I compiled Android from the source Nvidia made available on their website and, while the process went smoothly, after flashing the images to the device I experienced a large number of issues:
1. The tablet does not go to sleep (shown as Awake in the Battery settings) while not plugged in, leading to fast battery drain. If plugged in, the device is shown as not awake.
2. While not connected to Wifi, the screen would become completely unresponsive to touch input, after turning on the device (if it previously was on stand-by).
3. A lot of error messages of missing libraries and general files are shown in Logcat. A few examples:
E/phs:governor(14217): failed to load "libgov_gpucompute.so": No such file or directory
E/SoundPool(913): error loading /system/media/audio/ui/Lock.ogg
E/ConnectivityService(834): Unexpected mtu value: [email protected]
E/TLK_Daemon(176): IOCTL_FILE_NEW_REQ returned 61
E/wpa_supplicant(1435): wpa_driver_nl80211_driver_cmd: failed to issue private commands
I did scavange libgov_gpucompute.so from the factory image, but that did not fix any of the symptoms; the error message disappeared.
I built every version Nvidia made available and had the same results.
The recovery image does not suffer from these issues.
Any suggestion on how to get a handle on this would be greatly appreciated.
Related
At least I think it must be a resource leak. I have multiple WM6 devices that all behave the same way when my application is run. After a while, maybe 15 hours, they gradually deteriorate and refuse to start other applications. At first, it will just be an application like Opera that will not start. Eventually, things like File Explorer will also not start. When I say they won't start, I mean that I get the "not signed with a trusted certificate or one of its components cannot be found" message.
My first thought was that there was a memory leak, but according to the output of GlobalMemoryStatus, the system's memory use is not increasing over time. Then I thought that it might be storage space since I'm generating a big log file. But the storage space still sits above 60MB when this happens.
Restarting the device gets everything back to normal. So far, this is what I know:
1. I'm using the RIL. I noticed today that after about 12 hours, I stop getting RIL notifications.
2. I'm monitoring memory with GlobalMemoryStatus, but the available physical memory doesn't seem to be decreasing over time
3. The thread count remains constant for the life of the application
4. My storage space is decreasing, but there is still over 60MB available when everything starts to go wrong
5. In the end, the device winds up with the screen lock on, even though it is not configured.
It seems that there must be some kind of resource leak. The only other thing I can think of are kernel resources. I tried to rule out things like event handles through static code inspection, but maybe there's something I'm missing.
Does anyone have any suggestions as to how I would troubleshoot this further? I'm using VS2008 and a Tilt2 and an HTC Imagio (it happens on other devices as well).
I tracked this down to a registry handle leak. What bothers me is that I had to do it by static code inspection. I just looked for things like CreateEvent, RegOpenKeyEx, etc.
Since this didn't seem to show up as consumed physical memory, does anyone have any methods for inspecting kernel resource consumption on Windows Mobile? Do I have to rely on KITL and Platform Builder with the emulator? I'm hoping that there's some way I diagnose this kind of problem with a real device. From my perspective, the device just started to fail and there were no external indicators to warn me of the impending failure.
sbaker25 said:
I tracked this down to a registry handle leak. What bothers me is that I had to do it by static code inspection. I just looked for things like CreateEvent, RegOpenKeyEx, etc.
Since this didn't seem to show up as consumed physical memory, does anyone have any methods for inspecting kernel resource consumption on Windows Mobile? Do I have to rely on KITL and Platform Builder with the emulator? I'm hoping that there's some way I diagnose this kind of problem with a real device. From my perspective, the device just started to fail and there were no external indicators to warn me of the impending failure.
Click to expand...
Click to collapse
hi!can somebody to help me?
i have asus p535 and i reset "start""settings"default settings"and i lost all from device.after apear align screen and remain like this
i think i have to install window mobile.can you tell me step by step how?
10000 thanks
Sorry, accidental re-post. Don't think I can delete it altogether...
HI there,
I have written an app and on some devices, using the app freezes the system. My main problem is, that I do not get any log information, because as soon as the device is frozen, logging completely stops and obviously the crash-reports normally would be generated after the crash, which does not work, because the device is frozen.
I have attached the device to my laptop via adb (via USB) and logcat everything, but the logging just stops as the device freezes. I have also checked all possible log locations (/data/tombstones, /data/anr and what else is in /data) and could not find anything.
Since the app crashes only on about 10% of all devices (the devices are identical and have the same firmware) I need to test the app on the device and cannot use the emulator. However if the app crashes, I can relieable reproduce the crash using adb monkey. On the other 90% adb monkey won't crash the device.
