* Trent Piepho <tpiepho@xxxxxxxxx> [200924 06:31]: > On Wed, Sep 23, 2020 at 11:06 PM Tony Lindgren <tony@xxxxxxxxxxx> wrote: > > > > * Trent Piepho <tpiepho@xxxxxxxxx> [200924 05:49]: > > > On Wed, Sep 23, 2020 at 10:43 PM Tony Lindgren <tony@xxxxxxxxxxx> wrote: > > > > > > > > * Trent Piepho <tpiepho@xxxxxxxxx> [200924 01:34]: > > > > > On Tue, Sep 22, 2020 at 11:57 PM Tony Lindgren <tony@xxxxxxxxxxx> wrote: > > > > > > > > > > > > Also FYI, folks have also complained for a long time that the pinctrl-single > > > > > > binding mixes mux and conf values while they should be handled separately. > > > > > > > > > > > > > > > > Instead of combining two fields when the dts is generated they are now > > > > > combined when the pinctrl-single driver reads the dts. Other than > > > > > this detail, the result is the same. The board dts source is the > > > > > same. The value programmed into the pinctrl register is the same. > > > > > There is no mechanism currently that can alter that value in any way. > > > > > > > > > > What does combining them later allow that is not possible now? > > > > > > > > It now allows further driver changes to manage conf and mux separately :) > > > > > > The pinctrl-single driver? How will that work with boards that are > > > not am335x and don't use conf and mux fields in the same manner as > > > am335x? > > > > For those cases we still have #pinctrl-cells = <1>. > > If pincntrl-single is going to be am335x specific, then shouldn't it > be a different compatible string? Certainly different compatible strings can be used as needed. But pinctrl-single is not going to be am335x specific though :) We have quite a few SoCs using it: $ git grep pinctrl-single,function-mask arch/*/boot/dts/ | wc -l 41 > Are the driver changes something that can be not be done with the > pinconf-single properties? They all include a mask. Sure but in the long term we're better off with using #pinctrl-cells along the lines what we have for example for #interrupt-cells and #gpio-cells. Regards, Tony
