[JB] CM10 Android 4.1.2 for the Defy(+)

ccav2000 · Feb 21, 2013 at 8:04 AM

Did anybody encountered excessive battery drain in Quarx's 20/02? I seem to get excessive wakelocks from mediaserver.

---------- Post added at 10:04 AM ---------- Previous post was at 09:50 AM ----------

Also my logs seem to be spammed by one repeating entry:

Code:

AudioHardwareMot( 1204): Output bufSize from kernel = 8192

ktmbikerman · Feb 21, 2013 at 8:32 AM

ccav2000 said:
Did anybody encountered excessive battery drain in Quarx's 20/02? I seem to get excessive wakelocks from mediaserver.

---------- Post added at 10:04 AM ---------- Previous post was at 09:50 AM ----------

Also my logs seem to be spammed by one repeating entry:

Code:

AudioHardwareMot( 1204): Output bufSize from kernel = 8192

Same here...

Sent from my MB526 using xda premium

ccav2000 · Feb 21, 2013 at 9:56 AM

Managed to get rid of the Media Server drain by deleting my music files and other crap from my SD. My research shows that Media Server problems are very common, and it's only fixed in Android 4.2.2.

AnTerNoZ · Feb 21, 2013 at 9:58 AM

AW: [JB] CM10 Android 4.1.2 for the Defy(+)

hunyadia said:
Guys, I am lost little bit, please help. Can I use seeder for rngd entropy or the febr20 build include this feature ?
Thx

Sent from my MB526 using Tapatalk 2

seeder binary is now integrated in our CM10 builds, u don't need the app to control it. look in running processes with system tuner or some tool that show all services and u see system/bin/rngd

tombert.at · Aug 26, 2012 26 7 0

Battery drain

ccav2000 said:
Did anybody encountered excessive battery drain in Quarx's 20/02? I seem to get excessive wakelocks from mediaserver.

---------- Post added at 10:04 AM ---------- Previous post was at 09:50 AM ----------

Also my logs seem to be spammed by one repeating entry:

Code:

AudioHardwareMot( 1204): Output bufSize from kernel = 8192

Yes, after installing build 17-02 i got more than 5%/h over night ...
I am going to install the 20-02 build and see what happens.

pepcisko · Feb 2, 2007 838 627 0

yup, same here, mediaserver going crazy, dances the battery to death overnight

one more observation - I noticed there was some callrecording fix in the 20/02 build. But it still doesn't capture the whole conversation. It's cut, without capturing the end.

ktmbikerman · Feb 21, 2013 at 10:50 AM

New ringtone bug

Just found a new bug on the 2002 build... muting the ringtone via the lockscreen (either using the hardware button shortcut or holding the up and down volume buttons) doesn't work anymore. The statusbar indicates that vibration mode is enabled but the ringtone is still heard loud and clear when you get an incoming call.

YetAnotherForumUser · Feb 21, 2013 at 1:25 PM

New build w/JellyX, de-fanged fmradio

For those already pining for the days of 2.6.32.60 'goodness' (minus the 'ringtone' bug, maybe even minus the 'black screen' bug?) now that Quarx reverted back to 2.6.32.9 I have some good news. I'm currently running a build which combines Quarx's repo (synced 20130220) with Blechd0se's Jellyx kernel (synced 20130220). The download link is at the bottom of this post.

I patched:

pathmap.mk to get the fmradio stuff to build (apply in build/);

Code:

diff --git a/core/pathmap.mk b/core/pathmap.mk
index 91a7fd4..ab55fd6 100644
--- a/core/pathmap.mk
+++ b/core/pathmap.mk
@@ -101,6 +101,7 @@ FRAMEWORKS_BASE_SUBDIRS := \
            keystore \
            icu4j \
            voip \
+           fmradio \
         )
 
 #

With that patch (which I posted before but added here for completeness' sake) the fmradio code builds. Unfortunately the code is incomplete and leads to crashes, eg. when the phone is put in airplane mode. The following patch de-fangs the fmradio code by disabling the relevant services. This is a cleaner solution to the current problems than a complete revert of the branch merges which introduced this code. If you feel the urge to debug the fmradio code just build without this patch and watch systemserver crash on entering airplane mode...

SystemServer.java to exclude the fmradio stuff (apply in framworks/base):

Code:

diff --git a/services/java/com/android/server/SystemServer.java b/services/java/com/
index df36cd7..1d4d57d 100644
--- a/services/java/com/android/server/SystemServer.java
+++ b/services/java/com/android/server/SystemServer.java
@@ -74,8 +74,10 @@ import java.util.Timer;
 import java.util.TimerTask;
 
 import java.lang.reflect.Constructor;
+/* FDL20130221: disabled, not complete, causes segfaults, no hardware implementation
 import com.stericsson.hardware.fm.FmReceiverService;
 import com.stericsson.hardware.fm.FmTransmitterService;
+ */
 
 class ServerThread extends Thread {
     private static final String TAG = "SystemServer";
@@ -478,7 +480,7 @@ class ServerThread extends Thread {
             } catch (Throwable e) {
                 reportWtf("starting ThrottleService", e);
             }
-
+/* FDL20130221: disabled
             try {
                 Slog.i(TAG, "FM receiver Service");
                 ServiceManager.addService("fm_receiver",
@@ -494,6 +496,7 @@ class ServerThread extends Thread {
             } catch (Throwable e) {
                 Slog.e(TAG, "Failure starting FM transmitter Service", e);
             }
+ */
             try {
                 Slog.i(TAG, "UpdateLock Service");
                 ServiceManager.addService(Context.UPDATE_LOCK_SERVICE,

This build also contains the kernel logger and phone strings patches from sevenrock's post #15221.

