Database Users

Answer to bad battery life: phone never enters sleep state

Hi,
i noticed months ago the following on my IDEOS X5 (german) concerning bad battery life. It could be the reason why there are so many different reports regarding bettery lifetime of the U8800.
First - why did I start to investigate: because the phone sucked about 5-10% per hour while not beeing used / lying on my desk, so after max 20 hours the battery was empty. But this did/does not always happen....
What I did: it took some time to get used to android (its my first droid) and learn how to troubleshoot. A month after I got this phone I learned that the GSM-code *#*#4636#*#* will show you very useful information concenring battery/phone usage. I started to play around with the phone (factory reset, install/remove apps, enable/disable different options like 2G/3G/WLAN etc...) always keeping an eye on the 4636-infos.
Result: When I use WLAN the phone simetimes enters an "state" when it never again enters "sleep state" and the battery gets sucked as stated above. I switched off WLAN for a week and this did not occur while standby time increased to 48 hours and 4636 stated an "phone awake" percentage of below 20% (normal usage). Once I reenabled WLAN again it took about an day of WLAN usage to reproduce the bug and "phone awake" time increased to 100% while standby time decreased to below 20 hours. ONLY rebooting the device recovers from this "state". I further noticed that this does not happen if just WLAN is available but no 2G/3G (phone had no SIM). My conclusion is, that the
"You May Be Charged"-Popup is the culprit. It does lock the phone to this "state" for some reason, mybe when the WLAN connection is dropped while the phone is not used (traveling). Can someone reproduce this on his phone (maybe not german)?
Resolution: IMHO none. I send the device twice to reseller for service. First time the complete mainboard was changed without success regarding the battery drain. Second time nothing was done - they just returned the device to me.
Thanks.

This is a known problem, that already has a thread...
http://forum.xda-developers.com/showthread.php?t=1043760
This is affecting ALL Froyo Roms, if you want battery life, install Oxygen or CM7...which are based on Gingerbread.
There is no point in discussing this issue anymore, this is a very well known problem and Huawei doesn't care.
Sending the phone back to manufacturer won't help at all.

Thank you for the hint - it realy looks like the same issue. Knowing more about the nature of this issue could give people not willing to install custom roms a possible solution. Can someone move this thread over to that one?

katu2006 said:
This is a known problem, that already has a thread...
http://forum.xda-developers.com/showthread.php?t=1043760
This is affecting ALL Froyo Roms, if you want battery life, install Oxygen or CM7...which are based on Gingerbread.
There is no point in discussing this issue anymore, this is a very well known problem and Huawei doesn't care.
Sending the phone back to manufacturer won't help at all.
Click to expand...
Click to collapse
But Huawei is about to release official 2.3 in Q3 atleast I hope...

andreasha said:
But Huawei is about to release official 2.3 in Q3 atleast I hope...
Click to expand...
Click to collapse
Everyone is looking forward to official 2.3, but we'll have to wait and see if the "sleep" bug will be sorted.

katu2006 said:
Everyone is looking forward to official 2.3, but we'll have to wait and see if the "sleep" bug will be sorted.
Click to expand...
Click to collapse
I hope some more bugs are straightened out like headphones and internal mic at the same time. Headset button to work. Battery time!

none of my android phones could use the phone's mic when headphones were plugged in

I saw a solution for the mic-with-headphones-problem a few days ago, either here or at modaco (do not know if it is confirmed working though).
fajabird, I know that you want this to work with the ROMs provided by Huawei. Huawei obviously does not care too much about providing the community with updates (they do not even have gingerbread-support by now). This is the downside of this low priced phone.
Let me just say this: It took me about one hour to understand the ultimate-guide at modaco. It took me another hour to decide which ROM I want to use. If anything goes wrong at the installation I can easily start over again. Now I have a fully working phone with Android 2.3.5 with a battery that lasts at least 2 days, sometimes 3(if I do not play any games within that time). Of course I understand that you are probably very pissed at Huawei for providing a ROM with a bug that serious.

if you find that fix again, let us know where to look

darkado said:
if you find that fix again, let us know where to look
Click to expand...
Click to collapse
It is to be found at the Android-development-section, right here. Note that you cannot use the mic of your headset after applying this

too bad, it would be awesome to be able to use both

XphX said:
I saw a solution for the mic-with-headphones-problem a few days ago, either here or at modaco (do not know if it is confirmed working though).
fajabird, I know that you want this to work with the ROMs provided by Huawei. Huawei obviously does not care too much about providing the community with updates (they do not even have gingerbread-support by now). This is the downside of this low priced phone.
Let me just say this: It took me about one hour to understand the ultimate-guide at modaco. It took me another hour to decide which ROM I want to use. If anything goes wrong at the installation I can easily start over again. Now I have a fully working phone with Android 2.3.5 with a battery that lasts at least 2 days, sometimes 3(if I do not play any games within that time). Of course I understand that you are probably very pissed at Huawei for providing a ROM with a bug that serious.
Click to expand...
Click to collapse
Hi, thank you for sharing your experience with me. I always wanted to update to some custom roms but the time I started investigating this issue there was no stable one out there. On the other hand I thought there might be a dev interessted in more details regarding this issue helping him to sort out this nasty bug for whatever reason.
Can someone tell me, if his U8800 also has the ambient light sensor but like mine does: it won't reduce the brightness, just increase works. Thanks.

fajabird said:
Hi, thank you for sharing your experience with me. I always wanted to update to some custom roms but the time I started investigating this issue there was no stable one out there. On the other hand I thought there might be a dev interessted in more details regarding this issue helping him to sort out this nasty bug for whatever reason.
Can someone tell me, if his U8800 also has the ambient light sensor but like mine does: it won't reduce the brightness, just increase works. Thanks.
Click to expand...
Click to collapse
Mine reduces brightness but does it quite weird sometimes like adjusting several times without the light around me changing...

Mine always adjusts it way too low. If I am inside the brightness is always on minimum. I use the power-widget on the homescreen. I only change it at night, it is on max. for the whole day (the max. brightness is still kinda low, especially if it's sunny).

fajabird said:
Hi,
i noticed months ago the following on my IDEOS X5 (german) concerning bad battery life. It could be the reason why there are so many different reports regarding bettery lifetime of the U8800.
First - why did I start to investigate: because the phone sucked about 5-10% per hour while not beeing used / lying on my desk, so after max 20 hours the battery was empty. But this did/does not always happen....
What I did: it took some time to get used to android (its my first droid) and learn how to troubleshoot. A month after I got this phone I learned that the GSM-code *#*#4636#*#* will show you very useful information concenring battery/phone usage. I started to play around with the phone (factory reset, install/remove apps, enable/disable different options like 2G/3G/WLAN etc...) always keeping an eye on the 4636-infos.
Result: When I use WLAN the phone simetimes enters an "state" when it never again enters "sleep state" and the battery gets sucked as stated above. I switched off WLAN for a week and this did not occur while standby time increased to 48 hours and 4636 stated an "phone awake" percentage of below 20% (normal usage). Once I reenabled WLAN again it took about an day of WLAN usage to reproduce the bug and "phone awake" time increased to 100% while standby time decreased to below 20 hours. ONLY rebooting the device recovers from this "state". I further noticed that this does not happen if just WLAN is available but no 2G/3G (phone had no SIM). My conclusion is, that the
"You May Be Charged"-Popup is the culprit. It does lock the phone to this "state" for some reason, mybe when the WLAN connection is dropped while the phone is not used (traveling). Can someone reproduce this on his phone (maybe not german)?
Resolution: IMHO none. I send the device twice to reseller for service. First time the complete mainboard was changed without success regarding the battery drain. Second time nothing was done - they just returned the device to me.
Thanks.
Click to expand...
Click to collapse
Hey there friend, I've detailed much of this and been through a lot of troubleshooting. It did not appear to be the "You may be charged" popup as I disabled that whenever I changed ROMs. I never did get a solution for any of the Huawei ROMs, it just seemed like bad code as Dzo pointed out with the WiFi driver.
A very simple solution for me was installing one of the 2.3 ROMs, they are quite stable at the moment, especially Oxygen (although I use MIUI). You will find much better battery life with these.
Best of luck,
Ryan

My stock rom 2.2.1 Shows Android System 97% battery usage and battery drains 50% in 19minutes! With wi-fi, 3g, Bluetooth, everything is off and even i do not use it.

