Re: [net-next PATCH v4 1/4] dt-bindings: net: phy: Document new LEDs polarity property

[Christian Marangi <ansuelsmth@gmail.com>]

On Wed, Dec 20, 2023 at 09:22:09AM -0600, Rob Herring wrote:
> On Fri, Dec 15, 2023 at 10:22:41PM +0100, Christian Marangi wrote:
> > Document new LEDs polarity property to define what mode the LED needs to
> > be put to turn it on.
> > 
> > Currently supported modes are:
> > 
> > - active-low
> > - active-high
> > - active-low-tristate
> > - active-high-tristate
> 
> Why is having a polarity unique to LEDs on ethernet PHYs? It's not. We 
> already have 'active-low' established on several LED bindings. Please 
> move the definition to leds/common.yaml and extend it. I would simply 
> add an 'inactive-tristate' boolean property (if there's an actual user). 
>

Should I also drop the active-low from the current schema that have it?

Also we have led-active-low. (should we support both?)

On the marvell10g series we are discussing of using tristate or not. We
notice tristate might be confusing, would it be better to use
inactive-high-impedance ?

> I do worry this continues to evolve until we've re-created the pinctrl 
> binding...
> 

-- 
	Ansuel