linux-kernel - Re: [PATCH 1/2] gpiolib: call gcdev_unregister() sooner in the removal operations

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

Message-ID: <CAMRc=MdFuJGm2AUq45SkR8SpNDg-4Qe58fc7+ow0Oy+==aACnA@mail.gmail.com>
Date: Tue, 20 Feb 2024 14:47:06 +0100
From: Bartosz Golaszewski <brgl@...ev.pl>
To: Herve Codina <herve.codina@...tlin.com>
Cc: Kent Gibson <warthog618@...il.com>, Linus Walleij <linus.walleij@...aro.org>, 
	linux-gpio@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Luca Ceresoli <luca.ceresoli@...tlin.com>, Thomas Petazzoni <thomas.petazzoni@...tlin.com>
Subject: Re: [PATCH 1/2] gpiolib: call gcdev_unregister() sooner in the
 removal operations

On Tue, Feb 20, 2024 at 12:10 PM Herve Codina <herve.codina@...tlincom> wrote:
>
> When gpio chip device is removed while some related gpio are used by the
> user-space, the following warning can appear:
>   remove_proc_entry: removing non-empty directory 'irq/233', leaking at least 'gpiomon'
>   WARNING: CPU: 2 PID: 72 at fs/proc/generic.c:717 remove_proc_entry+0x190/0x19c
>   ...
>   Call trace:
>     remove_proc_entry+0x190/0x19c
>     unregister_irq_proc+0xd0/0x104
>     free_desc+0x4c/0xc4
>     irq_free_descs+0x6c/0x90
>     irq_dispose_mapping+0x104/0x14c
>     gpiochip_irqchip_remove+0xcc/0x1a4
>     gpiochip_remove+0x48/0x100
>   ...
>
> Indeed, the gpio cdev uses an IRQ but this IRQ is not released
> (irq_free() call) before the call to gpiochip_irqchip_remove().
>
> In order to give a chance to the gpio dev driver to release this
> irq before removing the IRQ mapping, notify the cdev driver about
> the gpio device removal before the gpiochip_irqchip_remove() call.
>
> Signed-off-by: Herve Codina <herve.codina@...tlin.com>
> ---
>  drivers/gpio/gpiolib.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> index 8b3a0f45b574..079181b9daa8 100644
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -1051,6 +1051,13 @@ void gpiochip_remove(struct gpio_chip *gc)
>
>         /* FIXME: should the legacy sysfs handling be moved to gpio_device? */
>         gpiochip_sysfs_unregister(gdev);
> +
> +       /*
> +        * Tell gcdev that the device is removing. If any gpio resources are in
> +        * use (irqs for instance), it's time for gcdev to release them.
> +        */
> +       gcdev_unregister(gdev);
> +
>         gpiochip_free_hogs(gc);
>         /* Numb the device, cancelling all outstanding operations */
>         gdev->chip = NULL;
> @@ -1085,7 +1092,6 @@ void gpiochip_remove(struct gpio_chip *gc)
>          * be removed, else it will be dangling until the last user is
>          * gone.
>          */
> -       gcdev_unregister(gdev);
>         up_write(&gdev->sem);

Please rebase it on top of the for-next branch of the GPIO tree. We've
had some significant rework recently, we no longer even have this
semaphore.

Bart

>         gpio_device_put(gdev);
>  }
> --
> 2.43.0
>