Flash latest official 2.2.2, wait for official 2.3, flash the custom 2.2.2 rom void or flash a custom 2.3 rom. I would recommend flashing the latest 138 or 162 to get the latest modem baseband or manually install the AMSS.bin.
Sent from my Ideos x5

andreasha said:
Flash latest official 2.2.2, wait for official 2.3, flash the custom 2.2.2 rom void or flash a custom 2.3 rom. I would recommend flashing the latest 138 or 162 to get the latest modem baseband or manually install the AMSS.bin.
Sent from my Ideos x5
Click to expand...
Click to collapse
I can't because phone do not use my external SD card somehow. In android i see files but in recovery-it does not use external sd.. I ask in forum why-but nobody answers

Backup ur data on it and format it. If that doesnt work ill send you a guide how i fixed mine, its just a partition error.
Sent from my Ideos x5

andreasha said:
Backup ur data on it and format it. If that doesnt work ill send you a guide how i fixed mine, its just a partition error.
Sent from my Ideos x5
Click to expand...
Click to collapse
I formated it via phone, then via PC - it works in powered on phone in android, but in recovery it doesn't it only sees sdcard as internal memory, if i try mount sd-ext shows error, can't mount.
Also i have problem if i use pc to copy some files to internal memory-phone switches storage mode OFF! and i can't copy files to internal 2GB memory..strange

Phone freezing in deep sleep; should I request an RMA?

I've been having a recurring problem where my phone will lock up in a deep sleep, leaving the phone completely off. In order to restart the phone, I have to hold Power and Vol Down which takes me to a charging logo, and then hit power again. This generally happens overnight when the phone is charging, but I've also had it happen once while making a VoIP call. I thought I resolved the issue by wiping the phone and flashing the stock 4.2 ROM. However, the issue just occurred last night after 4 days of bug-free usage.
My question is whether this is some sort of weird software bug or a hardware issue. I haven't seen others report identical behavior but I've seen lots of complaints about freezing and random reboots that seem to be due to software. Should I request an RMA?

+ 1 . I am having the same issue. I would say hang tight. Call the support team and let them know you are having issues. If atfer a few weeks the problem persists perhaps go through an rma. I suspect it is software related. Lets say you did RMA still had the issue then had a dead pixel or other problem? It would suck and you would have to go through the whole process again. rinse and repeat. Only if you want to return the device for a full refund, then I would say hurry up and make that call. My .02 Good luck.

NexUsandThem said:
+ 1 . I am having the same issue.
Click to expand...
Click to collapse
+1
Having the same issue. Although no clear pattern is visible. Sometimes I find the phone dead-locked after putting it aside for 10 minutes and sometimes it seems to run for a day or even two without issue. Running _motley kernel right now without any change on this front.
I rather think it's a software issue than hardware. So I am not going to RMA it.
Strangely when I power off the phone on such a lock-up and then reboot it I will see a substantial drop in battery after the reboot. So the phone seems to be in some state where it still drawns more battery than in normal operation. E.g. with _motley kernel and limited clocks I typically see 3-5% battery drop over night only. When it locks up during the night and I reset it in the morning then battery usually drops much more (e.g. 30-50% is gone). Maybe it's caused by some hanging/deadlocked driver.

If you full factory reset and flashed then i'd say some app of yours is causing the issue.

shotta35 said:
If you full factory reset and flashed then i'd say some app of yours is causing the issue.
Click to expand...
Click to collapse
In fact I got the device in dec 2012. Started to set it up and then decided to root it (including full factory reset). I only have a very minor amount of system-level applications installed. Actually the main reason to root was to use Voltage Control as I intended to limit clock rates and (if possible) voltages the same way I was used to do it on my Galaxy Tab 7 (P1000) where I was very pleased with 6-days battery lifetime.
Initially I found the minimum clock rate was set to 1GHz and I have reduced it to 304MHz (minimum with stock kernel). When I discovered first freezes/lock-ups then I reverted to default settings and found it still locking up occasionally. Meanwhile I am using _motley kernel which seems to include some voltage mod as well. Although Voltage Control does not show options to modify voltages. Then I found Trickster MOD which shows some voltage control but it seems to have no effect (at least I was used my galaxy tab to crash when undervolting too much and it's pretty "strange" to be able to set all clock voltages to 0.6V while still running stable; so I guess the settings won't have any effect right now).
I have also tried to remove all modifications and applications and/or profiles which could cause the issue without any difference. Even when run stock settings it occasionally completely locks-up the device.
Edit: I was wrong regarding voltages. In fact I found my device happily works at 1134MHz using 600mV. Since I am used to limit max clock rates to 810MHz I could simply put all voltages to 600mV without seeing a crash/reboot.
CAUTION: This does not necessarily have to apply to all devices. My one is "faster" binned Rev. 11 device. Some voltages of different users are collected here: https://docs.google.com/spreadsheet/ccc?key=0AnlqtEwFJwR6dE55WkZYSThEM3U4SGtKMVl2V1VqMmc#gid=0

