random and persistant hard kernel lock, 10.10 (old and newer kernel) and 12.04 (live

[dave.kts]

Seriously dissapointed, I dont know where to begin troubleshooting here. Getting hard lockups at random times, doing anything from web browsing to copying files to a thumb drive, just listening to music, I was even installing 12.04 Ubuntu Studio and it locked up on me during the install. (I am having an issue with anything 12.04 now, when I'm done installing and on first boot Im greeted with the "Your running on low graphics mode" wether I boot and install the live CD with "nomodeset" or not)

32 or 64 bit makes no difference. My old standby OS is super Ubuntu 10.10 32 bit. I dont remember having this problem with it until maybe a year ago with a kernel update and started getting hangs. Booted from old kernel, no issue. Now, it doesnt matter which kernel i boot.
I tried several newer distro's including solusOS 1.1 (got screwed up after an updated NvIdia driver or something), Ubuntu Studio 12.04 (kernel lockup near the end of the install) AVlinux (dont know if it would have locked up, never installed graphics correctly) and regular Ubuntu 12.04 64 bit, which would lock up running the live CD.

I was running solusOS for some time and dont remember any lockups at all with that. it just got messed up after a series of updates, and instead of trying to fix it i just started trying new distros. SolusOS 1.2 on the other hand had locked up on me tho shortly after install.
I was impressed by the firewire latency with my presonus audio interface with solusOS so I thought I'd try some multimedia distros.

one thing to mention, I almost always have my android phone plugged in and tether internet to this computer. I dont think its that tho, its been great on several computers of mine and no issues but here on this one.

When it happens I have no keyboard or mouse control, caps lock non functional, all video freezes but no distortion, audio that is playing plays a 1 or 2 second never ending seamless loop.

Hardware:
Intel DP43BF board, no on board vid
Q8300 core2 quad
4 GB Gskill ram, underclocked (bought it for faster MB that failed in 3 months)- memtested overnight, twice, no errors
PNY Nvidia quaddroFX550
2 SATA hard drives, Intel SSD, 2 DVD roms
stable, oversized 80Plus power supply
atheros wifi PCI, not being used but installed

This board requires a "pci=assign-busses" boot parameter. I thought once that a "nomodeset" parameter was needed for this board but i could be mistaken. Lockups happen either way.

I really want to smash this thing. Should I just smash it to little pieces? Hell I can't even enjoy videos without tearing or compiz effects without crashing any 3D program. XFCE= no tearing, but to me its ugly.
Enlighten me: why did Intel support warn me that this MB does not support linux??

any advice short of a geek hooking up to it via serial console and waiting for it to happen again?

[dave.kts]

hump day bump day

[Rexilion]

Ubuntu 12.04 LTS is using the 3.2 kernel, try upgrading to the newest kernel. Might fix your regression. Is there any change you can get us a dmesg when it hangs?

Furthermore, what video card do you have? If you install the proprietary drivers (if available), do the lookups occur as well? Do you know since what kernel version you experience this?

We could try a bisect if it's correctly reproducible...

[dave.kts]

As for using 12.04- I cant yet until I figure out how to get past "configure your graphics manually" screen (no GUI) after a fresh install. Never had to do this untill 12.04.

Not sure what specifically you want for dmesg, how far back does it save these if they are of use? Let me know what you want to see or what I should be doing to prepare for the next lockup.

My video card is a PNY Nvidia quaddroFX 550 and this has happened using both the Nouveau drivers and the proprietary drivers, My 10.10 system is using version 260.19.06 and the kernel will hang, and I've recently had SolusOS 1.2 installed with the most current Nvidia driver, and it locked up, as well as the live installer of Ubuntu studio, during install, using non-proprietary. I have heard of others having lockups using Nvidia cards with nouveau drivers, and there is a boot parameter that can fix that (havent tried it)

My 10.10 system is booting with 2.6.35.32 generic pae and have 2 other kernels available, has locked up using any of those. And whatever kernel version the 12.04 image has, locks up with that too. Sorry I cant be more specific on when this started happening, it may have always happened on this PC but thought it was something else, I had thought that "nomodeset" had cleared it up at one point, but doesnt help anymore, or never helped it to begin with (I just wasnt having any issues for awhile)

Im afraid its not reproducible, i can go days or weeks with no trouble, then next time i wake it, or boot it, it will hang- 5 min into a session, or maybe 2 hours.

If its worth mentioning, an XMBC session will lock on this as well.

One more thing (in 10.10 only), I have a quad core and sometimes ill have a constant 25+% CPU use even tho sys monitor does not show the culprit of the use. It speads itself out over all of the cores, but will remain at 25-28% CPU use forever until I suspend or reboot. System doesnt seem any slower because of this, but happens frequently, randomly, and seems unrelated to predicting a kernel hang.

[Rexilion]

Well that narrows it down to a regression between 2.6.35 and 2.6.36 which is a really good start. It seems to be CPU related.

If the machine locks-up, press (while waiting 5 seconds between each one):

SYSRQ+R
SYSRQ+E
SYSRQ+I
SYSRQ+S
SYSRQ+U
SYSRQ+B

Maybe that might save the logs. On next boot, check /var/log/dmesg* for errors/oops/bugs/warn etc.

[dave.kts]

In dmesg I do get these warnings, and I see that ACPI may cause lockups in some systems... I've never tried disabling it
ACPI Warning: 32/64 FACS address mismatch in FADT - two FACS tables! (20100428/tbfadt-369)
[ 0.000000] ACPI Warning: 32/64X FACS address mismatch in FADT - 0xDF624F40/0x00000000DF625E40, using 32 (20100428/tbfadt-486)