Re: [PATCH v4 3/3] arm64: dts: imx95: Add support for Data Modul i.MX95 eDM SBC

[Marek Vasut <marex@nabladev.com>]

On 9/17/26 8:46 PM, Frank Li wrote:

Hello Frank,

[...]

>>>>> +	reg_panel_vcc: regulator-panel-vcc {
>>>>> +		compatible = "regulator-fixed";
>>>>> +		regulator-min-microvolt = <3300000>;
>>>>> +		regulator-max-microvolt = <3300000>;
>>>>> +		regulator-name = "PANEL_VCC";
>>>>> +		gpios = <&i2c2_gpio_expander_20 17 0>;
>>>
>>> gpio third arg 0, means GPIO_ACTIVE_HIGH
>>>
>>> 	so need enable-active-high
>>>
>>> if your boards have invertor should use GPIO_ACTIVE_LOW
>>
>> I already went through all that, please see "Re: [PATCH 3/3] arm64: dts:
>> imx95: Add support for Data Modul i.MX95 eDM SBC" .
> 
> at https://lore.kernel.org/imx/f4dbeaac-ab10-4980-9fc5-533907be223c@nabladev.com/
> 
> <&i2c2_gpio_expander_20 17 0>;
>                             ^
> 
> 0 should use predefine MACRO, after I check 0 means GPIO_ACTIVE_HIGH.
> 
> so need enable-active-high for regulator.

This is extended via a DTO, so no, I must not use a macro because I do 
not know the polarity of this signal, that is why there is 0 here.

The enable-active-high DT property must explicitly NOT be here, because 
the DTO can not remove properties from the base DT. Adding 
enable-active-high here would make the DTO unable to operate this signal 
as active-low .

>>>>> +	tas2552: audio-codec@40 {
>>>>> +		compatible = "ti,tas2552";
>>>>> +		reg = <0x40>;
>>>>> +		#sound-dai-cells = <0>;
>>>>> +		avdd-supply = <&reg_1p8v>;
>>>>> +		iovdd-supply = <&reg_3p3v>;
>>>>> +		vbat-supply = <&reg_5p0v>;
>>>>> +		enable-gpio = <&i2c2_gpio_expander_20 23 GPIO_ACTIVE_HIGH>;
>>>>
>>>> [Severity: Low]
>>>> This isn't a bug, but should tas2552 be using the enable-gpios property name
>>>> instead of the deprecated enable-gpio? The patch changelog mentions replacing
>>>> gpio with gpios DT properties.
>>>
>>> Please fix this one.
>>
>> No, this is correct, see
>>
>> Documentation/devicetree/bindings/sound/ti,tas2552.yaml:  enable-gpio:
> 
> It is fine now, I suggest update bindings to allow both. which convert
> by Rob at Jan 2026. I think Rob may miss it.
> 
> kernel hack code covert -gpio to -gpios
I'll add that into my todo, but that is a separate fix from this DT.