Re: [RFC PATCH v3 4/5] arm64: dts: mediatek: mt8173-elm-hana: Mark touchscreens and trackpads as fail

[Doug Anderson]

Hi,

On Wed, Dec 6, 2023 at 2:02 AM AngeloGioacchino Del Regno
<angelogioacchino.delregno@xxxxxxxxxxxxx> wrote:
>
> Il 06/12/23 03:55, Chen-Yu Tsai ha scritto:
> > On Tue, Dec 5, 2023 at 6:22 PM AngeloGioacchino Del Regno
> > <angelogioacchino.delregno@xxxxxxxxxxxxx> wrote:
> >>
> >> Il 04/12/23 17:50, Doug Anderson ha scritto:
> >>> Hi,
> >>>
> >>> On Sun, Dec 3, 2023 at 10:59 PM Chen-Yu Tsai <wenst@xxxxxxxxxxxx> wrote:
> >>>>
> >>>> On Sat, Dec 2, 2023 at 8:58 AM Doug Anderson <dianders@xxxxxxxxxxxx> wrote:
> >>>>>
> >>>>> Hi,
> >>>>>
> >>>>> On Tue, Nov 28, 2023 at 12:45 AM Chen-Yu Tsai <wenst@xxxxxxxxxxxx> wrote:
> >>>>>>
> >>>>>> @@ -44,6 +46,7 @@ trackpad2: trackpad@2c {
> >>>>>>                   reg = <0x2c>;
> >>>>>>                   hid-descr-addr = <0x0020>;
> >>>>>>                   wakeup-source;
> >>>>>> +               status = "fail-needs-probe";
> >>>>>
> >>>>> While doing this, you could also remove the hack where the trackpad
> >>>>> IRQ pinctrl is listed under i2c4.
> >>>>
> >>>> Sure. I do think we can do away with it though. According to at least one
> >>>> schematic, the interrupt line has pull-ups on both sides of the voltage
> >>>> shifter.
> >>>>
> >>>> BTW, The touchscreen doesn't have pinctrl entries. This has pull-ups on
> >>>> both sides of the voltage shifter as well.
> >>>
> >>> I dunno if the convention is different on Mediatek boards, but at
> >>> least on boards I've been involved with in the past we've always put
> >>> pinctrl entries just to make things explicit. This meant that we
> >>> didn't rely on the firmware/bootrom/defaults to leave pulls in any
> >>> particular state. ...otherwise those external pull-ups could be
> >>> fighting with internal pull-downs, right?
> >>>
> >>
> >> MediaTek boards aren't special and there's no good reason for those to rely on
> >> firmware/bootrom/defaults - so there is no good reason to avoid declaring any
> >> relevant pinctrl entry.
> >
> > I think this should be migrated to use the proper GPIO bindings: the
> > GPIO_PULL_UP / GPIO_PULL_DOWN / GPIO_BIAS_DISABLE flags.
> >
> > But that's a different discussion.
> >
>
> 100% agreed.

I guess I'd need to see patches as an example to see how this looks,
but I'm at least slightly skeptical. In this case the GPIO is
indirectly specified via "interrupts". Would you add these flags to
the interrupts specifier, too? There's another potential pull as well
(PIN_CONFIG_BIAS_BUS_HOLD) as well as other pin configuration
(PIN_CONFIG_INPUT_DEBOUNCE, for instance). Do we try to fit all of
these into the GPIO / interrupt specifier?

Certainly I will admit that this is a complicated topic, but it seems
weird to say that we use pinctrl to specify pin configuration / pulls
for all pins except ones that are configured as GPIOs or GPIO
interrupts.

-Doug