Freeze on Toshiba NB300 without external interrupts

Problem still exists with kernel vmlinuz-2.6.35-21

+1, I too have a Toshiba N305-N310 am experiencing the same issue.

Problem persists with kernel vmlinuz-2.6.35-22
setting the nohz=off allows normal function but with a much shorter battery life.

I agree, still exists on 2.6.35-22.

Found that adding nolapic to the kernel options also works to allow the computer to boot normally. This also leads to shorter battery life.

I updated my BIOS to the latest revision (1.7) and the problem still exists.

Just found that changing the BIOS option "Dynamic CPU Frequency Mode" from "Dynamic" to "Always Low" allows the netbook to boot with the default kernel options. This is not ideal, but keeps the battery life up.

Did you read this page which suggests that setting it to "Always Low" might not be the best plan?

http://www.codon.org.uk/~mjg59/power/good_practices.html

In essence the machine can go to sleep for less time because it takes longer to do things.

However if you have actually found better power usage than nohz=off then I'll give it a go and see how much slower the system feels.

OK, I did a bit of quick comparisons of the different workarounds and the power usage as reported in powertop.

nolapic: 11.2W 97% of it's time in c4wait, 97% running at 1Ghz

nohz=off: 11.6W 98% of it's time polling, 100% running at 1GHz

clocksource=jiffies: 12.6W 99.5% of it's time polling, 50% running at 1.6GHz, 50% running at 1GHz

BIOS option "always low" 11.2W 59% of it's time in c1wait, 36% in c4wait (always running at 1GHz)

These were taken with simply booting, logging in and running powertop then letting the readings settle for about 5 minutes. I did not touch the system during this time. It would be better to run while doing specific tasks, but that is harder to replicate and I just wanted some down-n-dirty numbers. I think I will be sticking with the nolapic kernel option.

Hope this helps a bit and that we can find a fix to make this information obsolete. :)

I can confirm this bug.

Toshiba NB200 - Bios ver. 2.10
Kubuntu Maverick 10.10 Beta - completely updated as of today.

Problem still exists (and seems to have worsened for me)

One work around that was not mentioned: Setting Toshiba Bios setting : SATA Controller to Compatibility lessens the problem, but does not resolve.

The issue seems to arise from a Bios-level HDD parking mechanism that makes the harddisk park when the computer is 'idle' the only issue is that it does this when the computer is not idle. The alt-button or mouse movement tickles the HDD back into action.

Sounds like Bug #634702 try adding "intel_idle.max_cstate=0" as a boot parameter.

I can confirm that this workaround seems to have solved the issue for me.
Toshiba NB200 - Bios ver. 2.10
Kubuntu Maverick 10.10 Beta - completely updated as of today.

I spoke too soon, there is still occaisonal stuttering in music tracks, and videos.
Those are the only things where I find a hiccup is obvious.
The issues do seem to have lessened.

I have just tried this workaround (intel_idle.max_cstate=0) and it seems to work as it did in 10.04. Using my "highly scientific" power usage tests.. aka powertop, it is drawing only 7.8W spends ~80% in C4 and ~15% in C3 while running at 1GHz for 99% of the time.

It would be nice to get the intel idle option working to further improve the power savings, but I can live with this. Thank you Michael!

I will perform further testing to see if I experience any stuttering while playing music or videos.
Conor, for consistency's sake, were you streaming the music and videos, playing from the HDD or USB drive?
I would like to test under the same conditions to try to replicate the stuttering.

Hi Jason,

Videos streamed off Youtube.
Music played off the primary (windows) partition on my HDD. (linux is secondary).

In my testing to see if it was a pulseaudio problem, I disabled pulseaudio and the stuttering had decreased dramatically. However, it tends to be finicky and I can't replicate the results at everyboot, or every song, etc.

adding "intel_idle.max_cstate=0" seems to have solved the problem for me. Powertop indicates significantly less power usage than the previous kernel options yielded.

Thanks!
________________________________
Sophos Plc, The Pentagon, Abingdon Science Park, Abingdon, OX14 3YP, United Kingdom.
Company Reg No 2096520. VAT Reg No GB 348 3873 20.

Actually I also spoke too soon. There are still freezes. The machine seems to boot fine but often won't shutdown without keypresses to help it along. Also the sound regularly gets stuck in a loop unless you keep up constant activity like moving the mouse.

I tried the kernel build in bug 634702 (comment #27) but it doesn't help on my Toshiba hardware. (I had hoped it would since they are both N450 NM10 style machines).

So as far as I'm concerned this is still an outstanding issue.

Confirmed. Same symptoms even with the "intel_idle.max_cstate=0" option active.
They are infrequent, but still exist.

It sounds to me that with "intel_idle.max_cstate=0" the chance of system halt will decrease a lot, but still exist. Is it?

Yes. That about sums it up. Without intel_idle.max_cstate=0 the system is totally frozen unless you constantly provide keypresses, mouse movements, network activity, etc.

With intel_idle.max_cstate=0 the system will work but with regular enough freezes to be very frustrating. Playing music is a waste of time since it breaks up far too regularly.

