[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220907194754.nw4jpghsjzuz2jcu@soft-dev3-1.localhost>
Date: Wed, 7 Sep 2022 21:47:54 +0200
From: Horatiu Vultur <horatiu.vultur@...rochip.com>
To: Andy Shevchenko <andy.shevchenko@...il.com>
CC: "open list:GPIO SUBSYSTEM" <linux-gpio@...r.kernel.org>,
"Linux Kernel Mailing List" <linux-kernel@...r.kernel.org>,
Linus Walleij <linus.walleij@...aro.org>,
Microchip Linux Driver Support <UNGLinuxDriver@...rochip.com>
Subject: Re: [PATCH v2] pinctrl: ocelot: Fix interrupt controller
The 09/07/2022 11:36, Andy Shevchenko wrote:
> > + * handler.
> > + */
> > + if (active && !ack) {
> > + struct ocelot_irq_work *work;
> > +
> > + work = kmalloc(sizeof(*work), GFP_ATOMIC);
> > + if (!work)
> > + return;
> > +
> > + work->irq_desc = desc;
> > + INIT_WORK(&work->irq_work, ocelot_irq_work);
> > + queue_work(system_wq, &work->irq_work);
> > + }
>
> Here I see potential issues with the object lifetime. 1) The memory is
> allocated here and what does guarantee its freeing? 2) What does
> guarantee that work will be not scheduled if the driver or its parts
> are gone?
You are right, if the driver is removed once the work is queued, then
that object will not be freed or even worst get NULL pointers.
A solution to this would be not to use system_wq but allocate own workqueue
and once the driver is removed, make sure to destroy it.
In this way we make sure that all the work is done.
>
> --
> With Best Regards,
> Andy Shevchenko
--
/Horatiu
Powered by blists - more mailing lists