Database Develop

Compass and Accel sensors not working, Please Help!

I seem to be having an odd problem with my new G1.
In any Cupcake build (JF 1.51, TheDude's 1.2, Official ADP) I flash, neither the compass nor the accelerometer sensors work at all. I have flashed between them several times, with an Alt+W each time, so I know it's not a bad flash.
The strange thing is, when I jump back to an official RC29 build, both sensors work fine. I haven't tried RC30 or anything else, but I know the sensors are physically functional.
I have re-flashed the latest radio as well, as it is currently using 2.22.19.26I.
If anyone can help me with this problem, I would greatly appreciate it.

gigawatts said:
I seem to be having an odd problem with my new G1.
In any Cupcake build (JF 1.51, TheDude's 1.2, Official ADP) I flash, neither the compass nor the accelerometer sensors work at all. I have flashed between them several times, with an Alt+W each time, so I know it's not a bad flash.
The strange thing is, when I jump back to an official RC29 build, both sensors work fine. I haven't tried RC30 or anything else, but I know the sensors are physically functional.
I have re-flashed the latest radio as well, as it is currently using 2.22.19.26I.
If anyone can help me with this problem, I would greatly appreciate it.
Click to expand...
Click to collapse
I think your using the pre-cupcake apps. Did u download the apps after you updated or did you install them from a back up?
Best way to check, You should got to settings and then sounds&display and check off orientation. You phone should switch from portrait to landscape when you rotate any program (except homescreen). If that works then nothing is wrong with your phone. If it doesn't then something happened somewhere weird

any apps i have installed have been after i flashed a rom (after a wipe) and not from a backup. Plus, something as simple as Google Maps or the Browser App should just rotate without opening the keyboard.
Best way to check, You should got to settings and then sounds&display and check off orientation.
Click to expand...
Click to collapse
By that I assume you place the check mark (lit green) next to the option, not "check it off" (not lit green). I have tried both checked and unchecked in pre-installed apps like maps and browser, and still nothing.

Small update. I just flashed JF's v1.43_RC33 with root, keeping the same more recent radio, and both sensors work there, as they do in RC29. It seems to be ONLY cupcake based roms where they don't work.
Any other suggestions anyone has for me? I might try a hero build next, to see if they work there, but I'm not sure how stable they are yet.
Any help/suggestions would be greatly appreciated!

Accelerometer fix
http://forum.xda-developers.com/showpost.php?p=4081673&postcount=4885

I tried your suggestion, and it did not work. Unlike the other user reporting that the sensors would work for a short time after reflashing, mine has never worked after a cupcake flash. Only after a pre-cupcake flash.
Although ideas like that are what I am looking for, possible corrupt config files that need to be cleared or maybe a mod probe command of some sort to activate or wake up the sensors.

Upon some more inspection, I get the feeling something is wrong with my akmd process, or whatever executable generates the akmd_set.txt compass and accelerometer calibration data. I tried to follow some instuctions to purge the akmd_set.txt file to clear any bad config, but the file is no where to be found. My phone doesnt seem to generate one at all when it's booting.
The akmd process does remain running in the background, as here is its ps output:
Code:
# ps | grep akmd
compass 77 1 1320 328 c020d268 afe0c45c S /system/bin/akmd
This is the output I get from logcat when I try to manually run the /system/bin/akmd process or when it restarts automatically after being killed:
Code:
D/AKMD ( 362): akmd 1.6.4 START
D/AKMD ( 362): library version: 1.2.1.1129
D/AKMD ( 362): Compass OPEN
D/AKMD ( 362): No calibration data
D/AKMD ( 362): Can not initial compass parameters error: -5
and this is the output I get when I run a program like Orienteer:
Code:
I/ActivityManager( 90): Starting activity: Intent { action=android.intent.action.MAIN categories={android.intent.category.LAUNCHER} flags=0x10200000 comp={com.deafcode.android.Orienteer/com.deafcode.android.Orienteer.Orienteer} }
I/ActivityManager( 90): Start proc com.deafcode.android.Orienteer for activity com.deafcode.android.Orienteer/.Orienteer: pid=363 uid=10041 gids={}
D/SensorManager( 363): found sensor: AK8976A 3-axis Accelerometer, handle=0
D/SensorManager( 363): found sensor: AK8976A 3-axis Magnetic field sensor, handle=1
D/SensorManager( 363): found sensor: AK8976A Orientation sensor, handle=2
D/SensorManager( 363): found sensor: AK8976A Temperature sensor, handle=3
D/dalvikvm( 363): GC freed 920 objects / 64320 bytes in 92ms
D/LocationManager( 363): Constructor: service = [email protected]
D/Sensors ( 90): sensors=00000005, real=00000005
D/GpsLocationProvider( 90): setMinTime 2000
W/IInputConnectionWrapper( 145): showStatusIcon on inactive InputConnection
I/ActivityManager( 90): Displayed activity com.deafcode.android.Orienteer/.Orienteer: 1313 ms
D/dalvikvm( 145): GC freed 795 objects / 45896 bytes in 165ms
D/GpsLocationProvider( 90): TTFF: 7203
Maybe this will give someone some more insight as to the problem I am having.

hello,
i me french and i have the same problem.
No akmd_set.txt on my phone.
If you are a solution could you partage it ?
I think it's a driver problem but i'm not sure.
JEEP

Unfortuantly I have still not found a solution, but continue to look for one. I will let you know here if I do manage to solve it.

I am also having the same problem. Wiping does not fix it and the method stated by Cyanogen didn't work either. I don't know if this will be any help, but I do know the accelerometer issue started about 3 or 4 days ago for me. Maybe if I list things I've done to my phone in the last few days, it might help narrow down the cause if others with the same issue compare their experiences to each other.
I have a US T-Mobile G1 and I have been rooted since February. Since that time I have upgraded to (and currently use) the latest radio from HTC, Haykuro's SPL, and a class 6 4gb card. I've also been through more roms than I can count in that time, all without an accelerometer issue till now.
Now that I've got all that out of the way here's the more relevant stuff:
Within the last week I have upgraded to Cyanogen's 1.3.1 recovery boot and installed the latest JAC Hero build. I repartitioned my card into 3 partitions with the directions given in JustAnotherCrowd's thread using parted in the recovery's console.
Code:
#parted /dev/block/mmcblk0
rm 1
rm 2
mkpartfs primary fat32 0 3420
mkpartfs primary ext2 3420 3932
mkpartfs primary linux-swap 3932-3964
This is the first time I've ever used apps to sd. After using that rom for about a day, I switched to Cyanogen's build and wiped my user data without wiping the ext partition of my sd card and I continued to use the swap partition on my Cyanogen build. I'm fairly sure but not 100% certain that the accelerometer was still working at this point. Soon after installing this rom, Cyanogen released an update including the compcache modules to replace the linux-swap setup. I installed the custom userinit.sh script to system/sd to activate compcache. Then Cyanogen released another update (immediate bug fixes) almost as soon as I got this rom working. I flashed the update.zip, but then it started to boot loop. I wiped the phone (not the ext partition), still boot looped. I took out the sd, still boot looped. I put the sd card back in wiped again, reflashed, and then clicked "repair ext file system". STILL BOOT LOOPED! (Don't say I just wasn't giving it enough time to boot because I acctually watched the boot animation flicker in and out for a second and start over.) Next, I used fastboot to restore only the user data from a nandroid backup of an older Cyanogen rom. This time it worked. Next, I installed the audio resources to system/sd/media and linked them using the userinit.sh file. The next day, I realized that the accelerometer was no longer functioning.
I hope this description helps to figure this out.

