Re: [RFC PATCH v3 1/3] workqueue: Hold wq_pool_mutex while affining tasks to wq_unbound_cpumask

[Lai Jiangshan]

On Tue, Aug 2, 2022 at 4:42 PM Valentin Schneider <vschneid@xxxxxxxxxx> wrote:
>
> When unbind_workers() reads wq_unbound_cpumask to set the affinity of
> freshly-unbound kworkers, it only holds wq_pool_attach_mutex. This isn't
> sufficient as wq_unbound_cpumask is only protected by wq_pool_mutex.
>

Hello Valentin,

Updating wq_unbound_cpumask requires cpus_read_lock() and
unbind_workers() is in the CPU hotplug path and so it is sufficient to
access to wq_unbound_cpumask in unbind_workers().

The extra protection is only required when the logic is also moved to
destroy_worker().

Thanks
Lai