Date:   Thu, 16 Mar 2023 08:08:50 +0100
From:   Daniel Wagner <wagi@monom.org>
To:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc:     "Brennan Lamoreaux (VMware)" <brennanlamoreaux@gmail.com>,
        linux-rt-users@vger.kernel.org, linux-kernel@vger.kernel.org,
        blamoreaux@vmware.com, frederic.martinsons@gmail.com,
        srivatsa@csail.mit.edu, vsirnapalli@vmware.com,
        amakhalov@vmware.com, keerthanak@vmware.com, ankitja@vmware.com,
        bordoloih@vmware.com, srivatsab@vmware.com,
        Tejun Heo <tj@kernel.org>
Subject: Re: [PATCH 4.19-rt] workqueue: Fix deadlock due to recursive locking
 of pool->lock
Message-ID: <20230316070850.q3bj3g7likfbg7jl@beryllium.lan>
References: <20230228224938.88035-1-brennanlamoreaux@gmail.com>
 <20230313093641.pxvwFHCJ@linutronix.de>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20230313093641.pxvwFHCJ@linutronix.de>
Precedence: bulk

On Mon, Mar 13, 2023 at 10:36:41AM +0100, Sebastian Andrzej Siewior wrote:
> On 2023-02-28 14:49:38 [-0800], Brennan Lamoreaux (VMware) wrote:
> > Upstream commit d8bb65ab70f7 ("workqueue: Use rcuwait for wq_manager_wait")
> > replaced the waitqueue with rcuwait in the workqueue code. This change
> > involved removing the acquisition of pool->lock in put_unbound_pool(),
> > as it also adds the function wq_manager_inactive() which acquires this same
> > lock and is called one line later as a parameter to rcu_wait_event().
> 
> Daniel, I double checked and this patch is correct - the backport was
> faulty. Could you please pick it up and release an update?

Sure. I've updated the v4.19-rt branch and added this patch. Running local tests
now.