I'm having the exact same issue.

testing567, can you post the userinit.sh file you are currently using, and possibly you or someone else post their userinit.sh of a phone with a working accelerometer and compass here?
I am thinking that one of my init scripts may not be calling the akmd service properly, as in not with the correct permissions or something, and thus not calibrating itself on boot. I was planning on upgrading to the latest cyanogen update today, as i am running 3.6.5 right now. I was also considering upgrading my SPL to a dev SPL so i can use nandroid to backup (getting a little tired of loosing everything when attempting different roms to fix this issue).

I got my userinit.sh from this post.
http://forum.xda-developers.com/showpost.php?p=4140826&postcount=74
I also decided to take advantage of the media symlinking already in this script and pushed the audio resources to system/sd/media/.

I've always had the same issue when upgrading to a cupcake based rom. Nobody else until recently seemed to share my misfortune (sorry for everyone else). There is an exception though, JF 1.5 (T-Mobile) build the sensors work (although miscalibrated). I haven't found any other solutions or any other way to generate the settings file. My process output appears to be the exact same as you've posted.
I'll keep trying some things but I just want to keep this alive incase anyone else needs a fix and one of us comes up with a solution.

RemoteDJ, I may actually have a solution to your miscalibrated problem. In the process of going through and trying to get my sensors working, I created an empty akmd_set.txt file with chmod 777 permissions (just so it wouldnt ***** about some user permissions when trying to edit it), and then rebooted the phone. Upon looking at the file after reboot, it was filled with calibration data.
Although this has not fixed my problem, it may help you. Also, the application "Bubble" has a calibrate sensors menu option, which may help you too. I may have to try reverting back to JF 1.5 to see if I have better luck there. Was that JF 1.51 you were using? If not, please advise.

Yes it was JF 1.51, it's funny because I just got done making a blank chmod 777 akmd_set.txt lol. Of course i'm on the latest stable cm-3.6.8.1 currently and the sensor doesn't work at all.
Just so we stay up to date with each others findings, I also pulled the akmd app from Haykuro's ION, JF 1.43, JF 1.51, Official RC33 and the current CM build. I have only tried replacing the current one with the ION app (in recovery), so far no change, but I think the only file difference I noticed out of the files I pulled was that the JF 1.43 and Official RC33 seemd to be the same (haven't hashed them to verify) and different then the others which seem to be the same but since I haven't hashed any of them to verify yet I can't say for sure.
I am calling it a night right now but will let you know if I find anything tomorrow.
Just a followup with the hash info.. looks like I was correct:
517a87a4e6caa5e66f0520c68dcb7c0e - Official RC33
517a87a4e6caa5e66f0520c68dcb7c0e - JF 1.43
c3a80124ab072f9bf12cc9e4155a0f9f - JF 1.51
c3a80124ab072f9bf12cc9e4155a0f9f - ION
c3a80124ab072f9bf12cc9e4155a0f9f - CM 3.6.8.1
Another Edit: I am apparently not good at going to sleep.. Good News! (semi).
I left the akmd_set.txt in place (it got filled with data just like yours)
replaced ion/cm akmd with jf 1.43 and now the sensor appears to be working (although still miscalibrated for me, this might be a hardware issue.. lots of drops).
step by step.
copy jf1.43 akmd to sdcard (i just pulled it out of the update zip).
reboot recovery
mount /system
rm /system/bin/akmd
mount /sdcard
cp /sdcard/jf1.43_akmd /system/bin/akmd
then umount (probably not required)
reboot.
It still didn't work, but I deleted the akmd_set.txt via adb and then rebooted and voila!
Let me know if any of that works for you
Edit: I was able to fix my calibration by reverting to the rc30 akmd, I have absolutely no idea why my device is so picky, but it only has ever had correct calibration with accelerometer on the rc29/30 builds (this includes all official t-mobile ota releases). Anyway just wanted to through that out there incase anyone else uses the jf1.43 and is miscalibrated.

RemoteDJ,
Thank you so much for that suggestion! That totally worked for me! I flashed cm-3.6.8.1 tonight and replaced its /system/bin/akmd binary with the one from jf 1.42. After messing around with some permissions (I failed to realize the akmd binary was not set as executable when i moved it over ) i was able to create a blank akmd_set.txt, chmod 777 it, and then restart the akmd service (stop akmd, followed by start akmd).
I can not thank you enough for this suggestion! Guess it just took another similar brain bashing away at the same problem to solve it. I just wonder why this is not a problem for a lot of other people out there. It's the same hardware, so you would think everyone upgrading to cupcake would have an incompatible akmd binary. Strange...

Sweet, I'm really glad it worked for you!
I can't imagine for the life of me why other people don't have similar problems, but then again I've had a few problems that seemed indivualized to my device lol. Its only recently that I discovered a few other people were having problems with the accelerometor.
Oh well, all in all just glad this worked.
Maybe we can get Cyanogen to include this in his FAQ as an alternative for sensor repair for the people like us. I might reply on the latest build thread later, anyway have a good one.

So far I haven't had any luck with this. I'm actually tempted to send my unit back to HTC with the ota 1.5 cupcake on it and see what happens. I'm not sure if it was working when they sent it to me or not. It doeswork with the rc29 though.

I'm having the same issue, no compass apps seem to work, google sky maps doesn't work , the level app doesn't work, all are froze.
How can I fix it?

Software update once every 24 hours???

This is bullcrap. And it's an AT&T thing, no matter how much they want to deny it. And it's all handled locally. I performed this test to verify.
1) Perform a full Nandroid backup.
2) Check for software update via Settings > Software Update (of course I got a communication error)
3) Try again.. Oops, you have to wait 24 hours.
4) Restore Nandroid backup I created step 1.
5) Check for software update via Settings > Software Update (wow it works again)
This is how I think it works:
1) You tap "Check for update" in Settings > Software Update
2) The Settings application (or perhaps Software Update is a separate app) looks at some cache file SOMEWHERE on the filesystem (probably an XML) for a Unix timestamp
3) The Unix timestamp in this cache is compared to the current Unix timestamp. 4) If (current - previous) < 86400 then you are not allowed to check for updates.
5) If (current - previous) > 86400 then you are allowed to check for updates and the timestamp in the cache file is then updated to the current Unix timestamp.
My friend Don has a Vibrant and this restriction does not exist so I know it's an AT&T thing. I need to track this cache file down.

