[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <ZS-rCIejToOlJcqm@slm.duckdns.org>
Date: Tue, 17 Oct 2023 23:53:12 -1000
From: Tejun Heo <tj@...nel.org>
To: Frederic Weisbecker <frederic@...nel.org>
Cc: Lai Jiangshan <jiangshanlai@...il.com>,
LKML <linux-kernel@...r.kernel.org>,
"Paul E . McKenney" <paulmck@...nel.org>
Subject: Re: [PATCH] workqueue: Provide one lock class key per work_on_cpu()
callsite
On Sun, Sep 24, 2023 at 05:07:02PM +0200, Frederic Weisbecker wrote:
> All callers of work_on_cpu() share the same lock class key for all the
> functions queued. As a result the workqueue related locking scenario for
> a function A may be spuriously accounted as an inversion against the
> locking scenario of function B such as in the following model:
>
> long A(void *arg)
> {
> mutex_lock(&mutex);
> mutex_unlock(&mutex);
> }
>
> long B(void *arg)
> {
> }
>
> void launchA(void)
> {
> work_on_cpu(0, A, NULL);
> }
>
> void launchB(void)
> {
> mutex_lock(&mutex);
> work_on_cpu(1, B, NULL);
> mutex_unlock(&mutex);
> }
>
> launchA and launchB running concurrently have no chance to deadlock.
> However the above can be reported by lockdep as a possible locking
> inversion because the works containing A() and B() are treated as
> belonging to the same locking class.
Sorry about the delay. I missed this one. Applied to wq/for-6.7.
Thanks.
--
tejun
Powered by blists - more mailing lists