[Intel-gfx] [PATCH v3 2/2] drivers: i915: Default max backlight brightness value

[Olof Johansson]

Hi,

On Tue, Nov 8, 2011 at 3:11 PM, Matthew Garrett <mjg59 at srcf.ucam.org> wrote:
> On Tue, Nov 08, 2011 at 03:02:00PM -0800, Olof Johansson wrote:
>
>> How about a DMI table check that overrides whatever is setup (or not
>> setup) from the video bios? We know exactly what platforms need this
>> so that table would be easy to specify.
>
> dmi's horribly unscalable. It's much better to have a communication
> channel that doesn't require new code for new models of the same
> platform.

Yeah, agreed.

>> I'm not sure how well this would fit into our platform layer code, it
>> would be pretty nasty to have to export the default backlight variable
>> from the i915 driver and modify it from there as well, and I'm sure
>> noone wants to see any kind of chromeos-specific code paths in the 915
>> driver (myself included).
>
> Well right now this path is (effectively) chromeos-specific. Refactoring
> the code so we just have the register readback as a single information
> source and allow the existing platform-specific code to hook in would be
> conceptually cleaner. But then maybe this is grotesque over-engineering
> and we should just hack this case.


Actually, I didn't look closely enough to the original patches from
Simon for this and your original feedback.

Looks to me like filling in the original block of 'XXX add code'
fallback is the way to go here.

Getting the raw pch clock isn't hard, and setting a reasonable value
for the modulation frequency and the corresponding max duty cycle
based on that should solve our problem.


-Olof