Old news, work around is to set system date ahead one day and check again. AT&T did this to limit the hit on their update servers.
Sent from my SAMSUNG-SGH-I897 using XDA App

. .

Just curious how often you really need to check for an update considering there has only been one.
Sent telepathically using vulcan mind meld app.

This method is better than setting the clock ahead - http://forum.xda-developers.com/showpost.php?p=8394579&postcount=4

cappysw10 said:
Old news, work around is to set system date ahead one day and check again. AT&T did this to limit the hit on their update servers.
Sent from my SAMSUNG-SGH-I897 using XDA App
Click to expand...
Click to collapse
Not a solution. Next time I try to udpate it's 48, etc. Regardless of how often I should update, this is unacceptable. AT&T needs to get on the ball so the servers don't fail.
Furthermore, from a coding standpoint the update process should work like this if they MUST limit you to once every 24 (pseudo-code)...
upon tapping "Check for updates...."
$timestamp = check_timestamp()
if ((now() - $timestamp) > 86400) {
echo "Sorry, blah blah blah 24 hours"
} else {
check_update_sever()
]if (exit_code == 0) {
update_timetamp()[
}
If the exit code from the attempt to hit the servers is non-zero (failed to connect or what have you) then that timestamp should NOT be updated.

I tried this with little success..
* Full nandroid backup so I could restore it later
* Set clock ahead 20 years (so unix timestamp would be very different 19xxxx versus 12xxxx)
* Check update, of course it failed
* busybox find / -type f -exec busybox grep 1918 {} \;
I found a few files with the 2030 timestamp but none pertaining to the software update. They have it hidden well. I could also dissect their app and patch it.

i am wondering why you think its needed to be able to check for updates all the time from ATT?
As already stated, there has been exactly ONE update from ATT. what do you think you are going to get to update if you do connect?

[DEV][RC5] Team Nyx [Android 4.0.4] (ICSSGS)

Team Nyx
-- DemonWav -- Nushor -- Onecosmic -- PaulForde -- RaymondDull -- Trae32566 --​
This ROM is based directly on the source code straight from Google themselves, the way Android is meant to be. This ROM is produced by a collaboration of many different people, with the main development being done by Onecosmic himself. We strive to bring releases as often as possible and provide a stable AOSP ROM.
We are in the midst of many changes and will try to have our source code back up as soon as possible.
We do not support anything that is not produced by us!
Team ICSSGS is now Team Nyx!
We have made lots of changes, and part of that includes moving our base from XDA to our own website. All information, as well as downloads will be on our website, but please continue your discussions and we will continue our support here on the forums. Please check out our site at http://team-nyx.com/

I was wondering where this rom went

Glad to see the team back!

Great
Keeping simple and clean. My style. Thanks.

Downloading now thanks guys
Sent from my SGH-T959 using xda premium

So far I'm liking this one. Very fast and minimal. My GPS didn't lock but it hasn't for any ICS rom yet. Still gauging battery life but seems to be on par with Slim. Great job!

Was this something that is released on Rom Manager?
soulsurviva87 said:
Downloading now thanks guys
Sent from my SGH-T959 using xda premium
Click to expand...
Click to collapse
Where is the download link?
There is no link in the OP, no links on the external web site.
Neither here nor there has any screen shots either, by the way.
There is a listed release tab on the navigation on that site, some features on R5 but no links there or here.
Was this something that is released on Rom Manager?

serendipityguy said:
Where is the download link?
There is no link in the OP, no links on the external web site.
Neither here nor there has any screen shots either, by the way.
There is a listed release tab on the navigation on that site, some features on R5 but no links there or here.
Was this something that is released on Rom Manager?
Click to expand...
Click to collapse
It's on the Releases tab. Click on the word "Dowload" and it will automatically start the download.

I made a nandroid backup of my slim install and tried installing this. I selected custom, and choose vibrant. It went through the install process, but would only boot back to recovery. Once there, the volume buttons were reversed. I tried re installing the ROM, but it would not being up the initial boot sequence, just went to the screen where it said installing vibrant. Once again, it only booted back to recovery. I flashed slim fresh and retried the install on this, but it still would not start at the beginning of the install process and it went directly to the installing vibrant part and would only reboot to recovery. I restored my nandroid backup of slim and I'm back on it, but would like to try this one out.
I'm guessing that I shouldn't have selected custom the first time, but I can't get it to start at the beginning to just install the base.
Any ideas?
Sent from my SGH-T959 using xda premium

HeavyA said:
I made a nandroid backup of my slim install and tried installing this. I selected custom, and choose vibrant. It went through the install process, but would only boot back to recovery. Once there, the volume buttons were reversed. I tried re installing the ROM, but it would not being up the initial boot sequence, just went to the screen where it said installing vibrant. Once again, it only booted back to recovery. I flashed slim fresh and retried the install on this, but it still would not start at the beginning of the install process and it went directly to the installing vibrant part and would only reboot to recovery. I restored my nandroid backup of slim and I'm back on it, but would like to try this one out.
I'm guessing that I shouldn't have selected custom the first time, but I can't get it to start at the beginning to just install the base.
Any ideas?
Sent from my SGH-T959 using xda premium
Click to expand...
Click to collapse
It is supposed to start the install, flash a kernel, then reboot, and continue the install. Not sure why you had problems, but unfortunately I can't make a fool-proof install process, I can only get close. Flash back to a working rom and retry, you can choose custom if you want, that isn't a problem. I don't know why it didn't start the install automatically, but again, I can't make a fool-proof install process, sorry.

DemonWav said:
It is supposed to start the install, flash a kernel, then reboot, and continue the install. Not sure why you had problems, but unfortunately I can't make a fool-proof install process, I can only get close. Flash back to a working rom and retry, you can choose custom if you want, that isn't a problem. I don't know why it didn't start the install automatically, but again, I can't make a fool-proof install process, sorry.
Click to expand...
Click to collapse
No problem, I'll give it a go again later. just thought I'd let you know just in case. Not complaining, just informing. The installer is pretty slick though.
/edit
Gave it another shot, but it still jumps to the installing vibrant stage. Still can't get the installer to start over again at the beginning.
Starting to think I may need to Odin back to stock and try again in the future. For now, I'm staying on slim. I'll try again when I have more time to mess with it.
Sent from my SGH-T959 using xda premium

Onecosmic is in on this I'm in
Sent from my SGH-T959 using XDA

HeavyA said:
No problem, I'll give it a go again later. just thought I'd let you know just in case. Not complaining, just informing. The installer is pretty slick though.
/edit
Gave it another shot, but it still jumps to the installing vibrant stage. Still can't get the installer to start over again at the beginning.
Starting to think I may need to Odin back to stock and try again in the future. For now, I'm staying on slim. I'll try again when I have more time to mess with it.
Sent from my SGH-T959 using xda premium
Click to expand...
Click to collapse
Yeah, that's probably your best bet. But you will have to delete the stuff on your sdcard for it to start over from the beginning, it'll be clearly labeled with "teamnyx" (if everything goes correctly those things are deleted afterwards). But sorry for the inconvenience, man.

