain rebuild at the end of an operation.
>    * It can be set in
> @@ -219,7 +216,7 @@ static struct cpuset top_cpuset = {
>   		 BIT(CS_MEM_EXCLUSIVE) | BIT(CS_SCHED_LOAD_BALANCE),
>   	.partition_root_state = PRS_ROOT,
>   	.relax_domain_level = -1,
> -	.remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
> +	.remote_partition = false,
I forgot to notify you that this init is also not needed. Anyway, this 
is a minor issue.
>   };
>   
>   /*
> @@ -1572,7 +1569,7 @@ static int compute_trialcs_excpus(struct cpuset *trialcs, struct cpuset *cs)
>   
>   static inline bool is_remote_partition(struct cpuset *cs)
>   {
> -	return !list_empty(&cs->remote_sibling);
> +	return cs->remote_partition;
>   }
>   
>   static inline bool is_local_partition(struct cpuset *cs)
> @@ -1621,7 +1618,7 @@ static int remote_partition_enable(struct cpuset *cs, int new_prs,
>   
>   	spin_lock_irq(&callback_lock);
>   	partition_xcpus_add(new_prs, NULL, tmp->new_cpus);
> -	list_add(&cs->remote_sibling, &remote_children);
> +	cs->remote_partition = true;
>   	cpumask_copy(cs->effective_xcpus, tmp->new_cpus);
>   	spin_unlock_irq(&callback_lock);
>   	update_isolation_cpumasks();
> @@ -1651,7 +1648,7 @@ static void remote_partition_disable(struct cpuset *cs, struct tmpmasks *tmp)
>   	WARN_ON_ONCE(!cpumask_subset(cs->effective_xcpus, subpartitions_cpus));
>   
>   	spin_lock_irq(&callback_lock);
> -	list_del_init(&cs->remote_sibling);
> +	cs->remote_partition = false;
>   	partition_xcpus_del(cs->partition_root_state, NULL, cs->effective_xcpus);
>   	if (cs->prs_err)
>   		cs->partition_root_state = -cs->partition_root_state;
> @@ -3603,7 +3600,6 @@ cpuset_css_alloc(struct cgroup_subsys_state *parent_css)
>   	__set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
>   	fmeter_init(&cs->fmeter);
>   	cs->relax_domain_level = -1;
> -	INIT_LIST_HEAD(&cs->remote_sibling);
>   
>   	/* Set CS_MEMORY_MIGRATE for default hierarchy */
>   	if (cpuset_v2())
> @@ -3874,7 +3870,6 @@ int __init cpuset_init(void)
>   	nodes_setall(top_cpuset.effective_mems);
>   
>   	fmeter_init(&top_cpuset.fmeter);
> -	INIT_LIST_HEAD(&remote_children);
>   
>   	BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
>   
Reviewed-by: Waiman Long <longman@redhat.com>

Re: [PATCH -next v2 2/3] cpuset: remove global remote_children list

[Chen Ridong]

On 2025/11/11 22:56, Waiman Long wrote:
> On 11/11/25 8:24 AM, Chen Ridong wrote:
>> From: Chen Ridong <chenridong@huawei.com>
>>
>> The remote_children list is used to track all remote partitions attached
>> to a cpuset. However, it serves no other purpose. Using a boolean flag to
>> indicate whether a cpuset is a remote partition is a more direct approach,
>> making remote_children unnecessary.
>>
>> This patch replaces the list with a remote_partition flag in the cpuset
>> structure and removes remote_children entirely.
>>
>> Signed-off-by: Chen Ridong <chenridong@huawei.com>
>> ---
>>   kernel/cgroup/cpuset-internal.h | 10 +++++++---
>>   kernel/cgroup/cpuset.c          | 13 ++++---------
>>   2 files changed, 11 insertions(+), 12 deletions(-)
>>
>> diff --git a/kernel/cgroup/cpuset-internal.h b/kernel/cgroup/cpuset-internal.h
>> index 5cac42c5fd97..01976c8e7d49 100644
>> --- a/kernel/cgroup/cpuset-internal.h
>> +++ b/kernel/cgroup/cpuset-internal.h
>> @@ -158,6 +158,13 @@ struct cpuset {
>>       /* partition root state */
>>       int partition_root_state;
>>   +    /*
>> +     * Whether cpuset is a remote partition.
>> +     * It used to be a list anchoring all remote partitions — we can switch back
>> +     * to a list if we need to iterate over the remote partitions.
>> +     */
>> +    bool remote_partition;
>> +
>>       /*
>>        * number of SCHED_DEADLINE tasks attached to this cpuset, so that we
>>        * know when to rebuild associated root domain bandwidth information.
>> @@ -172,9 +179,6 @@ struct cpuset {
>>       /* Handle for cpuset.cpus.partition */
>>       struct cgroup_file partition_file;
>>   -    /* Remote partition silbling list anchored at remote_children */
>> -    struct list_head remote_sibling;
>> -
>>       /* Used to merge intersecting subsets for generate_sched_domains */
>>       struct uf_node node;
>>   };
>> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
>> index c90476d52f09..aff3ddc67393 100644
>> --- a/kernel/cgroup/cpuset.c
>> +++ b/kernel/cgroup/cpuset.c
>> @@ -94,9 +94,6 @@ static bool isolated_cpus_updating;
>>   static cpumask_var_t    boot_hk_cpus;
>>   static bool        have_boot_isolcpus;
>>   -/* List of remote partition root children */
>> -static struct list_head remote_children;
>> -
>>   /*
>>    * A flag to force sched domain rebuild at the end of an operation.
>>    * It can be set in
>> @@ -219,7 +216,7 @@ static struct cpuset top_cpuset = {
>>            BIT(CS_MEM_EXCLUSIVE) | BIT(CS_SCHED_LOAD_BALANCE),
>>       .partition_root_state = PRS_ROOT,
>>       .relax_domain_level = -1,
>> -    .remote_sibling = LIST_HEAD_INIT(top_cpuset.remote_sibling),
>> +    .remote_partition = false,
> I forgot to notify you that this init is also not needed. Anyway, this is a minor issue.

