Re: [PATCH v2 1/3] workqueue: Provide a handshake for canceling BH workers

[Lai Jiangshan]

On Tue, Sep 2, 2025 at 12:38 AM Sebastian Andrzej Siewior
<bigeasy@xxxxxxxxxxxxx> wrote:

> @@ -4222,17 +4251,18 @@ static bool __flush_work(struct work_struct *work, bool from_cancel)
>                     (data & WORK_OFFQ_BH)) {
>                         /*
>                          * On RT, prevent a live lock when %current preempted
> -                        * soft interrupt processing or prevents ksoftirqd from
> -                        * running by keeping flipping BH. If the BH work item
> -                        * runs on a different CPU then this has no effect other
> -                        * than doing the BH disable/enable dance for nothing.
> -                        * This is copied from
> -                        * kernel/softirq.c::tasklet_unlock_spin_wait().
> +                        * soft interrupt processing by blocking on lock which
> +                        * is owned by the thread invoking the callback.
>                          */
>                         while (!try_wait_for_completion(&barr.done)) {
>                                 if (IS_ENABLED(CONFIG_PREEMPT_RT)) {
> -                                       local_bh_disable();
> -                                       local_bh_enable();
> +                                       struct worker_pool *pool;
> +
> +                                       mutex_lock(&wq_pool_mutex);
> +                                       pool = get_work_pool(work);
> +                                       if (pool)
> +                                               workqueue_callback_cancel_wait_running(pool);
> +                                       mutex_unlock(&wq_pool_mutex);

The goal is to avoid using a potentially sleeping function in
__flush_work() for BH work items on PREEMPT_RT, but
mutex_lock(&wq_pool_mutex) is not appropriate in this context.

To obtain the pool of a work item, the preferred approach is to use
rcu_read_lock() together with get_work_pool(work), as is done in
start_flush_work().


>                                 } else {
>                                         cpu_relax();
>                                 }