06-19-2012 04:47 PM
44 12
tools
  1. Discovered's Avatar
    I do reboots, about every two days aswell, But its also hard for me to stick with a Rom.
    12-30-2011 10:49 AM
  2. LeslieAnn's Avatar
    Why!?

    What's the point of custom roms if it just introduces these problems..?
    No rom is perfect, not even the factory rom. Heck, you can't even get a perfect router which is far more simple than a cell phone.
    12-30-2011 03:34 PM
  3. Whyzor's Avatar
    For people that aren't experiencing this bug and have a CM7 based ROM how often do you reboot or even use the accelerometer?

    This bug occurs in ALL CM7 based roms... I found out the hard way.

    This is confirmed by JerryScript here. "acceleratometer bug still exists in all CM7 based ROMs"

    Also, BobZhome describes the symptoms as cumulative and the only fix is a reboot. Also, says it occurs on all CM7 roms here. "Sorry, the acceleratometer bug still exists in all CM7 based ROMs"

    So I've flashed CM7 roms correctly about a dozen times now and experienced it every damn time. Luckily, I made a stock nandroid backup, but that's so lame. ('bout to do it though just to prove to myself it's CM7)

    Anyway, accelerometer is officially broken for CM7 based roms on OV, in my, BobZhome, and Jerryscript's book at least.

    You may be referring to this issue: which according to BobZHome's comment was fixed, but I just did a quick check of inferiorhumanorgan's github and the fix is not in there. So Most builders probably don't have it. I would post in the official IHO thread and ask if it could be brought in.

    I just added this fix to my Motorola Triumph build since people were complaining of similar issue with it. Still waiting for feedback from testers.
    JerryScript likes this.
    12-30-2011 04:24 PM
  4. Schlidel's Avatar
    You may be referring to this issue: which according to BobZHome's comment was fixed, but I just did a quick check of inferiorhumanorgan's github and the fix is not in there. So Most builders probably don't have it. I would post in the official IHO thread and ask if it could be brought in.

    I just added this fix to my Motorola Triumph build since people were complaining of similar issue with it. Still waiting for feedback from testers.
    Thank you for finding that. Should I post this on github somewhere? Or in the IHO thread in this forum? I want to be sure the right people see this. Thanks again, that's the exact issue I'm experiencing.

    Edit: You must be talking about the IHO thread on xda. Ill check that out.
    12-30-2011 08:43 PM
  5. JerryScript's Avatar
    The issue is a cumulative one that gets worse over time till you reboot, and it does affect CM based ROMs.

    Most ROM devs reboot several times a day, so it goes out of sight out of mind. I can't tell you the last time I went more than a day without rebooting. You will have to forgive me, it is a known issue.
    notown likes this.
    12-31-2011 12:10 AM
  6. Schlidel's Avatar
    The issue is a cumulative one that gets worse over time till you reboot, and it does affect CM based ROMs.

    Most ROM devs reboot several times a day, so it goes out of sight out of mind. I can't tell you the last time I went more than a day without rebooting. You will have to forgive me, it is a known issue.
    You'll have to forgive me too for not knowing the hierarchy of the coding process. Where can I post this bug to make it official? Or where can I go to monitor the process already underway? Somewhere on github? I want to hit the right level so the fix gets filtered down properly. Thanks.
    12-31-2011 01:08 AM
  7. Schlidel's Avatar
    Btw, I'm using Harmonia_1.7 rom and the kernel from the fix black screen thread and the accelerometer seems to run great even after extended use. It's great.
    12-31-2011 01:31 AM
  8. Whyzor's Avatar
    You'll have to forgive me too for not knowing the hierarchy of the coding process. Where can I post this bug to make it official? Or where can I go to monitor the process already underway? Somewhere on github? I want to hit the right level so the fix gets filtered down properly. Thanks.
    I just commented on that issue on github it should be included in IHO. Any of the builders (including Jerry) can manually add the fix too.

    BTW as an update, one of the MT users reported it fixed his accelerometer problem with the linked I posted above.
    Schlidel and JerryScript like this.
    12-31-2011 02:19 AM
  9. BobZhome's Avatar
    I've had that fix in my build since October 30th: https://github.com/BobZhome/android_...0bec6e97b419f8
    Yet, Schlidel said he tested my rom and it didn't work...either the fix doesn't work or he never tested my rom!
    notown likes this.
    12-31-2011 07:46 AM
  10. Schlidel's Avatar
    About to download and test your rom again bobZhome.

    I think I may have been conflating two things into one.

    The first thing I've been testing after flashing rom is google sky. I didn't realize sometimes you have to calibrate the magnometer (OV does have one right?) with a figure-8 motion.

    I just figured it was the accelerometer and threw the baby out with the bath water.

    Anyway, about to test your rom again.
    12-31-2011 02:21 PM
  11. Schlidel's Avatar
    Nope, accelerometer problems still exist.

    I'll see if I can create steps to easily recreate the problem later today.

    I opened google sky and accelerometer seemed to be functioning, went back in after some things and I lost an axis. Opened paper airplane and no accelerometer at all.

    I'm back on my harmonia 1.7 nandroid where everything is cozy.

    Is it possible newer ov's with the hitachi screen also have other hardware changes?

