Marek On 4/2/19 11:17 AM, Marek Behun wrote: > On Tue, 2 Apr 2019 06:40:53 -0500 > Dan Murphy <dmurphy@xxxxxx> wrote: > >> I have had off line conversations with Jacek about this brightness model node. >> >> Your concern was actually one of my concerns as well. Not only millions of entries but also having a huge >> DT binary. >> >> We wanted to RFC this to get feedback. And this is why I have not added any support for this in the framework code. >> >> Dan >> > > What if we just stuck to color spaces and defined channel curves in the > device tree? > Vesa did a good summary of the current MC framework implementations that are being proposed The thread also documents Pavel's requirements https://lore.kernel.org/patchwork/patch/1037147/ And here is the series with a single brightness file and passing in multiple integers. I did not document this code as I was pretty certain it was not the way to go. https://patches.linaro.org/project/linux-leds/list/?series=18022 Dan > The curve could then be defined via an array of integer pairs, which > would be interpreted as points via which the curve passes and the curve > would be approximated by linear segemts... > > led@0 { > colorspace = <COLOR_SPACE_ID_RGB>; > red-curve = <0 0>, <255 255>; > green-curve = <0 0>, <255 128>; > blue-curve = <0 0>, <255 128>; > }; > -- ------------------ Dan Murphy
