On Mon, 11 May 2020 14:43:54 -0500 Rob Herring <robh@xxxxxxxxxx> wrote: > On Thu, Apr 23, 2020 at 08:50:59AM +0200, Marek Behún wrote: > > Add device-tree bindings documentation for Turris Omnia RGB LEDs. > > > > ... > > > > +properties: > > + compatible: > > + const: cznic,turris-omnia-leds > > + > > + reg: > > + description: I2C slave address of the microcontroller. > > + maxItems: 1 > > + > > +patternProperties: > > + "^led[0-9][0-9]?$": > > 'led-' is preferred. And numbering is usually hex. > > What's the numbering here based on? Might be better to use 'reg' (and a > unit-address) rather than led-sources. led-sources was for cases where > the control interface (defined by reg) could have differing regulators > connected to the LEDs. Jacek, Pavel, it seems that it would be better if I just rewrote the driver into Dan's multicolor API. It would solve Rob's complaint and could be merged once multicolor framework is... Marek