Great rom, runs very smooth only issue was aroma installer didn't install on its own it boot looped. Just pulled battery and booted into recovery and then it finished its install. GPS locks very fast only wish there was more customizations. Thanks for your hard work.
Sent from my SGH-T959 using XDA

Had the boot loop issue as well. Pull battery and forge on. My only issues so far are sound dropping in and out in mx video player because stock wouldn't play videos I recorded under slim and the gallery not syncing with picasa. Other than that it's a damn fine rom!
Sent from my SGH-T959 using XDA

dntipwn said:
Great rom, runs very smooth only issue was aroma installer didn't install on its own it boot looped. Just pulled battery and booted into recovery and then it finished its install. GPS locks very fast only wish there was more customizations. Thanks for your hard work.
Sent from my SGH-T959 using XDA
Click to expand...
Click to collapse
Yup, that's exactly the same thing that happened to me, it did a boot loop, pulled battery went back into CWM and re selected the zip file and finished installing. Have to say the slckest install ever!

For me install went without a hitch, great interface by the way. I didn't wipe data, and all apps worked great, but GPS did not work at all. I got this:
Code:
05-29 17:29:34.558 D/GpsLocationProvider( 344): set_capabilities_callback: 7
05-29 17:29:34.960 I/GPS ( 344): gps_init() : SAMSUNG GPS DRIVER INITIALIZED [Ver 2.0] 2011.03.31)
05-29 17:29:44.652 D/GpsLocationProvider( 344): NTP server returned: 1338323385557 (Tue May 29 17:29:45 BRT 2012) reference: 103594 certainty: 62 system time offset: 950
05-29 17:30:06.562 I/ActivityManager( 344): START {act=android.intent.action.MAIN cat=[android.intent.category.LAUNCHER] flg=0x10200000 cmp=com.chartcross.gpstestplus/.GPSTestPlus bnds=[240,402][360,552]} from pid 590
05-29 17:30:06.992 I/ActivityManager( 344): Start proc com.chartcross.gpstestplus for activity com.chartcross.gpstestplus/.GPSTestPlus: pid=1484 uid=10048 gids={}
05-29 17:30:08.644 I/GPS ( 344): gps_engine_status_update: called acquire_wakelock_cb
05-29 17:30:10.769 D/BootCompletedIntentReceiver( 1596): GPSMonitorService started
05-29 17:30:10.949 I/GPS ( 344): gps_engine_status_update: called release_wakelock_cb
Any ideas?
Tried again wiping data, same result.

looking forward to trying this out! thanks for supporting our device!

OK, I deleted the referenced teamnyx files from my internal memory and tried it again. The installer DID start over at the beginning!
However, no matter what option I choose (AOSP, Recommended, Customize) the screens that present the options for kernel and other applications are blank.
I re-downloaded the ROM and tried again, but all of selection screens were still blank as before.
------------------------------------------------
AHA! As a last ditch effort, I moved the zip file to the root of the internal SD card. (I normally copy them to a folder to make them easier to browse to in CWM)
Rerunning the installer now and all the options are showing up. I guess I inadvertently messed up the virtual folder mappings for the installer by not installing from the root of the internal memory.
You may want to add that to the notes about the installer.

Moronig said:
For me install went without a hitch, great interface by the way. I didn't wipe data, and all apps worked great, but GPS did not work at all. I got this:
Code:
05-29 17:29:34.558 D/GpsLocationProvider( 344): set_capabilities_callback: 7
05-29 17:29:34.960 I/GPS ( 344): gps_init() : SAMSUNG GPS DRIVER INITIALIZED [Ver 2.0] 2011.03.31)
05-29 17:29:44.652 D/GpsLocationProvider( 344): NTP server returned: 1338323385557 (Tue May 29 17:29:45 BRT 2012) reference: 103594 certainty: 62 system time offset: 950
05-29 17:30:06.562 I/ActivityManager( 344): START {act=android.intent.action.MAIN cat=[android.intent.category.LAUNCHER] flg=0x10200000 cmp=com.chartcross.gpstestplus/.GPSTestPlus bnds=[240,402][360,552]} from pid 590
05-29 17:30:06.992 I/ActivityManager( 344): Start proc com.chartcross.gpstestplus for activity com.chartcross.gpstestplus/.GPSTestPlus: pid=1484 uid=10048 gids={}
05-29 17:30:08.644 I/GPS ( 344): gps_engine_status_update: called acquire_wakelock_cb
05-29 17:30:10.769 D/BootCompletedIntentReceiver( 1596): GPSMonitorService started
05-29 17:30:10.949 I/GPS ( 344): gps_engine_status_update: called release_wakelock_cb
Any ideas?
Tried again wiping data, same result.
Click to expand...
Click to collapse
No GPS on this kernel. Tried the boot.img pulled from the ROM on other ROMs - and still nothing.

[ROM]Cyanogen Mod 9.0 Mod (For stock rom users)

Introduction
This is a port of Android 4.0.X for Streak7 [WIP]
The "official" CM9 port requires you to repartition your device to use it. This is a mod that allows it to run on standard partition devices.
Any bugs in the "official" CM9 port will likely carry over, and there is a possibility of the mod creating more bugs.
Do not bug DJ_Steve or Giveen about any bugs in this version.
I will be providing support/changes for using CM9 on devices with stock partition layouts (ie Stock 5xx, HS r7 and newer users)
This version is a mod of the CM9 build by DJ_Steve/Giveen for stock rom partition users.
IF YOU ARE CURRENTLY USING HONEYSTREAK R2-R6, "NORMAL" CM9 (ICS-B4), OR HAVE USED NVFLASH: THIS BUILD IS NOT FOR YOU
If you have installed the prerelease, you are STRONGLY RECOMMENDED to do a fac reset after installing.
Click to expand...
Click to collapse
How to flash a zip:
Entering recovery mode on device -XDA wiki
Flashing an update.zip with custom recovery - XDA wiki
Click to expand...
Click to collapse
Requirements:
A custom recovery:
Clockwork Mod 6.0.1.1 (or newer)
Team Win Recovery Project 2.2.1.4 (or newer)
A compatable device (see above)
Click to expand...
Click to collapse
Install instructions:
Download: update-S7-CM9-b4-mod-2.zip
Copy update zip to SD card
Reboot into custom recovery mod
Select "update from update.zip"
Navigate to update zip
Flash
REQUIRED: Flash Gapps
Finished
Click to expand...
Click to collapse
Changes:
Removal of gapps (which will be provided in a seperate zip)
(Semi) Working /sdcard and /sdcard2 (has some bugs still, but mostly works)
Minor polish
Credits:
DJ_Steve for porting CM9
Giveen for every random thing
Click to expand...
Click to collapse
Changelog:
Aug 23 2012: Updated install script
Jul 22 2012: Released full version
Jul 21 2012: Updated instructions
Jul 20 2012: Released prerelease 1
Click to expand...
Click to collapse
2char