**Ultimate Galaxy S3 Unusual Freezing Thread**

Latest Thread Update :
16/04/2013 ---- Added Rob's DFG trick Method for new comers to save searching for it .
04/03/2013 ---- Added XXEMB1 as Most reliable freeze free build .
01/03/2013 ---- added advice on kernels with SDS Fix .
26/02/2013 ---- Added links to LagFix (fstrim) premium ,and DFG.
Now that the Galaxy S3 Sudden Death Syndrome has supposedly been fixed by the Update 7 by Samsung patched for Kernel and Recovery ,
There is some poping up posts about the Galaxy S3 freezing with lockups,Lockscreen not responding that requires pulling out battery
and ending up with unsual rebooting and bootlooping which is a very awkward and annoying issue, especially the Q&A section which contains
a lot of angry S3 users reporting this problem ; Ex
Galaxy S 3 keeps freezing every 5 mins (50+ freezes a day)
Galaxy S3 keeps freezing
Well i decided to make this thread in the benefit of spotting the problem and to make sure we have accurate causes and eventually elaborate a solution to this disturbing problem.
Share your experiences down by posting the ROM, the Kernel, the Recovery and the Bootloader that you are on right now, so we can filter the answers and relate them with the freezing problems on The Galaxy S3.
Reported Issue Reasons :
1) - Facebook app for Android.
2) - Google Chrome Browser.
3) - Some Samsung Based Android 4.1.2 builds/Roms causing freezing.
4) - In Some cases, Media Scanner causing lockups.
5) - In Some cases, Swiftkey Flow Bêta causing freezing.
Possible Direct Reason :
Originally Posted by : Rob2222
The freezes are caused by the sudden death fix. On that kind of freezes the phone unfreeze itself after 5-25 minutes.
The phone freezes when writing data to an affected eMMC block.
An eMMC block is affected, when it's internal block pararameters (as f.e. write count for that block) are in such a state, that these parameters trigger a corrupted block without SD-fix (4.1.1) or trigger a freeze with SD-fix (4.1.2).
When a phone is hitting an affected block with a writiing operation is completely unpredictable.
So these freezes can occur on almost each situation on the smartphone when it writes data.
But you have indeed a higher chance to trigger a freeze when writing much data.
Click to expand...
Click to collapse
Possible Current Solution :
I. If the phone freezes, wait until it unfreezes itself. I think this is the important part.
II. The DFG "Dummy File Generator" is only used to trigger the freezes.
III. I think you could also just wait for the freezes to be triggered by all-day-use, but this would be very impractical to wait 20 minutes to
unfreeze when you need the phone.
IV. So we write dummy data with DFG to provoke the freezes.
Highly Recommended Advice :
The write access to an affected block without Sudden Death fix seems to damage the data on this block which _can_ lead to a sudden death.
Because of that I would recommend to stay with KERNELS that have the Sudden death fix included, even if it could cause freezes.
Reported Most Stable " Freeze Free " Build/Rom :
XXEMB1 and all Custom Roms based on this build.
Dummy File Generator :
Alternate Working Solution :
LagFix (fstrim) Premium
Rob's Dummy File Generator Method to solve persistant freezes on the Galaxy S3 :
Rob's DFG Method , Please take a moment to thank him for his contribution to solving this problem.
Let me repeat it again:
Whoever has problems with freezes, please try this at least and give positive or negative feedback here. We have good experience with this method to sightly reduce and/or completely eleminate the freezes.
No, this means that the problems occur when writing the eMMC, not reading.
Try to wait for the phone to unfreeze (5-25 minutes) it seems that it helps and the freezes get less.
If you have that kind of freezes, that are caused by the SD-Fix your phone will unfreeze after that time without reboot.
To provoke the freezes you can also empty your internal storage as much as possible and then use the dummy file generator (generate all) to fill the internal memory 1-2 times.
After each run you can simply delete the dummy files with the delete button.
https://play.google.com/store/apps/d...nomunomu.dummy
Free up 8GB or more to have enough free space on internal memory.
If you have 8 GB free internal space, write 2 times dummy files. 2x8GB are 16GB. The internal wear leveling algorithm spreads these writes over the whole 16GB physical chip area.
There is a good chance that writing EACH sector on the eMMC resets/averages out the internal block data (f.e. write counters) that trigger the Bug and the Fix (Freeze).
If the phone freezes while writing the dummy data (it propably will), just let the phone do what it wants. It will continue to work after 2-25 minutes for each freeze.
This unfreeze after 5-25 minutes seems to have a positive effect.
You can have more than one freeze while doing this. Just let the phone do what it wants. If the dummy file generator crashs cause of a freeze, just restart the App.
When your finally able to write a amount of 16GB data (2x8GB dummy files) in subsequent runs without freezes you have a good chance that your freezes are gone for now or maybe even forever.
You need to write a minimum amount of 16GB data (or more) with this procedure! If you write less data, you didn't even need to thest this DFG method.
The idea behind it:
The SD-Fix causes freezes when writing eMMC blocks in a specific eMMC block parameter state. There parameters are for example the block write counter. Just some few blocks are affected, but when they get written, the SD-Fix freezes the eMMC. Now it seems to have a positive ("healing") effect to wait for the eMMC to unfreeze itself, which happens after 5-25 minutes. For all day use this is not very practicable. So we write just some dummy data to the eMMC to trigger the freezes when we do have the time to wait for the unfreeze. Thats all.
-- From the forums I monitor I get 80-90% positive feedback that the S3 gets again long-time-stable (usable) with this DFG-method.
I got from 10-15 freezes/day to only 1 single freeze in the last 2 weeks without factory reset.
-- Some people needed to write 60-80GB of data until the phone became stable.
In the hardest case (only 1 case) I know, someone wrote about 1TB (1000GB) of data and then the phone became finally stable. He didn't got warranty so DFG was his last resort and after that it finally got stable.
In most cases 20-60GB written data should be enough to get back the phone on a stable state.
If youre familiar with odin, you can also flash XXEMB5 or newer firmware, cause it seems this includes a new, better SD-Fix that doesn't casue freezes anymore. DFG is not needed in this case.
PIT File :

nhariamine said:
Now that the Galaxy S3 Sudden Death Syndrome has supposedly been fixed by the latest update 7 by Samsung patch for Kernel and
Recovery ,There is some poping up posts about the Galaxy S3 freezing and ending up with unsual rebooting and bootlooping which is a
very awkward and weird and annoying issue, especially the Q&A section which contains a lot of angry S3 users reporting this problem.
Click to expand...
Click to collapse
It's a shame that this post has been mostly ignored in favour of the "SDS" one, but I do actually believe both issues are related. My I9300 is running stock - always has been - and suffered with "proper" SDS last November.
The initial symptoms were that my previously perfectly-behaved phone would would randomly FC apps that didn't before, and some of my photographs and video taken on the phone suddenly became corrupted and unusable.
Finally, one Saturday morning, I woke up to find the phone stuck in a boot loop, going from the "Samsung Galaxy S3" black-and-white boot scree to the first blue sweep of the Samsung logo, then reboot. It would power-off and power-on and I could enter ODIN or recovery modes fine, but it wouldn't boot any further.
I managed to get my photos and stuff off the phone through USB in recovery, and decided to wipe and flash the same stock ROM via ODIN.
This process failed - it couldn't partition the device correctly. None of the flash counters were reading anything so I popped it into my local Vodafone shop where they said it just needed the firmware popping back on and they'd have it ready in an hour.
I went back and they said they hadn't been able to load the software and it would have to go off for repair. *sigh*
It came back about 5 days later with a repair report stating that they'd replaced the "main board" - presumably because of the eMMC lockup bug.
It was only after this that the whole "Sudden Death" news came about and I realised what had happened to my phone. The eMMC check app says that my new controller is the same "faulty" version so I've been waiting with baited breath for "safe" official firmware.
The update system threw XXELL5 my way just before Christmas, and I thought that was that, but a coule of weeks ago the phone started hiccuping again. I knew I'd seen this before and my heart sank.
Last week another update was available, this time XXELLA. Since then, all hell has broken loose on my phone. It is locking up about 5 times a day. Initially, I was rebooting it manually (by holding in the power button). Tuesday morning I was woken by my wife saying "shouldn't you be up by now?" - glanced at my phone and it was frozen on the black-and-white boot logo - no alarm! Arrggh! It's dead!
Powered off, back on again and it booted fine. Hmmm, this isn't quite the same as before then. Having seen the advice about leaving it when it's locked, I've been doing that religiously and, so far, it's always woken up again by itself, but it's definitely indicative of a fault somewhere.
I've tried to check what's happening with adb logcat (I'm not rooted - fully stock - so can't see dmesg) and there is always some sort of I/O error when the freeze happens - often an sqlite database. Then, a couple of days ago, three photos and one video suddenly became corrupted.
I've realised that there are two issues, and only one of them has been "fixed".
The main problem, I believe, is that the flash memory is dying quite rapidly. We all know that flash memory has a limited write life, and wear levelling is supposed to extend that life beyond the typical lifespan of a device (say 5 years for a PC SSD drive?). For some reason, the flash in these devices is wearing out MUCH more rapidly.
The "faulty" eMMC controller obviously had a problem when dealing with faulty flash cells and would get stuck in a permanent loop, bricking the device. As far as I'm aware, the "fix" that has been applied prevents this permanent loop. The system still needs to try and handle disk errors with the flash memory, but it's not a permanent freeze any more and eventually (once it's finished dealing with flash faults - often unable to recover the problem) the system will break out of the freeze and carry on as before.
Sometimes this means that an app has bombed out as it couldn't read it's data correctly, but most people would either have forcibly rebooted their phone, or not notice as that app would just restart next time they used it. Sometimes, however, it means that something more important has crashed due to the disk I/O problem, resulting in the phone needing to spontaneously reboot itself.
What this boils down to is - yes, I believe that there has been a fix applied for the "faulty" eMMC controller getting stuck in a permanent, irretrievable loop. But no, I don't believe the actual original problem has been addressed - namely that the flash memory is dying at an unacceptably fast rate for ... well ... who knows what reason.
I do have a theory on that too. Recently a friend of mine was looking to change his laptop hard drive for an SSD. I investigated for him and decided that the Samsung 830 series would be his best bet, if he could get hold of them. And why not the newer 840 series? Well - that's purely because the "consumer" version of the 840 uses Samsung's latest triple level cell flash memory. Newer, faster, cheaper to manufacture almost certainly but ... more importantly in this case ... less resilient. How does 1000 write cycles grab you? They claim that the wear levelling algorithm in the drives mean that their lifetime is still pretty reasonable, but I'm not convinced in the real world.
Any idea what flash they use in the S3? I have no idea, but I betcha it's something (a) cheap and (b) new(ish).
I'm now at a bit of an impasse with my phone, however. When I returned it for repair last year it was properly frozen. No-one could do anything to fix it, the flash wouldn't format or write firmware - the only recourse was replacing internal hardware (or the entire phone).
This time, however, it works ... sort of. I'm occasionally losing photos I've taken and occasionally finding it's locked up and occasionally having to wait for it to recover. No repair centre has the time to "live with" my phone for half a day, waiting to see what happens with it - and even if they did, they'd blame a rogue app or "something left over from the upgrade" - they'll just factory wipe it, maybe flash the firmware and send it back again "repaired".
It's tempting to "lose" it and pay my insurance excess ... but even then, I won't know when the problem will strike again - but that's the thing ... I'm certain it would be back.
I used to love this phone. Now I can't rely on it. I can't rely on it storing the photos I take with it. I can't rely on it ringing when someone calls. I can't rely on texts getting through to me on time. I can't even rely on it waking me up in the morning.
It's sat next to me now,lying on the desk with the screen all shiny and black. I have no idea if it's fine, or frozen. Is someone calling me right this second? I have no idea. Oh, there we go - pressed the lock buton - it's awake. I shouldn't have to check my phone every ten minutes to see if it's working or not.
*sigh* sorry - rant over.
Suffice to say, that's my take on it. If I thought it would help diagnose something "fixable", I'd root and see what dmesg is doing, but I'm convinced this is hardware and I'm also certain that Samsung aren't going to take the blindest bit of notice.
I'm pretty sure there will be more and more people complaining that their phones are misbehaving over the next few months, but it's vague enough of a problem that the "repair" will be wiping - just long enough to take them over the first year warranty...

