On Sun, Feb 05, 2006 at 06:21:13AM +0200, Paul Sokolovsky wrote:
> Hello Shawn,
>
> > Couldn't get h1950 buttons driver to work either (as built-in) - it
> > kind of loads, but it's probe() function is not called so it doesn't
> > actually attach to keys. Dunno, maybe it can work only as the
> > module...
>
> Well, live & learn. h4000_buttons wasn't in platform_devices, so
> that's why it wasn't probed. Now it's working. Suspend also works! But
> the same LCD issue which I saw with loox420 patches is with your LCD
> code too. Even more, as your code actually queries GPIO to get
> the current power state, here's what we have:
>
> ~ # echo 0>/sys/class/backlight/pxa2xx-fb/power
>
> ~ # cat /sys/class/backlight/pxa2xx-fb/power
> 4
> ~ #
>
> Is it like GPIO state getting messed up when going to sleep mode,
> and is not reinitialized afterwards?
>
I tried to figure why this is happening, It seems like the suspend/resume
functions in pxafb.c aren't running.. but the ones in h4000_lcd.c are. I
didn't figure out why yet.
>
> Back to buttons, I propose to replace current h4000_buttons.c with
> one ported from h1900_buttons.c (the whole port is replacing h1900
> with h4000). Reasons are obvious: it supports all buttons and works
> fine, up to the current kernel coding standards and otherwise clean code,
> and of course common implementation with another model (next step
> would be to make them actally share the code).
>
>
> Going to post patch tomorrow for review.
>
ok, that sounds good.
cya,
sa
Received on Sun Feb 05 2006 - 02:45:18 EST
This archive was generated by hypermail 2.2.0 : Sun Feb 05 2006 - 02:45:23 EST