Re: [PATCH 1/4] arm64: dts: mediatek: mt8183-kukui-jacuzzi: Drop pp3300_panel voltage settings

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 04/11/24 14:47, Chen-Yu Tsai ha scritto:
> On Mon, Nov 4, 2024 at 9:19 PM AngeloGioacchino Del Regno
> <angelogioacchino.delregno@collabora.com> wrote:
>>
>> Il 04/11/24 14:00, Chen-Yu Tsai ha scritto:
>>> On Wed, Oct 30, 2024 at 3:02 PM Chen-Yu Tsai <wenst@chromium.org> wrote:
>>>>
>>>> The pp3300_panel fixed regulator is just a load switch. It does not have
>>>> any regulating capabilities. Thus having voltage constraints on it is
>>>> wrong.
>>>>
>>>> Remove the voltage constraints.
>>>>
>>>> Fixes: cabc71b08eb5 ("arm64: dts: mt8183: Add kukui-jacuzzi-damu board")
>>>> Signed-off-by: Chen-Yu Tsai <wenst@chromium.org>
>>>
>>> I see that the other three patches were merged and included in the pull
>>> request, but not this one. Were there any concerns?
>>>
>>
>> Sorry I forgot to actually provide an explanation for that - yes, I do have some
>> comment about this one.
>>
>> Despite this being a load switch, it's still switching power from regulator A to
>> target device X, so this is technically still providing 3.3V to device X.
>>
>> Think about how a "regular" full-fledged regulator works: you can (sometimes) set
>> a voltage, and then you can ENABLE the VOUT for said regulator (/rail): this kind
>> of "load switch" does exactly the same as the ENABLE switch for a full-fledged
>> regulator.
> 
> But it does not provide regulation. One cannot "set" the voltage on a load
> switch; one can only set it on its upstream supply, if that supply provides
> regulation.
> 
> IIRC Mark said some years ago that if a regulator doesn't regulate the
> voltage, then the voltage constraints should not be given. The constraints
> are then derived from its upstream supply.
> 
> That's the guideline I've followed for all the regulator related changes
> I've done over the years. Does that work for you?
> 

Ok, let's go with that then.

There's only one problem - patches 2 to 4 don't apply for whatever reason, is there
any dependency?

Cheers,
Angelo

>> So, this is switching on and off a power rail that is derived from a source rail,
>> practically creating... well, a "new" rail, with...
>>
>>    VIN=somewhere-3.3v,
>>    VOUT=somewhere-still-3.3v
>>
>> Any objections/doubts/etc? :-)
> 
> I agree with most of it, except the part that I laid out above about the
> load switch not providing regulation.
> 
>> P.S.: I'm writing fast, sorry if anything appears unclear, feel free to shoot more
>>         questions in case :-)
> 
> No, it's pretty clear, and I believe one of the common interpretations
> I see. Thank you for the quick response.
> 
> 
> Thanks
> ChenYu
> 
>> Cheers,
>> Angelo
>>
>>>
>>> ChenYu
>>>
>>>> ---
>>>>    arch/arm64/boot/dts/mediatek/mt8183-kukui-jacuzzi.dtsi | 2 --
>>>>    1 file changed, 2 deletions(-)
>>>>
>>>> diff --git a/arch/arm64/boot/dts/mediatek/mt8183-kukui-jacuzzi.dtsi b/arch/arm64/boot/dts/mediatek/mt8183-kukui-jacuzzi.dtsi
>>>> index 783c333107bc..7bbafe926558 100644
>>>> --- a/arch/arm64/boot/dts/mediatek/mt8183-kukui-jacuzzi.dtsi
>>>> +++ b/arch/arm64/boot/dts/mediatek/mt8183-kukui-jacuzzi.dtsi
>>>> @@ -35,8 +35,6 @@ pp1800_mipibrdg: pp1800-mipibrdg {
>>>>           pp3300_panel: pp3300-panel {
>>>>                   compatible = "regulator-fixed";
>>>>                   regulator-name = "pp3300_panel";
>>>> -               regulator-min-microvolt = <3300000>;
>>>> -               regulator-max-microvolt = <3300000>;
>>>>                   pinctrl-names = "default";
>>>>                   pinctrl-0 = <&pp3300_panel_pins>;
>>>>
>>>> --
>>>> 2.47.0.163.g1226f6d8fa-goog
>>>>
>>