debugging backlight issues

[Stefan Bader]

On 06/23/2010 02:38 AM, David Tombs wrote:
> On 06/17/2010 02:50 PM, Kamal Mostafa wrote:
>> On Thu, 2010-06-17 at 11:05 -0400, David Tombs wrote:
>>
>>>> param "acpi.brightness_switch_enabled=0" 
>>>
>>> Do you know of any documentation for this?
>>
>> No, I just observed it in the source -- It is present in Ubuntu kernel
>> source (Lucid onward, at least) but you're right, it does not in fact
>> disable acpi control of the backlight after all.
>>
>> As Stefan recommended, acpi_backlight=vendor might be interesting to
>> you.  I find that it does disable acpi's control of the backlight on one
>> of my laptops but has no effect on another laptop.  (The laptop where it
>> has no effect is likely already using a vendor-specific backlight
>> driver.)
>>
>>  -Kamal
> 
> Thanks Kamal,
> 
> I'm now trying to debug an issue where the backlight keys work but not
> properly (jumping steps, etc.). I think it's an issue where single
> keypresses are getting routed to multiple places and therefore the
> brightness is adjusted multiple times.
> 
> So, I ask now, what are the different routes in which brightness will
> get adjusted? I'm thinking there are three:
> 
> 1) Pure hw, like we discussed before.
> 
> 2) Pure kernel ACPI, which brightness_switch_enabled=0 would turn off.
> 
> 3) Userspace power manager like gnome-power-manager, which can be turned
> off by killing the process.
> 
> Is this correct?
> 
> Thanks in advance,
> David
> 

There also might be hw controlled plus acpi events in some odd cases. Or there
is a acpi driver which translates acpi events into input events and accidentally
still a script that is triggered by acpi events, too.

Stefan