RE: [PATCH v5 2/2] leds: ltc3220: Add Support for LTC3220 18 channel LED Driver

[Escala, Edelweise]

Hello Lee,

I would like to know your recommendation regarding this query before sending in a new version.
Right now what I've done is add comment which references the binding file.

> > > +	if (aggregated_led_found && num_leds > 1)
> > > +		return dev_err_probe(&client->dev, -EINVAL,
> > > +					"Aggregated LED must be the only LED
> > node\n");
> >
> > Must it?  Why?  Where does it say that?
> 
> Aggregated LED mode uses the hardware's "quick-write" feature which
> broadcasts writes to all 18 channels simultaneously. This is a hardware limitation
> - when quick-write mode is enabled, writing to LED channel 1 automatically
> updates ALL channels. Controlling LED individually is still possible however if LED
> 1 is changed all LED value will change.
> 
> The device tree binding currently supports two mutually exclusive modes:
> - Multiple independent LED nodes (quick-write disabled), OR
> - Single aggregated LED node with led-sources (quick-write enabled)
> 
> This aggregated LED approach was suggested in v2 review:
> https://lore.kernel.org/all/20260112-ltc3220-driver-v2-0-
> d043058fc4df@xxxxxxxxxx/
> 
> However, we'd like your recommendation on this design. Would it be better to:
> 1. Keep the aggregated LED mode with hardware quick-write2. Drop aggregated
> mode and let userspace control all 18 LEDs individually
>    (userspace can loop to set brightness if synchronized control is needed)
>

Thank you,
Edelweise Escala