[Intel-gfx] [PATCH] i915: Add native backlight support

[Matthew Garrett]

On Tue, Feb 23, 2010 at 08:40:35AM +1000, Dave Airlie wrote:
> On Tue, Feb 23, 2010 at 3:05 AM, Matthew Garrett <mjg at redhat.com> wrote:
> > Some machines do not provide a firmware or platform interface to backlight
> > control, and so it's necessary to drive the hardware directly. Add support
> > for a backlight driver to the drm. This will be flagged with the raw
> > backlight type, allowing userspace to determine whether there is a more
> > favourable backlight device available.
> >
> 
> So there is one backlight, and a few different methods exposed to userspace
> to use it? does the kernel not know which one it should expose? how does
> userspace know any better?

Modules may be loaded at any time. The kernel could have some sort of 
arbitration mechanism, but there's probably going to be some level of 
heuristics involved - we could potentially do that in future, but it's a 
moderate amount of extra complexity for no especially strong gain. 
Userspace ought to be able to work things out based on the backlight 
type attribute.

-- 
Matthew Garrett | mjg59 at srcf.ucam.org