Random kernel crash/freeze in Feisty

Bug #115275 reported by Ralf Hölzemer
8
Affects Status Importance Assigned to Milestone
linux-source-2.6.20 (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

Binary package hint: linux-source-2.6.20

The current installed kernel freezes/crashes at "random" times. The bug is triggered without any user interaction.
There are more people having a similar problem discussing the issue at http://ubuntuforums.org/showthread.php?t=412125

Things i observed while investigating are:

- It is definitely a 2.6.20 thing. Edgy doesn't crash on the same machine
- Seems to have nothing to do with proprietary NVIDIA drivers because the crash also happens with the "nv" driver enabled. Even a stock Feisty LiveCD crashed.
- Disabling every hardware component & feature in the BIOS, one at a time, has no effect. Crashes would still happen
- Booting with various kernel flags mentioned in the above thread didn't help ( noapic, assign-busses ... )
- The machine freezes/crashes 10-20 times per day!
- Once the machine didn't crash for longer than ~30min, it would run stable for the rest of the day
- I noticed that many people on launchpad have problems with their SAMSUNG (TSSTCorp) DVD-Burner, which i also have. One guy posted that he could fix the error with a firmware update of the burner ( https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/64587/comments/16 ). So today i updated my drives firmware to the latest version, but the crash is still there. I even disabled the drive in the BIOS to see if the crash is directly linked to it, but the situation is the same.

I managed to get some debug output via ALT+SysRq+1 & ALT+SysRq+T, which i will attach to this report.

Revision history for this message
Ralf Hölzemer (cheleb-deactivatedaccount) wrote :

The logfiles

Revision history for this message
Günter Hubner (guenter-hubner) wrote :

I've got the same problem on my laptop without using any DVD-burner. The system freezes within approximately 5 minutes after booting. It worked fine using 6.10.

Revision history for this message
Florian Kollmannsberger (florian-kollmannsberger) wrote :

I can confirm this bug,
sometimes the system works a few days without a freeze, sometimes it freezes several times a day.

Changed in linux-source-2.6.20:
status: Unconfirmed → Confirmed
Revision history for this message
napalm (geukes) wrote :

I have the same Problem.. (Total freeze)
I have installed follow kernels to testing

2.6.17-11-gerneric
2.6.17-50-gerneric

2.6.20-15-386

2.6.22-4-generic

With kernel 2.6.22-4 (gutsy-release) I have the same Problem....

It seems all Kernel since 2.6.20-15-generic have the problem.
In the german ubuntuusers forum, the people tell that this problem ist only under kernel 2.6.20-15-gerneric and not under kernel 2.6.20-15-386
Additional many people in the forum tell about acpi problems. the Url of the Forumtrade is http://forum.ubuntuusers.de/topic/88548/75/

escuse my english, i'am german

Revision history for this message
Sitsofe Wheeler (sitsofe) wrote :

Ralf:
Thank you for your bug report.

Which CPU/motherboard do you have? If it is an Athlon does disblaing cool 'n quiet in the BIOS make a difference?

Changed in linux-source-2.6.20:
status: Confirmed → Needs Info
Revision history for this message
Ralf Hölzemer (cheleb-deactivatedaccount) wrote :

Hi Sitsofe,

the CPU is an Intel P4 3Ghz HT. Motherboard is a Gigabyte 8KNXP Rev. 1. There is no cool 'n quiet function in the BIOS.

Revision history for this message
Günter Hubner (guenter-hubner) wrote :

My laptop is very old. The CPU is an Intel P3 700 Mhz.

Revision history for this message
Sitsofe Wheeler (sitsofe) wrote :

Ralf:
A bit of a long shot but does disabling powernowd as described in https://bugs.launchpad.net/ubuntu/+source/linux-restricted-modules-2.6.20/+bug/109643/comments/9 make any difference?

Revision history for this message
Ralf Hölzemer (cheleb-deactivatedaccount) wrote :

Okay. I did the change described above and all i can say for now is that the next boot after the change, the system came up without problems. It is also running stable for more than 30 minutes now.
Unfortunately that doesn't mean anything because sometime, after 10 - 20 crashes, if the system ran stable for more than ~30 minutes, it will run stable for the rest of the day.
So, to answer your question, it seems to have something to do with powernowd, because today i had only 3 crashes before i did the change. Compared to the usual behaviour, this is a suprising low crash rate!
Though i can only provide a correct answer in about 24 hours from now on.

It may be interesting to hear the results of other people in here.
Thanks for the suggestion. I keep my fingers crossed!

Revision history for this message
Sitsofe Wheeler (sitsofe) wrote :

Ralf:
If you can report your experiences back here in a few days time that would be great...

