• Introducing XDA Computing: Discussion zones for Hardware, Software, and more!    Check it out!

Official Cyanogenmod 12 Lollipop

Search This thread

PsychoGame

Senior Member
May 18, 2012
294
1,652
Koudum
Hello everybody,

I'm starting to grow a little bit tired of CyanogenMod Team.
I uploaded my workaround some time ago and spoke to CM Team dev which promised me he would push the workaround through.
Unfortunately he didn't keep his promise, and so the fix is still pending.
To be honest I don't think the workaround will be pulled in as there is to less interest in it by the CM Team devs in my opinion.
In the meantime I'm still working on the updated builds.
Hopefully CM will merge them in the end. (Finger crossed)
 

PsychoGame

Senior Member
May 18, 2012
294
1,652
Koudum
Hello everybody,

Today I was working with some gproj sources but in the corner of my eye I did see recognized CyanogenMod developer "rmcc" better known as Ricardo Cerqueira pushed a RIL change upstream which bypassed CM code review.
His patch can be found over here: https://github.com/CyanogenMod/andr...mmit/84e44a43ac2d0de4e3acd86322ffff43c38904e7
I'm not yet sure what impact his patch actually has but I think it's worth to keep an eye on this change.
This may be a workaround for the SIM Pin bug we are still working on.

Update 1:
I did some investigation on this pushed patch.
It should fix the constant RIL crashing which occurs on our device, due to RIL calling for an unimplemented function.
This may also fix the "deep sleep" issues, as I suspect the constantly crashing RIL may cause this but this claim I'm not sure about it's just a hypothesis which I had.
Currently I'm building CM-12.1 to try and find out the real implications of the patch and verify what has been improved exactly.

Update 2:
This is the second update.
The build I tried worked like a charm no RIL bug anymore!
It was just like I suspected due to constant crashing of the RIL,
I just suspected this due to outdated and old blobs in use, but "Ricardo Cerqueira" proved this was also fixable on the RIL side.
Unfortunately the "deep sleep" issue still hasn't been solved.
In the meantime I will just continue porting over KK blobs and try to make LG OG more CAF codebased.
This fix will assure you at least can enjoy stable CM-12.1 for now until I got everything sorted out and 100% working right.

Update 3:
I'm currently trying to find out what might cause problem of the phone not entering "deep sleep"
What I suspect currently is something with a service called qseecomd which is crashing.
When looking into kernel "dmesg" I found that the "power management" related stuff about exiting deep sleep is constantly right after qseecomd crashing.
Currently I'm busy trying to find a way to fix this issue.

Update 4:
I found out there are actually 2 wakelock issues which prevent the phone from entering "Deep Sleep".
First wakelock is: pm8xxx_adc_wakelock, which I already found a solution for. This patch I will try to upload to gerrit codereview today.
The second wakelock is a more difficult one to track down.
The name of the wakelock is "KeyEvents" but I really can't seem to find out what exactly is causing this.
What I know is that it's RIL/Radio related as when I put the phone in "Airplane Mode" this wakelock is released and phone goes into "Deep Sleep" without any problems.
I will keep investigating this problem and hopefully find a proper solution soon.
When someone has any suggestions please let me know.
And in the meantime I hope CM Team will merge my patch for the pm8xxx_adc_wakelock (Fingers Crossed)
The patch I uploaded can be found over here: http://review.cyanogenmod.org/#/c/113721/

Greetings PsychoGame
 
Last edited:

iakoboss7

Senior Member
Hello everybody,

Today I was working with some gproj sources but in the corner of my eye I did see recognized CyanogenMod developer "rmcc" better known as Ricardo Cerqueira pushed a RIL change upstream which bypassed CM code review.
His patch can be found over here: https://github.com/CyanogenMod/andr...mmit/84e44a43ac2d0de4e3acd86322ffff43c38904e7
I'm not yet sure what impact his patch actually has but I think it's worth to keep an eye on this change.
This may be a workaround for the SIM Pin bug we are still working on.

Update 1:
I did some investigation on this pushed patch.
It should fix the constant RIL crashing which occurs on our device, due to RIL calling for an unimplemented function.
This may also fix the "deep sleep" issues, as I suspect the constantly crashing RIL may cause this but this claim I'm not sure about it's just a hypothesis which I had.
Currently I'm building CM-12.1 to try and find out the real implications of the patch and verify what has been improved exactly.

