On Mon, Jul 9, 2018 at 5:37 PM Stephen Boyd <swboyd@xxxxxxxxxxxx> wrote: > Quoting Linus Walleij (2018-07-09 06:54:01) > > On Mon, Jul 2, 2018 at 7:56 PM Stephen Boyd <swboyd@xxxxxxxxxxxx> wrote: > > > > > I could do with some more clarity from Linus in the "Drivers needing > > > both pin control and GPIOs" section of > > > Documentation/driver-api/pinctl.rst but I read that section as stating > > > that the GPIO driver needs to mux the pin as a GPIO by requesting the > > > pinctrl backend to do so, unless the hardware overrides the muxed > > > function selection when the GPIO is used, without involving pinctrl > > > software. > > > > Yeah that text is especially terse :/ > > > > What it says (or what I meant to say) is that there is a choice > > between letting the pin control and GPIO functionality on the > > same pin be handled orthogonally or implementing these > > gpio_*() callbacks into the pin control backend, but in either case > > the two APIs must be used in sequence: > > pin control setting comes first, second the GPIO subsystem can > > request the GPIO line. > > > > I'll see if I can clarify. > > Ok. Is my interpretation correct though? The fundamental question here > is if gpio_request() should remux the GPIO for the GPIO function or if > drivers are expected to have pinmux settings to use their pin as a GPIO. It's an either/or situation. So there are two ways to do it, as the gpio_request() callback to pinctrl_gpio_request() etc are not compulsory to implement. For any one specific system, it is either done such that gpio_request() does it by calling down to pinctrl_gpio_request() and talking to the pinctrl back-end, OR the pin muxing is done as a side dish without any interaction with the GPIO subsystem. So pick one... I know this is not very consistent. Sorry for the inconvenience :( 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