Revision history for this message
Ralf Hölzemer (cheleb-deactivatedaccount) wrote :

Sistofe,

the system runs stable today! There was no single crash or freeze, which i didn't have for a long time now. I am sure the problem is powernowd.
Thank you very much for the tip. If you have any further questions or you want me to do something - shoot!

Revision history for this message
Sitsofe Wheeler (sitsofe) wrote :

Ralf:
Looks like you might have gotten lucky in finding the cause of the problem. Good luck! I'll close this bug and mark it a duplicate.

(PS if you want to help me out take a look at the recent nvidia binary bugs (https://bugs.launchpad.net/ubuntu/+bugs?field.searchtext=nvidia&orderby=-date_last_updated&search=Search&field.status%3Alist=Unconfirmed&field.status%3Alist=Confirmed&field.status%3Alist=In+Progress&field.status%3Alist=Needs+Info&field.status%3Alist=Fix+Committed&field.assignee=&field.bug_reporter=&field.omit_dupes=on&field.has_patch=&field.has_no_package= ) sometime and get a feel for the common problems. Read some of the bugs and the responses that people ask for and the typical patterns that crop up and if you feel like it try and help out other people having issues...)

Thanks for taking the time to report this bug and helping to make Ubuntu better. This particular bug has already been reported and is a duplicate of bug #109643 and is being marked as such. Please feel free to report any other bugs you may find.

Changed in linux-source-2.6.20:
status: Needs Info → Rejected
Revision history for this message
IVIisterX (doellies) wrote :

Hello Community!

I had the same problems written above; my machine did freeze at random times after restart.

My configuration is :

ASUS P4C800 Deluxe Mainboard BIOS-Version: 1019.002 AMI-BIOS

Intel P4 2,6 GHz

1024 MB RAM

ATI 9700 PRO

160 GB Hitachi SATA-HDD as Bootmedium

USB Keyboard and USB Maus from Logitech

In various tests I established, that there is an relation between the freezes and the Powermanagement-Settings:

BIOS:

APM enabled

ACPI 2.0 Support disabled

ACPI APIC Support enabled

BIOS -> AML ACPI Table enabled

Ubuntu 7.04 Alternate Installation -> there are freezes at random times. There is no dependency on the used Graphics-Driver; by the way I had the feeling that the accelerated Ati-Graphics-Driver is more stable.

BIOS:

APM enabled

ACPI 2.0 Support disabled

ACPI APIC Support enabled

BIOS -> AML ACPI Table enabled

Ubuntu:

Powernowd Service disabled

There are freezes at random times.

BIOS:

APM enabled

ACPI 2.0 Support disabled

ACPI APIC Support enabled

BIOS -> AML ACPI Table enabled

Ubuntu:

Powernowd Service disabled

Acpid Service disabled

Apmd Service disabled

It is impossible to work with the system; the machine freezes a few minutes after every restart.

BIOS:

APM disabled

ACPI 2.0 Support disabled

ACPI APIC Support disabled

BIOS -> AML ACPI Table disabled

Ubuntu:

Powernowd Service disabled

Acpid Service disabled

Apmd Service disabled

Now only Grub is starting; Ubuntu doesnt start anylonger.

BIOS:

APM disabled

ACPI 2.0 Support disabled

ACPI APIC Support enabled

Bios -> AML ACPI Table enabled

Ubuntu:

Powernowd Service disabled

Acpid Service disabled

Apmd Service disabled

Ubuntu starts again, but freezes at random times.

BIOS:

APM disabled

ACPI 2.0 Support enabled

ACPI APIC Support enabled

BIOS -> AML ACPI Table enabled

Ubuntu:

ACPID Service disabled

APMD Service disabled

Powernowd Service disabled

Ubuntu is more stable and freezes after approximately 30 minutes.

BIOS:

APM disabled

ACPI 2.0 Support enabled

ACPI APIC Support enabled

BIOS -> AML ACPI Table enabled

Ubuntu:

ACPID Service enabled

APMD Service enabled

Powernowd Service disabled

Kernel changed from 2.6.20-16 (generic) to 2.6.20-16-386

Now my machine is running without freezes since 5 days with many different progams and for many hours per day. I don't think that the Kernel-Changing is the reason for the stability, because another user (see also http://forum.ubuntuusers.de/topic/88548/180/) has an freeze with Kernel 2.6.20-16-386 two days ago. So I think that the "right" APM / APIC Settings are the reason for the stability; but I have not the knowledge what these settings mean. Perhaps there is an onther one who can explain the settings and perhaps see the reason for the freezes.