Update 2:
This is the second update.
The build I tried worked like a charm no RIL bug anymore!
It was just like I suspected due to constant crashing of the RIL,
I just suspected this due to outdated and old blobs in use, but "Ricardo Cerqueira" proved this was also fixable on the RIL side.
Unfortunately the "deep sleep" issue still hasn't been solved.
In the meantime I will just continue porting over KK blobs and try to make LG OG more CAF codebased.
This fix will assure you at least can enjoy stable CM-12.1 for now until I got everything sorted out and 100% working right.

Update 3:
I'm currently trying to find out what might cause problem of the phone not entering "deep sleep"
What I suspect currently is something with a service called qseecomd which is crashing.
When looking into kernel "dmesg" I found that the "power management" related stuff about exiting deep sleep is constantly right after qseecomd crashing.
Currently I'm busy trying to find a way to fix this issue.

Update 4:
I found out there are actually 2 wakelock issues which prevent the phone from entering "Deep Sleep".
First wakelock is: pm8xxx_adc_wakelock, which I already found a solution for. This patch I will try to upload to gerrit codereview today.
The second wakelock is a more difficult one to track down.
The name of the wakelock is "KeyEvents" but I really can't seem to find out what exactly is causing this.
What I know is that it's RIL/Radio related as when I put the phone in "Airplane Mode" this wakelock is released and phone goes into "Deep Sleep" without any problems.
I will keep investigating this problem and hopefully find a proper solution soon.
When someone has any suggestions please let me know.
And in the meantime I hope CM Team will merge my patch for the pm8xxx_adc_wakelock (Fingers Crossed)
The patch I uploaded can be found over here: http://review.cyanogenmod.org/#/c/113721/

Greetings PsychoGame

hello again :)
do you have any news? or we should drop cm for another rom?
 

king-j

New member
Nov 25, 2015
2
1
I've installed latest Nightly (20151124) and for me, there doesn't seem to be any bugs anymore.
Batterydrain is as normal as in CM12 (after 10h there's still about 63% left with moderate usage), but I'm not able to test the SIM-Bug because I don't know my PIN anymore, so someone else should test it.
I came from CM12, I simply made an update with the recovery. Everything is working.
I don't know how this comes, but if someone needs any logs, I will offer them.
 

king-j

New member
Nov 25, 2015
2
1
Well, CPU Deep Sleep Info Widget says that my Phone stayed about 70% of the time since it's turned on in Deep Sleep, and I think that's a good sign.
Since my phone has only about 22% of battery left now and I'm going to bed now, I will have to check battery drain over night next night..
But stay tuned!

Update:
After one day oft moderate usage, I'm still having 54% of battery. With Wakelock Detector I could not find any dramatic wakelockers.
I think everything is said here. The build is useable as a daily driver, and the batterydrain is as normal as it could be.
Update 2:
Sim-pin is working
 
Last edited:
  • Like
Reactions: iakoboss7

doubleagent23

Member
Oct 24, 2015
23
1
26
sargodha
Camera darkness

guys is there any fix for camera darkness.. i tried cm12 snapshot and cm12.1 nightly but they all have same camera darkness problem .. everything is working fine ..except this camera darkness. .. i have tried all apps of camera but still no usefull:rolleyes::rolleyes:
 

aminlord

New member
Oct 28, 2020
3
0
hello my friend
I flashed cm12 and I find Setupwizard.apk
I deleted it and reboot.
after installing rom correctly. setupwizard error again and I can't pass it
please help me :D
 
Last edited:

