power management

[Gopi P.M.]

Neat...
So do you have a proc/sysfs interface to easily initiate suspend?
How are you hooked up to  userspace?
Do you have plans to integrate DPM user interface apis? Or go with keep it
simple route and write a simple driver exposing some ioctls?

Assuming we have the basic suspend before sleep and resume on interrupts
from wakeup sources working, do you have any roadmap or next steps?

Next big thing is how to control initiation of suspend and how to  make this
automated based on various factors:

   - user inactivity should automatically initiate sleep/suspend
   - however if mp3/mpg movies are being played..the CPU may seem idle
   but system should not initiate suspend.
   - Certain devices would want to prevent suspend for genuine reasons
   such as USB data transfer, active wifi connection?? but as soon as the
   activity is complete the system should automatically try to sleep.
   - Some user apps may want to keep the system awake for good or for a
   period of time and hence may want to participate in the process of sleep
   initiation. Perhaps an api to increment sleep count from different user app
   processes may be needed. as long as the sleep count is  non-zero, the system
   is kept awake. Upon clearing of all sleep count,  system may attempt to go
   to sleep.
   - while you are constantly attempting to go to sleep when there is no
   activity, there may be situations when you just wakeup from sleep and you
   want to wait for a certain period before reinitiating sleep..to avoid waking
   up and immediately going back to sleep.
   - At the same time, if wakeup interrupts are generated accidentally,
   such as a key press by accident, you may still want to go back to sleep as
   soon as you can detect inactivity upon wakeup.

Can you think of any other optimizations or interfaces with userapps that
affect PM?


thanks,
Gopi





On 3/29/07, Jordan Crouse <jordan.crouse at amd.com> wrote:
>
> On 29/03/07 21:55 -0500, Gopi P.M. wrote:
> > I havnt setup my build environment yet, but hope to try this as soon as
> I
> > get to it.. I was wondering if anyone else has experienced the same? Any
> > thoughts?
>
> We've already gotten rid of them.  Thanks.
> Jordan
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.laptop.org/pipermail/devel/attachments/20070329/989c8c92/attachment.html>