Introduction
I'm gonna organize everything I'm currently aware of, I'm gonna update this post as people comment.
Wifi
[-] Overall: Works fine
[-] Speed: Same as on previous roms
[?] Untested: Wi-fi tethering
Bluetooth
[?] Overall: Untested
[?] Untested: File transfer
[?] Untested: BT Headset
[?] Untested: Keyboard
[?] Untested: Mouse
[?] Untested: BT Tethering
[?] Untested: Handsfree device
Cellular data
[!] Overall: Nothing works
GPS/Sensors
[-] Overall: Works fine
[?] Unknown: AGPS support (Needs confirmation)
[-] Works: Accelerometer
[-] Works: E-Compass
[!] Bug: FM Radio does not work
[!] Bug: GPS very inaccurate/does not work
[-] Works: Light Sensor
[?] Untested: Microphone
[-] Works: Vibrator
Dock and outputs
[-] Works: Mostly works
[-] Works: Dock detection
[-] Works: HDMI out
[?] Untested: USB tethering
Mass Storage and Inputs
[-] Overall: Mostly works
[?] BUG: Internal sdcard not presented does not work
[!] Bug: USB-Hosting does not work
Battery
[-] Overall: Airplane mode: Seems Fine
[?] Overall: Cellular on: Unknown
[-] Works: Charging
[!] Bug: Usage info not available
Graphics
[-] Overall: Works fine
Sound
[-] Overall: Works fine
[?] Untested: Headphone out
Buttons and leds
[-] Overall: Works fine
Camera
[!] Overall: Nothing works
[!] BUG: Rear cam doesnt work
[!] BUG: Front cam doesnt work
[!] BUG: Rear Flash doesnt work
Screen and UI
[-] Overall: Works fine
Apps
[-] Overall: Mostly works
[-] Not a bug: GAPPs isnt included (General consensus: stock based roms: including gapps is ok, source based roms: do not include gapps)
[!] BUG: Chrome settings button causes app to crash
Feature requests:
Phone mode UI
Virtual keys (in phone mode)
Overall summary:
[-] Wifi: Works fine
[?] Bluetooth: Untested
[?] Tethering: Untested
[!] Cellular: Nothing works
[?] GPS/Sensors: Untested
[-] Dock/Output: Mostly works
[-] Mass Storage/Input: Mostly works
[-] Battery: Good
[-] Graphics: Mostly works
[-] Sound: Works fine
[-] Buttons: Works fine
[!] Cameras: Nothing works
[-] Screen and UI: Mostly works
[-] Apps: Mostly works
Conclusion:
(placeholder)

Released a prerelease version, this version uses the B3 apks as I cannot extract the B4 ones.
Once someone is able to give me a copy of the B4 apks I'll release the full version of B4-mod

I installed but it's a bit empty.(I know there is no gApps in it) Also I noticed rotation not works by default until you switch it on from Trebuchet - General Settings. While I am switching between menus on wireless part, device becomes unresponsive sometimes.(T-Mobile 4G Model)
Camera and 3G/4G Radios are still same, not working.
On the other hand it boots up and works very fast. Good Job!
I'm new to Android system but I know a bit about Linux. Also I'm a software developer. Are there any guideline for development/environment? (I'm using Windows 7 and Ubuntu both) I wanna try to contribute.
Edit: This works for me very well: http://wiki.cyanogenmod.com/index.php?title=Latest_Version/Google_Apps
Second Edit: Switching airplane mode on, locks device completely.

Cell data was never supported, and airplane mode freezing is likely related to that.
Treb locking to landscape by default isnt really a bug, as there is a checkbox and it does remember which one you pick.
Camera is known to not work. All these are carry over from ICS B4
Gapps isnt ready yet, it will be soon.

Snarg is trying to get camera to work
Sent from my Nexus 7 using Tapatalk 2

VPN is not working too. It says "Unsucessful". I'm sure credentials and other setting are correct.
Bluetooth:
General Functions and pairing with bluetooth mouse is working.

I havnt changed anything at all regarding VPN,
it's either due to:
Using the B3 apks (since I dont have the B4 apks)
A bug in B4 (which I cant fix)
Keep in mind I'm only making minimal changes, I'm not actively updating the code.
But I'll take a look tomorrow and see if it's anything fixable

iboware said:
VPN is not working too. It says "Unsucessful". I'm sure credentials and other setting are correct.
Bluetooth:
General Functions and pairing with bluetooth mouse is working.
Click to expand...
Click to collapse
I cant reproduce as I dont have a vpn I can attempt to connect to.
I'm going to need a logcat of you attempting to connect to it to look at, otherwise I cant even look at what could be the issue.

It turns out that the B4's system.img is in fact 'damaged'
It's damaged to the point that ext2explore cant read it, but linux kernals can mount it.
I will be releasing the full version of CM9-mod soon
Edit: There's a really strange bug with the B4 apks.
You must also install Gapps or the home button will not respond.
It actually responds when you're in the clock app, but not anywhere else.
I dont know why installing gapps fixes it, as I've cleanly pulled gapps out of the rom.

