cpuset_cpus_allowed() uses a reader lock that is sleepable under RT,
which means it cannot be called inside raw_spin_lock_t context.

Introduce a new cpuset_cpus_allowed_locked() helper that performs the
same function as cpuset_cpus_allowed() except that the caller must have
acquired the cpuset_mutex so that no further locking will be needed.

Suggested-by: Waiman Long <longman@redhat.com>
Signed-off-by: Pingfan Liu <piliu@redhat.com>
Cc: Waiman Long <longman@redhat.com>
Cc: Tejun Heo <tj@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: "Michal Koutný" <mkoutny@suse.com>
Cc: linux-kernel@vger.kernel.org
To: cgroups@vger.kernel.org
---
 include/linux/cpuset.h |  9 +++++++-
 kernel/cgroup/cpuset.c | 51 +++++++++++++++++++++++++++++-------------
 2 files changed, 44 insertions(+), 16 deletions(-)

diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
index 2ddb256187b51..a98d3330385c2 100644
--- a/include/linux/cpuset.h
+++ b/include/linux/cpuset.h
@@ -74,6 +74,7 @@ extern void inc_dl_tasks_cs(struct task_struct *task);
 extern void dec_dl_tasks_cs(struct task_struct *task);
 extern void cpuset_lock(void);
 extern void cpuset_unlock(void);
+extern void cpuset_cpus_allowed_locked(struct task_struct *p, struct cpumask *mask);
 extern void cpuset_cpus_allowed(struct task_struct *p, struct cpumask *mask);
 extern bool cpuset_cpus_allowed_fallback(struct task_struct *p);
 extern bool cpuset_cpu_is_isolated(int cpu);
@@ -195,10 +196,16 @@ static inline void dec_dl_tasks_cs(struct task_struct *task) { }
 static inline void cpuset_lock(void) { }
 static inline void cpuset_unlock(void) { }
 
+static inline void cpuset_cpus_allowed_locked(struct task_struct *p,
+					struct cpumask *mask)
+{
+	cpumask_copy(mask, task_cpu_possible_mask(p));
+}
+
 static inline void cpuset_cpus_allowed(struct task_struct *p,
 				       struct cpumask *mask)
 {
-	cpumask_copy(mask, task_cpu_possible_mask(p));
+	cpuset_cpus_allowed_locked(p, mask);
 }
 
 static inline bool cpuset_cpus_allowed_fallback(struct task_struct *p)
diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
index 52468d2c178a3..7a179a1a2e30a 100644
--- a/kernel/cgroup/cpuset.c
+++ b/kernel/cgroup/cpuset.c
@@ -4116,24 +4116,13 @@ void __init cpuset_init_smp(void)
 	BUG_ON(!cpuset_migrate_mm_wq);
 }
 
-/**
- * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset.
- * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
- * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
- *
- * Description: Returns the cpumask_var_t cpus_allowed of the cpuset
- * attached to the specified @tsk.  Guaranteed to return some non-empty
- * subset of cpu_active_mask, even if this means going outside the
- * tasks cpuset, except when the task is in the top cpuset.
- **/
-
-void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
+/*
+ * Return cpus_allowed mask from a task's cpuset.
+ */
+static void __cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask)
 {
-	unsigned long flags;
 	struct cpuset *cs;
 
-	spin_lock_irqsave(&callback_lock, flags);
-
 	cs = task_cs(tsk);
 	if (cs != &top_cpuset)
 		guarantee_active_cpus(tsk, pmask);
@@ -4153,7 +4142,39 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
 		if (!cpumask_intersects(pmask, cpu_active_mask))
 			cpumask_copy(pmask, possible_mask);
 	}
+}
 
+/**
+ * cpuset_cpus_allowed_locked - return cpus_allowed mask from a task's cpuset.
+ * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
+ * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
+ *
+ * Similir to cpuset_cpus_allowed() except that the caller must have acquired
+ * cpuset_mutex.
+ */
+void cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask)
+{
+	lockdep_assert_held(&cpuset_mutex);
+	__cpuset_cpus_allowed_locked(tsk, pmask);
+}
+
+/**
+ * cpuset_cpus_allowed - return cpus_allowed mask from a task's cpuset.
+ * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
+ * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
+ *
+ * Description: Returns the cpumask_var_t cpus_allowed of the cpuset
+ * attached to the specified @tsk.  Guaranteed to return some non-empty
+ * subset of cpu_active_mask, even if this means going outside the
+ * tasks cpuset, except when the task is in the top cpuset.
+ **/
+
+void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
+{
+	unsigned long flags;
+
+	spin_lock_irqsave(&callback_lock, flags);
+	__cpuset_cpus_allowed_locked(tsk, pmask);
 	spin_unlock_irqrestore(&callback_lock, flags);
 }
 