Is there anything else I can log / test / check? Is there a way to get a memory dump, once the device is frozen? Any other idea how I can get hold of any information?
Although I don't think this is relevant: The device is a XORO 9718DR tablet and the device is rooted. The app is written by myself and not available in the market. It mainly uses only Buttons, TextView, ImageView, HttpClientConnection, Threads, AsyncTasks and other "normal" Android stuff. It even does not use the webview.
ToBe_HH said:
HI there,
I have written an app and on some devices, using the app freezes the system. My main problem is, that I do not get any log information, because as soon as the device is frozen, logging completely stops and obviously the crash-reports normally would be generated after the crash, which does not work, because the device is frozen.
I have attached the device to my laptop via adb (via USB) and logcat everything, but the logging just stops as the device freezes. I have also checked all possible log locations (/data/tombstones, /data/anr and what else is in /data) and could not find anything.
Since the app crashes only on about 10% of all devices (the devices are identical and have the same firmware) I need to test the app on the device and cannot use the emulator. However if the app crashes, I can relieable reproduce the crash using adb monkey. On the other 90% adb monkey won't crash the device.
Is there anything else I can log / test / check? Is there a way to get a memory dump, once the device is frozen? Any other idea how I can get hold of any information?
Although I don't think this is relevant: The device is a XORO 9718DR tablet and the device is rooted. The app is written by myself and not available in the market. It mainly uses only Buttons, TextView, ImageView, HttpClientConnection, Threads, AsyncTasks and other "normal" Android stuff. It even does not use the webview.
Click to expand...
Click to collapse
What you are saying is an App Not Responding situation where the app does too much work on the Main Thread causing a UI Freeze. You could analyse the ANR Trace file from your phone... I'm not quite sure ... But ... You could use the Below links to know more in detail :
investigating an ANR
How to debug Android ANR
Interpreting an ANR stack trace
No ANR
Thank you niranjan94 for your reply. But by "freeze" I mean that the complete device freezes. It is not accessible via USB or network any more. All apps are frozen (e.g. I had an app in the background, which recorded the logcat and that app stopped the same second the device freezes). Therefore there is no ANR trace. Unfortunately. :crying:
BTW: There is also no tombstone in /data/tombstones....
Hello,
I have "Pandigital Novel Black / Model R70E200 / which supposedly runs Android 1.5" which went through SD card update which did not finished correctly.
My knowledge is limited at this moment about android internals but I am coming from PC Linux world which could help in resolution of following situation.
Yesterday I decided to read some documentation on the e-Reader but found that touch screen did not respond (buttons and orientation worked fine). I did attempted to restart the e-Reader a few times but without any success -- touch screen did not respond (the e-Reader was not used for prolonged period of time).
And it seems that by accident I "hit" a combination of buttons which supposedly initiated some unknown process at boot time. After this incident I had no option as to attempt to re-flash firmware.
For some inexplicable reason the flash process did not went as I would expect. Now it does not matter what I do process verification of firmware never ends.
I've attempted to get into "hard reset" but no success so far -- I always end up in re-flash situation.
To get at least some information for debugging of the situation I installed "adb" and was able to retrieve a log message which states following
link_image[1714]: 1760 could not load needed library 'libm.so' for '/system/bin
/sh' (link_image[1703]: 1760 missing essential tables)CANNOT LINK EXECUTABLE
and similar message about libstdc++
These two messages could explain why re-flash never ends -- I am not sure what happened but some libraries or damaged or missing and I am not able to get "sh" run.
I was able to pull "sh" binary on my computer but in Linux "ldd" could not "decipher" location of missing libraries so that I could attempt to push them in proper location in the e-Reader.
I would be happy to hear any suggestions how the situation can be corrected.
I still keep some hope that the situation can be corrected as "RDU" is still running and e-Reader still "has some breath in it".
RDU Version 2.8.9a (current boot count: 0)
Firmware version S6410_PPP2_BBB_2010_11_18
At normal boot the e-Reader has following behavior:
1. Please be patient while system starts
2. On the screen "Pandigital novel image with Contains Reader(R) Mobile Technology by Adobe System Incorporated" strangely shifted about 1/5 of the screen toward right side
It is quite difficult to establish a connection between adb and e-Reader.
Thank you,
Andromeda
andromeda_2010 said:
Hello,
I have "Pandigital Novel Black / Model R70E200 / which supposedly runs Android 1.5" which went through SD card update which did not finished correctly.
My knowledge is limited at this moment about android internals but I am coming from PC Linux world which could help in resolution of following situation.