TheManii said:
I cant reproduce as I dont have a vpn I can attempt to connect to.
I'm going to need a logcat of you attempting to connect to it to look at, otherwise I cant even look at what could be the issue.
Click to expand...
Click to collapse
Code:
W/PrintK ( 68): dog, +tegra_wdt_set_timeout, wdt->timeout=130
V/PrintK ( 1323): <7>[ASOC-DBG]tegra_dapm_event_int_spk SND_SOC_DAPM_EVENT_ON ?
2
V/PrintK ( 1323):
I/keystore( 103): uid: 1000 action: s -> 1 state: 1 -> 1 retry: 4
I/keystore( 103): uid: 1000 action: s -> 1 state: 1 -> 1 retry: 4
I/keystore( 103): uid: 1000 action: s -> 1 state: 1 -> 1 retry: 4
D/dalvikvm( 2967): GC_CONCURRENT freed 318K, 23% free 6465K/8391K, paused 2ms+4m
s
E/BinaryDictionaryGetter( 1527): Could not find a dictionary pack
E/ActivityThread( 1527): Failed to find provider info for com.android.inputmetho
d.latin.dictionarypack
D/PrintK ( 1323): <6>wm8903_set_dai_trigger 1
D/dalvikvm( 1359): GC_EXPLICIT freed 745K, 22% free 9279K/11783K, paused 4ms+7ms
E/ActivityThread( 1527): Failed to find provider info for com.android.inputmetho
d.latin.dictionarypack
E/BinaryDictionaryGetter( 1527): Could not find a dictionary pack
D/dalvikvm( 1527): GC_CONCURRENT freed 474K, 21% free 7396K/9287K, paused 3ms+4m
s
I/keystore( 103): uid: 1000 action: i -> 1 state: 1 -> 1 retry: 4
I/Vpn ( 1359): Switched from [Legacy VPN] to [Legacy VPN]
I/Vpn ( 1359): Switched from [Legacy VPN] to [Legacy VPN]
V/LegacyVpnRunner( 1359): Waiting
V/LegacyVpnRunner( 1359): Executing
D/mtpd ( 3001): Waiting for control socket
I/LatinIME( 1527): InputType.TYPE_NULL is specified
W/LatinIME( 1527): Unexpected input class: inputType=0x00000000 imeOptions=0x000
00000
D/PrintK ( 1359): <6>wm8903_set_dai_trigger 0
D/PrintK ( 1323): <6>wm8903_set_dai_trigger 1
D/mtpd ( 3001): Received 20 arguments
I/mtpd ( 3001): Using protocol pptp
I/mtpd ( 3001): Connecting to ribbsede.getfoxyproxy.org port 1723 via wlan0
I/mtpd ( 3001): Connection established (socket = 11)
D/mtpd ( 3001): Sending SCCRQ
D/mtpd ( 3001): Received SCCRP -> Sending OCRQ (local = 39064)
I/mtpd ( 3001): Tunnel established
D/mtpd ( 3001): Received OCRQ (remote = 32984)
I/mtpd ( 3001): Session established
I/mtpd ( 3001): Creating PPPoX socket
F/mtpd ( 3001): Socket() Protocol not supported
D/dalvikvm( 2967): GC_CONCURRENT freed 336K, 23% free 6527K/8391K, paused 2ms+6m
s
I/LegacyVpnRunner( 1359): Aborting
I/LegacyVpnRunner( 1359): java.lang.IllegalStateException: mtpd is dead
I/LegacyVpnRunner( 1359): at com.android.server.connectivity.Vpn$LegacyVpn
Runner.execute(Vpn.java:569)
I/LegacyVpnRunner( 1359): at com.android.server.connectivity.Vpn$LegacyVpn
Runner.run(Vpn.java:447)
D/PrintK ( 0): <6>wm8903_set_dai_trigger 0
well, it looks like a common cm9 bug but I'm not sure : http://code.google.com/p/minicm/issues/detail?id=495
Another thread about this issue: http://code.google.com/p/android/issues/detail?id=23124
According to the posts I read, adding a kernel compatible tun.ko file will solve this problem.

It does seem to be the case, what rom were you on previously and did VPN work then?
None of our stock roms (by stock roms I mean for every dell device) have tun.ko (by default) and I havnt changed the kernel at all from ICS-B4.
If it worked in B4 then it might have been something from my end, if it didnt then there's nothing that can currently done.
Edit: I will be uploading the full version of CM9-mod soon, flash and fac reset and test against that.
As the prerelease is just that, it doesnt sound like something that would be fixed in the full version, but lets just check and confirm.

TheManii said:
It does seem to be the case, what rom were you on previously and did VPN work then?
None of our stock roms (by stock roms I mean for every dell device) have tun.ko (by default) and I havnt changed the kernel at all from ICS-B4.
If it worked in B4 then it might have been something from my end, if it didnt then there's nothing that can currently done.
Click to expand...
Click to collapse
I was on the latest stock HC and I have never installed any other rom.

If you need VPN support for anything critical (or do anything mission critical period), I would recommend staying on a stock rom.
Regardless of which version of CM9 you choose to install, they're all closer to beta then they are to release.
That said, the full version of CM9-mod is now live.
You are STRONGLY recommended to do a fac reset after installing it if you installed the prerelease.
You also MUST install Gapps as there's a strange bug that causes the home button to not work without it.

TheManii said:
If you need VPN support for anything critical (or do anything mission critical period), I would recommend staying on a stock rom.
Regardless of which version of CM9 you choose to install, they're all closer to beta then they are to release.
That said, the full version of CM9-mod is now live.
You are STRONGLY recommended to do a fac reset after installing it if you installed the prerelease.
You also MUST install Gapps as there's a strange bug that causes the home button to not work without it.
Click to expand...
Click to collapse
I have a timeout problem with marketplace on the full version. It says "loading" for ages and then "Connection timed out. [Retry]" (I did factory reset)

Are you able to connect to websites with the browser?

TheManii said:
Are you able to connect to websites with the browser?
Click to expand...
Click to collapse
Yes. I was able to connect. Today, I tried again and it's successful. I don't know why it could not connected yesterday maybe a DNS issue.
Edit: Shortcuts are displaying in wrong position(under the search icon) it needs alignment. Kindle is not working. I think pre-release version was more stable. Now, I'm back to the stock hc. I would like to try it if there is a new version.

Ive never seen a rom that you shouldnt use if you have Nvflashed before.. whats the cause, or what would happen if one did flash?

ssmr2t said:
Ive never seen a rom that you shouldnt use if you have Nvflashed before.. whats the cause, or what would happen if one did flash?
Click to expand...
Click to collapse
I was wondering the same thing.
sent from device using an app

joeclark said:
I was wondering the same thing.
sent from device using an app
Click to expand...
Click to collapse
Because nvflashed ones are changing the original partition table. if you can restore the partition table, than you can use this one as well, i think.

S10 Random Reboot // Kernel Panic

Hello,
It's been 2 weeks I have my S10.
It's a G9730, a Snapdragon but with magisk installed as it is one of the only snapdragon with bootloader unlock.
I rooted a first time and I experimented random reboot. I though I did something bad with system app so I Factory reseted and now I'm only using root for normal apps (Tasker, Adblock, Titanium etc...)
I don't have any mod or UI tweak.
I still have random reboot. Something 3 times an day, sometimes, it's OK for 2 days and then....
Sometime it's while I'm using the phone but most of the time, I find my phone waiting for startup.
I uploaded 2 pictures :
- Autolog after random reboot
- What I mean by waiting for startup
I find on Google I'm not the only one. People with any root have the same issue. But there are no fix!
Can you help please?