Surely this should be mentioned on the release notes for the release candidate. If people upgrade their Toshiba NB300/305/310 series laptops the systems won't boot anymore. The Lenovo S10-3 gets a section in the release notes but the Toshiba problem is surely just as serious (there just doesn't seem to be a fix on the cards as yet).

i have the same problem.. the nohz=off workaround worked for me on lucid. Were do i put it in maverick exactly? the 10_linux file is a little different please help. i have a nb200

I put it in /etc/default/grub at the end of the options line and then
ran update-grub as root.

Just to summarize the available options as I understand it.

Maverick with no options at all doesn't boot/function without constant keypresses.
With intel_idle.max_cstate=0 the system boots and functions but with enough freezes to make activities like playing music a waste of time and the system generally gets stuck during shutdown and doesn't power off.
With nohz=off or setting the CPU to 'Always Low' in the BIOS the system functions normally but with reduced battery life or slower speed and reduced battery respectively.

Anyone else have different/better options?

Got the same issue on my NB205. lucid worked just fine without any work-a-rounds. hope this is fixed soon. Really annoying. Other than that I like maverick - sound and bluetooth OOTB.

I can confirm too that this exact same problem concerns the toshiba NB205 too.

I created a bug report for Toshiba NB205 Bug #657990 that will probably be merged soon with this one.
Such a disappointment, Unity looks great but i can't even watch a video on my laptop :'(

I hope Ike Panhc could look into that as this problem must be quite similar to the Lenovo one.

Same on Acer Laptops. Tried 2 different ones.

It seems to be an Intel "restricted" issue.

Was trying all mentioned workarounds here and bug #657990, but so far no improvements.

Kinda sucks, its been two kernel releases and as far as i can see in the http://intellinuxgraphics.org/2010Q3.html list, there is no fix planed in the ToDoList or even the Beyond ToDoList.

Please *buntu Team, do something. I don't wanna switch to windows.

Same here, after upgrading from 10.04 to 10.10, system hangs everytime if I do not move the mouse or type anything.

My system:

Magalhães PC (Portuguese version of intel's Classmate)
CPU: Atom N270
Ram: 1G

Same comments for me as Marcos O; however, I'm using a Toshiba NB-205 w/ 2 GB RAM. (Using latest release 10.10, fully updated, fresh install.)

Well .. it says up there: add nohz=off to the command boot line.

example:
GRUB_CMDLINE_LINUX_DEFAULT="quiet splash nohz=off"

That actually works. It lets the system run smooth. It turns off the CPU Effects. Shorter battery life, higher risk of overheating the laptop is the downturn on that solution.

This bug also affects my NB200 (2GB RAM) running ubuntu 10.10 netbook edition.

Can I ask before I try it:

Exactly how much reduction in battery life (either in minutes or % of normal usage) does the nohz=off fix cause? And how does this compare with CPU 'always low' setting?
In other words, does anyone have real life experience of the tests noted in comment #10? In my case I could use my netbook with 10.04 on a daily train commute, 2h 20min total. If my battery charge with 10.10 now reduces below that figure I can't work (but then I can't work if it keeps freezing, either...).

I'm not asking about the "intel_idle.max_cstate=0" fix because as I understand it this doesn't actually put a stop to the freezes.

Thanks.

Further to comment #32, don't bother - I think there's a better workaround. See https://bugs.launchpad.net/linux/+bug/657990/comments/27

Paul Hirst, thank you for reporting this and helping make Ubuntu better. Maverick reached EOL on April 2012.
Please see this document for currently supported Ubuntu releases:
https://wiki.ubuntu.com/Releases

We were wondering if this is still an issue on a supported release? If so, can you try with the latest development release of Ubuntu? ISO CD images are available from http://cdimage.ubuntu.com/releases/ .

If it remains an issue in a supported release, could you run the following command from a Terminal (Applications->Accessories->Terminal). It will automatically gather and attach updated debug information to this report.

apport-collect -p linux <replace-with-bug-number>

Also, if you could test the latest upstream kernel available that would be great. It will allow additional upstream developers to examine the issue. Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Once you've tested the upstream kernel, please remove the 'needs-upstream-testing' tag. This can be done by clicking on the yellow pencil icon next to the tag located at the bottom of the bug description and deleting the 'needs-upstream-testing' text. Please let us know your results.

Thanks in advance.

[Expired for linux (Ubuntu) because there has been no activity for 60 days.]

This bug still affects me on my NB205-N310BN-G

https://bugzilla.kernel.org/show_bug.cgi?id=15289
and
https://bugzilla.redhat.com/show_bug.cgi?id=648837

appear to discuss a similar problem, but with different hardware. The kernel parameter "acpi_skip_timer_override" is suggested as a work around, will test.

Cody P Schafer, if you have a bug in Ubuntu, could you please file a new report by executing the following in a terminal:
ubuntu-bug linux

For more on this, please see the Ubuntu Kernel team article:
https://wiki.ubuntu.com/KernelTeam/KernelTeamBugPolicies#Filing_Kernel_Bug_reports

the Ubuntu Bug Control team and Ubuntu Bug Squad team article:
https://wiki.ubuntu.com/Bugs/BestPractices#X.2BAC8-Reporting.Focus_on_One_Issue

and Ubuntu Community article:
https://help.ubuntu.com/community/ReportingBugs#Bug_reporting_etiquette

When opening up the new report, please feel free to subscribe me to it.

Please note, not filing a new report may delay your problem being addressed as quickly as possible.

Thank you for your understanding.