suspend on 'idle'

[C. Scott Ananian]

On Thu, Aug 7, 2008 at 1:30 AM, Mikus Grinbergs <mikus at bga.com> wrote:
> Please don't forget to think about what 'idle' means.

Chris Ball is the power guru; if you can come up with a succinct test
that ohm can do to determine if there is 'real work' going on that
should inhibit idle, I'm sure he'd be interested.  On the other hand,
ohm has both a dbus and a filesystem interface to temporarily inhibit
suspend; long CPU intensive tasks like yours should probably be using
this to give ohm a hint -- that's what olpc-update does, for instance.
 --scott

--
 ( http://cscott.net/ )