kernel problem

English Install/Boot/Login
#1

Although this is not a new problem to me I decided to post it here. Hope it is not the wrong place.

My system is an athlon 64 X2 4800+ on an nforce4 chipset (MS****I k8n neo4-f) with 2Gb of ram.

Since some time (had the same issues with opensuse 12.1 32bit also) I have many instabilities and kernel panics (I would appreciate if you could tell me a way to retrieve the kernel panic log after I reboot). The problems appear when I boot normally, if I boot with nosmp flag the system runs smooth, of course with only one cpu.

Are there any changes in the kernel that can cause this issues? is it something known? maybe there is a flag (other than nosmp) that would let me use the system with both cpus?

Thank you very much

#2

On 2012-07-03 10:16, psaxioti wrote:
>
> Although this is not a new problem to me I decided to post it here. Hope
> it is not the wrong place.

It is, as you are using 11.1. You can post here if you use 12.2 till
September, then 12.2 will also be wrong here.

> Since some time (had the same issues with opensuse 12.1 32bit also) I
> have many instabilities and kernel panics (I would appreciate if you
> could tell me a way to retrieve the kernel panic log after I reboot).

provided the filesystem did not crash, it would be in /var/log/messages. If
not, take a photo or plug in a serial cable and do some
incantations.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

#3

i am not sure i can help, but i see no response in four+ hours, so:

  1. which kernel did you install in 12.1, and which in 12.2?

if you don’t know, copy/paste back to this thread the in and output of


uname -a

  1. there is no “kernel panic log” that i am aware of…but, there is a
    log /var/log/messages which generally contains messages from the kernel
    for the current and previous sessions…each line begins with a
    date/time stamp…

  2. your machine is 64 bit capable, and i wonder why you choose a 32 bit
    system—not that that is wrong (mine is also a 64 running 32)?

  3. while waiting for competent help to arrive you might want to do a
    little reading (but experiment ONLY carefully and after much
    thought…because not all are openSUSE, nor are all your exact hardware,
    nor etc etc etc…so, read a lot and don’t go down a passageway you can
    back out of, before help arrives!):

http://forums.opensuse.org/english/get-technical-help-here/hardware/450424-opensuse-11-1-intermittent-hangs.html

https://www.google.com/search?q=athlon+64+X2+4800%2B+linux+“kernel+panic”+nosmp

https://www.google.com/search?q=athlon+64+X2+4800%2B+linux+“kernel+panic”

–
dd

#4

Hello and thanks for your response.

I don’t know why you guess that I am running 11.1, but you are wrong. I am running 12.2 beta2 32bit. The thing is that I had the same issues with 12.1 32bit. Before that I had 11.4 64bit, where I don’t remember having any problems, and before that 11.3 32bit that I am almost 100% positive that it was working properly.
So for me, something seems to have changed in the kernels sometime after 11.3 or 11.4 (I know I report it a little bit late).
I am not blaming opensuse, because I have also tried ubuntu with the same instabilities, though I have the impression they are a little bit more rare.
I don’t know also if it is a hw problem, but under windows it seems ok (at least for as long I have checked), so I don’t think so.

As far as the kernel panic log. I have checked and never got anything in /var/log/messages
Usually I find the system frozen with a black screen, or if the thing happens while working, it just freezes. I am using nvidia drivers and it doesn’t go to the terminal with the kernel panic output. I can try uninstalling the nvidia drivers and check the output (I think without the drivers it falls back to the terminal).
For sure I dont know what to do with a serial cable.

Thank you very much

#5

Thanks for your replay.