I don't understand...
No random reboots for 3 days. I did not change anything and 4 random reboot since this morning!!!!

Hi,
As discovered by tulth, it's this bug that occurs. Why only so after rooting, I don't know. I think it's because of MagiskHide (which is doing nothing wrong, BTW) reading /proc/<PID>/status which triggers this kernel bug that Samsung didn't apply that patch to fix. As far as I know, Exynos models aren't affected because they have a slightly newer kernel; our .79 vs. their .85. As mentioned here, it occurs under "memory pressure", which is why it doesn't happen all the time. You can trigger this bug on demand and make your phone restart in the same way by doing this:
Code:
su
dd if=/dev/random of=/dev/shm # or any very large file (3-4 GB). This fills up the allocated space for shared memory
cat /proc/*/smaps_rollup
You're getting the exact same kernel panic as I do on my Snapdragon S10+/SM-G9750. My last_kmsg + disassembler output confirming that it's definitely the bug I linked to is here. I have a fix for the SM-G9750 but I'm waiting for a couple of days' uptime to ensure it doesn't mess anything else up on the phone (usually I get 16 hours idle or not - usually less - before my phone resets) before I open a thread with it. My phone currently does not reboot if I run the commands above, which wasn't the case before my fix.
What I could do with, before opening that thread, is the raw kernel image from a Snapdragon S10 to see if my patch needs adjusting for that model and, indeed, if it's even safe to apply on it. I don't want to download a 5GB firmware for a phone I don't have just to get a 35 MB file compressed.

Hi Qwerty!
Thank you for those really helpful information!
Do you want me to upload the 35Mb files for you?
I have the full ASF1 (last version on G9730) on my computer.
Note : I tried this command lines on terminal, nothing happend. See attached. Maybe I wrote it bad and should point to a large file as written?
(attachement)

Hi G-ThGraf,
Yes please, I only need the recovery.img.lz4 file, which should be inside the AP tarball
Note : I tried this command lines on terminal, nothing happend. See attached. Maybe I wrote it bad and should point to a large file as written?
Click to expand...
Click to collapse
Maybe /dev/random isn't enough, I must admit that I was dd'ing AP_G9750ZHU1ASF1_CL16082828_QB24224470_REV00_user_low_ship_MULTI_CERT_meta_OS9.tar.md5 from ADB because it happened to be on my phone because of the process to install Magisk...

qwerty12 said:
Hi,
As discovered by tulth, it's this bug that occurs. Why only so after rooting, I don't know. I think it's because of MagiskHide (which is doing nothing wrong, BTW) reading /proc/<PID>/status which triggers this kernel bug that Samsung didn't apply that patch to fix. As far as I know, Exynos models aren't affected because they have a slightly newer kernel; our .79 vs. their .85. As mentioned here, it occurs under "memory pressure", which is why it doesn't happen all the time. You can trigger this bug on demand and make your phone restart in the same way by doing this:
Code:
su
dd if=/dev/random of=/dev/shm # or any very large file (3-4 GB). This fills up the allocated space for shared memory
cat /proc/*/smaps_rollup
You're getting the exact same kernel panic as I do on my Snapdragon S10+/SM-G9750. My last_kmsg + disassembler output confirming that it's definitely the bug I linked to is here. I have a fix for the SM-G9750 but I'm waiting for a couple of days' uptime to ensure it doesn't mess anything else up on the phone (usually I get 16 hours idle or not - usually less - before my phone resets) before I open a thread with it. My phone currently does not reboot if I run the commands above, which wasn't the case before my fix.
What I could do with, before opening that thread, is the raw kernel image from a Snapdragon S10 to see if my patch needs adjusting for that model and, indeed, if it's even safe to apply on it. I don't want to download a 5GB firmware for a phone I don't have just to get a 35 MB file compressed.
Click to expand...
Click to collapse
I`ve got the same model here. Hope you can share your fix for s10+ ! I can be a tester I`ve unrooted because of these reboot problems.

Hi there,
@qwerty12, please find on this wetransfer the requested file : https://we.tl/t-HPxzXVTNwT
Is it OK ?
Also, I don't have any more random reboot since Saturday (27th) 3PM.
The only one thing I modify since saturday morning, I fully debloated the rom : https://forum.xda-developers.com/galaxy-s10/how-to/galaxy-s10-s10-debloat-bloatware-t3912073
I removed each of those app. (expect car mode)

Hi,
FlatOutRU said:
I`ve got the same model here. Hope you can share your fix for s10+ ! I can be a tester I`ve unrooted because of these reboot problems.
Click to expand...
Click to collapse
Will do, I'm just waiting until tomorrow for the two-day mark to pass... I'll be sure to send you something to test
My fix is simply this patch applied to the existing Samsung kernel on the device (the kernels I did manage to build from source would not boot on my phone) through, basically, a hex editor.
G-ThGraf said:
Hi there,
@qwerty12, please find on this wetransfer the requested file : https://we.tl/t-HPxzXVTNwT
Is it OK ?
Click to expand...
Click to collapse
Great, thank you! I'll use this to see if the kernel binary patch I have for my phone is safe to apply outright for your model, and fix it if not.
Also, I don't have any more random reboot since Saturday (27th) 3PM.
The only one thing I modify since saturday morning, I fully debloated the rom : https://forum.xda-developers.com/galaxy-s10/how-to/galaxy-s10-s10-debloat-bloatware-t3912073
I removed each of those app. (expect car mode)
Click to expand...
Click to collapse
You may not be seeing it any more because there's now less memory being sucked up by these vampiric programs, or one of these programs was going through /proc/ and reading smaps_rollup and status, which can trigger this kernel bug...

Hope there's some progress... i wish i can also do the fix.
i also using 9750 and magisk rooted... and yes, random reboot maybe about 2-3 times a day.
Hope a help too..

