Just had a strange one. I got out of the shower and my battery was pretty low but not in the red zone, as I hadn't plugged it in all day. I woke the phone to see if I had any messages, and when I didn't, I put the phone to sleep again. Then a few seconds later it rebooted to the Google screen and lingered there for far too long (at least 45 seconds, maybe more?). I got scared and pulled the battery. Decided that if any recovery would have to happen, I'd put my spare charged battery in so I don't have to worry about charge.

Started up the phone again, and it lingered on the Google screen longer than usual but did successfully boot.

I tried going in with adb to a root shell and typing cat /proc/last_kmsg > /sdcard/last_kmsg.txt
but I just got: sh: cat: /proc/last_kmsg: No such file or directory

So I guess there is no last kmsg.

Originally Posted by cmstlist

Just had a strange one. I got out of the shower and my battery was pretty low but not in the red zone, as I hadn't plugged it in all day. I woke the phone to see if I had any messages, and when I didn't, I put the phone to sleep again. Then a few seconds later it rebooted to the Google screen and lingered there for far too long (at least 45 seconds, maybe more?). I got scared and pulled the battery. Decided that if any recovery would have to happen, I'd put my spare charged battery in so I don't have to worry about charge.

Started up the phone again, and it lingered on the Google screen longer than usual but did successfully boot.

I tried going in with adb to a root shell and typing cat /proc/last_kmsg > /sdcard/last_kmsg.txt
but I just got: sh: cat: /proc/last_kmsg: No such file or directory

So I guess there is no last kmsg.

every so often a fsck will freak people out
you can't have a last_kmsg of a battery pull
check your dmesg for signs of the fsck in the current kernel messages

EDIT: That is a prime example of what to share in Trinity thread and get info there from many

Originally Posted by morfic

every so often a fsck will freak people out
you can't have a last_kmsg of a battery pull
check your dmesg for signs of the fsck in the current kernel messages

EDIT: That is a prime example of what to share in Trinity thread and get info there from many

Ok but shouldn't there have been a last_kmsg from the random reboot?

I had a random reboot today that I didn't even notice when it happened. Pulled the phone out of my pocket in a grocery store and noticed the media scanner was running, which usually only happens after a music sync or a reboot. Checked battery info and found that the time since boot was only at 6 minutes. Retrieved the last_kmsg immediately and saved it for later.

Here is a pastebin of the last 30% or so of the 1.5 MB last_kmsg:
http://pastebin.com/12YD5Tab

In particular here is the last section which seems to be where it starts to go astray:

[18797.706176] suspend: exit suspend, ret = 0 (2012-09-27 00:40:12.520211054 UTC)
[18798.650695] PM: Syncing filesystems ... 
[18798.655670] EXT4-fs error (device mmcblk0p12): ext4_read_inode_bitmap:117: comm Sidekick_AsyncF: Cannot read inode bitmap - block_group = 60, inode_bitmap = 1966081
[18798.701019] Aborting journal on device mmcblk0p12-8.
[18798.760925] Kernel panic - not syncing: EXT4-fs (device mmcblk0p12): panic forced after error
[18798.760955] 
[18798.761535] Backtrace: invalid frame pointer 0x00008180
[18798.762176] CPU1: stopping
[18798.762359] Backtrace: 
[18798.762969] [<c784df74>] (0xc784df74) from [<c0055408>] (__irq_svc+0x48/0xe0)
[18798.763153] Backtrace aborted due to bad frame pointer <c784df90>
[18798.763549] Rebooting in 5 seconds..
[18800.040130] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18800.551147] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18801.061248] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18801.573181] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18802.083190] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18802.594055] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18803.104248] [MIPI-HSI] wrong state : 0000000e, recv_step : 12, size : 13
[18803.111694] 
[18803.111907] Restarting Linux version 3.0.38-04149-g1d1c555-dirty (morfic@othris) (gcc version 4.7.1 20120531 (prerelease) (Linaro GCC 4.7-2012.06) ) #666 SMP PREEMPT Mon Jul 23 00:23:38 CDT 2012
[18803.111938] 

No errors detected
Boot info:
Last reset was warm software reset (PRM_RSTST=0x2)

So... uh... my filesystem is corrupted?

Bump. I guess this is why I was told to check the main thread

Sent from my Galaxy Nexus using Tapatalk 2

Originally Posted by cmstlist

Bump. I guess this is why I was told to check the main thread

Sent from my Galaxy Nexus using Tapatalk 2

exactly

btw, have you tried the newest yet? alpha63 is very nice

Still on the stable release. Don't really want to try anything that'll destabilize it.

Originally Posted by cmstlist

Still on the stable release. Don't really want to try anything that'll destabilize it.

many of the alphas are as stable or more stable then the release kernels. a63 is one of them. plus it includes any of the latest fixes/revisions in the source code. if you look inn the thread, you will see many good opinions on a63. plus, at the very worst, you can reflash 35 if you dont like 63.

I'm also getting some random reboots that are soft and don't generate a last kmsg. I can identify them by a gap in the battery graph:



Any thoughts on debugging those?

Based on a screenshot time of 1:32 PM, the random reboot happened around 11:59 AM when I wasn't noticing. Moreover, the total time of 3h28m brings us back to 9:59 AM which is a full 80 minutes after I unplugged from my car charger when I got to work... so that suggests I also had a hard reboot before 10AM. But if that generated a last_kmsg it must have been obliterated during the soft reboot.

I see that 63 is now labelled STABLE 63 so I guess I should give it a try...

I'm using basically default settings except for interactive/sio.

Today I had a random reboot again, and the last_kmsg is basically a massive string of lines that look like this:

[46116.176422] 4c60: ???????? ???????? ???????? ???????? ???????? ???????? ???????? ????????
[many many many many lines omitted]
[46118.078643] ee80: ???????? ???????? ???????? ???????? ???????? ???????? ???????? ????????

Followed by:

4 Corrected bytes, 1 unrecoverable blocks
Boot info:
Last reset was MPU Watchdog Timer reset (PRM_RSTST=0x8)

I noticed one of the Smart Reflex options is MPU. If I am not choosing any overclock options, only the CPU governor and I/O scheduler, is there a reason I should or shouldn't be disabling the Smart Reflex the way I would if I were over/under clocking/volting?