[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHp75VfqT0FWX5TPjiDcP=w48HRJGZ-xK55FuatU1XZVOx8tKw@mail.gmail.com>
Date: Mon, 26 Jun 2023 09:53:49 +0300
From: Andy Shevchenko <andy.shevchenko@...il.com>
To: YE Chengfeng <cyeaa@...nect.ust.hk>
Cc: "linus.walleij@...aro.org" <linus.walleij@...aro.org>,
"brgl@...ev.pl" <brgl@...ev.pl>,
"andy@...nel.org" <andy@...nel.org>,
"linux-gpio@...r.kernel.org" <linux-gpio@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] gpio: gpiolib-cdev: Fix potential &lr->wait.lock deadlock issue
On Sun, Jun 25, 2023 at 5:45 PM YE Chengfeng <cyeaa@...nect.ust.hk> wrote:
>
> linereq_put_event is called from both interrupt context (e.g.,
> edge_irq_thread) and process context (process_hw_ts_thread).
> Therefore, interrupt should be disabled before acquiring lock
> &lr->wait.lock inside linereq_put_event to avoid deadlock when
> the lock is held in process context and edge_irq_thread comes.
>
> Similarly, linereq_read_unlocked running in process context
> also acquies the same lock. It also need to disable interrupt
acquires
> otherwise deadlock could happen if the irq edge_irq_thread
> comes to execution while the lock is held.
>
> Fix the two potential deadlock issues by spin_lock_irqsave.
Sounds legit to me, but
1) do you have any warning/oops/etc to show the real case?
2) shouldn't we annotate with respective lockdep asserts this code?
--
With Best Regards,
Andy Shevchenko
Powered by blists - more mailing lists