[PATCH 14/33] cpuset: Update HK_TYPE_DOMAIN cpumask from cpuset

[Chen Ridong]

On 2025/12/24 21:45, Frederic Weisbecker wrote:
> Until now, HK_TYPE_DOMAIN used to only include boot defined isolated
> CPUs passed through isolcpus= boot option. Users interested in also
> knowing the runtime defined isolated CPUs through cpuset must use
> different APIs: cpuset_cpu_is_isolated(), cpu_is_isolated(), etc...
> 
> There are many drawbacks to that approach:
> 
> 1) Most interested subsystems want to know about all isolated CPUs, not
>   just those defined on boot time.
> 
> 2) cpuset_cpu_is_isolated() / cpu_is_isolated() are not synchronized with
>   concurrent cpuset changes.
> 
> 3) Further cpuset modifications are not propagated to subsystems
> 
> Solve 1) and 2) and centralize all isolated CPUs within the
> HK_TYPE_DOMAIN housekeeping cpumask.
> 
> Subsystems can rely on RCU to synchronize against concurrent changes.
> 
> The propagation mentioned in 3) will be handled in further patches.
> 
> Signed-off-by: Frederic Weisbecker <frederic at kernel.org>
> ---
>  include/linux/sched/isolation.h |  7 +++
>  kernel/cgroup/cpuset.c          |  3 ++
>  kernel/sched/isolation.c        | 76 ++++++++++++++++++++++++++++++---
>  kernel/sched/sched.h            |  1 +
>  4 files changed, 81 insertions(+), 6 deletions(-)
> 
> diff --git a/include/linux/sched/isolation.h b/include/linux/sched/isolation.h
> index 109a2149e21a..6842a1ba4d13 100644
> --- a/include/linux/sched/isolation.h
> +++ b/include/linux/sched/isolation.h
> @@ -9,6 +9,11 @@
>  enum hk_type {
>  	/* Revert of boot-time isolcpus= argument */
>  	HK_TYPE_DOMAIN_BOOT,
> +	/*
> +	 * Same as HK_TYPE_DOMAIN_BOOT but also includes the
> +	 * revert of cpuset isolated partitions. As such it
> +	 * is always a subset of HK_TYPE_DOMAIN_BOOT.
> +	 */
>  	HK_TYPE_DOMAIN,
>  	/* Revert of boot-time isolcpus=managed_irq argument */
>  	HK_TYPE_MANAGED_IRQ,
> @@ -35,6 +40,7 @@ extern const struct cpumask *housekeeping_cpumask(enum hk_type type);
>  extern bool housekeeping_enabled(enum hk_type type);
>  extern void housekeeping_affine(struct task_struct *t, enum hk_type type);
>  extern bool housekeeping_test_cpu(int cpu, enum hk_type type);
> +extern int housekeeping_update(struct cpumask *isol_mask, enum hk_type type);
>  extern void __init housekeeping_init(void);
>  
>  #else
> @@ -62,6 +68,7 @@ static inline bool housekeeping_test_cpu(int cpu, enum hk_type type)
>  	return true;
>  }
>  
> +static inline int housekeeping_update(struct cpumask *isol_mask, enum hk_type type) { return 0; }
>  static inline void housekeeping_init(void) { }
>  #endif /* CONFIG_CPU_ISOLATION */
>  
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index 5e2e3514c22e..e13e32491ebf 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -1490,6 +1490,9 @@ static void update_isolation_cpumasks(void)
>  	ret = tmigr_isolated_exclude_cpumask(isolated_cpus);
>  	WARN_ON_ONCE(ret < 0);
>  
> +	ret = housekeeping_update(isolated_cpus, HK_TYPE_DOMAIN);
> +	WARN_ON_ONCE(ret < 0);
> +
>  	isolated_cpus_updating = false;
>  }
>  
> diff --git a/kernel/sched/isolation.c b/kernel/sched/isolation.c
> index 83be49ec2b06..a124f1119f2e 100644
> --- a/kernel/sched/isolation.c
> +++ b/kernel/sched/isolation.c
> @@ -29,18 +29,48 @@ static struct housekeeping housekeeping;
>  
>  bool housekeeping_enabled(enum hk_type type)
>  {
> -	return !!(housekeeping.flags & BIT(type));
> +	return !!(READ_ONCE(housekeeping.flags) & BIT(type));
>  }
>  EXPORT_SYMBOL_GPL(housekeeping_enabled);
>  
> +static bool housekeeping_dereference_check(enum hk_type type)
> +{
> +	if (IS_ENABLED(CONFIG_LOCKDEP) && type == HK_TYPE_DOMAIN) {
> +		/* Cpuset isn't even writable yet? */
> +		if (system_state <= SYSTEM_SCHEDULING)
> +			return true;
> +
> +		/* CPU hotplug write locked, so cpuset partition can't be overwritten */
> +		if (IS_ENABLED(CONFIG_HOTPLUG_CPU) && lockdep_is_cpus_write_held())
> +			return true;
> +
> +		/* Cpuset lock held, partitions not writable */
> +		if (IS_ENABLED(CONFIG_CPUSETS) && lockdep_is_cpuset_held())
> +			return true;
> +
> +		return false;
> +	}
> +
> +	return true;
> +}
> +
> +static inline struct cpumask *housekeeping_cpumask_dereference(enum hk_type type)
> +{
> +	return rcu_dereference_all_check(housekeeping.cpumasks[type],
> +					 housekeeping_dereference_check(type));
> +}
> +
>  const struct cpumask *housekeeping_cpumask(enum hk_type type)
>  {
> +	const struct cpumask *mask = NULL;
> +
>  	if (static_branch_unlikely(&housekeeping_overridden)) {
> -		if (housekeeping.flags & BIT(type)) {
> -			return rcu_dereference_check(housekeeping.cpumasks[type], 1);
> -		}
> +		if (READ_ONCE(housekeeping.flags) & BIT(type))
> +			mask = housekeeping_cpumask_dereference(type);
>  	}
> -	return cpu_possible_mask;
> +	if (!mask)
> +		mask = cpu_possible_mask;
> +	return mask;
>  }
>  EXPORT_SYMBOL_GPL(housekeeping_cpumask);
>  
> @@ -80,12 +110,46 @@ EXPORT_SYMBOL_GPL(housekeeping_affine);
>  
>  bool housekeeping_test_cpu(int cpu, enum hk_type type)
>  {
> -	if (static_branch_unlikely(&housekeeping_overridden) && housekeeping.flags & BIT(type))
> +	if (static_branch_unlikely(&housekeeping_overridden) &&
> +	    READ_ONCE(housekeeping.flags) & BIT(type))
>  		return cpumask_test_cpu(cpu, housekeeping_cpumask(type));
>  	return true;
>  }
>  EXPORT_SYMBOL_GPL(housekeeping_test_cpu);
>  
> +int housekeeping_update(struct cpumask *isol_mask, enum hk_type type)
> +{
> +	struct cpumask *trial, *old = NULL;
> +
> +	if (type != HK_TYPE_DOMAIN)
> +		return -ENOTSUPP;
> +