The kernel configuration differs from 'stock JellyX' in the following ways:

Using the SLUB allocator (default is SLAB)
Only interactive/boosted/performance/convervative/powersave/ondemand governors included (default includes a host of fancy governors)
debugfs enabled
sweep2wake disabled (doesn't work yet, once it works it'll probably cause extra battery drain)

BBS tells me the average battery use is 0.7% per hour (data off, Wifi on). In airplane mode it goes down to about 0.3% per hour. All these values are of course related to overclock/voltage and signal strength parameters, YMMV.

Download: cm-10-20130216-jellyx-UNOFFICIAL-mb526.zip (151.6 MB)
md5sum: [FONT="Monospace, Courier New"]49675f07f97a4072bac083edf50bccc1[/FONT]

pepcisko · Feb 21, 2013

YetAnotherForumUser said:
For those already pining for the days of 2.6.32.60 'goodness' (minus the 'ringtone' bug, maybe even minus the 'black screen' bug?) now that Quarx reverted back to 2.6.32.9 I have some good news. I'm currently running a build which combines Quarx's repo (synced 20130220) with Blechd0se's Jellyx kernel (synced 20130220). The download link is at the bottom of this post.

is the link correct? build.prop says

Code:

ro.build.date=Sat Feb 16 20:56:42 CET 2013

and you mentioned the repo was synchronized on 2/20

Blechd0se · Feb 21, 2013 at 2:30 PM

YetAnotherForumUser said:
The kernel configuration differs from 'stock JellyX' in the following ways:

Using the SLUB allocator (default is SLAB)

Only interactive/boosted/performance/convervative/powersave/ondemand governors included (default includes a host of fancy governors)

debugfs enabled

sweep2wake disabled (doesn't work yet, once it works it'll probably cause extra battery drain)

SLUB allocator could cause problems with wifi, in fact the only allocator which worked fine for me was SLAB, even my backported SLQB caused problems. If you haven't such problems, tell me please what patch did you applied?

Sweep2wake is far away from working at the currently stage. I need somehow the right display sizes and if you enable it at the moment the touchscreen can behave very wired.

YetAnotherForumUser · Feb 21, 2013 at 2:32 PM

pepcisko said:
is the link correct? build.prop says

Code:

ro.build.date=Sat Feb 16 20:56:42 CET 2013

It has said that for quite a few builds now. Check the kernel info...

pawelsz · Feb 21, 2013 at 2:32 PM

Quarx 20130220 battery life very good, but after 1day of using ROM, after each call phone app crashing

YetAnotherForumUser · Feb 21, 2013 at 2:33 PM

Blechd0se said:
SLUB allocator could cause problems with wifi, in fact the only allocator which worked fine for me was SLAB, even my backported SLQB caused problems. If you haven't such problems, tell me please what patch did you applied?
Sweep2wake is far away from working at the currently stage. I need somehow the right display sizes and if you enable it at the moment the touchscreen can behave very wired.

I have been using SLUB for as long as I've built Android kernels. I have never had any problems with Wifi... maybe those problems are not (directly) related to the allocator?

frullewulle · Feb 21, 2013 at 2:33 PM

YetAnotherForumUser said:
BBS tells me the average battery use is 0.7% per hour (data off, Wifi on). In airplane mode it goes down to about 0.3% per hour. All these values are of course related to overclock/voltage and signal strength parameters, YMMV.

What app do you mean with BBS?

pepcisko · Feb 21, 2013 at 2:36 PM

he meant BetterBatteryStats

ktmbikerman · Feb 21, 2013 at 2:37 PM

frullewulle said:
What app do you mean with BBS?

Better Battery Stats... do a search on XDA for a free version, or support the developer by getting it in the Playstore

ccav2000 · Feb 21, 2013 at 3:25 PM

Has anyone noticed that the compass doesn't work in Google Maps anymore since a couple of months? It appears to be working fine in my Nexus 7 so there must be something wrong with our ROM.

Sent from my Nexus 7 using xda app-developers app

Kayant · Feb 21, 2013 at 3:40 PM

YetAnotherForumUser said:
It has said that for quite a few builds now. Check the kernel info...

Yh it uses the same build.prop as the last build each time you recompile expect you delete the build.prop before recompiling or you can do a make clean or make installclean before using the make or brunch command to recompile.....

rubjonny · Feb 21, 2013 at 3:50 PM

ccav2000 said:
Has anyone noticed that the compass doesn't work in Google Maps anymore since a couple of months? It appears to be working fine in my Nexus 7 so there must be something wrong with our ROM.

Sent from my Nexus 7 using xda app-developers app

compas works fine for me on 20/02 quarx build, I had to calibrate it first though

gkkovacs · Feb 21, 2013

Testing YAFU's 20120220 build (why is it named 0216?) with JellyX kernel.

Battery drain problems still apparent.
Before it was called Android OS, now it's called *wakelock*...

The fact remains: out of 25 minutes uptime, something kept the (otherwise idle) phone woken up for 23.

How can I help investigate this?

*** EDIT *** It only showed *wakelock* for the first half hour of uptime, now it's back to Android OS and the rest.

[JB] CM10 Android 4.1.2 for the Defy(+)

Senior Member

Senior Member

Senior Member

Senior Member

Member

Senior Member

Senior Member

Senior Member

Senior Member

Inactive Recognized Developer

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member

Senior Member