FS#41556 - [linux] Kernel 3.16 Kernel Panics & System Lockup/Corruption

Attached to Project: Arch Linux
Opened by Michael Murphy (mmstick) - Wednesday, 13 August 2014, 23:09 GMT
Last edited by Doug Newgard (Scimmia) - Thursday, 15 October 2015, 01:34 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Thomas Bächler (brain0)
Architecture x86_64
Severity Critical
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 6
Private No

Details

Problem:

Linux 3.16 has been causing problems with my system, and now that Arch is including it in stable repos, I'm currently forced to run linux-lts. Shortly after booting, a kernel panic occurs and programs begin locking up due to no longer being able to perform read/write tasks -- which also means I cannot get a text log of the problem. At one point it was actually so bad that it corrupted my root partition and I had to redo my system.

System Specs:

ASUS Sabertooth 990FX
FX-8120 & 16GB DDR3
240GB Seagate SSD
6TB HDD + 3TB HDD
Radeon HD 7950

Partitioning:

/ is installed on a 20 GiB btrfs partition on the SSD, and /home is installed on a btrfs partiton across the entire 6TB drive.
This task depends upon

Closed by  Doug Newgard (Scimmia)
Thursday, 15 October 2015, 01:34 GMT
Reason for closing:  Fixed
Additional comments about closing:  User requested: Bug no longer occurs in more recent kernels.
Comment by jb (jb.1234abcd) - Thursday, 14 August 2014, 04:16 GMT
Removed.
Comment by AndrzejL (AndrzejL) - Thursday, 14 August 2014, 04:31 GMT
I am not sure if its related but my pentium 2 machine thinkpad 600E is freezing right after kernel is suppose to be loaded. It says loading and hangs indefinitely - I was forced to downgrade to 3.15.8 which has fixed the issue.

Cheers.

Andrzej
Comment by AndrzejL (AndrzejL) - Thursday, 14 August 2014, 04:32 GMT
I am sure that You can deduce this from the "pentium 2" bit but I will add it anyway - its a 32 bit install...

Cheers.

Andrzej
Comment by Leonidas Spyropoulos (inglor) - Friday, 15 August 2014, 07:27 GMT
I got very similar hardware and I am able to use the 3.16 without issues.
Asus Crosshair V
AMD 8120 FX
SSD Intel 120
Nvidia 670

In BIOS can you make sure the IOACPI is "Enabled" ? I got Kernel panic when it is "Disabled" - could be similar.
Comment by Alim Gokkaya (alimg) - Saturday, 16 August 2014, 06:29 GMT
3.16 crashes every time when I run a virtualbox vm. I've got the attached journal log using command `journalctl -n 20000 _TRANSPORT=kernel >journal.log`. journal_full file contains several crashes.
   journal.log (14.8 KiB)
   journal_full (1.61 MiB)
Comment by patrick (potomac) - Monday, 18 August 2014, 12:08 GMT
@Alim Gokkaya : maybe it's the same bug like  FS#41597 

can you confirm ?

since kernel 3.16.x I get a kernel panic in host after running virtualbox ( 5~15 minutes after launching virtualbox, a random time ),

your journal.log about the kernel crash seems similar to mine
Comment by Alim Gokkaya (alimg) - Monday, 18 August 2014, 16:59 GMT
I can confirm that the crash occurs at a random time as you mention but as far as I see, stack trace is different each time it occurs. //which seem to me like a concurrency bug in somewhere//
Comment by patrick (potomac) - Monday, 18 August 2014, 20:39 GMT
@Alim Gokkaya

Can you tell me if your virtual machine is a "linux guest", or a "windows guest" ?

because it seems that the bug occurs only on "linux guest",

my virtual machines are only "linux OSes" ( archlinux, debian ) and all can trigger the bug

Comment by Alim Gokkaya (alimg) - Monday, 18 August 2014, 21:15 GMT
I encounter this when running archlinux guest with kde running. I realized that it usually crashes right after desktop is loaded. I've run a centos guest with no gui and a win7 guest for a while but there wasn't any problem that I see.
Comment by Michael Murphy (mmstick) - Monday, 18 August 2014, 21:29 GMT
People, please file bug reports instead of commenting on this one. These comments are all completely unrelated to the original bug report that I filed.

@inglor: I don't have an IOAPCI option in the UEFI BIOS for the Sabertooth 990FX. This motherboard's firmware has always been rather fickle with Linux. In example, If I disable IOMMU, I get a lot of USB read/write errors.
Comment by Mike Cloaked (mcloaked) - Wednesday, 20 August 2014, 09:22 GMT
I have only just seen this report, and I have had some major issues with kernel oops on both bootup and shutdown on my Lenovo Y510p laptop. I reported upstream to the kernel bug list at:

https://bugzilla.kernel.org/show_bug.cgi?id=82711

I don't know if my bug above is related to this one, and I am not using btrfs, but it is possibly not as a result of the filesystem but another kernel bug since it also occurs randomly (though very frequently) and suggests it is a quite serious bug in the new kernel. However by persevering I was able to get some logs for diagnostics which are in the report I referenced above.
Comment by Mike Cloaked (mcloaked) - Saturday, 04 October 2014, 14:59 GMT
I have just updated my Y510p laptop and now every time I shutdown I get a delayed shutdown with errors in the journal log for nouveau and xorg related messages as in the attached log.

My system has
[2014-10-04 14:52] [PACMAN] upgraded xf86-video-nouveau (1.0.11-1 -> 1.0.11-2)
[2014-10-04 14:52] [PACMAN] upgraded xf86-video-intel (2.99.916-1 -> 2.99.916-3)
[2014-10-04 14:52] [PACMAN] upgraded xorg-server-common (1.16.0-6 -> 1.16.1-1)
[2014-10-04 14:52] [PACMAN] upgraded xorg-server (1.16.0-6 -> 1.16.1-1)

and running kernel 3.16.3-1 x86_64.

The graphics hardware is:
00:02.0 VGA compatible controller: Intel Corporation 4th Gen Core Processor Integrated Graphics Controller (rev 06)
01:00.0 VGA compatible controller: NVIDIA Corporation GK107M [GeForce GT 750M] (rev a1)
   nouveau-shutdown-journal.log (260.5 KiB)
Comment by Michael Murphy (mmstick) - Thursday, 15 October 2015, 01:19 GMT
I've noticed that this bug is still on the list but I haven't had this problem in a very, very long time with more recent kernels. It should be marked as resolved.