Re: [PATCH 2/3] gpio: Add Intel USBIO GPIO driver

[Sakari Ailus <sakari.ailus@linux.intel.com>]

Hi Hans,

On Mon, Aug 11, 2025 at 11:23:50AM +0200, Hans de Goede wrote:
> Hi Sakari,
> 
> Thank you for your review, note this
> is not a full reply.
> 
> On 11-Aug-25 9:07 AM, Sakari Ailus wrote:
> 
> ...
> 
> >> +static int usbio_gpio_probe(struct auxiliary_device *adev,
> >> +		const struct auxiliary_device_id *adev_id)
> >> +{
> >> +	struct usbio_gpio_bank_desc *bank_desc;
> >> +	struct device *dev = &adev->dev;
> >> +	struct usbio_gpio *gpio;
> >> +	int bank, ret;
> >> +
> >> +	bank_desc = dev_get_platdata(dev);
> >> +	if (!bank_desc)
> >> +		return -EINVAL;
> >> +
> >> +	gpio = devm_kzalloc(dev, sizeof(*gpio), GFP_KERNEL);
> >> +	if (!gpio)
> >> +		return -ENOMEM;
> > 
> > All devm-allocated memory is released once the device is gone but you may
> > still have users. Could you release this in the adev's release callback
> > instead?
> 
> All GPIO users go through gpio_descriptors which are seperately
> refcounted. The GPIO core can handle GPIO controllers being
> hot-unplugged since the end of 2023 and guarantees that no
> calls will be made to the gpiochip driver after calling
> gpiochip_remove().

Thanks, I wasn't aware of that. This simplifies the client driver
implementation nicely.

-- 
Regards,

Sakari Ailus