QorbeQ said:
It's a shame that this post has been mostly ignored in favour of the "SDS" one, but I do actually believe both issues are related. My I9300 is running stock - always has been - and suffered with "proper" SDS last November.
The initial symptoms were that my previously perfectly-behaved phone would would randomly FC apps that didn't before, and some of my photographs and video taken on the phone suddenly became corrupted and unusable.
Finally, one Saturday morning, I woke up to find the phone stuck in a boot loop, going from the "Samsung Galaxy S3" black-and-white boot scree to the first blue sweep of the Samsung logo, then reboot. It would power-off and power-on and I could enter ODIN or recovery modes fine, but it wouldn't boot any further.
I managed to get my photos and stuff off the phone through USB in recovery, and decided to wipe and flash the same stock ROM via ODIN.
This process failed - it couldn't partition the device correctly. None of the flash counters were reading anything so I popped it into my local Vodafone shop where they said it just needed the firmware popping back on and they'd have it ready in an hour.
I went back and they said they hadn't been able to load the software and it would have to go off for repair. *sigh*
It came back about 5 days later with a repair report stating that they'd replaced the "main board" - presumably because of the eMMC lockup bug.
It was only after this that the whole "Sudden Death" news came about and I realised what had happened to my phone. The eMMC check app says that my new controller is the same "faulty" version so I've been waiting with baited breath for "safe" official firmware.
The update system threw XXELL5 my way just before Christmas, and I thought that was that, but a coule of weeks ago the phone started hiccuping again. I knew I'd seen this before and my heart sank.
Last week another update was available, this time XXELLA. Since then, all hell has broken loose on my phone. It is locking up about 5 times a day. Initially, I was rebooting it manually (by holding in the power button). Tuesday morning I was woken by my wife saying "shouldn't you be up by now?" - glanced at my phone and it was frozen on the black-and-white boot logo - no alarm! Arrggh! It's dead!
Powered off, back on again and it booted fine. Hmmm, this isn't quite the same as before then. Having seen the advice about leaving it when it's locked, I've been doing that religiously and, so far, it's always woken up again by itself, but it's definitely indicative of a fault somewhere.
I've tried to check what's happening with adb logcat (I'm not rooted - fully stock - so can't see dmesg) and there is always some sort of I/O error when the freeze happens - often an sqlite database. Then, a couple of days ago, three photos and one video suddenly became corrupted.
I've realised that there are two issues, and only one of them has been "fixed".
The main problem, I believe, is that the flash memory is dying quite rapidly. We all know that flash memory has a limited write life, and wear levelling is supposed to extend that life beyond the typical lifespan of a device (say 5 years for a PC SSD drive?). For some reason, the flash in these devices is wearing out MUCH more rapidly.
The "faulty" eMMC controller obviously had a problem when dealing with faulty flash cells and would get stuck in a permanent loop, bricking the device. As far as I'm aware, the "fix" that has been applied prevents this permanent loop. The system still needs to try and handle disk errors with the flash memory, but it's not a permanent freeze any more and eventually (once it's finished dealing with flash faults - often unable to recover the problem) the system will break out of the freeze and carry on as before.
Sometimes this means that an app has bombed out as it couldn't read it's data correctly, but most people would either have forcibly rebooted their phone, or not notice as that app would just restart next time they used it. Sometimes, however, it means that something more important has crashed due to the disk I/O problem, resulting in the phone needing to spontaneously reboot itself.
What this boils down to is - yes, I believe that there has been a fix applied for the "faulty" eMMC controller getting stuck in a permanent, irretrievable loop. But no, I don't believe the actual original problem has been addressed - namely that the flash memory is dying at an unacceptably fast rate for ... well ... who knows what reason.
I do have a theory on that too. Recently a friend of mine was looking to change his laptop hard drive for an SSD. I investigated for him and decided that the Samsung 830 series would be his best bet, if he could get hold of them. And why not the newer 840 series? Well - that's purely because the "consumer" version of the 840 uses Samsung's latest triple level cell flash memory. Newer, faster, cheaper to manufacture almost certainly but ... more importantly in this case ... less resilient. How does 1000 write cycles grab you? They claim that the wear levelling algorithm in the drives mean that their lifetime is still pretty reasonable, but I'm not convinced in the real world.
Any idea what flash they use in the S3? I have no idea, but I betcha it's something (a) cheap and (b) new(ish).
I'm now at a bit of an impasse with my phone, however. When I returned it for repair last year it was properly frozen. No-one could do anything to fix it, the flash wouldn't format or write firmware - the only recourse was replacing internal hardware (or the entire phone).
This time, however, it works ... sort of. I'm occasionally losing photos I've taken and occasionally finding it's locked up and occasionally having to wait for it to recover. No repair centre has the time to "live with" my phone for half a day, waiting to see what happens with it - and even if they did, they'd blame a rogue app or "something left over from the upgrade" - they'll just factory wipe it, maybe flash the firmware and send it back again "repaired".
It's tempting to "lose" it and pay my insurance excess ... but even then, I won't know when the problem will strike again - but that's the thing ... I'm certain it would be back.
I used to love this phone. Now I can't rely on it. I can't rely on it storing the photos I take with it. I can't rely on it ringing when someone calls. I can't rely on texts getting through to me on time. I can't even rely on it waking me up in the morning.
It's sat next to me now,lying on the desk with the screen all shiny and black. I have no idea if it's fine, or frozen. Is someone calling me right this second? I have no idea. Oh, there we go - pressed the lock buton - it's awake. I shouldn't have to check my phone every ten minutes to see if it's working or not.
*sigh* sorry - rant over.
Suffice to say, that's my take on it. If I thought it would help diagnose something "fixable", I'd root and see what dmesg is doing, but I'm convinced this is hardware and I'm also certain that Samsung aren't going to take the blindest bit of notice.
I'm pretty sure there will be more and more people complaining that their phones are misbehaving over the next few months, but it's vague enough of a problem that the "repair" will be wiping - just long enough to take them over the first year warranty...
Click to expand...
Click to collapse
I'm really sorry for the issues you've been living with mate, sure thing is that what you said about memory lockups and blocking writing
firmware is totally true except mine doesn't freeze which is strange, mine is the affected model and is shown in eMMC brick bug check ;
-VTU00M
-06/2012
-Yes, Insane Chip.
The poll i made is permanant, but you have to post down Rom and kernel and recovery and bootloader to be more accurate to spot the problem and maybe some DEV can fix it as soon as possible.
Thanks for sharing ur experience with us mate.

