Changes in the handling of sleep in idlesuspend (rtcwake?) for F11 based builds?

Martin Langhoff martin.langhoff at gmail.com
Thu Nov 5 08:44:25 EST 2009


On Wed, Nov 4, 2009 at 9:21 PM, Martin Langhoff
<martin.langhoff at gmail.com> wrote:
> More details at http://dev.laptop.org/ticket/8177 -- perhaps we were
> waiting on something in kernel land that may now be there...?

Quozl made some tests but using rtcwake in that mode just sends the
machine to sleep. It's not what we want.

What we have here is a cronjob that we want to try ever ~30 minutes,
but with some "jitter" to avoid hitting the XS with hundreds of backup
requests.

So we start a shell script and "/bin/sleep" it for a random time.
Chris pointed out there's a good chance the machine will go into deep
sleep (as in S1) and not wake up to return.

In my view, it doesn't matter. If /bin/sleep returns only once a user
action has awoken the XO, that's fine. /bin/skleep never promised to
be exact.

OTOH, if we're to find a powermgmt-friendly /bin/sleep I think we
could combine calling "rtcwake --mode no --seconds 90" (which
schedules a wakeup and returns, without actually going to sleep) with
a 'sleep 90'.

In summary -- not a problem. I think we can punt on this.



m
-- 
 martin.langhoff at gmail.com
 martin at laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff


More information about the Devel mailing list