Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <ccd04d8a-154b-543e-e1c3-84bc655508d1@I-love.SAKURA.ne.jp>
Date:   Sat, 12 Feb 2022 14:31:53 +0900
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 6.3; Win64; x64; rv:91.0) Gecko/20100101
 Thunderbird/91.6.0
Subject: Re: [syzbot] possible deadlock in worker_thread
Content-Language: en-US
To:     Bart Van Assche <bvanassche@acm.org>,
        syzbot <syzbot+831661966588c802aae9@syzkaller.appspotmail.com>,
        jgg@ziepe.ca, linux-kernel@vger.kernel.org,
        linux-rdma@vger.kernel.org, syzkaller-bugs@googlegroups.com,
        Tejun Heo <tj@kernel.org>,
        Lai Jiangshan <jiangshanlai@gmail.com>
References: <0000000000005975a605d7aef05e@google.com>
 <8ea57ddf-a09c-43f2-4285-4dfb908ad967@acm.org>
From:   Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
In-Reply-To: <8ea57ddf-a09c-43f2-4285-4dfb908ad967@acm.org>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2022/02/12 3:59, Bart Van Assche wrote:
> On 2/10/22 11:27, syzbot wrote:
>> ======================================================
>> WARNING: possible circular locking dependency detected
>> 5.17.0-rc2-syzkaller-00398-gd8ad2ce873ab #0 Not tainted
>> ------------------------------------------------------
> 
> Since the SRP initiator driver is involved, I will take a look.
> However, I'm not sure yet when I will have the time to post a fix.
> 
> Thanks,
> 
> Bart.
> 

This problem was already handled by commit bf23747ee0532090 ("loop:
revert "make autoclear operation asynchronous"").

But this report might be suggesting us that we should consider
deprecating (and eventually getting rid of) system-wide workqueues
(declared in include/linux/workqueue.h), for since flush_workqueue()
synchronously waits for completion, sharing system-wide workqueues
among multiple modules can generate unexpected locking dependency
chain (like this report).

If some module needs flush_workqueue() or flush_*_work(), shouldn't
such module create and use their own workqueues?

Tejun, what do you think?