-- 
2.49.0

On 2025/11/19 17:55, Pingfan Liu wrote:
> cpuset_cpus_allowed() uses a reader lock that is sleepable under RT,
> which means it cannot be called inside raw_spin_lock_t context.
> 
> Introduce a new cpuset_cpus_allowed_locked() helper that performs the
> same function as cpuset_cpus_allowed() except that the caller must have
> acquired the cpuset_mutex so that no further locking will be needed.
> 
> Suggested-by: Waiman Long <longman@redhat.com>
> Signed-off-by: Pingfan Liu <piliu@redhat.com>
> Cc: Waiman Long <longman@redhat.com>
> Cc: Tejun Heo <tj@kernel.org>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: "Michal Koutný" <mkoutny@suse.com>
> Cc: linux-kernel@vger.kernel.org
> To: cgroups@vger.kernel.org
> ---
>  include/linux/cpuset.h |  9 +++++++-
>  kernel/cgroup/cpuset.c | 51 +++++++++++++++++++++++++++++-------------
>  2 files changed, 44 insertions(+), 16 deletions(-)
> 
> diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
> index 2ddb256187b51..a98d3330385c2 100644
> --- a/include/linux/cpuset.h
> +++ b/include/linux/cpuset.h
> @@ -74,6 +74,7 @@ extern void inc_dl_tasks_cs(struct task_struct *task);
>  extern void dec_dl_tasks_cs(struct task_struct *task);
>  extern void cpuset_lock(void);
>  extern void cpuset_unlock(void);
> +extern void cpuset_cpus_allowed_locked(struct task_struct *p, struct cpumask *mask);
>  extern void cpuset_cpus_allowed(struct task_struct *p, struct cpumask *mask);
>  extern bool cpuset_cpus_allowed_fallback(struct task_struct *p);
>  extern bool cpuset_cpu_is_isolated(int cpu);
> @@ -195,10 +196,16 @@ static inline void dec_dl_tasks_cs(struct task_struct *task) { }
>  static inline void cpuset_lock(void) { }
>  static inline void cpuset_unlock(void) { }
>  
> +static inline void cpuset_cpus_allowed_locked(struct task_struct *p,
> +					struct cpumask *mask)
> +{
> +	cpumask_copy(mask, task_cpu_possible_mask(p));
> +}
> +
>  static inline void cpuset_cpus_allowed(struct task_struct *p,
>  				       struct cpumask *mask)
>  {
> -	cpumask_copy(mask, task_cpu_possible_mask(p));
> +	cpuset_cpus_allowed_locked(p, mask);
>  }
>  
>  static inline bool cpuset_cpus_allowed_fallback(struct task_struct *p)
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index 52468d2c178a3..7a179a1a2e30a 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -4116,24 +4116,13 @@ void __init cpuset_init_smp(void)
>  	BUG_ON(!cpuset_migrate_mm_wq);
>  }
>  
> -/**
> - * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset.
> - * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
> - * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
> - *
> - * Description: Returns the cpumask_var_t cpus_allowed of the cpuset
> - * attached to the specified @tsk.  Guaranteed to return some non-empty
> - * subset of cpu_active_mask, even if this means going outside the
> - * tasks cpuset, except when the task is in the top cpuset.
> - **/
> -
> -void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
> +/*
> + * Return cpus_allowed mask from a task's cpuset.
> + */
> +static void __cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask)
>  {
> -	unsigned long flags;
>  	struct cpuset *cs;
>  
> -	spin_lock_irqsave(&callback_lock, flags);
> -
>  	cs = task_cs(tsk);
>  	if (cs != &top_cpuset)
>  		guarantee_active_cpus(tsk, pmask);
> @@ -4153,7 +4142,39 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
>  		if (!cpumask_intersects(pmask, cpu_active_mask))
>  			cpumask_copy(pmask, possible_mask);
>  	}
> +}
>  
> +/**
> + * cpuset_cpus_allowed_locked - return cpus_allowed mask from a task's cpuset.
> + * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
> + * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
> + *
> + * Similir to cpuset_cpus_allowed() except that the caller must have acquired
> + * cpuset_mutex.
> + */
> +void cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask)
> +{
> +	lockdep_assert_held(&cpuset_mutex);
> +	__cpuset_cpus_allowed_locked(tsk, pmask);
> +}
> +
> +/**
> + * cpuset_cpus_allowed - return cpus_allowed mask from a task's cpuset.
> + * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
> + * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
> + *
> + * Description: Returns the cpumask_var_t cpus_allowed of the cpuset
> + * attached to the specified @tsk.  Guaranteed to return some non-empty
> + * subset of cpu_active_mask, even if this means going outside the
> + * tasks cpuset, except when the task is in the top cpuset.
> + **/
> +
> +void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
> +{
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&callback_lock, flags);
> +	__cpuset_cpus_allowed_locked(tsk, pmask);
>  	spin_unlock_irqrestore(&callback_lock, flags);
>  }
>  

