Re: [PATCH] cgroup/cpuset: avoid unneeded cpuset_mutex re-lock

[Waiman Long <longman@redhat.com>]

On 7/4/23 08:03, Miaohe Lin wrote:
> cpuset_mutex unlock and lock pair is only needed when transferring tasks
> out of empty cpuset. Avoid unneeded cpuset_mutex re-lock when !is_empty
> to save cpu cycles.
>
> Signed-off-by: Miaohe Lin <linmiaohe@huawei.com>
> ---
>   kernel/cgroup/cpuset.c | 9 ++++-----
>   1 file changed, 4 insertions(+), 5 deletions(-)
>
> diff --git a/kernel/cgroup/cpuset.c b/kernel/cgroup/cpuset.c
> index 601c40da8e03..e136269c152c 100644
> --- a/kernel/cgroup/cpuset.c
> +++ b/kernel/cgroup/cpuset.c
> @@ -3521,17 +3521,16 @@ hotplug_update_tasks_legacy(struct cpuset *cs,
>   	is_empty = cpumask_empty(cs->cpus_allowed) ||
>   		   nodes_empty(cs->mems_allowed);
>   
> -	mutex_unlock(&cpuset_mutex);
> -
>   	/*
>   	 * Move tasks to the nearest ancestor with execution resources,
>   	 * This is full cgroup operation which will also call back into
>   	 * cpuset. Should be done outside any lock.
>   	 */
> -	if (is_empty)
> +	if (is_empty) {
> +		mutex_unlock(&cpuset_mutex);
>   		remove_tasks_in_empty_cpuset(cs);
> -
> -	mutex_lock(&cpuset_mutex);
> +		mutex_lock(&cpuset_mutex);
> +	}
>   }
>   
>   static void
Reviewed-by: Waiman Long <longman@redhat.com>