On Jan 25, Gabriel Burt <gburt@xxxxxxxxxx> wrote: > On Sun, 2009-01-25 at 17:44 +0100, Marco d'Itri wrote: > > People usually want to use pm-utils to suspend the system instead of > > the raw kernel interface, so I added an option to just exit after > > configuring the wakeup time. > Isn't this what the "on" option did already? No, "on" keeps waiting until the alarm fires. Even worse, if you do something like "rtcwake -m on -s 60 & pm-utils suspend" then rtcwake will miss the event and stay around forever keeping the rtc fd open and preventing other applications from accessing it. I am not attached to these changes, but something is needed to make rtcwake actually useful. I can't see how it could be used with pm-utils in its present form nor the use cases for directly accessing the kernel suspend interface. > And why would you not want > to sync() before exiting? You want to make sure the timer is fully setup > before exiting and suspending via pm-utils, I would think. pm-utils itself runs sync (/usr/lib/pm-utils/bin/pm-action). -- ciao, Marco
Attachment:
signature.asc
Description: Digital signature
