Re: [PATCH 4/4] arm64: dts: meson: add libretech-pc boards support

From: Martin Blumenstingl
Date: Sun Dec 08 2019 - 13:05:47 EST


Hi Jerome,

this is looking good overall - I have some questions / nit-picks below

On Fri, Dec 6, 2019 at 11:02 AM Jerome Brunet <jbrunet@xxxxxxxxxxxx> wrote:
[...]
> + adc_keys {
on most boards we use "adc-keys" instead of "adc_keys"

[...]
> + button-onoff {
> + label = "On/Off";
> + linux,code = <KEY_VENDOR>;
based on the label I assumed that the code is KEY_POWER
why is KEY_VENDOR the better choice here?

[...]
> + cvbs-connector {
> + compatible = "composite-video-connector";
> + status = "disabled";
is there CVBS on the board? if I remember correctly the VPU driver
works fine when omitting the CVBS connector
so if the board doesn't have it you may drop the whole node instead of
keeping it disabled

[...]
> + leds {
> + compatible = "gpio-leds";
> +
> + green {
> + label = "librecomputer:green:disk";
you can use the "function" and "color" properties instead of the (now
deprecated) "label"

[...]
> +&external_mdio {
> + external_phy: ethernet-phy@0 {
> + reg = <0>;
it would be great to have a comment above which PHY is used on this board

> + max-speed = <1000>;
> + reset-assert-us = <10000>;
> + reset-deassert-us = <30000>;
> + reset-gpios = <&gpio GPIOZ_14 GPIO_ACTIVE_LOW>;
> + interrupt-parent = <&gpio_intc>;
a comment like /* MAC_INTR on GPIOZ_15 */ would be great here
> + interrupts = <25 IRQ_TYPE_LEVEL_LOW>;

[...]
> +&pinctrl_periphs {
> + /*
> + * Make sure the reset pin of the usb HUB is driven high to take
> + * it out of reset.
> + */
> + usb1_rst_pins: usb1_rst_irq {
> + mux {
> + groups = "GPIODV_3";
> + function = "gpio_periphs";
> + bias-disable;
> + output-high;
> + };
> + };
on other boards (like Odroid-C2) we use a GPIO hog for this. I'm not
sure which one is better

[...]
> +&pinctrl_periphs {
> + /*
> + * Make sure the irq pin of the TYPE C controller is not driven
> + * by the SoC.
is this because the SoC default configuration pulls the IRQ line LOW,
which then generates "phantom" IRQs?

[...]
> + fusb302@22 {
typec-portc@22

[...]
> + interrupt-parent = <&gpio_intc>;
> + interrupts = <59 IRQ_TYPE_LEVEL_LOW>;
a comment above with the GPIO number would be great


Martin