linux-kernel - Re: [PATCH] gpio: tpmx86: fix uninitialized variable girq

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 16 Sep 2022 08:56:23 +0100
From:   Marc Zyngier <maz@...nel.org>
To:     Dongliang Mu <mudongliangabcd@...il.com>
Cc:     Dongliang Mu <dzm91@...t.edu.cn>,
        Linus Walleij <linus.walleij@...aro.org>,
        Bartosz Golaszewski <brgl@...ev.pl>,
        linux-gpio@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] gpio: tpmx86: fix uninitialized variable girq

On Thu, 15 Sep 2022 07:08:10 +0100,
Dongliang Mu <mudongliangabcd@...il.com> wrote:
> 
> On Wed, Sep 14, 2022 at 3:36 PM Marc Zyngier <maz@...nel.org> wrote:
> >
> > On Wed, 14 Sep 2022 06:18:42 +0100,
> > Dongliang Mu <dzm91@...t.edu.cn> wrote:
> > >
> > > From: Dongliang Mu <mudongliangabcd@...il.com>
> > >
> > > The commit 924610607f19 ("gpio: tpmx86: Move PM device over to
> > > irq domain") adds a dereference of girq that may be uninitialized.
> > >
> > > Fix this by initializing girq and checking irq before invoking
> > > irq_domain_set_pm_device.
> > >
> > > Fixes: 924610607f19 ("gpio: tpmx86: Move PM device over to irq domain")
> > > Signed-off-by: Dongliang Mu <mudongliangabcd@...il.com>
> > > ---
> > >  drivers/gpio/gpio-tqmx86.c | 6 ++++--
> > >  1 file changed, 4 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/drivers/gpio/gpio-tqmx86.c b/drivers/gpio/gpio-tqmx86.c
> > > index fa4bc7481f9a..bdef182c11c2 100644
> > > --- a/drivers/gpio/gpio-tqmx86.c
> > > +++ b/drivers/gpio/gpio-tqmx86.c
> > > @@ -231,7 +231,7 @@ static int tqmx86_gpio_probe(struct platform_device *pdev)
> > >       struct device *dev = &pdev->dev;
> > >       struct tqmx86_gpio_data *gpio;
> > >       struct gpio_chip *chip;
> > > -     struct gpio_irq_chip *girq;
> > > +     struct gpio_irq_chip *girq = NULL;
> > >       void __iomem *io_base;
> > >       struct resource *res;
> > >       int ret, irq;
> > > @@ -315,7 +315,9 @@ static int tqmx86_gpio_probe(struct platform_device *pdev)
> > >               goto out_pm_dis;
> > >       }
> > >
> > > -     irq_domain_set_pm_device(girq->domain, dev);
> > > +     if (girq) {
> > > +             irq_domain_set_pm_device(girq->domain, dev);
> > > +     }
> > >
> > >       dev_info(dev, "GPIO functionality initialized with %d pins\n",
> > >                chip->ngpio);
> >
> > I wonder if a better fix wouldn't be to directly hoist the
> > irq_domain_set_pm_device() call into the 'if (irq > 0)' block.
> 
> If irq_domain_set_pm_device has no dependency on
> devm_gpiochip_add_data, we can directly move irq_domain_set_pm_device
> to the if block.

There is no dependency there, and you can convince yourself by looking
at the (trivial) code.

	M.

-- 
Without deviation from the norm, progress is not possible.