Re: [PATCH] drm/kms: fix fbdev blanking regression

[Alex Deucher <alexdeucher@gmail.com>]

On Wed, Jan 13, 2010 at 2:48 PM, James Simmons <jsimmons@infradead.org> wrote:
>
>> > > On 01/07/2010 12:42 AM, Johan Hovold wrote:
>> > > >> Yeap. The fix uncovered a bug in your driver. I haven't heard of
>> > > >> problems with the other drm drivers.
>> > > >>
>> > > >>> The backlight is handled via the DRI driver I assume. At least
>> > > >>> i9xx_crtc_dpms is called on powerdown.
>> > > >>
>> > > >> Can you post your dmesg and kernel config.
>> > >
>> > > [snip]
>> > >
>> > > Adding the Intel DRM people in CC as well. I have the same issue
>> > > with my GM45.
>> >
>> > Okay I looked at the code to figure out what is happening and why
>> > only this driver has problems. The problem is that the framebuffer
>> > layer expects the backlight to be a seperate device. The reason being
>> > is that some embedded systems will use a gpio backlight. That way
>> > power management for a graphics card/backlight has 3 seperate states.
>> > Currently the intel DRM driver treats the backlight as being apart of
>> > the encoder. Jesse do you have objections to having the intel driver
>> > expose a backlight device. The bonus of that is the user can also set
>> > the backlight levels.
>>
>> On Intel we usually expect the backlight to be exposed by ACPI or a
>> platform driver.  On recent platforms, the ACPI driver will actually
>> send requests to the gfx driver to do the actual register writes to
>> adjust the backlight, but it's still ACPI driven.
>>
>> Maybe we just need to wire up the fb backlight hooks appropriately?
>
> Yeap. I can have patch ready for you this weekend. I lack the hardware to
> test it tho. Never been able to find a intel pci card that is not built
> into the motherboard.

They don't exist, other than the old i740s.

Alex