Hard lockup after upgrade to 3.2.12-1

cfr · 2012-03-22 02:29:44

I upgraded this evening (pacman -Syu) and everything seemed to go smoothly. The new kernel rebooted fine and all seemed well.

However, shortly afterwards, the system locked up. I thought it might be related to https://bbs.archlinux.org/viewtopic.php?id=137330 but reading that thread carefully, I'm not sure it is the same problem.

Essentially, the machine became apparently entirely unresponsive. None of ctrl+alt+f1, ctrl+alt+f2, ctrl+alt+backspace had any effect. The only way to escape was to hold down the power, forcing reboot. I rebooted to the recovery prompt and examined the logs. Nothing seemed to have been logged at all.

I don't have any leds on my keyboard so there is no way to know if, for example, caps lock is responding or not. Before I forced the reboot, I pressed the key combo for sleep as I wondered if that might be caught and register in the logs. (I didn't think it would sleep and it didn't.) That did register in the logs and the logs show a message saying the system is readying for sleep. The last three lines from kernel.log:

Mar 21 22:54:46 localhost kernel: [  138.398703] kfilemetadatare[4003]: segfault at 10 ip 00007f318d225141 sp 00007fffa12a5e60 error 4 in li
bkdecore.so.5.8.0[7f318d0b1000+2c9000]
Mar 22 01:51:19 localhost kernel: [10729.750818] PM: Syncing filesystems ... done.
Mar 22 01:51:19 localhost kernel: [10729.759083] PM: Preparing system for mem sleep

On reboot, KDE Activities Manager crashed immediately. This made me realise that this had stopped for a while. It was happening on every boot but this is the first time in a bit. Definitely the indexing etc. is playing up again but that hasn't caused anything this serious before.

Memtest was run a couple of weeks ago on the current memory after a previous module went bad.

toad · 2012-03-25 17:55:14

Sort of the same but system locks after miraculously after 15, sometimes 20 minutes.

I was checking top and iotop but there was nothing worth mentioning. Here the last entries from my kernel.log:

Mar 25 20:31:59 localhost kernel: [ 1831.487631] sr 0:0:0:0: [sr0]  ASC=0x11 ASCQ=0x0
Mar 25 20:31:59 localhost kernel: [ 1831.487634] sr 0:0:0:0: [sr0] CDB: cdb[0]=0x28: 28 00 00 08 6a e1 00 00 01 00
Mar 25 20:31:59 localhost kernel: [ 1831.487639] end_request: I/O error, dev sr0, sector 2206596
Mar 25 20:31:59 localhost kernel: [ 1831.487643] Buffer I/O error on device sr0, logical block 551649
Mar 25 20:32:02 localhost kernel: [ 1834.533851] sr 0:0:0:0: [sr0] Unhandled sense code
Mar 25 20:32:02 localhost kernel: [ 1834.533854] sr 0:0:0:0: [sr0]  Result: hostbyte=0x00 driverbyte=0x08
Mar 25 20:32:02 localhost kernel: [ 1834.533856] sr 0:0:0:0: [sr0]  Sense Key : 0x3 [current] 
Mar 25 20:32:02 localhost kernel: [ 1834.533859] Info fld=0x373a70
Mar 25 20:32:02 localhost kernel: [ 1834.533860] sr 0:0:0:0: [sr0]  ASC=0x11 ASCQ=0x0
Mar 25 20:32:02 localhost kernel: [ 1834.533863] sr 0:0:0:0: [sr0] CDB: cdb[0]=0x28: 28 00 00 37 3a 70 00 00 01 00
Mar 25 20:32:02 localhost kernel: [ 1834.533868] end_request: I/O error, dev sr0, sector 14477760
Mar 25 20:32:02 localhost kernel: [ 1834.533872] Buffer I/O error on device sr0, logical block 3619440
Mar 25 20:32:04 localhost kernel: [ 1836.613626] sr 0:0:0:0: [sr0] Unhandled sense code
Mar 25 20:32:04 localhost kernel: [ 1836.613629] sr 0:0:0:0: [sr0]  Result: hostbyte=0x00 driverbyte=0x08
Mar 25 20:32:04 localhost kernel: [ 1836.613631] sr 0:0:0:0: [sr0]  Sense Key : 0x3 [current] 
Mar 25 20:32:04 localhost kernel: [ 1836.613634] Info fld=0x373a71
Mar 25 20:32:04 localhost kernel: [ 1836.613635] sr 0:0:0:0: [sr0]  ASC=0x11 ASCQ=0x0
Mar 25 20:32:04 localhost kernel: [ 1836.613637] sr 0:0:0:0: [sr0] CDB: cdb[0]=0x28: 28 00 00 37 3a 71 00 00 01 00
Mar 25 20:32:04 localhost kernel: [ 1836.613643] end_request: I/O error, dev sr0, sector 14477764
Mar 25 20:32:04 localhost kernel: [ 1836.613647] Buffer I/O error on device sr0, logical block 3619441

I thought sr0 is the DVD drive which was not in use at the time so slightly confused here. What does your kernel log say?