Yesterday I decided to read some documentation on the e-Reader but found that touch screen did not respond (buttons and orientation worked fine). I did attempted to restart the e-Reader a few times but without any success -- touch screen did not respond (the e-Reader was not used for prolonged period of time).
And it seems that by accident I "hit" a combination of buttons which supposedly initiated some unknown process at boot time. After this incident I had no option as to attempt to re-flash firmware.
For some inexplicable reason the flash process did not went as I would expect. Now it does not matter what I do process verification of firmware never ends.
I've attempted to get into "hard reset" but no success so far -- I always end up in re-flash situation.
To get at least some information for debugging of the situation I installed "adb" and was able to retrieve a log message which states following
link_image[1714]: 1760 could not load needed library 'libm.so' for '/system/bin
/sh' (link_image[1703]: 1760 missing essential tables)CANNOT LINK EXECUTABLE
and similar message about libstdc++
These two messages could explain why re-flash never ends -- I am not sure what happened but some libraries or damaged or missing and I am not able to get "sh" run.
I was able to pull "sh" binary on my computer but in Linux "ldd" could not "decipher" location of missing libraries so that I could attempt to push them in proper location in the e-Reader.
I would be happy to hear any suggestions how the situation can be corrected.
I still keep some hope that the situation can be corrected as "RDU" is still running and e-Reader still "has some breath in it".
RDU Version 2.8.9a (current boot count: 0)
Firmware version S6410_PPP2_BBB_2010_11_18
At normal boot the e-Reader has following behavior:
1. Please be patient while system starts
2. On the screen "Pandigital novel image with Contains Reader(R) Mobile Technology by Adobe System Incorporated" strangely shifted about 1/5 of the screen toward right side
It is quite difficult to establish a connection between adb and e-Reader.
Thank you,
Andromeda
Click to expand...
Click to collapse
Hello,
Thanks for using XDA Assist.
There's no dedicated forum for your device here in XDA.You can post your query in Android Q&A,Help and Troubleshooting.Experts there may be able to help you
___
v7
XDA Assist
Hi all. I have Nokia 8.1 for a few months now. Updated to Android 10 immediately and it has been rock stable. Until today.
At about 11am I heard an incoming message notification but was busy and decided to look at it later.
At 11:55am I picked the phone but could not wake it up - tapping didn't work, fingerprint didn't work. Black screen. So I long-pressed the power button and it booted up.
It is working fine since but I'm somewhat concerned about the issue.
I collected a zip log file through "adb bugreport" command.
As I understand, it's tricky to find the cause after the reboot because all log files are generated fresh after boot. I looked in \FS\data\tombstones folders and found three records there, all complaining about the following issue:
Code:
signal 6 (SIGABRT), code -1 (SI_QUEUE), fault addr --------
Abort message: 'FORTIFY: %n not allowed on Android'
But they are almost a month old and I don't remember anything weird happening at that time; except maybe some Android system update and manual reboot.
Is there any place in that log zip or somewhere else on the device (without root) where I could find some info about what happened right before the unexpected shutdown? Kernel panics? Crashes?
Hmm, nobody? Does it mean that everybody else is just ignoring random shutdowns and have never attempted to debug the problem?
AFAIK logcat output isn't persisted over reboot, and maybe that's one reason nobody replied to this thread. But you can always write log to a file every time your phone boots (via shell commands etc) and just check out that file once phone rebooted.
Regarding kernel log, /proc/last_kmsg or /sys/fs/pstore/console-ramoops might be helpful.
Problems about "dev/block/loop", I/O error, and iosetup failed 1 Android-x86
So, I've been using Android-x86 for years without any problems, until recently.
A while ago, while I was playing some games on my previous Android-x86, my wifi died all of a sudden, and since I don't have anything else to do without an internet connection to it, I turned off my laptop and reboot when the wifi is back up.
While rebooting, I was then stuck in an infinite loop on the splash screen (which had happened a few times already), prompting me to reinstall Android-x86 (as per usual) because I can't think of any other method to fix it.
Normally, everything would be just fine, I just need to enable native bridge to be able to play the games I usually play, but this time something was off, and almost every games I used to play crash right away as if native bridge hadn't been enabled (which it was).
So I looked up to try finding a solution, and came across a guide to manually install houdini.sfs through terminal console, and the following kept happening:
mount: dev/block/loop(xx) I/O error
or
mount: iosetup failed 1
I absolutely have no idea what's causing this, and I can't find anything that talks about this error, and I hope I can find an answer here
Thank you for your time in advance