Re: [PATCH 02/11] arm64: dts: qcom: sc8280xp-arcata: Enable backlight

[Jérôme de Bretagne <jerome.debretagne@gmail.com>]

On 5/20/26 18:40, Jérôme de Bretagne via B4 Relay wrote:
> From: Jérôme de Bretagne <jerome.debretagne@gmail.com>
> 
> Add backlight nodes and enable backlight so that it can be controlled
> with the corresponding buttons found on Surface Pro Type Cover keyboards.
> 
> The nodes are almost identical to the ones from the ThinkPad X13s with
> a different enable GPIO for the vreg_edp_bl node.
> 
> The difference was found by looking at the output of /sys/kernel/debug/gpio
> on the ThinkPad X13s to see the properties of its enable gpio9:
> 
>    gpiochip1: GPIOs 742-751, parent: platform/c440000.spmi:pmic@1:gpio@8800, c440000.spmi:pmic@1:gpio@8800:
>     ...
>     gpio9 : out  high normal  vin-1 pull-down 10uA              push-pull  medium  atest-1 dtest-0
>     ...
> 
> The same gpio9 on the Surface Pro 9 5G has different properties:
> 
>    gpiochip0: 10 GPIOs, parent: platform/c440000.spmi:pmic@1:gpio@8800, c440000.spmi:pmic@1:gpio@8800:
>     ...
>     gpio9 : in   high normal  vin-1 pull-up 30uA                push-pull  low     atest-1 dtest-0
>     ...
> 
> Looking for a match on the Surface Pro 9 5G pointed to this only one:
> 
>    gpiochip2: 10 GPIOs, parent: platform/c440000.spmi:pmic@3:gpio@8800, c440000.spmi:pmic@3:gpio@8800:
>     ...
>     gpio6 : out  high normal  vin-1 pull-down 10uA              push-pull  medium  atest-1 dtest-0
>     ...
> 
> Signed-off-by: Jérôme de Bretagne <jerome.debretagne@gmail.com>
> ---
>   .../boot/dts/qcom/sc8280xp-microsoft-arcata.dts    | 49 ++++++++++++++++++++++
>   1 file changed, 49 insertions(+)
> 
> diff --git a/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts b/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts
> index f2b4470d4407fb5b6a3dbac8bc972c010c31bd06..14287092fb07ea15e527fd318f7ea2155448b26c 100644
> --- a/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts
> +++ b/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts
> @@ -46,6 +46,16 @@ wcd938x: audio-codec {
>   		#sound-dai-cells = <1>;
>   	};
>   
> +	backlight: backlight {
> +		compatible = "pwm-backlight";
> +		pwms = <&pmc8280c_lpg 3 1000000>;
> +		enable-gpios = <&pmc8280_1_gpios 8 GPIO_ACTIVE_HIGH>;
> +		power-supply = <&vreg_edp_bl>;
> +
> +		pinctrl-0 = <&edp_bl_en>, <&edp_bl_pwm>;
> +		pinctrl-names = "default";
> +	};
> +


Investigating the feedback I got from an AI bot review [1], I plan to 
update this patch in v2. Let me know what you think.

I was skeptical at first but upon review I've tried the following patch:


diff --git a/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts 
b/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts
index f938e68c69ea..82e0a01276a8 100644
--- a/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts
+++ b/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts
@@ -198,7 +198,7 @@ vreg_edp_bl: regulator-edp-bl {
                 regulator-min-microvolt = <3600000>;
                 regulator-max-microvolt = <3600000>;

-               gpio = <&pmc8280c_gpios 6 GPIO_ACTIVE_HIGH>;
+               gpio = <&pmc8280_2_gpios 6 GPIO_ACTIVE_HIGH>;
                 enable-active-high;

                 pinctrl-0 = <&edp_bl_reg_en>;
@@ -1050,6 +1050,11 @@ vol_up_n: vol-up-n-state {
  };

  &pmc8280_2_gpios {
+       edp_bl_reg_en: edp-bl-reg-en-state {
+               pins = "gpio6";
+               function = "normal";
+       };
+
         wwan_sw_en: wwan-sw-en-state {
                 pins = "gpio1";
                 function = "normal";
@@ -1061,11 +1066,6 @@ edp_bl_pwm: edp-bl-pwm-state {
                 pins = "gpio8";
                 function = "func1";
         };
-
-       edp_bl_reg_en: edp-bl-reg-en-state {
-               pins = "gpio6";
-               function = "normal";
-       };
  };

  &pmr735a_gpios {



and ... backlight still works! It can also be enabled/disabled with:
    /sys/class/backlight/backlight/bl_power
as expected.

I have no clue why both gpio6 on pmc8280_2 and gpio6 on pmc8280c can 
enable and disable the backlight. Is it a common scenario?

Anyway, if I follow my own investigation fully based on the X13s 
reference and the properties of its backlight enable GPIO, I should 
switch to pmc8280_2 to use the better match.

I will update this patch in v2.

Jérôme


[1] https://lore.kernel.org/all/20260520165709.E485C1F00893@smtp.kernel.org/