[Intel-gfx] Legacy backlight control with KMS and BACKLIGHT property

[Jesse Barnes]

On Mon, 17 Aug 2009 15:56:50 +0800
Zhenyu Wang <zhenyuw at linux.intel.com> wrote:

> On 2009.08.14 19:54:05 +0200, Matthias Hopf wrote:
> > What we're supposing is
> > 
> > - a kernel driver for the /sys/class/backlight/ api for the legacy
> > method
> 
> yeah, I think /sys/class/backlight/ should be the only backlight
> interface for any user space apps or X RandR property creation. So we
> need to add backlight device inside i915 KMS driver,
> when !acpi_video_backlight_support().
> 
> Current KMS driver like LVDS only do PWM operation, maybe combo will
> be better if it's supposed to be, Jesse?

We'll want PWM+LBB and plain LBB at a minimum; not sure if we need a
plain PWM variant.

> Current x-v-i driver in KMS could setup any property for all
> connectors, like attached testing patch which will hook the property
> up. But if we just go with opening /sys/class/backlight/ in xorg
> driver, we simply don't need it...

Right, we can just expose whatever /sys/class/backlight interface we
find through x-v-i for xrandr/xbacklight.  No need for the
BACKLIGHT_CONTROL property.

> So for X BACKLIGHT property creation, we'll have a checklist for
> acpi_video, platform_driver, and new backlight device maybe like
> 'intel_lvds'.

If we see more than one interface in /sys/class/backlight we should
also consider it a kernel bug (unless they control different displays
somehow; but the /sys/class/backlight API would need work for that).

-- 
Jesse Barnes, Intel Open Source Technology Center