Kernel Panic? Maybe Xorg Failure?

woodape · 2016-07-05 22:59:59

Hi All. I've been running Arch on my Samsung Series 9 laptop for over a year now and occasionally it would crash, I would ignore it and a few weeks later it would crash again. Recently, say within the last month or two, the rte of panics has increased. Yesterday and today it's crashed about 4 times so now it seems serious. I believe its a kernel panic; X freezes up, whatever audio is playing starts to loop over about a second, and the caps lock indicator light flashes.

Things I've tried:

I ran memtest86 overnight last night with 10 passes and no reported errors, so I'm hoping I can reliably say its not a hardware error.
I've installed the intel microcode and updated grub.
I've changed my Xorg to use "uxa" instead of "sna" for Acceleration
I've turned off acceleration
I've tried different setups entirely: Cinnamon, Openbox + compton, bspwm + compton, bspwm no compositing
I've wiped the hard drive and reinstalled a very basic bspwm setup without the cruft I had accumulated over a year and a half of use

None of these things have stop the panicking.

This has been happening for a while so it isn't specific to my current kernel, but here is my info:

Kernel 4.6.3-1-ARCH
CPU Intel i5-3317U @1.70GHz

Can anyone help diagnose what is going on?

Last edited by woodape (2016-08-01 20:28:03)

jasonwryan · 2016-07-05 23:11:03

I have the same machine and haven't had any issues. Are you UEFI booting or using compatibility mode?

FWIW, I uninstalled the intel drivers some weeks ago and have just been using the kernel's native modesetting driver and it has worked flawlessly.

woodape · 2016-07-05 23:23:40

I'm using compatibility mode. Some years ago when I bought the machine I read a few articles about people bricking the machine because of UEFI, so I disabled it before installing and have never turned it back on.

I've been growing suspicious of the Intel drivers. I can try uninstalling them and see if I survive a few days without a panic.

I suppose I should add that more often then not the panics occur when I'm playing video or audio, though not uniquely.

EDIT:
Should I install any additional packages after removing xf86-video-intel?

Last edited by woodape (2016-07-05 23:25:52)

jasonwryan · 2016-07-05 23:43:49

No, just remove the intel package and any conf file you have in /etc/X11/xorg.conf.d/

Yeah, I read about the bricking thing when I installed (probably three years ago now), but just ploughed on anyway. Apart from some brutal panics when trying to install (in compatibility mode), the machine has run like a dream since.

woodape · 2016-07-21 14:10:02

UPDATE:

It's been 2 weeks since I removed the xf86-video-intel drivers and I haven't had a kernel panic since.

The first day I opened 4 videos in different vlc players and had them on loop and chromium open to a streaming video site for about 6 hours without issue. Since then I've been on vacation so I've mainly been watching a few videos here and there and using chromium to do normal browsing.

So far so good, I'll update again in another two weeks, but I think the issue has been solved. Thanks for the help jasonwryan.

woodape · 2016-08-01 20:30:00

UPDATE 2:

After a while emulating a SNES, I got the kernel panic again, even with the Intel drivers uninstalled. I'm back at the drawing board trying to figure out what's up. For now this is back to unsolved.

woodape · 2016-08-02 07:51:59

I enabled "mcelog.service", the kernel panicked some time later and now in my journalctl I have this:

Aug 02 09:03:15 potato mcelog[287]: Hardware event. This is not a software error.
Aug 02 09:03:15 potato mcelog[287]: MCE 0
Aug 02 09:03:15 potato mcelog[287]: CPU 0 BANK 4
Aug 02 09:03:15 potato mcelog[287]: TIME 1470121391 Tue Aug  2 09:03:11 2016
Aug 02 09:03:15 potato mcelog[287]: MCG status:
Aug 02 09:03:15 potato mcelog[287]: MCi status:
Aug 02 09:03:15 potato mcelog[287]: Uncorrected error
Aug 02 09:03:15 potato mcelog[287]: Error enabled
Aug 02 09:03:15 potato mcelog[287]: Processor context corrupt
Aug 02 09:03:15 potato mcelog[287]: MCA: Internal unclassified error: 402
Aug 02 09:03:15 potato mcelog[287]: Running trigger `unknown-error-trigger'
Aug 02 09:03:15 potato mcelog[287]: STATUS b200000000100402 MCGSTATUS 0
Aug 02 09:03:15 potato mcelog[287]: MCGCAP c07 APICID 0 SOCKETID 0
Aug 02 09:03:15 potato mcelog[287]: CPUID Vendor Intel Family 6 Model 58
Aug 02 09:03:15 potato mcelog[292]: CPU 0 on socket 0 received unknown error
Aug 02 09:03:15 potato mcelog[287]: <27>Aug  2 09:03:15 mcelog: CPU 0 on socket 0 received unknown error
Aug 02 09:03:15 potato mcelog[293]: Location: CPU 0 on socket 0
Aug 02 09:03:15 potato mcelog[287]: <27>Aug  2 09:03:15 mcelog: Location: CPU 0 on socket 0

Anybody have any clues on what I can do about this?

Arch Linux

#1 2016-07-05 22:59:59

Kernel Panic? Maybe Xorg Failure?

#2 2016-07-05 23:11:03

Re: Kernel Panic? Maybe Xorg Failure?

#3 2016-07-05 23:23:40

Re: Kernel Panic? Maybe Xorg Failure?

#4 2016-07-05 23:43:49

Re: Kernel Panic? Maybe Xorg Failure?

#5 2016-07-21 14:10:02

Re: Kernel Panic? Maybe Xorg Failure?

#6 2016-08-01 20:30:00

Re: Kernel Panic? Maybe Xorg Failure?

#7 2016-08-02 07:51:59

Re: Kernel Panic? Maybe Xorg Failure?

Board footer