PPCGeeks

PPCGeeks (http://forum.ppcgeeks.com/index.php)
-   Android On TP2 Development (http://forum.ppcgeeks.com/forumdisplay.php?f=319)
-   -   [TESTING] - NAND OMGB (1.2.3) (http://forum.ppcgeeks.com/showthread.php?t=147260)

arrrghhh 12-13-2011 11:05 AM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by mbellot (Post 2160038)
Nope. Spontaneous reboot within minutes. Logs attached again.

Time to wipe data or ???

Hrm. I would say yes, time to pillage - I would wipe everything. But perhaps there's some other interim troubleshooting step to try first that I'm not aware of... I would pull as many logs as you can, and pillage myself.

Just curious - what state is the phone in? Sleeping, running an app? You say spontaneous, so I assume there's nothing in common with them, and you're not doing anything with the phone when it happens...?

[ACL] 12-13-2011 11:50 AM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by arrrghhh (Post 2160071)
Hrm. I would say yes, time to pillage - I would wipe everything. But perhaps there's some other interim troubleshooting step to try first that I'm not aware of... I would pull as many logs as you can, and pillage myself.

Just curious - what state is the phone in? Sleeping, running an app? You say spontaneous, so I assume there's nothing in common with them, and you're not doing anything with the phone when it happens...?

odd, after the kenrel upgrade there is no more mdp crashes but there isnt much either.

the last kmsg isnt showing much.

mbellot 12-13-2011 12:41 PM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by arrrghhh (Post 2160071)
Hrm. I would say yes, time to pillage - I would wipe everything. But perhaps there's some other interim troubleshooting step to try first that I'm not aware of... I would pull as many logs as you can, and pillage myself.

Just curious - what state is the phone in? Sleeping, running an app? You say spontaneous, so I assume there's nothing in common with them, and you're not doing anything with the phone when it happens...?

The phone is probably "doing things" (email sync, contact sync, etc). I was not actively using it, but I can't say for sure if it was asleep or in a momentary wake doing some sync operation.

The freezes (mdp and flinger) are definitely awake and "active use" crashes. Possibly worth noting is that (IIRC) the two times it has happened it was returning to the home screen.

Quote:

Originally Posted by [ACL] (Post 2160073)
odd, after the kenrel upgrade there is no more mdp crashes but there isnt much either.

the last kmsg isnt showing much.

Not surprising, I rebooted into recovery, updated the kernel, went back into OMGB and in a minute or two "blamo", it rebooted. I pulled that last_kmsg midway through the boot, basically as soon as adb recognized the phone, hoping it would still hold info about what caused the reboot.


Further odd developments... It went through the night charging without a reboot and has run fine since the last spontaneous reboot over 10 hours ago - without wiping the data partition.

I know signal levels at my home are in the toilet, is it possible that it's the radio/RIL that is causing these weird reboots (unrelated to the mdp/flinger problem)? I am running the VZW 2.37 radio in my RHOD400, I updated to it shortly before moving to OMGB.

I'm going to keep using it "as is" (no data wipe) to see how it holds together and to see if I can generate any more logs. If it gets unusable I'll do a wipe.

[ACL] 12-13-2011 01:48 PM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by mbellot (Post 2160077)
The phone is probably "doing things" (email sync, contact sync, etc). I was not actively using it, but I can't say for sure if it was asleep or in a momentary wake doing some sync operation.

The freezes (mdp and flinger) are definitely awake and "active use" crashes. Possibly worth noting is that (IIRC) the two times it has happened it was returning to the home screen.



Not surprising, I rebooted into recovery, updated the kernel, went back into OMGB and in a minute or two "blamo", it rebooted. I pulled that last_kmsg midway through the boot, basically as soon as adb recognized the phone, hoping it would still hold info about what caused the reboot.


Further odd developments... It went through the night charging without a reboot and has run fine since the last spontaneous reboot over 10 hours ago - without wiping the data partition.

I know signal levels at my home are in the toilet, is it possible that it's the radio/RIL that is causing these weird reboots (unrelated to the mdp/flinger problem)? I am running the VZW 2.37 radio in my RHOD400, I updated to it shortly before moving to OMGB.

