Re: [PATCH RFC 3/4] Input: gpio-keys - add regulator to gpio_keys

[Mark Brown <broonie@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1386 bytes --]

On Fri, May 15, 2026 at 01:41:38PM +0200, Griffin Kroah-Hartman wrote:
> On 5/8/26 15:44, Mark Brown wrote:
> > On Fri, May 08, 2026 at 02:53:15PM +0200, Griffin Kroah-Hartman wrote:

> > > +		if (fwnode_property_present(child, "vdd-supply")) {
> > > +			button->regulator = devm_fwnode_regulator_get_optional(dev, child, "vdd");
> > > +			if (IS_ERR(button->regulator)) {

> > As well as the issue I mentioned on a prior thread with this assigning a
> > non-physical "vdd" name to the single supply that these components can
> > have (which has had issues in the past

> Our Hall Effect sensor IC does have a named "vdd" pin, but we are totally
> open to changing this to power-supply or whatever best follows the standard.

The problem is that there is no standard here, this is not specific to a
particular consumer device.  Your device has one supply called VDD but
some other device might have some other name, and another might have two
supplies.

> > painful) the fact that this is fwnode means that this opens up support
> > for using this with ACPI which is very problematic given that ACPI has a
> > strong model of how regulators should work which is that they should not
> > be OS visible at all.

> Would it be more appropriate to drop the devm_fwnode_regulator_get() and
> replace it with a type-casted devm_of_regulator_get()?

I'd feel a lot more comfortable with that.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]