The same reboot bug on A50 device ?
Hi Qwerty!
You wrote in another discussion on this topic
qwerty12 said:
Yes:
I think you're right and that applying ASB-2019-01-05_4.9 would stop these KPs - just look at where smaps_pte_range+0x29c points to...
Magisk Canary from about four days ago has been causing many, many abrupt restarts in a day on my S10+; I had to downgrade to ianmacd-7.1.2-2019042401 (thanks Fznwolf), which has been stable (I get maybe one reset a day), but using old Magisk versions isn't really a viable long-term solution.
I could be wrong (I'm not a good programmer in the slightest), but from my quick glances over the kernel code, I think ASB-2019-01-05_4.9 could be worked-around via a module. However, the Samsung S10 kernel is built with CONFIG_MODULE_SIG_FORCE enabled, so loading any module not signed with Samsung's private key fails. A cursory look at load_module() seems to indicate patching the enforcement code out on disk (like Magisk does) or in memory would be trivial.
On a tangent, Samsung kernels are built with "clang version 6.0.10 for Android NDK", but the Qualcomm site only offers 6.0.9 to us mere mortals. Clang/LLVM 6.0.10 doesn't appear to be a thing, so it seems like .10 is some internal Qualcomm or Samsung toolchain. Does anybody know more about this? I don't want to replace the Samsung-provided kernel on my phone, which is why I'd like to be as close as I can be to Samsung's kernel-building environment, even if it seems impossible in this case...
Click to expand...
Click to collapse
I have an EXYNOS A50 device
And I have many times restarted with the same message
But for me the location is
Code:
PC is at smaps_pte_range+0x138/0x42c
LR is at smaps_pte_range+0xb0/0x42c
the kernel version is
Linux version 4.14.62-15891145 ([email protected]) (Android (4639204 based on r316199) clang version 6.0.1 (https://android.googlesource.com/toolchain/clang 279c0d3a962121a6d1d535e7b0b5d9d36d3c829d) (https://android.googlesource.com/toolchain/llvm aadd87ffb6a2eafcb577913073d46b20195a9cdc) (based on LLVM 6.0.1svn)) #1 SMP PREEMPT Thu Apr 25 23:38:19 KST 2019
Click to expand...
Click to collapse
And that's part of the log
Code:
<0>[ 28.558408] [0: android.bg: 4477] Unable to handle kernel read from unreadable memory at virtual address 000000f0
<2>[ 28.558434] [0: android.bg: 4477] sec_debug_set_extra_info_fault = KERN / 0xf0
<1>[ 28.558499] [0: android.bg: 4477] Mem abort info:
<1>[ 28.558512] [0: android.bg: 4477] Exception class = DABT (current EL), IL = 32 bits
<1>[ 28.558522] [0: android.bg: 4477] SET = 0, FnV = 0
<1>[ 28.558532] [0: android.bg: 4477] EA = 0, S1PTW = 0
<1>[ 28.558542] [0: android.bg: 4477] Data abort info:
<1>[ 28.558551] [0: android.bg: 4477] ISV = 0, ISS = 0x00000005
<1>[ 28.558560] [0: android.bg: 4477] CM = 0, WnR = 0
<1>[ 28.558571] [0: android.bg: 4477] user pgtable: 4k pages, 39-bit VAs, pgd = ffffffc87f702000
<1>[ 28.558580] [0: android.bg: 4477] [00000000000000f0] *pgd=00000008ff6fe003, *pud=00000008ff6fe003, *pmd=0000000000000000
<0>[ 28.558596] [0: android.bg: 4477] Internal error: Oops: 96000005 [#1] PREEMPT SMP
<4>[ 28.558607] [0: android.bg: 4477] Modules linked in:
<0>[ 28.558622] [0: android.bg: 4477] debug-snapshot: core register saved(CPU:0)
<0>[ 28.558632] [0: android.bg: 4477] CPUMERRSR: 0000000001100301, L2MERRSR: 0000000010045220
<0>[ 28.558641] [0: android.bg: 4477] debug-snapshot: context saved(CPU:0)
<6>[ 28.558681] [0: android.bg: 4477] debug-snapshot: item - log_kevents is disabled
<6>[ 28.558703] [0: android.bg: 4477] TIF_FOREIGN_FPSTATE: 0, FP/SIMD depth 0, cpu: 89
<4>[ 28.558715] [0: android.bg: 4477] CPU: 0 PID: 4477 Comm: android.bg Not tainted 4.14.62-15891145 #1
<4>[ 28.558725] [0: android.bg: 4477] Hardware name: Samsung A50 LTN OPEN rev04 board based on Exynos9610 (DT)
<4>[ 28.558735] [0: android.bg: 4477] task: ffffffc8454cd280 task.stack: ffffff800b2b0000
<4>[ 28.558750] [0: android.bg: 4477] PC is at smaps_pte_range+0x138/0x42c
<4>[ 28.558761] [0: android.bg: 4477] LR is at smaps_pte_range+0xb0/0x42c
<4>[ 28.558770] [0: android.bg: 4477] pc : [<ffffff8008449ad4>] lr : [<ffffff8008449a4c>] pstate: 40400145
Is this also a fault of ASB-2019-01-05_4.9 ?

Can I help try to resolve the problem
Hi @qwerty12
I would like to find a solution to this problem,
Because I have some devices that I rooted for people
And they have the same problem,
I can help check it out on many new Samsung models,
This happens most in my sample devices A50 but also occurs in A30 S10E A70 etc.
Is the process of editing in X-EDITOR simple?
Could I possibly do this myself?
Because I don't know how to find this operator's location inside the kernel file,
I just tried to open it with BLESS-HEX-EDITOR
In the past I fixed a lot of binaries with it, but only things that were defined as strings and not an operator that was compiled with the code.

Hullo,
aharonshor said:
Is this also a fault of ASB-2019-01-05_4.9 ?
Click to expand...
Click to collapse
Given your kernel version, I would surmise so, yes. (I can't actually give you a definitive answer without looking at a disassembly of your kernel, which isn't something I wish to do. If you want to verify it yourself, get this working.)
4.14.81 is the first version of the LInux kernel 4.14 branch with that fix included; quite evidently, 4.14.62 is a long way off from there. I wouldn't know what was the first version of 4.14 to introduce the problematic code.
S10 Exynos versions weren't affected by this because they were shipped with 4.14.85, I think. We Snapdragon users got .78...
aharonshor said:
I would like to find a solution to this problem,
Because I have some devices that I rooted for people
And they have the same problem,
Click to expand...
Click to collapse
I can empathise, because I once did the same, but I only root my own phones and not others' these days. Taking on the burden of technical support and being the person to blame if something unrelated causes a problem isn't worth it, unless you're being recompensed for your time.
Is the process of editing in X-EDITOR simple?
Click to expand...
Click to collapse
I mean, you might use a hex editor to apply a resulting patch, but you'd want to use a disassembler. A hex editor on a 40 MB binary to find something isn't happening, unless you can turn random bytes of hex into ARM instructions in your head.
Could I possibly do this myself?
Click to expand...
Click to collapse
As I told an Internet friend, if you know ARM assembly, you don't need my help at all, and if not, well talk about a case of the blind leading the blind... I'm not a programmer and do not have any business teaching bad habits. In any case, I don't, and will not for the foreseeable future, the time to write a guide.
I'd recommend just building your own kernel with this patch applied. That process is far simpler, and unless you know what instructions you should be looking for inside the ARM Reference Manual and how to calculate the offset of the check_shmem_swap member inside the mss struct, you would need to build a kernel anyway first to have the compiler do all of that work.
I haven't looked at this in a long time and do not plan to revisit the topic; my own kernels are good enough for me, if rather basic.