On 26/08/2020 10:13:04+0200, Bastian Krause wrote: > >> Are you okay with that? > >> > > > > I agree boolean should be avoided in RTC drivers because we need a way > > to express "don't change this value". > > Alright. > > >> Some more context: > >> > >> I originally tried to add a chargeable flag for rx8130. Prior to this > >> patch, there was no need to set "trickle-diode-disable" for this, > >> because the driver did not pass the chargeable flag to the RTC. With the > >> patch the default would have been to charge as long as > >> "trickle-diode-disable" is not there. So there's a change in behavior. > >> > > > > Yes, IIRC, my point was simply to move the documentation for > > aux-voltage-chargeable to the generice rtc binding documentation, > > Documentation/devicetree/bindings/rtc/rtc.yaml > > > > For now, you sould keep support for trickle-diode-disable but it has to be > > superseded by aux-voltage-chargeable if present. Is that more clear? > > Yes, thanks for the clarification. > > Should I set the deprecated flag for trickle-diode-disable in the > dt-binding yaml? > That's a good idea, yes. -- Alexandre Belloni, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
