arm64: dts: mt8183: Incorrect mediatek,pull-*-adv values

Hsin-Yi Wang hsinyi at chromium.org
Tue Jul 27 04:06:00 PDT 2021


On Mon, Jul 26, 2021 at 5:48 PM Enric Balletbo Serra
<eballetbo at gmail.com> wrote:
>
> Hi Chen-Yu and Hsin-Yi,
>
>
> Missatge de Chen-Yu Tsai <wenst at chromium.org> del dia dl., 26 de jul.
> 2021 a les 10:58:
> >
> > On Mon, Jul 26, 2021 at 4:50 PM Hsin-Yi Wang <hsinyi at chromium.org> wrote:
> > >
> > > On Mon, Jul 26, 2021 at 4:20 PM Chen-Yu Tsai <wenst at chromium.org> wrote:
> > > >
> > > > Hi,
> > > >
> > > > I was looking at MTK pinctrl stuff upstream, and it seems there are a few
> > > > `mediatek,pull-*-adv` entries that have invalid values:
> > > >
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-down-adv = <10>;
>
> The confusion comes probably because the binding says that this value
> is valid, see
>
>    Documentation/devicetree/bindings/pinctrl/pinctrl-mt8183.txt
>
> it'd be nice to convert that binding and the others to yaml format and
> remove the possibility to specify it in binary format.
>
>   Documentation/devicetree/bindings/pinctrl/pinctrl-mt65xx.txt
>   Documentation/devicetree/bindings/pinctrl/pinctrl-mt7622.txt
>   Documentation/devicetree/bindings/pinctrl/pinctrl-mt6797.txt
>
> Should be pretty easy as there is already an example:
> Documentation/devicetree/bindings/pinctrl/mediatek,mt6779-pinctrl.yaml
>

Hi Enric,

I've sent a patch for converting these into yaml here:
https://lore.kernel.org/patchwork/patch/1468449/

Thanks

> Thanks,
>   Enric
>
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-up-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-up-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-kukui.dtsi:
> > > > mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts:
> > > >  mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts:
> > > >  mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts:
> > > >  mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts:
> > > >  mediatek,pull-up-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts:
> > > >  mediatek,pull-down-adv = <10>;
> > > > arch/arm64/boot/dts/mediatek/mt8183-pumpkin.dts:
> > > >  mediatek,pull-up-adv = <10>;
> > > >
> > > > According to the bindings, the values should 0~3. <10> is probably a
> > > > incorrect attempt at using binary representation, which is wrong. This
> > > > probably leads to the pull-up/down getting disabled or ignored.
> > > >
> > > > Cound people still working on these two devices take a look?
> > > >
> > > Thanks for pointing this out. It's an incorrect value but 10=0b1010 so
> > > the result is same as 2, since the driver test the value by checking
> > > the last 2 bit. We should still fix this in dts.
> >
>
> I see that the are
>
>
>
> > That probably explains why no one noticed.



More information about the Linux-mediatek mailing list