Last edited by toad (2012-03-25 18:07:06)

akspecs · 2012-03-26 08:42:34

Hey man, at least it's not this. This hit me and evidently a few others:

https://bbs.archlinux.org/viewtopic.php?id=138105

Long story short, somethings wrong with this new kernel

Lesson learned in the end, backup stable working kernels to use as fallback.

toad · 2012-03-26 09:03:58

Just booted 3.2.13-1 so hope that will bring you relief, akspecs.

Anyway, I've got a 2TB NTFS disk for the home network hanging on this box and it is riddled with errors. Since I've commented it out in the fstab things have run smoothly under the old kernel.

cfr · 2012-03-26 22:37:32

It is very odd. I ran for quite a long time today with just restarting kdm once. Admittedly, I wouldn't usually call this a success but it is all relative.

When I got home, I found my laptop had unsuspended itself and was boiling itself to bits. I cooled it down and then tried to get a response. Just got a blinking white underscore in top left. Forced to hard reboot. Rebooted and very quickly got another freeze/lockup. [Note: I checked the temps at this point and they were a bit higher than usual but nowhere near critical - 56ish rather than 45ish with 100 being critical. So I think it had cooled enough for it not to be that. Plus, it behaved in just the same way as the other freezes.]

Have not rebooted using lts kernel and am hoping this will work better.

For what it is worth, I don't even *have* an optical drive...

Something is definitely wrong - whether the kernel or something else. (I guess the kernel or it should presumably kill off anything behaving this badly?)

Whatever it is, it gets on especially badly with Firefox.

When I rebooted into the lts kernel, my machine was snail like. It seems to be improving a bit now, though. I've noticed it has been very slow since the kernel update (but I'm also running with less RAM than usual so it is hard to be sure I'm being fair). Don't know why that would affect reboot into the lts kernel, though.

Does anybody know what the new kernel is like? I noticed the upgrade but I'm too scared to try it!

Last edited by cfr (2012-03-26 22:39:54)

cfr · 2012-03-28 00:51:20

Fwiw: the lts kernel is so far working out much better. Not only am I not getting the hard lockups, I'm not seeing the graphics corruption either and the system is more responsive. Whatever is in the new species of kernel, it doesn't agree with my machine!

akspecs · 2012-03-28 01:25:55

cfr wrote:

Fwiw: the lts kernel is so far working out much better. Not only am I not getting the hard lockups, I'm not seeing the graphics corruption either and the system is more responsive. Whatever is in the new species of kernel, it doesn't agree with my machine!

Long story short: stay on 3.2.11

linux : http://schlunix.org/archlinux/core/os/x … pkg.tar.xz
linux-headers : http://schlunix.org/archlinux/core/os/x … pkg.tar.xz

Last edited by akspecs (2012-03-28 01:28:37)

cfr · 2012-03-29 00:48:09

akspecs wrote:

Hey man, at least it's not this. This hit me and evidently a few others:
https://bbs.archlinux.org/viewtopic.php?id=138105

That could be because I've already got pcie_aspm=force in my kernel line. Although I booted to a recovery shell earlier, which doesn't have the option set, and didn't get a panic. My pci bridge is very like the one listed in that thread - just mine is one revision earlier.

Arch Linux

#1 2012-03-22 02:29:44

Hard lockup after upgrade to 3.2.12-1

#2 2012-03-25 17:55:14

Re: Hard lockup after upgrade to 3.2.12-1

#3 2012-03-26 08:42:34

Re: Hard lockup after upgrade to 3.2.12-1

#4 2012-03-26 09:03:58

Re: Hard lockup after upgrade to 3.2.12-1

#5 2012-03-26 22:37:32

Re: Hard lockup after upgrade to 3.2.12-1

#6 2012-03-28 00:51:20

Re: Hard lockup after upgrade to 3.2.12-1

#7 2012-03-28 01:25:55

Re: Hard lockup after upgrade to 3.2.12-1

#8 2012-03-29 00:48:09

Re: Hard lockup after upgrade to 3.2.12-1

Board footer