Re: [PATCH v2 4/5] arm64: dts: qcom: add device tree for SDM845-HDK

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
- [High] dts: usb_1_qmpphy swaps vdda-phy and vdda-pll voltage supplies
- [Medium] dts: home_pin_a pinctrl incorrectly configures gpio6 instead of gpio5
- [Medium] dts: impossible 1.8V vin-supply for 3.3V fixed regulator vreg_sys_bob_3p3
--

commit 48f1c407e24f3899e533d5dddbd59a77b4901a02
Author: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>

arm64: dts: qcom: add device tree for SDM845-HDK

This patch adds the device tree for the SDM845-HDK board, a development
platform utilizing the modem-less SDA845 SoC. It introduces initial support
for features like HDMI display, uSD, UFS, USB, and various onboard peripherals.

> diff --git a/arch/arm64/boot/dts/qcom/sdm845-hdk.dts b/arch/arm64/boot/dts/qcom/sdm845-hdk.dts
> new file mode 100644
> index 0000000000000..4dd426912d20f
> --- /dev/null
> +++ b/arch/arm64/boot/dts/qcom/sdm845-hdk.dts

[ ... ]

> +	vreg_sys_bob_3p3: regulator-sys-bob {
> +		compatible = "regulator-fixed";
> +		regulator-name = "sys_bob";
> +		regulator-min-microvolt = <3300000>;
> +		regulator-max-microvolt = <3300000>;
> +
> +		regulator-always-on;
> +		regulator-boot-on;
> +
> +		vin-supply = <&vreg_s4a_1p8>;
> +	};

[Severity: Medium]
Is vreg_s4a_1p8 the correct vin-supply for this regulator?

vreg_s4a_1p8 provides 1.8V, but vreg_sys_bob_3p3 is defined as a 3.3V fixed
regulator. A fixed regulator cannot step up a 1.8V input to 3.3V.

Could this create an incorrect power dependency tree, potentially causing the
1.8V supply to be kept awake unnecessarily or turned off while the 3.3V rail
is still needed?

[ ... ]

> +&pm8998_gpios {
> +	home_pin_a: home-active-state {
> +		pins = "gpio6";
> +		function = "normal";
> +		input-enable;
> +		bias-pull-up;
> +		qcom,drive-strength = <PMIC_GPIO_STRENGTH_NO>;
> +	};

[Severity: Medium]
Does home_pin_a incorrectly configure gpio6 instead of gpio5?

Earlier in the file, the key-home node is defined to use gpio5:

		key-home {
			label = "Home";
			linux,code = <KEY_HOMEPAGE>;
			gpios = <&pm8998_gpios 5 GPIO_ACTIVE_LOW>;
		};

Configuring gpio6 here leaves gpio5 without its required pull-up configuration,
which could break the Home button. Additionally, since vol_up_pin_a also uses
gpio6, this results in gpio6 being configured twice.

[ ... ]

> +&usb_1_qmpphy {
> +	vdda-phy-supply = <&vreg_l26a_1p2>;
> +	vdda-pll-supply = <&vreg_l1a_0p88>;
> +
> +	status = "okay";
> +};

[Severity: High]
Are the voltage supplies for vdda-phy and vdda-pll swapped?

The QMP PHY expects a ~0.88V core supply and a 1.2V PLL supply. Assigning
vreg_l26a_1p2 (1.2V) to vdda-phy-supply and vreg_l1a_0p88 (0.88V) to
vdda-pll-supply applies excessive voltage to the 0.88V rail and insufficient
voltage to the 1.2V rail.

Could this reversed configuration lead to hardware failure or potential damage?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261014-sdm845-hdk-v2-0-6d0b4b42f267@oss.qualcomm.com?part=4