    Anyway, later today I'll reflash bobZhome and create steps. I'll say one thing though this bug doesn't seemed to be linked to uptime but rather how much data the accelerometer has generated? Which obviously the longer the uptime the more data its created but if you use accelerometer intensive apps this problem seems to creep up fast.
    12-31-2011 03:31 PM
  12. JerryScript's Avatar
    I'm begining to think you may have a hardware issue. The bug never manifested itself that fast for anyone else, usually takes at least a couple of days. I do use both MyPaperPlane3d (addictive) and SkyMaps, I never had the bug show up. I'm pretty sure if it's occuring for you that fast, it is probably more than just a software issue.
    12-31-2011 06:36 PM
  13. JerryScript's Avatar
    OK, I was just about to apply the patch to BACKside, and I noticed I've already included it since the 1207 build:
    https://github.com/JerryScript/andro.../sensorservice

    Since I haven't been receiving reports about this bug for a while, and the commit fixes the issue on more than one device, I'm leaning more towards some issue other than the ROM. I may have already asked this, but did you try a complete wipe/format before flashing one of the IHO builds?
    12-31-2011 07:35 PM
  14. Schlidel's Avatar
    Yes, I do a complete wipe every time
    I probably even go overboard. Factory reset, cache, mounts, dalvik, the whole nine.

    I'm on bobZhome (reflashed it since this morning) as I type this and at this very moment the accelerometer works flawlessly on paper airplane yet my launcher, settings pages, opera don't.

    Ok, I just reopened paper airplane and its dead again. Argh. But a minute ago paper airplane was working while launcher, etc weren't. As soon as the accelerometer failed to work for my launcher, settings, opera etc I opened PA and PA worked fine. Close it out, wont work anymore.

    I really doubt this is a hardware problem, I don't experience any issues with harmonia 1.7 (froyo I believe). Plus reboot fixes it.

    My mother has an OV bought at the same time, I'll put bobZhome's latest on there and see what happens. I'm willing to bet the accelerometer will freeze on me.
    12-31-2011 10:08 PM
  15. Schlidel's Avatar
    Just rebooted and it's normal again. I'm still leaning towards software issue.
    12-31-2011 10:15 PM
  16. Schlidel's Avatar
    Settings, launcher, opera not working (orientation) at the same time paper airplane is working flawlessly. Opened it twice and it worked perfect both times. Opera which I'm using right now to type this, broken.
    12-31-2011 10:29 PM
  17. Schlidel's Avatar
    Ok, it's been running fine now as long as I don't open Google Sky.

    I googled google sky orientation freeze or something similar and read about people experiencing similar issues.

    Someone said it can be caused by other apps that use the compass like earth.

    Anyway sky is a cool app, but at least on iho it causes screen lock. That's my theory and my phone is fine now.
    JerryScript likes this.
    01-01-2012 09:18 PM
  18. galexand123's Avatar
    Well, sorry to say, I have to say I have similar symptoms to Schlidel. I have an LG Optimus V (LG-VM670), baseband VM670ZV5. I purchased it in October 2011, so it is a fairly recent edition. I used the stock rom up until a month ago, and google sky map worked reliably with the phone for 6 months with the stock rom (I used it more than once a week). Also, the accelerometer would reliably change the screen orientation properly.

    Recently I flashed with CyanogenMod-7.2.0-RC1-BACKside-IHO eng.jerryscript.20120502.032319. I installed it because I wanted to work with a bluetooth keyboard. Mostly trouble free, and definitely an upgrade over the stock rom (which is surprising, I thought the stock rom was pretty good, all things considered). Thanks, JerryScript et al!

    Anyways, since I flashed it, the accelerometer is flaky. I have two apps now that use it: PicaSim (which is awesome, btw), and Sky Map, and of course the system screen rotation. If I load either of those apps, then it seems to select at random 2 possibilities: (1) works perfectly, (2) locks one of the axes of the accelerometer (and mangles the other axis?). Once the accelerometer is messed up, it is generally unavailable for all purposes (i.e., often my screen will even be stuck in a funny orientation). But once I have had it spontaneously start working again, too. Force killing google sky map & picasim doesn't change anything.

    It seems to happen at random whether the phone has been on for a long time or not. I do not reboot very often, but I rebooted yesterday because of exactly this issue, and it is already not working again.

    Killing ami304d (which automatically respawns) also doesn't fix it. So I suspect it is a kernel problem, but maybe ami304d keeps some calibration data somewhere else? Anyways, when I get a chance I will try to prove whether it is a kernel problem or not.

    Sorry for the newbie question, but: is there kernel source for the specific kernel (with config and patches) that was used for this backside release? Like a git repository somewhere or directions for patching a stock 2.6.32.9 kernel? I'm probably too lazy, but I might even make a patch...since the trouble isn't reproducing for the developers, as they say: if it is to be, it is up to me.

    Thanks!
    - Greg
    06-19-2012 03:52 PM
  19. galexand123's Avatar
    Sorry to keep talking to myself, but I'm just going to post stuff as I discover it.

    It is definitely not a purely kernel problem. Killing system_server "fixes" it without rebooting the kernel.

    When it fails, this is in logcat:

    D/Stardroid.LocationController(12750): LocationController stop
    D/Stardroid.LocationController(12750): LocationController -stop
    D/Stardroid.SensorOrientationController(12750): Unregistering sensor listeners
    and then a series of:

    E/AmitSensor(12176): Sensor is closed. Can't set delay

    When it is working /sys/devices/platform/i2c-gpio.5/i2c-5/5-0019/x (y and z too) always have new values in them. But once "Sensor is closed" happens, they always contain the same stale data.

    So my theory is that this is a power management fail. When sky map (Stardroid) releases its handle on the orientation sensors, the KR3DH gets powered down (or something), and stops updating. Then when the sensor is needed again, it doesn't get properly "opened" again. Sky map pretends that one axis is still working wonkily because it is still getting enough information from the magnetometer to know that the phone is changing orientation, even though the accelerometer is AWOL.

    I'm going to put this down for the moment, I swear. Maybe this will ring a bell for someone and I won't get a chance to use baksmali and grep tomorrow?
    06-19-2012 04:47 PM
44 12
LINK TO POST COPIED TO CLIPBOARD