Re: [PATCH v7 2/7] arm64: dts: imx8mm: imx8mp: Add LVDS DTOs for Data Modul i.MX8M Mini and Plus eDM SBC

On 7/30/26 5:31 PM, Frank Li wrote:

Hello Frank,

>>> +&backlight {
>>> +	power-supply = <&reg_panel_bl_supply>;
>>> +};
>>
>> [Severity: Medium]
>> Will this result in the wrong GPIO being toggled for the native i.MX8MP LVDS
>> interface?
>>
>> The shared imx8mm-data-modul-edm-sbc-overlay-lvds-g215hvn011.dtsi file hardcodes
>> enable-gpios to &gpio3 0 and defines a pinctrl. However, the native i.MX8MP
>> LVDS interface uses the IO expander's &gpiowifi 0 (via reg_panel_bl in
>> edm-sbc-imx8mp-lvds.dtsi).
>>
>> Since this overlay only overrides power-supply and does not remove the
>> inapplicable pinctrl or override enable-gpios, won't the backlight driver
>> toggle the incorrect baseboard adapter GPIO pin instead of the intended one?
> 
> I am not sure what's bad impact to wrong toggle enable-gpios. need check
> this,
> 
> change exist ones to keep consistenet, all use enable-gpios or power-supply.

Maybe the AI is confused, that display works, I have it locally.

[...]

>> [Severity: Medium]
>> Will this clock configuration be silently ignored by the clock framework?
>>
>> This overlay (and several others for the native LVDS interface) defines
>> assigned-clock-rates without providing or inheriting an accompanying
>> assigned-clocks array.
>>
>> The base board file imx8mp-data-modul-edm-sbc.dts does not include
>> imx8mp-nominal.dtsi, where the assigned-clocks array is typically defined.
>>
>> When assigned-clocks is missing, of_parse_phandle_with_args() returns -ENOENT,
>> and __set_clk_rates() silently skips setting the rates. Should assigned-clocks
>> be explicitly defined here so the panel clock frequency is correctly set
> 
> This is most like false alarm. But for better read purpose, suggest
> write whole group property
> 
> assign-clocks
> assign-clock-parents
> assign-clock-rates
> 
> Incase some change order in soc dtsi and forget update here and cause
> wrong.
> 
> I met similar issue before.
> 
> the similar case is clocks and clock-names, should appear in pair at
> every where.
This is the result of deduplication that was requested in V1->V2. I 
think it is fine to keep it deduplicated.