> > > > Should probably also document the gpios as the binding should attempt to be as complete > > as possible independent of what the driver currently supports. > > > > Lots of complex options for this device and the use of those pins, so maybe something > > that can be left for now - but the patch description should then mention that is intentional. > > The device has so many options for connecting stuff... It's indeed possible to > (also) use it as a GPIO expander and as a clock source and more... > > I'll put it in the patch description that the definition is incomplete by design. > > The main reason I'm submitting is that this is about the third time I've > written a driver for this chip, and I'm sure that other companies are writing > their own as well. I'm hoping this will result in some joint effort to > properly support it... > Fair enough. Incomplete bindings aren't ideal, but definitely better than no bindings in my view. Jonathan