I am totally stock jellybean not rooted etc. Was having bad problems with phone freezing and lagging so started uninstalling things to try and fix. I found the problem was chrome, I had chrome, chrome to phone and chrome beta on took them all off now running silky smooth again.
Sent from my GT-I9300T using Tapatalk 2

kiwiflasher said:
I am totally stock jellybean not rooted etc. Was having bad problems with phone freezing and lagging so started uninstalling things to try and fix. I found the problem was chrome, I had chrome, chrome to phone and chrome beta on took them all off now running silky smooth again.
Sent from my GT-I9300T using Tapatalk 2
Click to expand...
Click to collapse
First ever user to report Chrome as one source of the problem, thanks mate, if any other symptoms occur and have anything to do with freezing, let us know ok !

I made a similar thread here:
http://forum.xda-developers.com/showthread.php?t=2127443
TL;DR?
Phone started to freeze too much,every 5-10 mins.
Unrooted went to total stock after a megawipe.
Phone died after 2 days.
GOt 16 GB motherboard replaced with 32 gb.

I dont get it, you have a 2 year warranty. If it happens, and im sure the vast majority wont have any issues, take it for warranty.
Sent from my GT-I9300 using xda app-developers app
---------- Post added at 06:50 PM ---------- Previous post was at 06:49 PM ----------
eggman89 said:
I made a similar thread here:
http://forum.xda-developers.com/showthread.php?t=2127443
TL;DR?
Phone started to freeze too much,every 5-10 mins.
Unrooted went to total stock after a megawipe.
Phone died after 2 days.
GOt 16 GB motherboard replaced with 32 gb.
Click to expand...
Click to collapse
This though seems vere interesting. Ive read that a few mention their board got replaced by a 32gb. Which prolly doesnt have the bug. So im guessing samsung is very aware of this bug.
Sent from my GT-I9300 using xda app-developers app

This though seems vere interesting. Ive read that a few mention their board got replaced by a 32gb. Which prolly doesnt have the bug. So im guessing samsung is very aware of this bug.
Click to expand...
Click to collapse
I suppose that freezing problem is occured when a device is severly damaged even when on safe kernel and recovery, as much as i hate to admit it i think freezing S3 is the last symptom before SDS. Let's hope it's not.

nhariamine said:
I suppose that freezing problem is occured when a device is severly damaged even when on safe kernel and recovery, as much as i hate to admit it i think freezing S3 is the last symptom before SDS. Let's hope it's not.
Click to expand...
Click to collapse
I got 16gb Motherboard replaced with 32gb.
So it may not be as bad as it seems :cyclops:

nhariamine said:
I suppose that freezing problem is occured when a device is severly damaged even when on safe kernel and recovery, as much as i hate to admit it i think freezing S3 is the last symptom before SDS. Let's hope it's not.
Click to expand...
Click to collapse
I don't know if this is true in every case (freezing S3 is the last symptom before SDS) since I have 4 or 5 freezes on 2 consecutive days about 2 months ago. After that the phone runs normally, not a single freeze since then. And another thing my SGS is the 32 Gb model which is supposed to be safe from SDS (it has a different version of eMMC)

thank goodness for this thread.
this phone has been killing itself up to 3 times in 10 minutes. i really dont know what the hell is going on with this.
i will try uninstalling chrome and see if it helps..
my phone is bone stock running 4.1.2 with the latest firmware updated just today.

data connection is maybe the problem
Hey guys,
i have these freezes too since i updated to 4.1.2.
but in my case it's only when i have mobile data connection or w-lan connection turned on. If i turn off any data connection, it works the whole day without freezing!
It's for two weeks now and today i decided to install CyanogenMod 10.1 to try if this solves my problem.
I'm about to synchronisize my apps (does it automatically when entering google account and choosing "restoren from my account") and the first freeze occured (W-Lan on). But this time the light of the touch keys is full functional. When i touch the screen, the background light of the touch keys turns on. But power button, volume up/down and the touchscreen are frozen.
What I've done:
- rooted and installed CWM
- Cleared Dalvik cache
- wiped all other cache partition and stuff
- installed CyanogenMod 10.1 nightly build 20130208
- factory reset and so on like described in the manuals for installing CFs
Original data of my phone before installing CyanogenMod:
Galaxy S3
AP: I9300XXELL4
CP: I9300XXELL4
CSC: I9300DBTELL1
Android Version:
4.1.2
Kernel-Version:
3.0.31-566833
[email protected]#1
No Branding.
So it is a bit confusing that this only happens with data connection turned on.
I've also read some threads with analytics of the update code and the SD fix so i am well informed now and think it's because of the eMMC bug where it tries to repair the corrupted sectors and freezes while doing this (cause after ~20 minutes everything is fine like nothing happened).
Has anyone an idea or a solution now besides sending it to the local dealer and change it for warranty reasons?
Thanks,
Dark Smile

Odp: **Ultimate Galaxy S3 Unusual Freezing Thread**
It happened to me once, about 2 weeks ago, just after the release of the new firmware LM2. So I have flashed back with PC Odin to the previous LLA but the problem persisted. The phone was freezing after about 5 minutes of use. Forced reboot was needed.
So then I flashed some older 3 files-low level firmware (can't remember what one) in PC Odin nd then back to LM2 again, and everything is alright since then.
Strange...
Wysyłane z mojego Nexus 7 za pomocą Tapatalk 2

After 2 Freezes it's running stable for the moment... i'll watch it over the weekend and give feedback.

