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: <20201226025117.2770-1-jiangshanlai@gmail.com> <20201226025117.2770-4-jiangshanlai@gmail.com>
 <20210104135649.GO3021@hirez.programming.kicks-ass.net>
In-Reply-To: <20210104135649.GO3021@hirez.programming.kicks-ass.net>
From:   Lai Jiangshan <jiangshanlai@gmail.com>
Date:   Tue, 5 Jan 2021 10:41:07 +0800
Message-ID: <CAJhGHyB_MUHG8GGANcN9sQbjY7M5m8WPHQgXp-PmkGK481M5Tg@mail.gmail.com>
Subject: Re: [PATCH -tip V3 3/8] workqueue: introduce wq_online_cpumask
To:     Peter Zijlstra <peterz@infradead.org>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        Valentin Schneider <valentin.schneider@arm.com>,
        Qian Cai <cai@redhat.com>,
        Vincent Donnefort <vincent.donnefort@arm.com>,
        Dexuan Cui <decui@microsoft.com>,
        Lai Jiangshan <laijs@linux.alibaba.com>,
        Tejun Heo <tj@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Jan 4, 2021 at 9:56 PM Peter Zijlstra <peterz@infradead.org> wrote:
>
> On Sat, Dec 26, 2020 at 10:51:11AM +0800, Lai Jiangshan wrote:
> > From: Lai Jiangshan <laijs@linux.alibaba.com>
> >
> > wq_online_cpumask is the cached result of cpu_online_mask with the
> > going-down cpu cleared.
>
> You can't use cpu_active_mask ?


When a cpu is going out:
(cpu_active_mask is not protected by workqueue mutexs.)

create_worker() for unbound pool  |  cpu offlining
check cpu_active_mask             |
                                  |  remove bit from cpu_active_mask
                                  |  no cpu in pool->attrs->cpumask is active
set pool->attrs->cpumask to worker|
and hit the warning


And when a cpu is onlining, there may be some workers which were just created
after the workqueue hotplug callback is finished but before cpu_active_mask
was updated. workqueue has not call back after cpu_active_mask updated and
these workers' cpumask is not updated.

For percpu workers, these problems can be handled with the help of
POOL_DISASSOCIATED which is protected by workqueue mutexs and the
help of sched/core.c which doesn't warn when per-cpu-kthread.

For unbound workers, the way to handle it without using wq_online_cpumask
is much more complex when a cpu is going out.