Re: [PATCH v7 2/5] dt-bindings: leds: Add LED_COLOR_ID_MOONLIGHT definitions

[Gene Chen]

Jacek Anaszewski <jacek.anaszewski@xxxxxxxxx> 於 2020年11月20日 週五 上午6:26寫道：
>
> On 11/19/20 10:57 PM, Pavel Machek wrote:
> > On Thu 2020-11-19 22:03:14, Jacek Anaszewski wrote:
> >> Hi Pavel, Gene,
> >>
> >> On 11/18/20 10:37 PM, Pavel Machek wrote:
> >>> Hi!
> >>>
> >>>> From: Gene Chen <gene_chen@xxxxxxxxxxx>
> >>>>
> >>>> Add LED_COLOR_ID_MOONLIGHT definitions
> >>>
> >>> Why is moonlight a color? Camera flashes are usually white, no?
> >>>
> >>> At least it needs a comment...
> >>
> >> That's my fault, In fact I should have asked about adding
> >> LED_FUNCTION_MOONLIGHT, it was evidently too late for me that evening...
> >
> > Aha, that makes more sense.
> >
> > But please let's call it "torch" if we do that, as that is already
> > used in kernel sources... and probably in the interface, too:
>
> I'd say that torch is something different that moonlight,
> but we would need more input from Gene to learn more about
> the nature of light emitted by ML LED on his device.
>
> Please note that torch is usually meant as the other mode of
> flash LED (sometimes it is called "movie mode"), which is already
> handled by brightness file of LED class flash device (i.e. its LED class
> subset), and which also maps to v4l2-flash TORCH mode.
>

It's used to front camera fill light.
More brightness than screen backlight, and more soft light than flash.
I think LED_ID_COLOR_WHITE is okay.

> > ./arch/arm/mach-pxa/ezx.c:                      .name = "a910::torch",
> >
> > Best regards,
> >                                                       Pavel
> >
>
> --
> Best regards,
> Jacek Anaszewski