Re: [PATCH 4/4] arm64: dts: allwinner: add Teclast P85T tablet

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] dts: dummy 1.8V regulator used for 3.3V IO banks, causing wrong bias configuration
- [Low] dts: inbound USB-C power mapped to outbound VBUS regulator in usbphy
--

commit 62a9921e9859ecb056d3f2383913082baa7f4de1
Author: Mikhail Kalashnikov <iuncuim@gmail.com>

arm64: dts: allwinner: add Teclast P85T tablet

The Teclast P85T is an 8-inch tablet that was announced in 2023
based on the Allwinner A523 (sun55i) SoC.

> diff --git a/arch/arm64/boot/dts/allwinner/sun55i-a523-teclast-p85t.dts b/arch/arm64/boot/dts/allwinner/sun55i-a523-teclast-p85t.dts
> new file mode 100644
> index 0000000000000..38a9c98ecbec6
> --- /dev/null
> +++ b/arch/arm64/boot/dts/allwinner/sun55i-a523-teclast-p85t.dts
[ ... ]
> +	reg_pio18: pio-18 {
> +		compatible = "regulator-fixed";
> +		regulator-name = "pio-18";
> +		regulator-min-microvolt = <1800000>;
> +		regulator-max-microvolt = <1800000>;
> +		regulator-always-on;
> +	};
[ ... ]
> +&pio {
> +	vcc-pc-supply = <&reg_pio18>;
> +	vcc-pe-supply = <&reg_pio18>;
> +	vcc-pf-supply = <&reg_pio18>;
> +	vcc-pg-supply = <&reg_pio18>;
> +};

[Severity: Medium]
Will configuring these IO bank supplies with a dummy 1.8V fixed regulator cause
incorrect IO bias configuration?

For instance, the PF bank is used by mmc0 (SD card), which physically
requires 3.3V to operate (as correctly identified by its
vmmc-supply = <&reg_cldo3>;).

The Allwinner pinctrl driver reads the voltage from the assigned supply to
configure the IO pad bias registers. Providing a fake 1.8V regulator for a
bank that is physically powered at 3.3V causes the driver to write incorrect
bias values, which might lead to signal integrity issues or pad damage.

Could the real supplies (e.g., reg_cldo3 for 3.3V and reg_cldo1 for 1.8V
from the axp717 PMIC) be used directly? The pinctrl driver requests regulators
lazily, so this shouldn't cause a circular dependency.

[ ... ]
> +			reg_cldo1: cldo1 {
> +				/* codec CPVIN, SD/eMMC 1.8V IO */
> +				regulator-min-microvolt = <1800000>;
> +				regulator-max-microvolt = <1800000>;
> +				regulator-name = "vcc-codec-sd";
> +			};
[ ... ]
> +&usbphy {
> +	usb0_vbus-supply = <&reg_vcc5v>;

[Severity: Low]
Does this correctly assign the inbound board power supply as the outbound
VBUS supply?

reg_vcc5v represents the inbound 5V supply coming from the USB-C connector.
However, usb0_vbus-supply is the regulator that the PHY enables to provide
power out to external devices when operating in host mode.

Since the board is configured purely as a peripheral (no VBUS drive capability),
assigning its main inbound power supply as the PHY's VBUS output seems reversed.

Should usb0_vbus-supply be omitted here since the driver uses other
properties for sensing inbound VBUS?

> +	status = "okay";
> +};

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260918165414.1129076-1-iuncuim@gmail.com?part=4