Re: [PATCH next] gpiolib: check for parent device in devprop_gpiochip_set_names()

From: Andy Shevchenko
Date: Wed Sep 16 2020 - 16:45:54 EST


On Wed, Sep 16, 2020 at 03:43:27PM +0200, Bartosz Golaszewski wrote:
> From: Bartosz Golaszewski <bgolaszewski@xxxxxxxxxxxx>
>
> It's possible for a GPIO chip to not have a parent device (whose
> properties we inspect for 'gpio-line-names'). In this case we should
> simply return from devprop_gpiochip_set_names(). Add an appropriate
> check for this use-case.

Ah, nice!
Can we also add a small comment in the code, b/c w/o it I would stumble over
and eager to remove looks-as-unneeded check?
Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>

> Fixes: 7cba1a4d5e16 ("gpiolib: generalize devprop_gpiochip_set_names() for device properties")
> Reported-by: Anders Roxell <anders.roxell@xxxxxxxxxx>
> Signed-off-by: Bartosz Golaszewski <bgolaszewski@xxxxxxxxxxxx>
> ---
> drivers/gpio/gpiolib.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> index b7b608ef9..ac2c77edc 100644
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -375,6 +375,9 @@ static int devprop_gpiochip_set_names(struct gpio_chip *chip)
> int ret, i;
> int count;
>
> + if (!dev)
> + return 0;
> +
> count = device_property_string_array_count(dev, "gpio-line-names");
> if (count < 0)
> return 0;
> --
> 2.26.1
>

--
With Best Regards,
Andy Shevchenko