Mine is freezing lots now.
First time it went to o2 repair (UK) they claimed they had fixed it, pretty sure they just did a factory reset (which I had done already 3 times). So when I collected it, it froze within 2 minutes.
I'm tracking this repair publicly at http://myo2repair.wordpress.com so you can check out how this sort of thing pans out.
Problem is it can be intermittent, so you could get a couple of hours without a freeze. So it's gone back to repair and they say no fault found. So a bit stuck now.
It will definitely happen again when it's returned. They say after 3 attempts they will give me a refurb device. Samsung themselves give you a brand new sealed device as a replacement. Recommend going through Samsung if in same predicament.

Yes mine also got replaced by a 32 gb one from samsung india today
Sent from my GT-I9300 using xda app-developers app

good news.
after updating today (and uninstalling google chrome) my SGS3 is no longer freezing itself every few minutes!
will report if it starts doing it again. If you dont see me post, assume all is good now.

freeze after 4.1.2
I've had constant freezes when I upgraded to 4.1.2. I tried many versions. all the same.
I rolled back to 4.1.1 and the phone is freeze free for one week now.

Mine started to behave strangely and then started to freeze constantly. Didn't happen after an upgrade, just started happening out of the blue.
So I wiped it completely and flashed it with the latest wanamlite rom (I was on a rooted stock rom before).
Now it only freezes once every few days (still it NEVER used to freeze before, ever).
I think it might have some connection to the facebook app. I've noticed it only freezes shortly after I have been using the facebook app.

SiggiJarl said:
Mine started to behave strangely and then started to freeze constantly. Didn't happen after an upgrade, just started happening out of the blue.
So I wiped it completely and flashed it with the latest wanamlite rom (I was on a rooted stock rom before).
Now it only freezes once every few days (still it NEVER used to freeze before, ever).
I think it might have some connection to the facebook app. I've noticed it only freezes shortly after I have been using the facebook app.
Click to expand...
Click to collapse
So far, users are reporting Chrome and Facebook as one of the sources of the freezing in Galaxy s3.More to come, Share ur experiences down

Nexus 10 turning off in sleep mode

Hello!
I bought Nexus 10 at nearly two weeks, and since the beginning i have noticed an issue i thought would be Software related. As the device is on idle, everything works pretty well. Watching movies, web surfing, etc. Doesn't matter if it's for 5mins or 5 hours. But when it goes to sleep mode (automatically or forced by me), it just turns off consequently. I would say 90% of the times. An then Nexus just turn back to life after a long power button press, like 10 secs or more.
First i read about some people who have problems caused by autosleep function related with some cases. When i remove my tablet from the Poetic case, it keeps going off.
Then i read about Location services. It's turned off too. Such as the wi-fi in sleep mode. Didn't solve either.
Any idea?
PS: I'm from Portugal and my English is not the best. Sorry about that.

nando_pt said:
Hello!
I bought Nexus 10 at nearly two weeks, and since the beginning i have noticed an issue i thought would be Software related. As the device is on idle, everything works pretty well. Watching movies, web surfing, etc. Doesn't matter if it's for 5mins or 5 hours. But when it goes to sleep mode (automatically or forced by me), it just turns off consequently. I would say 90% of the times. An then Nexus just turn back to life after a long power button press, like 10 secs or more.
First i read about some people who have problems caused by autosleep function related with some cases. When i remove my tablet from the Poetic case, it keeps going off.
Then i read about Location services. It's turned off too. Such as the wi-fi in sleep mode. Didn't solve either.
Any idea?
PS: I'm from Portugal and my English is not the best. Sorry about that.
Click to expand...
Click to collapse
You can view my profile for my posts I've been working on that problem for over month. Sorry this is short reply - pressed for time right now.

bigmatty said:
You can view my profile for my posts I've been working on that problem for over month. Sorry this is short reply - pressed for time right now.
Click to expand...
Click to collapse
Thanks for the response anyway.
Is that thread about freezing? My Nexus is not freezing.

Hello! Yes, that is one of the threads Im talking about. I found it to hard to find people who posted with as good of detail as you did, so I had to "latch on" to the biggest group that was vocal about their problems. But mine manifests itsself almost always as sleep of death, and NOT reboot - just like yours. Here are some links:
Google: https://groups.google.com/a/googleproductforums.com/d/msg/mobile/-/GFnACv1nC0MJ
On of the XDA: http://forum.xda-developers.com/showthread.php?t=1998496&page=39
As I say on the Google thread:
My thoughts on the error:
1) The most logical explanation I've heard yet is the memory leak that B has been posting about
2) I think this memory leak/problem can manifest its self as crash, reboot, or sleep of death
Right now I am trying Stock 4.2.2, rooted, using KTmanta kernal. This is the second day Im on a non-stock kernal, too early to post any useful info.
Good luck on this,
bm
---------- Post added at 02:09 PM ---------- Previous post was at 02:05 PM ----------
say, quick question since you are in same boat as me - does your device's serial # start with R32C or R32D?

My device is R32D.
I guess i will call the seller tomorrow and figure out if it's possible to return the tablet. At this point, it's completely unusable.

nando_pt said:
My device is R32D.
I guess i will call the seller tomorrow and figure out if it's possible to return the tablet. At this point, it's completely unusable.
Click to expand...
Click to collapse
I agree that the inability of this device to reliably stay powered on is a deal breaker, at least in a work scenario.
I actually bought mine "just for games", but have been really concerned about trying to resolve this issue to verify its not a hardware issue and I miss my warranty period. I cant honestly say I think its a harware issue, so I am really trying to find a ROM/Kernal/settings combo that resolves this.
I have a lot of faith in the dev community for ROMs and kernels, but I think its absolutely scandalous and horrible that I have to click past a warning that says 'may void your warranty' to install softare on my device to make it perform AS IT SHOULD HAVE direct from Google.
If not for all these issues I would have NEVER gone custom recovery, root, and new kernel. Im glad this has forced me to do so because Ive loved finally having a justification for "risking" my expensive device, but again that learning experience is NOT what Google is taking $500 payments for here https://play.google.com/store/devices/details?id=nexus_10_32gb so Google should be SHAMED!
Thanks for the info on your serial number, that actually relieves a concern of mine that somehow the 'newer' units werent having this problem. Mine's a C.
bm

just bought my nexus 10 yesterday. it also turning off when sleepmode. but when i update it to 4.2.2 its gone.

jamir5 said:
just bought my nexus 10 yesterday. it also turning off when sleepmode. but when i update it to 4.2.2 its gone.
Click to expand...
Click to collapse
I hope it stays that way for you. It will take a few more days with the device to know for sure. If the problem comes back it'd be worth noting here for troubleshooting. :good:

I called back the seller, and they will do the return. Really bad situation...

i left my n10 in my house and its idle with 41% on battery i just got home and open the device it consume only 1% without random reboots. hope itll stay this way ill root this and install crom maybe nextweek.

So - now I am 2.5 days into testing 4.2.2 OS with a different kernel. And so far, NO sleep of death! AND I've been using the tablet a lot more than my "minimal use" in my other tests! This of course is the ultimate test - I'm posting/talking about it on the internet so certainly if its going to blow up in my face, now would be the time. But as of right now - it looks like the alternate kernal is better for my device than Google's...
I am using the KTmanta kernel.

