Comment 43 for bug 1970069

I now expect the kernel patch would change before it gets proposed to dri-devel. In particular the primary delay doesn't need to be configurable (or even finite?), and the secondary delay probably doesn't need to be more than zero. But the feature would still be gated on the "splash" kernel parameter which is Plymouth-specific and may cause pushback from kernel developers.

It's starting to make more sense to implement the console disabling using ioctls in plymouthd itself. Because I think plymouthd is running early enough to do that, it's only refusing to render for several seconds so fbcon wins the race to the framebuffer.