diff mbox series

arm64: dts: qcom: sm6125-seine: Clean up gpio-keys (volume down)

Message ID 20221217100455.52593-1-marijn.suijten@somainline.org (mailing list archive)
State Superseded
Headers show
Series arm64: dts: qcom: sm6125-seine: Clean up gpio-keys (volume down) | expand

Commit Message

Marijn Suijten Dec. 17, 2022, 10:04 a.m. UTC
- Remove autorepeat (leave key repetition to userspace);
- Remove unneeded status = "okay" (this is the default);
- Allow the interrupt line for this button to be disabled;
- Use a full, descriptive node name;
- Set proper bias on the GPIO via pinctrl;
- Sort properties.

Fixes: 82e1783890b7 ("arm64: dts: qcom: sm6125: Add support for Sony Xperia 10II")
Signed-off-by: Marijn Suijten <marijn.suijten@somainline.org>
---
 .../qcom/sm6125-sony-xperia-seine-pdx201.dts  | 20 ++++++++++++++-----
 1 file changed, 15 insertions(+), 5 deletions(-)

Comments

Konrad Dybcio Dec. 17, 2022, 3:04 p.m. UTC | #1
On 17.12.2022 11:04, Marijn Suijten wrote:
> - Remove autorepeat (leave key repetition to userspace);
> - Remove unneeded status = "okay" (this is the default);
> - Allow the interrupt line for this button to be disabled;
> - Use a full, descriptive node name;
> - Set proper bias on the GPIO via pinctrl;
> - Sort properties.
> 
> Fixes: 82e1783890b7 ("arm64: dts: qcom: sm6125: Add support for Sony Xperia 10II")
> Signed-off-by: Marijn Suijten <marijn.suijten@somainline.org>
> ---
>  .../qcom/sm6125-sony-xperia-seine-pdx201.dts  | 20 ++++++++++++++-----
>  1 file changed, 15 insertions(+), 5 deletions(-)
> 
> diff --git a/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts b/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts
> index 1b9e40d3d269..8c7c9331fd21 100644
> --- a/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts
> +++ b/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts
> @@ -41,17 +41,17 @@ extcon_usb: extcon-usb {
>  	};
>  
>  	gpio-keys {
> -		status = "okay";
>  		compatible = "gpio-keys";
> -		autorepeat;
> +		pinctrl-0 = <&gpio_keys_state>;
> +		pinctrl-names = "default";
>  
> -		key-vol-dn {
> +		key-volume-down {
>  			label = "Volume Down";
>  			gpios = <&tlmm 47 GPIO_ACTIVE_LOW>;
> -			linux,input-type = <1>;
>  			linux,code = <KEY_VOLUMEDOWN>;
> -			gpio-key,wakeup;
>  			debounce-interval = <15>;
> +			linux,can-disable;
> +			gpio-key,wakeup;
>  		};
>  	};
>  
> @@ -270,6 +270,16 @@ &sdhc_1 {
>  
>  &tlmm {
>  	gpio-reserved-ranges = <22 2>, <28 6>;
> +
> +	gpio_keys_state: gpio-keys-state {
> +		key-volume-down-pins {
I see no need for defining a wrapper node.
The other changes look good!

Konrad
> +			pins = "gpio47";
> +			function = "gpio";
> +			drive-strength = <2>;
> +			bias-disable;
> +			input-enable;
> +		};
> +	};
>  };
>  
>  &usb3 {
Marijn Suijten Dec. 18, 2022, 10:18 a.m. UTC | #2
On 2022-12-17 16:04:17, Konrad Dybcio wrote:
> On 17.12.2022 11:04, Marijn Suijten wrote:
> > [..]
> > @@ -270,6 +270,16 @@ &sdhc_1 {
> >  
> >  &tlmm {
> >  	gpio-reserved-ranges = <22 2>, <28 6>;
> > +
> > +	gpio_keys_state: gpio-keys-state {
> > +		key-volume-down-pins {
> I see no need for defining a wrapper node.
> The other changes look good!

I did the same for sm6350-lena, which we should flatten out then too.

For these uses I'm not sure when it's clearer/better to use:

    thing@x {
        pinctrl-0 = <&thing_state>;
        ...
    };

    thing_state: thing-state {
        specific-pin {
            ...
        };

        other-specific-pin ...
        ...
    };

Or separate out the pins with their own state and instead use:

    thing@x {
        pinctrl-0 = <&specific_pin1_state &specific_pin2_state>;
        ...
    };

If I had to guess the former groups related pins together (as we finally
do now for SDC...) which should all be toggled at once.  In this
specific gpio-keys case, irrespective of whether it has one or more
keys, the pins aren't related apart from representing keys, and should
thus better be individual pinctrl nodes and individually referenced in
pinctrl-X.

Did I sympathize that correctly?

(side-note: the SDC pinctrl groups typically get extended with a
 card-detect pin in board DTS or in some likely-erroneous cases directly
 in SoC DTSI.  This may also count as unrelated pins being grouped
 together only because that is how the hardware/DTS node consumes them,
 but it is rather concise/readable/convenient though...)