It is important to note that the issue described in this thread does NOT involve the device turning itself off. Rather the issue is that the device is unresponsive when in sleep mode. You can tell this is the case in a few ways, but two obvious ones are that the notification light continues to function and the device requires a 10-second power button reset (the procedure for forcibly rebooting the device).
Personally, I am on CM10.1 night lies with Ktmanta kernel and still experiencing this problem.
Bigmatty, can you tell me what governor, i/o scheduler and clock settings you are using? Thanks.

neatchee said:
It is important to note that the issue described in this thread does NOT involve the device turning itself off. Rather the issue is that the device is unresponsive when in sleep mode. You can tell this is the case in a few ways, but two obvious ones are that the notification light continues to function and the device requires a 10-second power button reset (the procedure for forcibly rebooting the device).
Personally, I am on CM10.1 night lies with Ktmanta kernel and still experiencing this problem.
Bigmatty, can you tell me what governor, i/o scheduler and clock settings you are using? Thanks.
Click to expand...
Click to collapse
Well, the majority of the complaint threads do not specifically or mainly focus on the device turning its self off, but this thread actually does.
My sleep- of- death problem on my old unit persisted on any scheduler/clock settings I tried, and I tried a lot w/ KTmanta. But it was RMA my device to get a new one that fixed my SoD.
Oh, and the LED flashing does not mean the unit is actually on, you can test that by plugging in a charger before attempting a long (power on) press = if you get instant battery charge indicator your unit was actually off.

bigmatty said:
Well, the majority of the complaint threads do not specifically or mainly focus on the device turning its self off, but this thread actually does.
My sleep- of- death problem on my old unit persisted on any scheduler/click settings I tried, and I tried a lot w/ KTmanta. But it was RMA my device to get a new one that fixed my SoD.
Click to expand...
Click to collapse
That's REALLY weird, because based on a number of findings by a number of people, it turns out that this issue is related to a bug in the GPU driver that is causing a massive memory leak. You can check by opening a Terminal Emulator and running the command "ps su" This will show the memory usage for the "surfaceflinger" application, which is where the memory leak manifests. Memory usage should never go above ~70mb for that application, but on the N10 it balloons up above ~300mb before the system locks up.
Apparently we're all waiting on Samsung to fix the problem :\

neatchee said:
That's REALLY weird, because based on a number of findings by a number of people, it turns out that this issue is related to a bug in the GPU driver that is causing a massive memory leak. You can check by opening a Terminal Emulator and running the command "ps su" This will show the memory usage for the "surfaceflinger" application, which is where the memory leak manifests. Memory usage should never go above ~70mb for that application, but on the N10 it balloons up above ~300mb before the system locks up.
Apparently we're all waiting on Samsung to fix the problem :\
Click to expand...
Click to collapse
After my two months of trouble shooting I feel strongly that there are two main issues: memory leak and "something to do with power". Likely a software/driver/OS problem on the memory leak like the majority says. BUT - not everyone has the power problems. I think the power problems are hardware related.
Only a couple other users have talked about experiencing the power problems...
If I was going to place a bet, I'd say based on your description of symptoms you too have the power problem. Ask the others in threads you are following if they have the LED thing, I'm guessing they don't.
Rxnelson on these forms does. He's seen it on two physical units. For me, it didn't appear on my second physical unit.
Good luck!!

"LED thing?" The LED thing is just the notification light continuing to flash even when surfaceflinger has locked the system. Don't really see how that's related to power in any way. The way that LED notification lights work is that the system sends an "on" signal to the LED control chip which then just runs the notification light pattern in a loop until it receives the "off" signal from the system. If you don't have any notifications prior to surfaceflinger going ballistic, or if you have the notification light disabled (many people do) you'd never notice this particular symptom.

Also please note the tip I shared about diagnosing if it's off or on via the charger I posted above, handy trick another user taught me.
---------- Post added at 01:51 PM ---------- Previous post was at 01:48 PM ----------
neatchee said:
"LED thing?" The LED thing is just the notification light continuing to flash even when surfaceflinger has locked the system. Don't really see how that's related to power in any way. The way that LED notification lights work is that the system sends an "on" signal to the LED control chip which then just runs the notification light pattern in a loop until it receives the "off" signal from the system. If you don't have any notifications prior to surfaceflinger going ballistic, or if you have the notification light disabled (many people do) you'd never notice this particular symptom.
Click to expand...
Click to collapse
Not quite what it seems. If the unit was on, plugging in the charger would not instantly,w/out delay, display the powered off charging indicator.

Ahh, okay, bit of a miscommunication. I'm not referring the power-on notification light. I'm referring to the notification light flashing that occurs when you have a notification waiting (new email, new gTalk message, etc). The charging light is something different, and not what I was referencing
NOTE: This is a function of CyanogenMOD 10.1 that is not present in the stock ROM, I think, so that may be where the confusion lies. Sorry about that

neatchee said:
Ahh, okay, bit of a miscommunication. I'm not referring the power-on notification light. I'm referring to the notification light flashing that occurs when you have a notification waiting (new email, new gTalk message, etc). The charging light is something different, and not what I was referencing
NOTE: This is a function of CyanogenMOD 10.1 that is not present in the stock ROM, I think, so that may be where the confusion lies. Sorry about that
Click to expand...
Click to collapse
Yes dude, I understand. Just test what I am saying. When I saw my white gmail led flash, but then couldn't turn on I thought the exact same thing you typed. Then while discussing w/power users someone told me to try that. Next time it happened I tried it, was SHOCKED to see what happened, knew I had a problem (that was not just shell reboot.)

My tablet just started doing this as well. It's never done it before, it started about a week ago. What could be causing this? I just factory reset to see if it'll fix the problem. I'll report back in a few days to see if this fixes the issue.
Sent from my Nexus 10 using xda app-developers app

[Q] Full stock n9005 - Hotplugging misbehaving, cores stuck online (2 or 4)

Hi,
Anyone been keeping an eye on hotplugging on the n9005 ?
Mine is still unrooted, so I'm talking 100% vanilla Samsung configuration...
I use Android Tuner's online core notifications to keep an eye on this, there are other apps that can to this as well.
At some point, after a few hours of uptime, sometimes only after quite a many (20h+), I just notice that it will never drop below 2 or 4, never had it lock on 3 minimum, so can't comment on that scenario.
Once this happens, it's impossible to get back to single core even on very low load.
Also when leaving the device doing nothing, freq. drops to 300MHz and still 4 cores remain online.
There is obviously an impact on battery drain as well.
Only way to get back to normal is to reboot...
JP.

I've noticed the opposite issue, two cores just going offline forever until I reboot the device.

So that would point to obvious hotplugging issues, except if you ran into thermal throttling, maybe.
Were you running some heavy stuff when this happened ?
JP.

i tested it out of curiosity using the widget in system monitor app.
i have exactly the same issue as SentinelBorg above. the two cores go offline and never come back online until the next restart.
very wierd.

Well for the very least we can conclude hotplugging is flawed in the current firmware ...
As soon as we get a proper recovery (nandroid restore that don't bootloop) I'll start working on an improved kernel, this is just not acceptable for a top notch 700€ pricetagged device
JP.
Send from my Note 3 (n9005)

Latest case ...
Send from my Note 3 (n9005)

Again ...
Send from my Note 3 (n9005)

You're on full stock. Doesn't that mean you have no options regarding CPU government?

nakedtime said:
You're on full stock. Doesn't that mean you have no options regarding CPU government?
Click to expand...
Click to collapse
Correct, so I cannot do anything about this misbehaving, except reboot.
I want to know whether this is something others observe as well.
The thing is, if you're not looking at it, you won't know it's happening, cause who really keeps an eye on online cores all the time ? I do as I do kernel coding and as such also work on governors, which make me look at stuff like that.
If this happens of yours, you'll just have a hotter Note 3 with less lasting battery without knowing why, or worse without knowing battery easily could last longer if that was fixed.
This clearly is a bug in the current firmware, since as you correctly stated, being full stock unrooted, I can't be influencing this in any other way than applying load to the CPU which is clearly not the case as you can see in the screenshots... and it just happens randomly.
Once I start on a Note 3 kernel that will be in the top list of things to take care of.
JP.
Sent from my SlimBean 4.3 build 1.8 / Yank555.lu CM10.2 kernel v1.6g-beta4 (Linux 3.0.99) powered Galaxy S3 i9300

I would have gone with the N9000 if I knew how to develop a kernel. BIG.little sounds like so much fun to dev on.
PS
Have you found any Info about the thermal throttling? I'd love to be able to manipulate the temps for less heat when over clocking.