LGTM

Reviewed-by: Chen Ridong <chenridong@huawei.com>

-- 
Best regards,
Ridong

On 11/19/25 4:55 AM, Pingfan Liu wrote:
> cpuset_cpus_allowed() uses a reader lock that is sleepable under RT,
> which means it cannot be called inside raw_spin_lock_t context.
>
> Introduce a new cpuset_cpus_allowed_locked() helper that performs the
> same function as cpuset_cpus_allowed() except that the caller must have
> acquired the cpuset_mutex so that no further locking will be needed.
>
> Suggested-by: Waiman Long <longman@redhat.com>
> Signed-off-by: Pingfan Liu <piliu@redhat.com>
> Cc: Waiman Long <longman@redhat.com>
> Cc: Tejun Heo <tj@kernel.org>
> Cc: Johannes Weiner <hannes@cmpxchg.org>
> Cc: "Michal Koutný" <mkoutny@suse.com>
> Cc: linux-kernel@vger.kernel.org
> To: cgroups@vger.kernel.org
> ---
>   include/linux/cpuset.h |  9 +++++++-
>   kernel/cgroup/cpuset.c | 51 +++++++++++++++++++++++++++++-------------
>   2 files changed, 44 insertions(+), 16 deletions(-)
>
> diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h
> index 2ddb256187b51..a98d3330385c2 100644
> --- a/include/linux/cpuset.h
> +++ b/include/linux/cpuset.h
> @@ -74,6 +74,7 @@ extern void inc_dl_tasks_cs(struct task_struct *task);
>   extern void dec_dl_tasks_cs(struct task_struct *task);
>   extern void cpuset_lock(void);
>   extern void cpuset_unlock(void);
> +extern void cpuset_cpus_allowed_locked(struct task_struct *p, struct cpumask *mask);
>   extern void cpuset_cpus_allowed(struct task_struct *p, struct cpumask *mask);
>   extern bool cpuset_cpus_allowed_fallback(struct task_struct *p);
>   extern bool cpuset_cpu_is_isolated(int cpu);
> @@ -195,10 +196,16 @@ static inline void dec_dl_tasks_cs(struct task_struct *task) { }
>   static inline void cpuset_lock(void) { }
>   static inline void cpuset_unlock(void) { }
>   
> +static inline void cpuset_cpus_allowed_locked(struct task_struct *p,
> +					struct cpumask *mask)
> +{
> +	cpumask_copy(mask, task_cpu_possible_mask(p));
> +}
> +
>   static inline void cpuset_cpus_allowed(struct task_struct *p,
>   				       struct cpumask *mask)
>   {
> -	cpumask_copy(mask, task_cpu_possible_mask(p));
> +	cpuset_cpus_allowed_locked(p, mask);
>   }
>   
>   static inline bool cpuset_cpus_allowed_fallback(struct task_struct *p)
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index 52468d2c178a3..7a179a1a2e30a 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -4116,24 +4116,13 @@ void __init cpuset_init_smp(void)
>   	BUG_ON(!cpuset_migrate_mm_wq);
>   }
>   
> -/**
> - * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset.
> - * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
> - * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
> - *
> - * Description: Returns the cpumask_var_t cpus_allowed of the cpuset
> - * attached to the specified @tsk.  Guaranteed to return some non-empty
> - * subset of cpu_active_mask, even if this means going outside the
> - * tasks cpuset, except when the task is in the top cpuset.
> - **/
> -
> -void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
> +/*
> + * Return cpus_allowed mask from a task's cpuset.
> + */
> +static void __cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask)
>   {
> -	unsigned long flags;
>   	struct cpuset *cs;
>   
> -	spin_lock_irqsave(&callback_lock, flags);
> -
>   	cs = task_cs(tsk);
>   	if (cs != &top_cpuset)
>   		guarantee_active_cpus(tsk, pmask);
> @@ -4153,7 +4142,39 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
>   		if (!cpumask_intersects(pmask, cpu_active_mask))
>   			cpumask_copy(pmask, possible_mask);
>   	}
> +}
>   
> +/**
> + * cpuset_cpus_allowed_locked - return cpus_allowed mask from a task's cpuset.
> + * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
> + * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
> + *
> + * Similir to cpuset_cpus_allowed() except that the caller must have acquired
> + * cpuset_mutex.
> + */
> +void cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask)
> +{
> +	lockdep_assert_held(&cpuset_mutex);
> +	__cpuset_cpus_allowed_locked(tsk, pmask);
> +}
> +
> +/**
> + * cpuset_cpus_allowed - return cpus_allowed mask from a task's cpuset.
> + * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
> + * @pmask: pointer to struct cpumask variable to receive cpus_allowed set.
> + *
> + * Description: Returns the cpumask_var_t cpus_allowed of the cpuset
> + * attached to the specified @tsk.  Guaranteed to return some non-empty
> + * subset of cpu_active_mask, even if this means going outside the
> + * tasks cpuset, except when the task is in the top cpuset.
> + **/
> +
> +void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
> +{
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&callback_lock, flags);
> +	__cpuset_cpus_allowed_locked(tsk, pmask);
>   	spin_unlock_irqrestore(&callback_lock, flags);
>   }
>   
Reviewed-by: Waiman Long <longman@redhat.com>