Re: [PATCH RFC 4/4] arm64: dts: qcom: milos-fairphone-fp6: add supply for Hall Effect sensor

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Removing `regulator-always-on` from the Hall sensor's supply causes interrupt storms, spurious events, and broken system wakeup when the input device is closed.
--

commit 35bf0d08eef4da2139153f5425e351a93d9d591d
Author: Griffin Kroah-Hartman <griffin.kroah@fairphone.com>

arm64: dts: qcom: milos-fairphone-fp6: add supply for Hall Effect sensor

This commit adds a vdd-supply property to the Hall Effect sensor and removes
the regulator-always-on property from vreg_l10b, relying on the gpio-keys
driver to manage power.

> diff --git a/arch/arm64/boot/dts/qcom/milos-fairphone-fp6.dts b/arch/arm64/boot/dts/qcom/milos-fairphone-fp6.dts
> index c1899db46e714..ae6900c3f75c6 100644
> --- a/arch/arm64/boot/dts/qcom/milos-fairphone-fp6.dts
> +++ b/arch/arm64/boot/dts/qcom/milos-fairphone-fp6.dts
> @@ -32,13 +32,13 @@ gpio-keys {
>  		pinctrl-0 = <&volume_up_default>, <&hall_sensor_default>;
>  		pinctrl-names = "default";
>  
> -		/* Powered by the always-on vreg_l10b */
>  		event-hall-sensor {
>  			label = "Hall Effect Sensor";
>  			gpios = <&tlmm 70 GPIO_ACTIVE_LOW>;
>  			linux,input-type = <EV_SW>;
>  			linux,code = <SW_LID>;
>  			linux,can-disable;
> +			vdd-supply = <&vreg_l10b>;

When the device is opened, the driver reads the GPIO state immediately after
enabling the regulator. Without a startup-delay-us property on the regulator
to allow the voltage to stabilize, could this read an uninitialized state and
generate spurious lid events?

>  			wakeup-source;
>  		};
>  
[ ... ]
> @@ -326,8 +326,6 @@ vreg_l10b: ldo10 {
>  			regulator-min-microvolt = <1800000>;
>  			regulator-max-microvolt = <1800000>;
>  			regulator-initial-mode = <RPMH_REGULATOR_MODE_HPM>;
> -			/* Hall sensor VDD */
> -			regulator-always-on;
>  		};

Could removing this property cause an interrupt storm?

The gpio-keys driver requests and unmasks the GPIO interrupt during probe, but
only enables the vdd-supply regulator when the input device is explicitly
opened by userspace.

If the device is closed, the regulator is disabled. Cutting power to the Hall
sensor while its edge-triggered interrupt remains active might cause the GPIO
line to float. A floating line can pick up electrical noise and trigger a
continuous interrupt storm.

Additionally, does this break the system's ability to wake up?

If the system suspends while the device is closed, the sensor remains
unpowered. Even though the wake IRQ is armed due to the wakeup-source
property, an unpowered sensor wouldn't be able to detect lid events to wake
the system.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260508-gpiokeys-vdd-supply-v1-0-0bb32e8e6428@fairphone.com?part=4