Hmm, I should have noticed that. This could be fixed if we send a related patch in the future, but I
don't think it's worth submitting a standalone patch.

>>   };
>>     /*
>> @@ -1572,7 +1569,7 @@ static int compute_trialcs_excpus(struct cpuset *trialcs, struct cpuset *cs)
>>     static inline bool is_remote_partition(struct cpuset *cs)
>>   {
>> -    return !list_empty(&cs->remote_sibling);
>> +    return cs->remote_partition;
>>   }
>>     static inline bool is_local_partition(struct cpuset *cs)
>> @@ -1621,7 +1618,7 @@ static int remote_partition_enable(struct cpuset *cs, int new_prs,
>>         spin_lock_irq(&callback_lock);
>>       partition_xcpus_add(new_prs, NULL, tmp->new_cpus);
>> -    list_add(&cs->remote_sibling, &remote_children);
>> +    cs->remote_partition = true;
>>       cpumask_copy(cs->effective_xcpus, tmp->new_cpus);
>>       spin_unlock_irq(&callback_lock);
>>       update_isolation_cpumasks();
>> @@ -1651,7 +1648,7 @@ static void remote_partition_disable(struct cpuset *cs, struct tmpmasks *tmp)
>>       WARN_ON_ONCE(!cpumask_subset(cs->effective_xcpus, subpartitions_cpus));
>>         spin_lock_irq(&callback_lock);
>> -    list_del_init(&cs->remote_sibling);
>> +    cs->remote_partition = false;
>>       partition_xcpus_del(cs->partition_root_state, NULL, cs->effective_xcpus);
>>       if (cs->prs_err)
>>           cs->partition_root_state = -cs->partition_root_state;
>> @@ -3603,7 +3600,6 @@ cpuset_css_alloc(struct cgroup_subsys_state *parent_css)
>>       __set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
>>       fmeter_init(&cs->fmeter);
>>       cs->relax_domain_level = -1;
>> -    INIT_LIST_HEAD(&cs->remote_sibling);
>>         /* Set CS_MEMORY_MIGRATE for default hierarchy */
>>       if (cpuset_v2())
>> @@ -3874,7 +3870,6 @@ int __init cpuset_init(void)
>>       nodes_setall(top_cpuset.effective_mems);
>>         fmeter_init(&top_cpuset.fmeter);
>> -    INIT_LIST_HEAD(&remote_children);
>>         BUG_ON(!alloc_cpumask_var(&cpus_attach, GFP_KERNEL));
>>   
> Reviewed-by: Waiman Long <longman@redhat.com>
> 

Thanks.

-- 
Best regards,
Ridong

[PATCH -next v2 3/3] cpuset: remove need_rebuild_sched_domains

[Chen Ridong]

From: Chen Ridong <chenridong@huawei.com>

Previously, update_cpumasks_hier() used need_rebuild_sched_domains to
decide whether to invoke rebuild_sched_domains_locked(). Now that
rebuild_sched_domains_locked() only sets force_rebuild, the flag is
redundant. Hence, remove it.

Signed-off-by: Chen Ridong <chenridong@huawei.com>
Reviewed-by: Waiman Long <longman@redhat.com>
---
 kernel/cgroup/cpuset.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index aff3ddc67393..daf813386260 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -2184,7 +2184,6 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
 {
 	struct cpuset *cp;
 	struct cgroup_subsys_state *pos_css;
-	bool need_rebuild_sched_domains = false;
 	int old_prs, new_prs;
 
 	rcu_read_lock();
@@ -2348,15 +2347,12 @@ static void update_cpumasks_hier(struct cpuset *cs, struct tmpmasks *tmp,
 		if (!cpumask_empty(cp->cpus_allowed) &&
 		    is_sched_load_balance(cp) &&
 		   (!cpuset_v2() || is_partition_valid(cp)))
-			need_rebuild_sched_domains = true;
+			cpuset_force_rebuild();
 
 		rcu_read_lock();
 		css_put(&cp->css);
 	}
 	rcu_read_unlock();
-
-	if (need_rebuild_sched_domains)
-		cpuset_force_rebuild();
 }
 
 /**
-- 
2.34.1

Re: [PATCH -next v2 0/3] cpuset: code cleanups

[Tejun Heo]

> Chen Ridong (3):
>   cpuset: simplify node setting on error
>   cpuset: remove global remote_children list
>   cpuset: remove need_rebuild_sched_domains

Applied 1-3 to cgroup/for-6.19.

Thanks.

--
tejun