Re: [PATCH] dt-bindings: regulator: fcs,fan53555: Remove vsel-gpios property

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1746 bytes --]

On Mon, Apr 27, 2026 at 10:09:53PM +0300, Alexander Shiyan wrote:
> Hello Comor.
> 
> > On Mon, Apr 27, 2026 at 02:54:43PM +0300, Alexander Shiyan wrote:
> > > The vsel-gpios property is currently documented in the binding but
> > > is not used by the driver. The FAN53555 family of regulators supports
> > > two voltage selector registers (VSEL0/VSEL1), and the selection between
> > > them is intended to be controlled by an external hardware pin (VSEL).
> > > However, the driver does not support dynamic toggling of this pin via
> > > a GPIO, it only uses the fcs,suspend-voltage-selector property to
> > > statically assign which register is used for runtime voltage and which
> > > for suspend voltage.
> > > Remove the vsel-gpios property from the binding to prevent incorrect DT
> > > usage and to reflect the actual hardware description supported by the
> > > driver.
> >
> > From the wording/justification here, I disagree with this patch. The
> > binding should document what the hardware can do, not what the driver
> > can.
> >
> > Maybe instead you should make fcs,suspend-voltage-selector mutually
> > exclusive with vsel-gpios?
> 
> The main problem here is that this feature (vsel-gpios) was never
> implemented in the driver. So, the patch consists solely of removing a
> non-existent property.

I'm not disputing that, but if the functionality is supported by the
hardware the existence of the property is fine. Bindings should be
complete, even if the driver in linux doesn't use something.

Rob's 2023 patch that added the property even says "Add the undocumented
'vsel-gpios' property used to control the VSEL pin.", so having it was
an intentional decision by the main binding maintainer!

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]