On 2025-01-15 02:56, Simon Ser wrote: > Is this "ignore" something we could do at the core DRM level, instead > of doing it in all drivers? e.g. by silently ignoring user-space requests > to set the property? > I think it'd be better to reject setting the property. The problem is that a client without the COLOR_PIPELINE client cap might have set these properties and we'll want to ignore that when we are programming a commit with a color pipeline. Since it's the drivers that decide on the programming based on the states it's up to each driver to ignore the deprecated properties. > It sounds like this codepath still resets the colorspace to sRGB, which > is later overwritten by colorops pulled in the atomic state anyways. We didn't yet define a colorop for COLOR_RANGE and COLOR_ENCODING so the color pipeline currently only supports RGB content. The intention is to set a reasonable default but the HW programming code should be ignoring this for RGB content. Harry
