Hi,
On Mon, Oct 28, 2024 at 2:14 AM Chen-Yu Tsai <wenst@xxxxxxxxxxxx> wrote:
>
> > > @@ -35,6 +37,7 @@ touchscreen@40 {
> > > hid-descr-addr = <0x0001>;
> > > interrupt-parent = <&pio>;
> > > interrupts = <88 IRQ_TYPE_LEVEL_LOW>;
> > > + status = "fail-needs-probe";
> >
> > It's a little weird that there's no pinctrl definition for the
> > touchscreens but there is one for the trackpad, but that predates your
> > patch and is unlikely to be a big deal.
>
> To be honest I'm in favor of getting rid of pinctrl settings that
> do nothing more than mux in a GPIO, as mentioned in my talk at ELCE.
> Such settings are already implied by the interrupts or gpios properties.
> The fact that the OS doesn't enforce exclusiveness between the
> subsystems is not something the DT should deal with.
One could also argue that the fact that the Linux kernel happens to
auto-mux pins to GPIO is not something that the device tree should
assume. Personally I have never liked the "auto-mux" behavior of Linux
and I've found that it can get in the way when you need to do more
advanced pinmuxing, like when a driver needs to sometimes use a pin in
"special function" mode and sometimes in GPIO mode. The auto-muxing
happens behind the back of the driver which then needs to account for
this fact and work around it. Just sayin. :-P
