stargo cm-11-201402126 build doesn't deep sleep as much as it should

Search This thread

stashcroft

Member
Dec 1, 2013
17
5
(I'd post this to the development thread but I don't have enough posts yet.)

I seem to have problems getting the phone into deep sleep since the build of the 23rd.
Better battery stats is showing the problem is the 'PowerManagerSevice.WakeLocks' wakelock.
I have that 11.3% of the time which at the moment is around 3hrs.
However, the sum of the partial wakelocks is under 3 minutes.

How can I find out why the wakelock was held for the other 2hrs 57min?

Thanks,
Scott
 

stargo

Inactive Recognized Developer
Jan 7, 2011
538
1,718
Fürth
Hi,

I seem to have problems getting the phone into deep sleep since the build of the 23rd.
Better battery stats is showing the problem is the 'PowerManagerSevice.WakeLocks' wakelock.
I have that 11.3% of the time which at the moment is around 3hrs.
However, the sum of the partial wakelocks is under 3 minutes.

How can I find out why the wakelock was held for the other 2hrs 57min?

How much time was the suspend_backoff (kernel) wake_lock active?

If this is an huge amount of time, the you are affected by the USBHS problem which can then only be (temporarily) cured by a reboot.

If it's "just" PowerManagerSevice.WakeLocks then it's most probably an application responsible for your problem.

Regards,
Michael
 
  • Like
Reactions: stashcroft

stashcroft

Member
Dec 1, 2013
17
5
Hi,



How much time was the suspend_backoff (kernel) wake_lock active?

If this is an huge amount of time, the you are affected by the USBHS problem which can then only be (temporarily) cured by a reboot.

If it's "just" PowerManagerSevice.WakeLocks then it's most probably an application responsible for your problem.

Regards,
Michael

Not seeing any time in suspend_backoff. I think the fix for that on the 25th has worked.
I'll keep digging around to see if I can figure out why I'm suddenly seeing time in PowerManagerSevice.WakeLocks.
I've not installed any new apps.

Cheers,
Scott
 

stashcroft

Member
Dec 1, 2013
17
5
Not seeing any time in suspend_backoff. I think the fix for that on the 25th has worked.
I'll keep digging around to see if I can figure out why I'm suddenly seeing time in PowerManagerSevice.WakeLocks.
I've not installed any new apps.

Cheers,
Scott

Looks like it is something in bluetooth. I've got bluetooth turned on but not connected and the process com.android.bluetooth is running.
Running 'dumpsys power' every few seconds shows that the StartingObexMapTransaction partial wakelock is always taken.
BBS doesn't see this and says that lock has only been taken for a few minutes where as using dumpsys shows it taken for hours.

So I wonder if there's a path in the bluetooth app which doesn't release the lock or a bug in the kernel where the bluetooth app calls to release the lock, BBS sees that but in fact the lock isn't released.

Cheers,
Scott
 

cooper_s1

Senior Member
Aug 23, 2005
88
12
Looks like it is something in bluetooth. I've got bluetooth turned on but not connected and the process com.android.bluetooth is running.
Running 'dumpsys power' every few seconds shows that the StartingObexMapTransaction partial wakelock is always taken.
BBS doesn't see this and says that lock has only been taken for a few minutes where as using dumpsys shows it taken for hours.

So I wonder if there's a path in the bluetooth app which doesn't release the lock or a bug in the kernel where the bluetooth app calls to release the lock, BBS sees that but in fact the lock isn't released.

Cheers,
Scott
I've got exactly this issue on ProBam ROM. Is there any fix for this as its a battery waster... ;)
 
  • Like
Reactions: stashcroft

stashcroft

Member
Dec 1, 2013
17
5
I've got exactly this issue on ProBam ROM. Is there any fix for this as its a battery waster... ;)

I've still not managed to figure out what is going on.
At the moment I just reboot the phone or kill the com.android.bluetooth process which stop it until I use bluetooth again.

I'll try to see if I can get some useful logging for the bluetooth process.

It is almost as if when the bluetooth connection ends (e.g. I walk away from my car) the phone tries to reopen a connection but it fails and the failure path doesn't release the wakelock.

Thanks for confirming it isn't just me. :)

Cheers,
Scott
 
  • Like
Reactions: cooper_s1

cooper_s1

Senior Member
Aug 23, 2005
88
12
I've still not managed to figure out what is going on.
At the moment I just reboot the phone or kill the com.android.bluetooth process which stop it until I use bluetooth again.

I'll try to see if I can get some useful logging for the bluetooth process.

It is almost as if when the bluetooth connection ends (e.g. I walk away from my car) the phone tries to reopen a connection but it fails and the failure path doesn't release the wakelock.

Thanks for confirming it isn't just me. :)

Cheers,
Scott

Can't wait to see your logs. I feel the same about your theory.

Verstuurd vanaf mijn GT-N7100 met Tapatalk
 

cooper_s1

Senior Member
Aug 23, 2005
88
12
Is there any movement on this issue? I've flashed the latest version of ProBam, 1.2.9, and still have this issue...
The only workaround seems to stop bluetooth and reboot.
 

stashcroft

Member
Dec 1, 2013
17
5
This seems to be fixed by the raft of bluetooth changes in cm-11-20140226-NIGHTLY-maserati.

Cheers,
Scott
 

Top Liked Posts

  • There are no posts matching your filters.
  • 1
    Hi,

    I seem to have problems getting the phone into deep sleep since the build of the 23rd.
    Better battery stats is showing the problem is the 'PowerManagerSevice.WakeLocks' wakelock.
    I have that 11.3% of the time which at the moment is around 3hrs.
    However, the sum of the partial wakelocks is under 3 minutes.

    How can I find out why the wakelock was held for the other 2hrs 57min?

    How much time was the suspend_backoff (kernel) wake_lock active?

    If this is an huge amount of time, the you are affected by the USBHS problem which can then only be (temporarily) cured by a reboot.

    If it's "just" PowerManagerSevice.WakeLocks then it's most probably an application responsible for your problem.

    Regards,
    Michael
    1
    Looks like it is something in bluetooth. I've got bluetooth turned on but not connected and the process com.android.bluetooth is running.
    Running 'dumpsys power' every few seconds shows that the StartingObexMapTransaction partial wakelock is always taken.
    BBS doesn't see this and says that lock has only been taken for a few minutes where as using dumpsys shows it taken for hours.

    So I wonder if there's a path in the bluetooth app which doesn't release the lock or a bug in the kernel where the bluetooth app calls to release the lock, BBS sees that but in fact the lock isn't released.

    Cheers,
    Scott
    I've got exactly this issue on ProBam ROM. Is there any fix for this as its a battery waster... ;)
    1
    I've got exactly this issue on ProBam ROM. Is there any fix for this as its a battery waster... ;)

    I've still not managed to figure out what is going on.
    At the moment I just reboot the phone or kill the com.android.bluetooth process which stop it until I use bluetooth again.

    I'll try to see if I can get some useful logging for the bluetooth process.

    It is almost as if when the bluetooth connection ends (e.g. I walk away from my car) the phone tries to reopen a connection but it fails and the failure path doesn't release the wakelock.

    Thanks for confirming it isn't just me. :)

    Cheers,
    Scott