[oe] [RFC] ipaq-sleep backlight control

[Paul Sokolovsky]

Hello,

On Thu, 20 Mar 2008 07:40:27 -0500
"Andrew Paulsen" <andrew.paulsen at gmail.com> wrote:

> I did a code cleanup on ipaq-sleep and added the option to use your
> own backlight control script instead of gpe-conf to adjust brightness.
> 
> http://bugs.openembedded.org/show_bug.cgi?id=4101
> 
> I have since heard that all 2.6 devices use the standard
> /sys/class/backlight/*/ method to control brightness and that we don't
> need/want to support others.
> 
> So I have three options and I'd like some opinions:
> 1) Leave ipaq-sleep the way it is and ship a standard backlight
> control script that can be overridden/changed per machine.  This still
> allows the legacy gpe-conf method (the default right now)
> 2) Have ipaq-sleep write directly to /sys/class/backlight/*/ to tweak
> the brightness and skip this whole mess
> 3) Have the option to select between 1&2 via conf file entry

IMHO, ipaq-sleep should go to /dev/null. Long-term, it should be
replaced by that mythical "machine daemon" which sometimes appear in
blogs and ML conversations. Short-time, I wrote another, sane
and efficient backlight control daemon, packages/backsaver/ in OE. That
still leaves question on suspending device after timeout open though.
My plan was to do gradual conversion (disable backlight control in
ipaq-sleep, use backsaver for that, with it still using ipaq-sleep's
config), but I had to put that on hold, as there're just too many
things I have in works already...

> 
> Thanks!
> Andrew

[]


-- 
Best regards,
 Paul                          mailto:pmiscml at gmail.com