Insomniac screen can't dpms suspend/standby/off - wakes instantly

[Zorael]

Perhaps this is not as KDE-specific as I had imagined. Should I take
it to some other mailing list?

On Sun, Oct 25, 2009 at 10:35 PM, Zorael <zorael at gmail.com> wrote:
> Hello,
>
>
> I have an MSI Wind U100 netbook clone, which exhibited the brightness
> bug that recently urged Felix Geyer (debfx)'s powerdevil patch into
> kdebase-workspace (https://bugs.launchpad.net/bugs/415023). The screen
> of the same machine never sleeps, and it's been doing this for a good
> while now, possibly throughout Karmic but worked in Jaunty.
>
> Whenever it tries to dpms suspend/standby/off, the screen properly
> goes dark, and then wakes instantly. When it does this, the brightness
> OSD shows. Running dbus-monitor --system at the time, it reveals that
> three brightness key events are apparently firing; down,down,down if
> at 0% brightness, down,down,up otherwise.
>
> debfx postulates (at the bug report linked above) that "[the] MSI Wind
> seems to send brightness up/down key events on every brightness
> change", and from what I gather, that the kernel seems to simulate a
> keypress at the same time. This caused it to zoom up to 100%
> brightness and then down to 0% whenever you tried to change brightness
> (with keys or via the applet), or to oscillate between brightness
> settings - all the while swamping the system with dbus messages with
> soaring cpu usage.
>
> Could this be an extension of the same issue? Could the hardware (or
> kernel) also be sending/simulating keypresses when the screen turns
> off, immediately causing it to wake again? Could powerdevil be told to
> ignore brightness key events when listening for activity to break
> dpms, perhaps across the board or at least if
> laptop_panel.brightness_in_hardware is set to true? I was informed
> that wouldn't be easy, so perhaps it's the wrong approach.
>
> [18:21] <debfx> there is nothing powerdevil can do to prevent key events
> [18:23] <debfx> hmm well yeah it could probably ignore them
> [18:23] <debfx> i'll have a look at the code
> [18:38] <debfx> doesn't seem that easy, by default powerdevil uses the
> xsync extension to get notified on user activity
>
> Thankfully it's a netbook so I can close the lid, otherwise the screen
> will just unsuccessfully try to doze off every n minutes (as per
> powerdevil settings) in perpetuity. The machine model/series itself is
> plagued by various issues in Karmic, notably uvcvideo causing USB to
> stop working (https://bugs.launchpad.net/bugs/435352), and no USB is
> sort of a showstopper for a netbook. Both that bug and the brightness
> bug have been listed as "also affects ubuntu-release-notes" in LP, so
> hopefully it'll show up in Known Issues at release.
>
>
> Thanks in advance,
> JR
>