Received-SPF: pass (google.com: domain of linux-kernel+bounces-11680-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Sender: Tejun Heo <htejun@gmail.com>
Date: Wed, 27 Dec 2023 05:12:40 +0900
From: Tejun Heo <tj@kernel.org>
To: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: linux-kernel@vger.kernel.org, Naohiro.Aota@wdc.com,
	kernel-team@meta.com
Subject: Re: [PATCH 08/10] workqueue: Introduce struct wq_node_nr_active
Message-ID: <ZYszuHEX6Pb-K4Au@mtj.duckdns.org>
References: <20231220072529.1036099-1-tj@kernel.org>
 <20231220072529.1036099-9-tj@kernel.org>
 <CAJhGHyC_9JemZdv4Gp4XG7BqLc72bS1uM8xv9vpQtSsyhefZvQ@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: 8bit
In-Reply-To: <CAJhGHyC_9JemZdv4Gp4XG7BqLc72bS1uM8xv9vpQtSsyhefZvQ@mail.gmail.com>

On Tue, Dec 26, 2023 at 05:14:18PM +0800, Lai Jiangshan wrote:
> On Wed, Dec 20, 2023 at 3:26 PM Tejun Heo <tj@kernel.org> wrote:
> 
> > @@ -1498,12 +1539,21 @@ static bool pwq_tryinc_nr_active(struct pool_workqueue *pwq)
> >  {
> >         struct workqueue_struct *wq = pwq->wq;
> >         struct worker_pool *pool = pwq->pool;
> > -       bool obtained;
> > +       struct wq_node_nr_active *nna = wq_node_nr_active(wq, pool->node);
> > +       bool obtained = false;
> >
> >         lockdep_assert_held(&pool->lock);
> >
> > -       obtained = pwq->nr_active < wq->max_active;
> > +       if (!nna) {
> > +               /* per-cpu workqueue, pwq->nr_active is sufficient */
> > +               obtained = pwq->nr_active < wq->max_active;
> > +               goto out;
> > +       }
> 
> For unbound workqueue, it is not checked against wq->max_active anymore
> and it is increased unconditionally.  Is it by design?

Ah, I made a mistake while splitting the patches. This gets added by a later
patch but this step should have an explicit check against wq->max_active.
Lemme add a check for the unbound path.

Thanks.

-- 
tejun