I'm going to keep using it "as is" (no data wipe) to see how it holds together and to see if I can generate any more logs. If it gets unusable I'll do a wipe.

sorry to use you as a guinea pig. Anyway the last set of logs you posted really have nothing. The last kmsg is clean and the radio looks clean. The mdp issue should lock up your phone, not reboot it. Its a kernel panic so thats why i want to squich that bug. THe reboots however we will just need to keep collecting logs and see who else suffers from this.

you could always wipe "cache" and not data. The odd thing about cache is that its on a nand location we really dont know much about. i know it can be over written by the winmo boot loader. We have nothing that really writes here, but that wont stop some apps from writing on there. Also have a backup handy incase you need to rewipe. THanks good.

mbellot 12-13-2011 03:52 PM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by [ACL] (Post 2160086)
sorry to use you as a guinea pig.

No need to apologize, that's why I'm doing this. :D

Quote:

Originally Posted by [ACL] (Post 2160086)
Anyway the last set of logs you posted really have nothing. The last kmsg is clean and the radio looks clean. The mdp issue should lock up your phone, not reboot it. Its a kernel panic so thats why i want to squich that bug. THe reboots however we will just need to keep collecting logs and see who else suffers from this.

The mdp lockup happened before I switched to the new test kernel and has not reared it's head since, but it only happens rarely anyway. It's just really weird the way I had so many reboots almost one right after another and now it's fine with no intervention.

EDIT: Is there any way to get the "old" logs (pre-reboot) before they get over-written by the new ones from the most recent boot?

Quote:

Originally Posted by [ACL] (Post 2160086)
you could always wipe "cache" and not data. The odd thing about cache is that its on a nand location we really dont know much about. i know it can be over written by the winmo boot loader. We have nothing that really writes here, but that wont stop some apps from writing on there. Also have a backup handy incase you need to rewipe. THanks good.

I'll take a look at wiping cache. I know I can do it per application with the settings, so I might try that first.

I'm not too worried about having a backup, rebuilding the phone takes less than an hour, but I agree it's handy.

[ACL] 12-13-2011 05:20 PM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by mbellot (Post 2160102)
EDIT: Is there any way to get the "old" logs (pre-reboot) before they get over-written by the new ones from the most recent boot?

thats what lastkmsg is. THe last kernel log. I dont think there is a way to get a last logcat tho. for that we just need to debug the tombstones directory and see if there is anything there.

mbellot 12-14-2011 12:51 AM

Re: [TESTING] - NAND OMGB (1.2.3)
 
1 Attachment(s)
I had what appeared to be a SOD (green LED, no response to power button).

These are the logs that I pulled after the phone rebooted.

Oddly enough there was no /proc/last_kmsg file after the reboot.

Also no /data/tombstone files - just checked.

[ACL] 12-14-2011 01:12 AM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Quote:

Originally Posted by mbellot (Post 2160186)
I had what appeared to be a SOD (green LED, no response to power button).

These are the logs that I pulled after the phone rebooted.

Oddly enough there was no /proc/last_kmsg file after the reboot.

Also no /data/tombstone files - just checked.

everything looks kosher. Not sure what to do lol.

the /proc/last_kmsg will only survive a reset i think. A battery pull will wipe everything. If the LED was green the the device was asleep with it went down. Silent killer. lol.

i havent been able to replicate the mdp crash myself after a full day of messing with the camera, so i'll leave that patch for the next release. I'm just going to wait back for more reports.:)

mbellot 12-14-2011 01:23 AM

Re: [TESTING] - NAND OMGB (1.2.3)
 
No problem. I'll keep pulling what I can when the phone acts up in the hopes that something can be found.

I did a cache wipe from recovery this evening (after the SOD) to see if it helps at all.

I haven't had a waking freeze since upgrading to the test kernel, but it's only been about 24 hours and that seems to happen much less frequently.

Good to know about battery pull killing /proc/last_kmsg - I'll avoid doing that unless absolutely necessary.

And to be quite honest, even with the reboots and SODs OMGB still beats the snot out of WinMo... Keep up the awesome work.

starling100 12-14-2011 10:22 AM

Re: [TESTING] - NAND OMGB (1.2.3)
 
Hi ACL

I appreciate for your hard working on Touch pro2.

My phone is Rhod100.

Thanks for your newest Rhod100 Experimental Kernel part 3, it fixed the previous Rhod100 kernel problem that the Panel wont power back up after sleep.
It's working well.

Problem:
Microphone is not working.
Now the main problem of Rhod100 is that when you call somebody you can only hear his/her voice but they can't hear yours. If i use the original headset, it's working.


Hope we can use it soon.

I will test your new development on Rhod100 and give feedback on time.

Any help if I can provide please let me know.


All times are GMT -4. The time now is 10:09 AM.

Powered by vBulletin® ©2000 - 2025, Jelsoft Enterprises Ltd.
©2012 - PPCGeeks.com


Content Relevant URLs by vBSEO 3.6.0