Top Liked Posts

  • There are no posts matching your filters.
  • 10
    Hello everybody,

    Currently my developing time is limited due to some graduation assignments which I need to finish for my study.
    To fix the RIL problems we're having, in my opinion and also some recognized CM developers' opinion, is to get the binary blobs (at least for the RIL) up-to-date with the KK blobs or preferably higher (mako blobs for example).
    This opinion is backed by the knowledge that for some reason the mako based geehrc builds related to the gproj build do not experience the RIL/SIM PIN bugs we are experiencing.
    The blobs which are currently in use in the Jenkins build process are still from the Jellybean release which the Optimus G was originally released with onto the market.
    In my porting process I focused on using MAKO blobs as they still get updated, unlike the official Optimus G blobs.
    With a small kernel change I was able to get a lot of mako blobs working already for the E975.
    The kernel change which need to be made is the following:
    Replace drivers/usb/misc/ks_bridge.c from the gproj kernel source with mako version of ks_bridge: ks_bridge.c
    The ks_bridge.c from mako initializes the kernel sockets ks_bridge and efs_bridge which were used in gproj under ks_hsic_bridge and efs_hsic_bridge which are the new sockets used in the mako builds.
    To initialize these sockets with the right permissions for android to work with some uevent.geehrc.rc changes are necessary as well.
    Code:
    /dev/ks_bridge              0660  system        system
    /dev/efs_bridge            0660  system        system
    [COLOR="Orange"]Needs to be replaced with:[/COLOR]
    /dev/ks_hsic_bridge              0660  system        system
    /dev/efs_hsic_bridge            0660  system        system
    This change is necessary to be able to use the mako efsks an ks binaries.
    When these changes are made the original gee efsks and ks binaries are no longer usable.
    This is no problem since the mako ks and efsks also work with the JB RIL I found out.
    In the list below I highlight in green which blobs are currently working in the e975 builds I built.
    As everybody can see still a lot of files blobs need to be made compatible but at least there is some start.
    Most of the blobs which still do not work are directly related to qmuxd
    What I know is that the following list of blobs are all RIL related. I left out libril.so and the rild binary because when possible I prefer source build rild and libril.so. For now we don't know if this is possible for as long the remaining blobs still need to be ported

    Update 1:
    Using Mako blobs apparently was a mistake. A lot of kernel stuff needs to be changed that way. Maybe in the future this will be a project I will look into, but for now I switched over to E975 V20A Kitkat blobs for porting. This porting process still involves the replacement of the file ks_bridge.c from the gproj kernel source with the ks_bridge.c from the V20A kernel source available from LG Opensource Release Center. I updated the list of blobs which are already in my current rom and still prove to deliver a rom with functioning RIL (still sim/pin bug occurring so don't ask for test builds please.). Once the project is finished I make sure it gets pushed into CM github so everybody can enjoy it.

    Binaries:

    • [*]efsks
      [*]ks
      [*]netmgrd
      [*]qcks
    • qmuxd
      [*]rmt_storage
    Libraries:

    • [*]libdiag.so
    • libDiagService.so
    • libdsi_netctrl.so
      [*]libdsutils.so
      [*]libidl.so
      [*]libnetmgr.so
    • libqcci_legacy.so
    • libqdi.so
    • libqdp.so
    • libqmi.so
    • libqmiservices.so
    • libqmi_cci.so
    • libqmi_client_qmux.so
    • libqmi_common_so.so
    • libqmi_csi.so
    • libqmi_csvt_srvc.so
    • libqmi_encdec.so
    • libril-qc-qmi-1.so
    • libril-qcril-hook-oem.so

    Hopefully I've given everybody some insight in what I found out and what I've been doing. When somebody still has some questions please don't hesitate to ask, and hopefully some other developers are willing to help me getting the remaining blobs ported over to gproj.

    Greetings PsychoGame
    10
    Hello everybody,

    Today I was working with some gproj sources but in the corner of my eye I did see recognized CyanogenMod developer "rmcc" better known as Ricardo Cerqueira pushed a RIL change upstream which bypassed CM code review.
    His patch can be found over here: https://github.com/CyanogenMod/andr...mmit/84e44a43ac2d0de4e3acd86322ffff43c38904e7
    I'm not yet sure what impact his patch actually has but I think it's worth to keep an eye on this change.
    This may be a workaround for the SIM Pin bug we are still working on.

    Update 1:
    I did some investigation on this pushed patch.
    It should fix the constant RIL crashing which occurs on our device, due to RIL calling for an unimplemented function.
    This may also fix the "deep sleep" issues, as I suspect the constantly crashing RIL may cause this but this claim I'm not sure about it's just a hypothesis which I had.
    Currently I'm building CM-12.1 to try and find out the real implications of the patch and verify what has been improved exactly.

