You can track who holds and releases wake_lock. I looked into this a bit sometimes back, to reduce this exact delay, but shortening the delay is not a quick hack it seems.
The delay between 'key-press' and 'actual-suspend' is mostly dicated by user side, which are mostly due to wakelock held by below components -system activity wakelock management is at ./frameworks/base/services/java/com/android/server/PowerManagerService.java -key events handling from ./frameworks/base/libs/ui/EventHub.cpp its also very possible that a kernel driver also holding a wakelock for some timeout period.. On Tue, Dec 7, 2010 at 11:26 AM, Arun Joseph <[email protected]> wrote: > Hi, > > I am trying to get suspend/resume enabled in OMAP3 EVM with rowboat > sources. For this I enabled power management feature (set hw.nopm to > false). When I press power key, early-suspend handlers are getting > executed and surfaceflinger is down. But It takes 6-7 seconds to enter > into linux suspend procedures (enter_state). > > What would be the reasons for there is a delay between > request_suspend_state and enter_state invokations? > > I see an new wake lock introduced to the system: lock name: alarm lock > type WAKE_LOCK_SUSPEND timeout 5.0 seconds after I press power key. > > Can someone help me in understanding where this wakelock is getting > introduced from? Is this wakelock delaying the suspend procedure? > > Thanks, > Arun > > -- > unsubscribe: > [email protected]<android-porting%[email protected]> > website: http://groups.google.com/group/android-porting > -- Regards, Deva www.bittoggler.com -- unsubscribe: [email protected] website: http://groups.google.com/group/android-porting
