Re: [TESTING] Updated 2/2/11 New Kernel (system_server Debug Help Needed!)
 

No problems, waiting patiently for it!

Re: [TESTING] Updated 2/2/11 New Kernel
 

I have a SoD SYSTEM_LAST_KMESG for you. However, it's real weird looking. Not sure how, but it seems to be corrupted. I recopied the file several times, and each time it looks like this. I copied the KMESG before that, and everything is crystal clear. Any ideas on what could cause this?

Anyway, this happened while plugged in. Light was orange. No response whatsoever from phone to power button. No response or lights on opening keyboard. Pushed all the face buttons to no response. No apparent backlight.

Re: [TESTING] Updated 2/2/11 New Kernel
 

I don't know that we have a 100% complete explanation, but a theory was thrown out that SYSTEM_LAST_KMSG is written from a particular memory address after the boot following the issue, the memory is still there from before, when the issue occurred. WinMo may be using some or all of this memory, so by the time you get back into Android a portion of that memory has been overwritten with what looks like garbage. Call it an educated guess.

Re: [TESTING] Updated 2/2/11 New Kernel
 

Got another Sleep of Death tonight. Put it in my pocket with the light green, and battery around 80%. Pulled it out of my pocket 3-4 hours later. Phone was hot, light was orange, backlight was off. No response whatsoever to pressing buttons, or opening keyboard. Pulled out battery and restarted into WinMo, where it informed me battery was less than 10% remaining.

Log is pretty corrupted again. Anyone get a Sleep of Death where it's not corrupted?

Re: [TESTING] Updated 2/2/11 New Kernel
 

Wow, that log is really messed up. I've had some messed up logs, but that is pretty much useless... Looks like you were on the newest kernel too.

Very odd, I haven't had any SoD's. Are you still mounting the libc? Seems the /proc/last_kmsg might be more useful in certain cases - although when I've had kernel panics (reboots to WinMo) I usually get flawless logging output in that dropbox folder. I dunno why, but WisTilt2 seems to think he has SoD licked. I'm wondering if this kernel that I have posted is the newest that he has - I've been testing a lot of random crap for him, but nothing that he's deemed for release. I haven't had any SoD's yet, just a few kernel panics (but I'm also running a funky system image... more on that later hopefully.)

Re: [TESTING] Updated 2/2/11 New Kernel
 

Nope, should I be mounting that still? I stopped mounting it after the first post was updated with the info on the dropbox logging method. Is the libc way more resistant to corruption?

Re: [TESTING] Updated 2/2/11 New Kernel
 

I don't know if it is more resistant to corruption, it's just another place to potentially get logs - and with SoD we need anything we can get.

I guess I should add them back so both are available - in theory what is placed in dropbox will be exactly the same as the last_kmsg in /proc, but I'm not positive if that's how it works in practice - everyone seems a little fuzzy on how this is actually achieved.

Re: [TESTING] Updated 2/2/11 New Kernel
 

Does Wistilt still want logs? I thought by your earlier post that he was set on that, so I've been waiting for another kernel to drop and set up a clean FRX04 to test it on (as long as the freaking weather here cooperates a bit...). Do you know what kernel Wistilts test one is based on, i.e. 1253 that has usb headphones, etc? I've been having reboots and SOD's lately on just about every build I'm running, but haven't tried the test kernel with FRX04.

Re: [TESTING] Updated 2/2/11 New Kernel
 

No more logs necessary on the system_server bug, but SoD logs and kernel panic logs (reboots to WinMo basically - unplanned soft reset) are always welcome!

Also, his kernel is based on 1253.

Re: [TESTING] Updated 2/2/11 New Kernel
 

Excellent. As long as wifee leaves me alone for a bit tonight I'll set this beotch up and see what happens!