Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <3837a2ee-f3d5-2732-7d15-7e9616319f05@huawei.com>
In-Reply-To: <3837a2ee-f3d5-2732-7d15-7e9616319f05@huawei.com>
From:   Lai Jiangshan <jiangshanlai@gmail.com>
Date:   Fri, 20 Nov 2020 09:45:34 +0800
Message-ID: <CAJhGHyBBFebdb4=G35KaHWYCs+ASox+br7o-=nabXEm9Y1bqJA@mail.gmail.com>
Subject: Re: [PATCH v2] workqueue: Kick a worker based on the actual
 activation of delayed works
To:     Yunfeng Ye <yeyunfeng@huawei.com>
Cc:     Tejun Heo <tj@kernel.org>, LKML <linux-kernel@vger.kernel.org>,
        Shiyuan Hu <hushiyuan@huawei.com>,
        Hewenliang <hewenliang4@huawei.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, Nov 19, 2020 at 2:21 PM Yunfeng Ye <yeyunfeng@huawei.com> wrote:
>
> In realtime scenario, We do not want to have interference on the
> isolated cpu cores. but when invoking alloc_workqueue() for percpu wq
> on the housekeeping cpu, it kick a kworker on the isolated cpu.
>
>   alloc_workqueue
>     pwq_adjust_max_active
>       wake_up_worker
>
> The comment in pwq_adjust_max_active() said:
>   "Need to kick a worker after thawed or an unbound wq's
>    max_active is bumped"
>
> So it is unnecessary to kick a kworker for percpu's wq when invoking
> alloc_workqueue(). this patch only kick a worker based on the actual
> activation of delayed works.
>
> Signed-off-by: Yunfeng Ye <yeyunfeng@huawei.com>

Reviewed-by: Lai Jiangshan <jiangshanlai@gmail.com>

> ---
>  kernel/workqueue.c | 14 +++++++++++---
>  1 file changed, 11 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/workqueue.c b/kernel/workqueue.c
> index c41c3c17b86a..b3c9d6ef7c69 100644
> --- a/kernel/workqueue.c
> +++ b/kernel/workqueue.c
> @@ -3725,17 +3725,25 @@ static void pwq_adjust_max_active(struct pool_workqueue *pwq)
>          * is updated and visible.
>          */
>         if (!freezable || !workqueue_freezing) {
> +               bool kick = false;
> +
>                 pwq->max_active = wq->saved_max_active;
>
>                 while (!list_empty(&pwq->delayed_works) &&
> -                      pwq->nr_active < pwq->max_active)
> +                      pwq->nr_active < pwq->max_active) {
>                         pwq_activate_first_delayed(pwq);
> +                       kick = true;
> +               }
>
>                 /*
>                  * Need to kick a worker after thawed or an unbound wq's
> -                * max_active is bumped.  It's a slow path.  Do it always.
> +                * max_active is bumped. But in realtime scenario, kick a
> +                * worker always will have interference on the isolated
> +                * cpu cores. So do it only based on the actual activation
> +                * of delayed works.
>                  */
> -               wake_up_worker(pwq->pool);
> +               if (kick)
> +                       wake_up_worker(pwq->pool);
>         } else {
>                 pwq->max_active = 0;
>         }
> --
> 2.18.4