Re: [PATCH 3/4] arm64: dts: rockchip: add Forlinx OK3568-C

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
- [Medium] dts: missing device-wakeup, interrupt, and reset properties in bluetooth node
- [Medium] dts: orphaned modem_5g_reset pinctrl leaves 5G modem reset unhandled
- [Medium] dts: orphaned wifi_host_wake pinctrl disables out-of-band WiFi wake
--

commit ebd497a35c265bc281369082ab62727d0e435e1b
Author: Dmitry Yashin <dmt.yashin@gmail.com>

arm64: dts: rockchip: add Forlinx OK3568-C

This commit adds support for the Forlinx OK3568-C board, which is the
carrier board for the FET3568-C System on Module. It maps the peripherals
and components for the board including connectivity modules like 5G, WiFi,
and Bluetooth.

> diff --git a/arch/arm64/boot/dts/rockchip/rk3568-ok3568-c.dts b/arch/arm64/boot/dts/rockchip/rk3568-ok3568-c.dts
> new file mode 100644
> index 0000000000000..d292df546d419
> --- /dev/null
> +++ b/arch/arm64/boot/dts/rockchip/rk3568-ok3568-c.dts
[ ... ]
> +&pinctrl {
[ ... ]
> +	modem {
> +		modem_5g_pwr: modem-5g-pwr {
> +			rockchip,pins = <1 RK_PB1 RK_FUNC_GPIO &pcfg_pull_none>;
> +		};
> +
> +		modem_5g_reset: modem-5g-reset {
> +			rockchip,pins = <1 RK_PB2 RK_FUNC_GPIO &pcfg_pull_none>;
> +		};

[Severity: Medium]
The modem_5g_reset pinctrl node is defined here but does not appear to be
referenced in any pinctrl-0 property or assigned to a device node,
regulator, or gpio-hog.

Without being actively driven, could this prevent the 5G modem from
initializing and enumerating properly if it requires this reset signal?

[ ... ]
> +	wifi {
> +		wifi_reg_on: wifi-reg-on {
> +			rockchip,pins = <3 RK_PD4 RK_FUNC_GPIO &pcfg_pull_none>;
> +		};
> +
> +		wifi_host_wake: wifi-host-wake {
> +			rockchip,pins = <0 RK_PD4 RK_FUNC_GPIO &pcfg_pull_none>;
> +		};
> +	};
> +};

[Severity: Medium]
Similar to the modem reset, the wifi_host_wake pinctrl group configures
a GPIO for handling out-of-band host wake interrupts from the WiFi
module, but it is completely isolated. It isn't referenced by the
sdmmc2 node's pinctrl-0, and the corresponding interrupts property is
missing.

Does this missing configuration break Wake-on-WLAN and sleep power
management for the WiFi module during system suspend?

[ ... ]
> +/* BT part of the AW-CM358SM connected with UART */
> +&uart8 {
> +	dma-names = "tx", "rx";
> +	pinctrl-names = "default";
> +	pinctrl-0 = <&uart8m0_xfer &uart8m0_ctsn &uart8m0_rtsn>;
> +	uart-has-rtscts;
> +	status = "okay";
> +
> +	bluetooth {
> +		compatible = "nxp,88w8987-bt";
> +		pinctrl-names = "default";
> +		pinctrl-0 = <&bt_reg_on &bt_wake_host &host_wake_bt>;
> +		vcc-supply = <&vcc1v8>;
> +		wakeup-source;
> +	};
> +};

[Severity: Medium]
While pinctrl-0 sets up the hardware pins as GPIOs, the corresponding DT
properties (device-wakeup-gpios, interrupts, reset-gpios) required by
the nxp,88w8987-bt binding are missing from this node.

Without device-wakeup-gpios and the wakeup interrupt, will the
btnxpuart driver be unable to perform power management signaling and
break sleep/wake functionality? Also, without a driver managing the
bt_reg_on GPIO, could the chip remain in reset and fail to initialize?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260930-ok3568-c-v1-0-a0ffd55f5c95@gmail.com?part=3