I was actually not going to get any more Samsung device because of crappy open source support they (in fact don't) give
I ended up going for the note 3 because of the specs and exactly because it was the Qualcom and not the Exynos version and a one time offer I could not refuse, but I will go nowhere Samsung's own super secret Exynos CPUs ever again unless they change their policies, which I doubt.
I just do hope I won't regret my move...
JP.
Sent from my SlimBean 4.3 build 1.8 / Yank555.lu CM10.2 kernel v1.6g-beta4 (Linux 3.0.99) powered Galaxy S3 i9300

exactly the same here, after a reboot everything is back to normal
Sent from my SM-N9005 using Tapatalk 2

Happened again after ~10h uptime while watching a live video stream...
Cores went to a steady 4 online, load was unchanged over a longer period (~45 mins of video stream over WiFi).
This makes little sense, hotplugging is flawed.
JP.
PS: Still stock unrooted.
Send from my Note 3 (n9005)

This is still happening for me, reboot solves it. Running 4.4.2 and thought this should be solved by now.
Sent from my SM-N9005 using XDA Premium 4 mobile app

Stock kitkat XXUENB3, unrooted. I have a similar problem when all of the sudden all cores are stuck online with core 1 and 2 stuck at 1958 mhz and core 0 and 3 scaling normally just not going offline. A reboot fixes the issue untill it happens again randomly. This is my last Android device and I had gs2,gnex,n4 and now n9005.

I've rooted and when the issue happens, disabling mp decision and enabling it again fixes the issue so this means that hotplug doesn't play well. Maybe a custom kernel will fix this issue.

ciprianruse88 said:
I've rooted and when the issue happens, disabling mp decision and enabling it again fixes the issue so this means that hotplug doesn't play well. Maybe a custom kernel will fix this issue.
Click to expand...
Click to collapse
Yes, long time I've open this thread
Indeed, on TW I used to do just that, "pkill -9 mpdecision", which killed the daemon and it respawned and all was back to usual.
I've since then moved to CM11, and this never happened again ever since, so I expect this to be a TW related issue, since mpdecision had the same MD5 signature in CM11 as on my stock 4.3 TW ROM.
And it looks like if the problem still persists on KK TW...
But it's certainly not a kernel related issue, as I've use my own kernel on JB TW, where this happened as well, once I went CM11, I ported my kernel to CM, and as I mentioned, never happened ever since.
Might be TW messing with DVFS (for benchmark cheating and maybe other things...).
Maybe worth trying to use Wanam's Xposed Module and disable ROM DVFS.
JP.

2 cores always online and the other 2 always offline= Dual Core
Ok guys I have experienced the bug with 2 cores being always offline which basically makes your Note 3 a dual core phone many times in the past untill I found what caused this! Even firing up Antutu the cores are still 2 and not 4 online! So yeah at first a reboot was an easy and effective sollution but that was not enough so I devoted 2 days trying to find out what caused this. Guess what? The problem was inside one of Samsung apps, specifically the bloody S-Finder app! I can reproduce the bug 100% whenever I want so I am very sure about it! Also I have tested this on 2 other Note 3s (n9005) and I noticed the same bug , with 4.3 and unfortunately now on 4.4.2!
Steps to reproduce it:1) Fire up S-finder 2) Search any app and launch it or even without launching just only by typing on the search bar the phone enters into dual core mode(!) after few seconds 3)Now this is the critical step, remember now the phone is already having only 2 cores enabled so if you press the home button then you are stuck with the buggy dual core mode!!! So this is how you end up with only 2 out of 4 cores working!
The sollution is of course to reboot the phone but another more faster and specific sollution is to close the S-Finder with only this (correct) way: while you are on the S-finder app close it by simply pressing the back button untill it exits completely and you are out to the menu. Incase you are not in the S-finder app then you must check if the program is shown in the recently used apps(longs pressing the home button) then tap on it to open it and then close it by hitting the back button." Don't close/kill it from the task manager or from the recently used apps, only close it when you are inside of the app and hitting the back button!"
Synopsis All in all, the culprit at least on my case is the S-finder app in which when it is used it somehow forces the phone to get into dual core mode and it stays like this because you forget to properly close it after the search is done. The sollution is simply closing the S-Finder gently/properly by using the back button from the inside of the app and not with any other way.
I don't know if this is truly a bug of Samsung or just a sneaky way to increase/improve the battery life at the expense of performance but I surely know that many people have this issue without even realising it!
I said too many things so that I can give a detailed review on this bug, I hope I have expressed it clearly and not to confuse you!
I think it is a very serious issue and it should be fixed as well as inform people about it!

SAVVAS. said:
Ok guys I have experienced the bug with 2 cores being always offline which basically makes your Note 3 a dual core phone many times in the past untill I found what caused this! Even firing up Antutu the cores are still 2 and not 4 online! So yeah at first a reboot was an easy and effective sollution but that was not enough so I devoted 2 days trying to find out what caused this. Guess what? The problem was inside one of Samsung apps, specifically the bloody S-Finder app! I can reproduce the bug 100% whenever I want so I am very sure about it! Also I have tested this on 2 other Note 3s (n9005) and I noticed the same bug , with 4.3 and unfortunately now on 4.4.2!
Steps to reproduce it:1) Fire up S-finder 2) Search any app and launch it or even without launching just only by typing on the search bar the phone enters into dual core mode(!) after few seconds 3)Now this is the critical step, remember now the phone is already having only 2 cores enabled so if you press the home button then you are stuck with the buggy dual core mode!!! So this is how you end up with only 2 out of 4 cores working!
The sollution is of course to reboot the phone but another more faster and specific sollution is to close the S-Finder with only this (correct) way: while you are on the S-finder app close it by simply pressing the back button untill it exits completely and you are out to the menu. Incase you are not in the S-finder app then you must check if the program is shown in the recently used apps(longs pressing the home button) then tap on it to open it and then close it by hitting the back button." Don't close/kill it from the task manager or from the recently used apps, only close it when you are inside of the app and hitting the back button!"
Synopsis All in all, the culprit at least on my case is the S-finder app in which when it is used it somehow forces the phone to get into dual core mode and it stays like this because you forget to properly close it after the search is done. The sollution is simply closing the S-Finder gently/properly by using the back button from the inside of the app and not with any other way.
I don't know if this is truly a bug of Samsung or just a sneaky way to increase/improve the battery life at the expense of performance but I surely know that many people have this issue without even realising it!
I said too many things so that I can give a detailed review on this bug, I hope I have expressed it clearly and not to confuse you!
I think it is a very serious issue and it should be fixed as well as inform people about it!
Click to expand...
Click to collapse
Yes. I've did the exact same thing and cpu was dual core but after I've reset mp decision I couldn't reproduce it anymore. Good catch!

Yank555 said:
I use Android Tuner's online core notifications to keep an eye on this, there are other apps that can to this as well.
Click to expand...
Click to collapse
How did you enable the online core notifications in Android Tuner?