Kernel freeze, new Dell laptop (intel graphics), hard reset only solution

[veroslav]

Hi,

I am running Kubuntu 11.10, 64-bit, fresh installation on a three week old Dell 17r laptop with the following specifications:

Intel i7-2670QM (2.20 Ghz, 6 MB, 4 cores)
nVidia GeForce GT 525M (2 GB)
6 GB DDR3 SDRAM

It comes with hybrid graphics, (nVidia + Intel HD 3000) but I've disabled the nVidia card with Bumblebee and am running on Intel only.

During the last week, I experienced two freezes (kernel freezes I believe) from which I couldn't recover with neither of the following:

Ctrl+Alt+F1..F6
Ctrl+Alt+Del
Ctrl+Alt+Backspace
Alt+SysRq+REISUB

The screen was completely frozen, and I couldn't use neither mouse nor keyboard. The only thing that helped was a hard reboot. I am starting to get worried as I know that a hard reboot is not good for a computer and is not a longterm solution for my problem.

I haven't managed to pinpoint anything special I was doing that caused the freeze (the first time I went to click on a link in Firefox and it froze while hoovering over the link with the mouse, the other time I tried to click on a button in kdenlive, also while hoovering with the mouse over this button).

Does anybody have any clues as to what might have caused the freeze? I am willing to put in as much effort as needed in order to resolve this issue, how should I start debugging?

Thank you in advance.

Kind Regards,
Veroslav

[SteveRiley]

It comes with hybrid graphics, (nVidia + Intel HD 3000) but I've disabled the nVidia card with Bumblebee and am running on Intel only.

Does Bunblebee really disable the hardware completely? If that's your goal, I'd recommend instead changing the BIOS/UEFI setting. That's what I've done with my ThinkPad T520.

The screen was completely frozen, and I couldn't use neither mouse nor keyboard. The only thing that helped was a hard reboot. I am starting to get worried as I know that a hard reboot is not good for a computer and is not a longterm solution for my problem.

If it's any consolation, I experience similar freezes on rare occasion. I don't know why -- I can never see anything interesting in syslog after a power cycle. But also I'm not aware of any long-term damage this can do to a computer, either. In their default configurations, Linux file systems have strong data protection; I've not lost a single file because of a lock-up.

[vw72]

That seems to be a common problem with the 17r in both 32bit and 64bit. You might check out the actual Ubuntu forums, there has been discussion there regarding kernel parameters to set at boot. You might also google: dell 17r linux freeze as it appears that the problem is not k/ubuntu specific.

As for the hard reboot, if there were to be damage to the filesystem, it would have occurred at the point the system froze, not the actual reboot.

[veroslav]

Thank you both for replying!

SteveRiley,
not really sure whether Bumblebee does this, I will have a look in my BIOS and try to disable it from there as well. I was also unable to find anything of value in the syslog around the time the freeze occured unfortunately. The only "ERROR" line I found that I thought might be of slight interest was this one:

[ 1.801520] [drm:intel_dsm_platform_mux_info] *ERROR* MUX INFO call failed
[ 1.801545] [drm:intel_dsm_platform_mux_info] *ERROR* MUX INFO call failed

These two occured ca 40 min before the freeze occured. I looked up these on google and found out that they are related to another bug linked to the Sandy Bridge processor architecture, and usually manifests itself in the computer not being able to boot (when the message is logged). No such issues for me though.

vw72,
I've tried searching for this issue on ubuntu forums and though I found a few threads mentioning lock-ups, ubuntu 11.10 and Dell laptops, none of these came to any solution or conclusion unfortunately.
Will keep a close eye though. The kernel parameters I was experimenting with earlier (while getting the freeze) were these:

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash pcie_aspm=force i915.i915_enable_rc6=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1"

I noticed quite a decrease in the power consumption when using the above parameters, however, I suspected that these might be the ones that were causing the freeze so I removed them today again in hope that it MIGHT fix the freeze issue. Also, I have switched to Ubuntu 11.10 on my other partition in order to find out whether the freeze is occuring there as well.

I also ran the Dell Diagnostics utility from the boot menu and let it perform an extended check of the hardware and it came out ok.

Thanks for the input. I will keep an eye on this thread in case any more tips come up. I am really hoping that the kernel available in (K)ubuntu 12.04 will at least fix some of these issues.

Kind Regards,
Veroslav

[SteveRiley]

not really sure whether Bumblebee does this, I will have a look in my BIOS and try to disable it from there as well

If you really want to avoid using the nVidia (which I do, because it's a power hog and Intel graphics are perfectly fine for what I do on my T520) then BIOS/UEFI disabling is the way to go. That will remove the nVidia chip from PCI bus and therefore eliminate any power consumption.

GRUB_CMDLINE_LINUX_DEFAULT="quiet splash pcie_aspm=force i915.i915_enable_rc6=1 i915.i915_enable_fbc=1 i915.lvds_downclock=1"

First, get rid of "quiet splash" unless you really like to see pretty boot screens. It's mostly a personal preference, but I like seeing the reports of what my computer is doing while it boots. It's comforting

Enabling rc6 is known to cause a number of problems. On my T520 it introduces sufficient instability that I can't use it. But on my ThinkPad X1, with the exact same HD 3000 graphics, rc6 works without any problems. Odd.

I noticed quite a decrease in the power consumption when using the above parameters, however, I suspected that these might be the ones that were causing the freeze so I removed them today again in hope that it MIGHT fix the freeze issue.

Of those three parameters, enabling rc6 contributes the most to lowering power consumption. You might try removing only that one (because it's also the most troublesome) and then gauge system stability against power consumption with the other two parameters left in place.

[veroslav]

Does Bunblebee really disable the hardware completely? If that's your goal, I'd recommend instead changing the BIOS/UEFI setting. That's what I've done with my ThinkPad T520.

I just checked my BIOS settings and it appears that I dont have the option to disable the nVidia card from there. I read that they had the option on the first laptops that appeared with Optimus technology but that they later removed this option. This looks to be the case for me.

[veroslav]

Of those three parameters, enabling rc6 contributes the most to lowering power consumption. You might try removing only that one (because it's also the most troublesome) and then gauge system stability against power consumption with the other two parameters left in place.

I will try that SteveRiley. It will take a few days to verify whether there is any effect though. The two lock-ups I've got were one week apart.

Thank you for helping me, I appreciate it!