- Marijn
Konrad Dybcio Dec. 19, 2022, 10 a.m. UTC | #3
On 18.12.2022 11:18, Marijn Suijten wrote:
> On 2022-12-17 16:04:17, Konrad Dybcio wrote:
>> On 17.12.2022 11:04, Marijn Suijten wrote:
>>> [..]
>>> @@ -270,6 +270,16 @@ &sdhc_1 {
>>>  
>>>  &tlmm {
>>>  	gpio-reserved-ranges = <22 2>, <28 6>;
>>> +
>>> +	gpio_keys_state: gpio-keys-state {
>>> +		key-volume-down-pins {
>> I see no need for defining a wrapper node.
>> The other changes look good!
> 
> I did the same for sm6350-lena, which we should flatten out then too.
> 
> For these uses I'm not sure when it's clearer/better to use:
> 
>     thing@x {
>         pinctrl-0 = <&thing_state>;
>         ...
>     };
> 
>     thing_state: thing-state {
>         specific-pin {
>             ...
>         };
> 
>         other-specific-pin ...
>         ...
>     };
> 
> Or separate out the pins with their own state and instead use:
> 
>     thing@x {
>         pinctrl-0 = <&specific_pin1_state &specific_pin2_state>;
>         ...
>     };
> 
> If I had to guess the former groups related pins together (as we finally
> do now for SDC...) which should all be toggled at once.  In this
> specific gpio-keys case, irrespective of whether it has one or more
> keys, the pins aren't related apart from representing keys, and should
> thus better be individual pinctrl nodes and individually referenced in
> pinctrl-X.
> 
> Did I sympathize that correctly?
I think so.

> 
> (side-note: the SDC pinctrl groups typically get extended with a
>  card-detect pin in board DTS or in some likely-erroneous cases directly
>  in SoC DTSI.  This may also count as unrelated pins being grouped
>  together only because that is how the hardware/DTS node consumes them,
>  but it is rather concise/readable/convenient though...)
8450 has:

pinctrl-0 = <&sdc2_default_state &sdc2_card_det_n>;

which seems like a sane application of what you described.

Konrad
> 
> - Marijn
Marijn Suijten Dec. 19, 2022, 7:25 p.m. UTC | #4
On 2022-12-19 11:00:10, Konrad Dybcio wrote:
> 
> 
> On 18.12.2022 11:18, Marijn Suijten wrote:
> > On 2022-12-17 16:04:17, Konrad Dybcio wrote:
> >> On 17.12.2022 11:04, Marijn Suijten wrote:
> >>> [..]
> >>> @@ -270,6 +270,16 @@ &sdhc_1 {
> >>>  
> >>>  &tlmm {
> >>>  	gpio-reserved-ranges = <22 2>, <28 6>;
> >>> +
> >>> +	gpio_keys_state: gpio-keys-state {
> >>> +		key-volume-down-pins {
> >> I see no need for defining a wrapper node.
> >> The other changes look good!
> > 
> > I did the same for sm6350-lena, which we should flatten out then too.
> > 
> > For these uses I'm not sure when it's clearer/better to use:
> > 
> >     thing@x {
> >         pinctrl-0 = <&thing_state>;
> >         ...
> >     };
> > 
> >     thing_state: thing-state {
> >         specific-pin {
> >             ...
> >         };
> > 
> >         other-specific-pin ...
> >         ...
> >     };
> > 
> > Or separate out the pins with their own state and instead use:
> > 
> >     thing@x {
> >         pinctrl-0 = <&specific_pin1_state &specific_pin2_state>;
> >         ...
> >     };
> > 
> > If I had to guess the former groups related pins together (as we finally
> > do now for SDC...) which should all be toggled at once.  In this
> > specific gpio-keys case, irrespective of whether it has one or more
> > keys, the pins aren't related apart from representing keys, and should
> > thus better be individual pinctrl nodes and individually referenced in
> > pinctrl-X.
> > 
> > Did I sympathize that correctly?
> I think so.

Ack, will respin like this for V2.

> > (side-note: the SDC pinctrl groups typically get extended with a
> >  card-detect pin in board DTS or in some likely-erroneous cases directly
> >  in SoC DTSI.  This may also count as unrelated pins being grouped
> >  together only because that is how the hardware/DTS node consumes them,
> >  but it is rather concise/readable/convenient though...)
> 8450 has:
> 
> pinctrl-0 = <&sdc2_default_state &sdc2_card_det_n>;
> 
> which seems like a sane application of what you described.

Glad to hear we (I and sm8450 dts writer(s)) came to the same conclusion
independently.  Not sure if it's worth retroactively cleaning up
existing DTS, but feel free to.  There are still DT's out there that
define all pins individually, too...

- Marijn
diff mbox series

Patch

diff --git a/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts b/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts
index 1b9e40d3d269..8c7c9331fd21 100644
--- a/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts
+++ b/arch/arm64/boot/dts/qcom/sm6125-sony-xperia-seine-pdx201.dts
@@ -41,17 +41,17 @@  extcon_usb: extcon-usb {
 	};
 
 	gpio-keys {
-		status = "okay";
 		compatible = "gpio-keys";
-		autorepeat;
+		pinctrl-0 = <&gpio_keys_state>;
+		pinctrl-names = "default";
 
-		key-vol-dn {
+		key-volume-down {
 			label = "Volume Down";
 			gpios = <&tlmm 47 GPIO_ACTIVE_LOW>;
-			linux,input-type = <1>;
 			linux,code = <KEY_VOLUMEDOWN>;
-			gpio-key,wakeup;
 			debounce-interval = <15>;
+			linux,can-disable;
+			gpio-key,wakeup;
 		};
 	};
 
@@ -270,6 +270,16 @@  &sdhc_1 {
 
 &tlmm {
 	gpio-reserved-ranges = <22 2>, <28 6>;
+
+	gpio_keys_state: gpio-keys-state {
+		key-volume-down-pins {
+			pins = "gpio47";
+			function = "gpio";
+			drive-strength = <2>;
+			bias-disable;
+			input-enable;
+		};
+	};
 };
 
 &usb3 {