On Tue, Mar 17, 2015 at 5:56 PM, Andrew Bresticker <abrestic@xxxxxxxxxxxx> wrote: [Me] >> Sorry if I don't really know how things work now... :( >> It seems like a logical way to me. (...) > 4) The ->set_mux() op must set the proper function for the pin. > 5) The ->set_mux() op must also disable the GPIO function for the pin. > To disable the GPIO function, the pinctrl driver must map the pin to a > GPIO bank/offset and disable the GPIO via the GPIO bank's GPIO_EN > register. That sounds like the"GPIO" registers are actually involved in any muxing usecase, meaning there is not really a clean split between the pinctrl and GPIO hardware, the case I refer to as "GPIO mode pitfalls" in Documentation/pinctrl.txt. In such cases both halves of the driver(s) need to be aware of the other, and that is what you seem to be wanting to achieve. So I was wrong in thinking the GPIO device could be a separate subdevice, the two parts are too dependent on each other. So keep a single probe() function and let the two driver halves poke into each others' registers. Sorry for the fuzz... Yours, Linus Walleij -- To unsubscribe from this list: send the line "unsubscribe linux-gpio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html