Re: [PATCH v5 02/31] drm/i915: Don't register backlight when another backlight should be used (v2)

[Matthew Garrett <mjg59@srcf.ucam.org>]

On Thu, Oct 27, 2022 at 10:51:45AM +0200, Hans de Goede wrote:

> In their backlight register paths and this has been present since
> circa 2015.
> 
> So both before and after my 6.1 refactor vendor is only preferred
> on devices which don't implement the ACPI video bus control method.

Sorry, yes, that's the case I meant.

> Just because a vendor interface is present does not mean that it will
> work. Unfortunately for none of the 3 main native/acpi_video/vendor
> backlight control methods the control method being present also guarantees
> that it will work. Which completely sucks, but it is the reality we
> have to deal with.

But traditionally that's been logic that we've encoded into the vendor 
drivers, which can take other factors into account when determining 
whether the exposed interface works. You've now discarded that 
knowledge. The only way you can maintain the degree of functionality 
that 6.0 had is to move that determination into core code, or 
alternatively support dynamic reattachment of backlight interfaces based 
on vendor drivers loading later. An alternative would be to just revert 
all of this, and instead only use this logic for the output property 
interface (which would still result in different outcomes, but only for 
userland that's choosing to use the new interface, so that's a different 
problem).