I look forward that my comments will help to solve these bug. Please apologize my bad English.

IVIisterX

Revision history for this message
IVIisterX (doellies) wrote :

Hello altogether!

After 12 days without any freezes I wanted to test wether the Kernel or the Bios-settings are responsible for the stability.

So I used again Kernel 2.6.20-16-generic. After a few minutes working in Firefox the machine freezed again.

Until I am writing this post the machine uses Kernel 2.6.20-16-386 again.

Also I enabled the service powernowd again. I will tell you next week which consequences this modification has.

Best regards

IVIr.X

Revision history for this message
IVIisterX (doellies) wrote :

Hello altogether!

To enable the powernowd service was a flop; there were freezes until watching an avi-video with the mplayer.

I found out, that in the case of my mainboard (ASUS P4C800 Deluxe) the ACPI 2.0 Support had to be enabled and in Ubuntu I had to install the ATI driver 8.39.4 for my ATI 9700 PRO and to disable the powernowd service to get more stability.

With these settings I have no freezes anymore; but if I only change one setting Ubuntu freezes at once.

Best regards

IVIr.X

Revision history for this message
IVIisterX (doellies) wrote :

Hello alltogether!

After another freeze a few days before I checked out the solution discribed in the following link:

http://ubuntuforums.org/showpost.php?p=3134000&postcount=544

The summary:

Various freezes with different programs with these settings.

But since I took back the additional kernel parameters "noapic nolapic irqpoll ht=on" my machine runs stable without freezes with the kernels 2.6.20-15, 2.6.20-16 and 2.6.20-386.

So I come to the conclusion that the maintrigger for the freezes is the powernowd-Service. In the case of freezes first checkout wether the remove from powernowd brings more stability.

Best regards

IVIr.X

Revision history for this message
IVIisterX (doellies) wrote :

Hello alltogether!

I have some new knowledges in the case of the freezes.

Without any changes at the "Ubuntu Feisty Fawn Alternate"-installation my machine freezes with the Kernel 2.6.20-15-generic, 2.6.20-16-generic and 2.6.20-16-386.

To get more stability I have to change

1. The Bios-Setting ACPI 2.0 support from disabled to enabled.
2. Uninstall powernowd
3. Install the ATI-driver 8.39.4

With these settings and Kernel 2.6.20-15-generic my machine runs without any freezes; with the kernels 2.6.20-16-generic and 2.6.20-16-386 my machine freezes within 30 minutes max.

If I add the additional kernel parameters noapic nolapic irqpoll ht=on my machine freezes a few minutes after reboot with various programs.
If I uninstall the restricted graphics driver my machine freezes a few minutes after disabling and with all kernels above-mentioned and various programs.

While I am writing this comment I test the "sudo /etc/init.d/powernowd stop" with kernel 2.6.20-16-generic - it still works; this trick was posted from Riccardo Lucchese at
https://bugs.launchpad.net/ubuntu/+bug/131973
respectively at
http://ubuntuforums.org/showthread.php?t=412125&page=60

Best regards

IVIr.X

Revision history for this message
IVIisterX (doellies) wrote :

Hello altogether!

I made the kernel update to version 16.31 and updated the Ati-driver by Envy to version 8.40.4; since this time my system runs pretty stable.
But after 10 days without any freezes by multiple starts every day i had two freezes at friday evening. The first 8 minutes after start; the second 9 minutes after reset. Since another reset my system is working without any additional freezes up to the moment.
I guess furthermore, that there are some periodical processes which are responsible for the freezes. Is there perhaps anything else like the periodical harddisc checking? Does anyone else know something about periodical actions?

IVIr.X

Revision history for this message
IVIisterX (doellies) wrote :

Hi, everyone!

Because my machine has no freezes with kernel 2.6.20-16.32-386 and MESA with indirect rendering, I made the following interesting test:

On the page [url]http://www.unixboard.de/vb3/showthread.php?t=33850[/url] I found an instructions manual to get direct rendering with MESA. After configuring and restarting my system I had really direct rendering with MESA. But both kernels 2.6.20-16.32-generic and 2.6.20-16.32-386 freeze a few minutes after restart and using firefox. So I had to reinstall the ATI driver 8.40.4 with envy.

In this context I have a question:

If I change the AGPmode in the xorg.conf file to "4", the ATI Catalyst Control Center shows an 8x AGPmode under kernel 2.6.20-16.32-generic and an 0x AGPmode under kernel 2.6.20-16.32-386. Why does the driver overwrite the xorg.conf settings and where is perhaps a possibility to change this setting?

Thanks and Greetz

IVIr.X

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.