    Update 2:
    This is the second update.
    The build I tried worked like a charm no RIL bug anymore!
    It was just like I suspected due to constant crashing of the RIL,
    I just suspected this due to outdated and old blobs in use, but "Ricardo Cerqueira" proved this was also fixable on the RIL side.
    Unfortunately the "deep sleep" issue still hasn't been solved.
    In the meantime I will just continue porting over KK blobs and try to make LG OG more CAF codebased.
    This fix will assure you at least can enjoy stable CM-12.1 for now until I got everything sorted out and 100% working right.

    Update 3:
    I'm currently trying to find out what might cause problem of the phone not entering "deep sleep"
    What I suspect currently is something with a service called qseecomd which is crashing.
    When looking into kernel "dmesg" I found that the "power management" related stuff about exiting deep sleep is constantly right after qseecomd crashing.
    Currently I'm busy trying to find a way to fix this issue.

    Update 4:
    I found out there are actually 2 wakelock issues which prevent the phone from entering "Deep Sleep".
    First wakelock is: pm8xxx_adc_wakelock, which I already found a solution for. This patch I will try to upload to gerrit codereview today.
    The second wakelock is a more difficult one to track down.
    The name of the wakelock is "KeyEvents" but I really can't seem to find out what exactly is causing this.
    What I know is that it's RIL/Radio related as when I put the phone in "Airplane Mode" this wakelock is released and phone goes into "Deep Sleep" without any problems.
    I will keep investigating this problem and hopefully find a proper solution soon.
    When someone has any suggestions please let me know.
    And in the meantime I hope CM Team will merge my patch for the pm8xxx_adc_wakelock (Fingers Crossed)
    The patch I uploaded can be found over here: http://review.cyanogenmod.org/#/c/113721/

    Greetings PsychoGame
    7
    Hello everybody,

    Sorry for being quite for a while.
    Status on the bug is unfortunately still the same.
    I'm a bit stuck in porting the Mako RIL over to the gproj based devices.
    To speed things up a little, together with Seth Shelnutt from CM-Team we decided it would be wise to first just port over the KK rom as it's an official released rom for the Optimus G.
    We are still working on this issue, but things are a little more complicated than I anticipated in the beginning so it just doesn't go that fast.
    When CM Roms are working again with stable RIL, I'll continue on updating the device code and trying to rebase a lot of stuff over to mako source and binaries.
    But that's a project for later on when there's time to experiment.

    Greetings PsychoGame
    5
    Hello everybody,

    I'm starting to grow a little bit tired of CyanogenMod Team.
    I uploaded my workaround some time ago and spoke to CM Team dev which promised me he would push the workaround through.
    Unfortunately he didn't keep his promise, and so the fix is still pending.
    To be honest I don't think the workaround will be pulled in as there is to less interest in it by the CM Team devs in my opinion.
    In the meantime I'm still working on the updated builds.
    Hopefully CM will merge them in the end. (Finger crossed)
    5
    we can add problems with autorotate to bug list. it's weird. I noticed that it works only once for YouTube. hangouts rotating only chat, not main window and other (but maybe not every) apps rotating always..
    @PsychoGame how is your work going on? have you got your repos on gh?

    Sent from LG Optimus G

    Hello Dadi11,

    I'm not working through github, as I make a lot of changes during the day so uploading them everytime would result in a lot of commits which may not do anything.
    Once I fixed something I'll upload the full changes at once to gerrit code review of CM.
    Currently I'm trying to update the CM e975 codebase as much as possible with the mako code as the codebase which is build upon still originates from mako 4.3_r2.1 while mako's code is on 5.1.1 already.
    Currently the RIL unfortunately in the latest nightly builds is completely broken and isn't able to get any mobile network at all (at least in my case, I don't know if you guys experience the same problems?).
    I'm still investigating this RIL bug, and it has my top priority at this moment as this is a serious bug which renders the CM12.1 roms completely useless at this moment.
    The RIL stopped fully working between the build 09/06/2015 and 11/06/2015, so a change in CM code between these two builds should be responsible for this problem.
    Currently I'm bisecting it to find out exactly which commit.

    Greetings PsychoGame