Nit:

The current if statement indicates that we only support modifying the cpumask for HK_TYPE_DOMAIN,
which makes the type argument seem unnecessary. This seems to be designed for better scalability.
However, when a new type needs to be supported in the future, this statement would have to be
removed. Also, the use of cpumask_andnot below is not a general operation.

Anyway, looks good to me.

> +	trial = kmalloc(cpumask_size(), GFP_KERNEL);
> +	if (!trial)
> +		return -ENOMEM;
> +
> +	cpumask_andnot(trial, housekeeping_cpumask(HK_TYPE_DOMAIN_BOOT), isol_mask);
> +	if (!cpumask_intersects(trial, cpu_online_mask)) {
> +		kfree(trial);
> +		return -EINVAL;
> +	}
> +
> +	if (!housekeeping.flags)
> +		static_branch_enable(&housekeeping_overridden);
> +
> +	if (housekeeping.flags & BIT(type))
> +		old = housekeeping_cpumask_dereference(type);
> +	else
> +		WRITE_ONCE(housekeeping.flags, housekeeping.flags | BIT(type));
> +	rcu_assign_pointer(housekeeping.cpumasks[type], trial);
> +
> +	synchronize_rcu();
> +
> +	kfree(old);
> +
> +	return 0;
> +}
> +
>  void __init housekeeping_init(void)
>  {
>  	enum hk_type type;
> diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h
> index 475bdab3b8db..653e898a996a 100644
> --- a/kernel/sched/sched.h
> +++ b/kernel/sched/sched.h
> @@ -30,6 +30,7 @@
>  #include <linux/context_tracking.h>
>  #include <linux/cpufreq.h>
>  #include <linux/cpumask_api.h>
> +#include <linux/cpuset.h>
>  #include <linux/ctype.h>
>  #include <linux/file.h>
>  #include <linux/fs_api.h>

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

-- 
Best regards,
Ridong