  1. The kernel I had was always the official opensuse kernel, usually the desktop flavor. Right now (it just updated) I have the 3.4.4-2.4-desktop. With 12.2 I experimented with kernels from the a repository (I think 3.1.x, 3.2.x), but with the same results.
  2. I have checked, but nothing relevant in there, at least to my understanding.
  3. I didn’t like the 64bit as was a mixture of 32 and 64 bit. I guess it has become better by now but … Also this is a pc I use for work also and I have some programs that I think run better under 32bit (I am a noob and I cannot support what I am saying, it is just a feeling. I can try again the 64bit architecture).
  4. I will do some more reading hoping to find something

Thank you very much

#6

On 07/03/2012 03:16 PM, psaxioti wrote:

> 1. The kernel I had was always the official opensuse kernel, usually
> the desktop flavor. Right now (it just updated) I have the
> 3.4.4-2.4-desktop.

yes, there was a ‘bad’ kernel pushed out though the update repo which
caused problems for lots of folks…

> With 12.2 I experimented with kernels from the a
> repository (I think 3.1.x, 3.2.x), but with the same results.

so, what that means to me is that you are having similar problems with a
variety of kernels, from a variety of versions of openSUSE and
Ubuntu…so it is probably not a kernel fault, but something else
interacting with the kernel, or just upsetting the whole system:

  • maybe a hardware problem, or

  • maybe a software conflict of some kind

  • maybe something else

so, lets go back and learn more: you wrote “many instabilities and
kernel panics” and “Usually I find the system frozen with a black
screen, or if the thing happens while working, it just freezes.”

each of those FOUR (unstable, panic, black screen & freeze) could be
caused by four very different problems

“kernel panic” is normally accompanied by flashing keyboard LEDs,
completely unresponsive to keyboard input AND tell-tale info in
/var/log/message

“kernel panic” is often cause by bad RAM: have you run memtest (seen as
“Memory Test” on every openSUSE install disk…and, it is VERY possible
that because of the way Linux uses memory a small problem which would
not upset Windows will KILL your system (linux uses ALL of the RAM and
therefore exercises sections of the RAM stick that Windows may never
ever touch)…so run memtest…not less than overnight…really, it take
a long time to test 2GB of memory…one pass does NOT do it…takes
many many passes…

“instabilities” can be cause by about a million (maybe a billion)
things…even just using the same /home from 11.3 to 11.4 to 12.1 and
12.2 AND with Ubuntu could introduce enough crud into the configs in
your /home that there is no telling what might be in there…so, try
this add a new test user, and log into that test user and do a lot of
browsing, whatever and see if you ever have a problem…(yes i know, all
of “your stuff” is not at your finger tips anymore…

“black screen” could be the result of either a video card problem or
simply a misset power savings setting…

“freeze” could be due conflicting software…

so, please show us the terminal input/output from


zypper lr -d

copy/paste the in/output back to this thread using the instructions
here: http://goo.gl/i3wnr

and any of the four could be prompted by a weak or failing power
supply, cracked motherboard, loose connection anywhere in the machine,
faulty ground, leaking capacitor(s), heat problems, etc etc etc etc etc…

> 2. I have checked, but nothing relevant in there, at least to my
> understanding.

the next time it happens look at your most accurate clock, and write
down the exact time…convert that exact local time to UTC time…and,
then look into /var/log/messages and copy paste from five minutes prior
to the freeze/panic/instability/black screen up the gap in time caused
by the delay while you rude killed the machine, and rebooted

and paste that log to paste.opensuse.org and reference the URL in your
next post to this thread…

> 3. I didn’t like the 64bit as was a mixture of 32 and 64 bit. I guess
> it has become better by now but … Also this is a pc I use for work
> also and I have some programs that I think run better under 32bit (I am
> a noob and I cannot support what I am saying, it is just a feeling. I
> can try again the 64bit architecture).

ok, i do not think moving to 64 will help…unless you do a complete
format and install without bring forward any problems which may (or
may not) exist in the configs in your /home

i’ve very happy with my 32 bit system in a 64 bit machine…

> 4. I will do some more reading hoping to find something

at this point it is kinda like looking for a needle in a haystack, in a
hurricane:

-do the RAM check

-tell us more about your hardware–like age…power supplies weaken over
time and can (WILL) cause all kinds of seemingly unrelated
problems…and age can allow dust (and chicken bones) to build up in
critical areas…or or or or

-show us the “zypper lr -d”

OH: if you boot from a live CD and just run and run and run and run, do
you ever (or never) have one (or all of) these four problems?

well… i can tell you this: i have had NONE of those problems since
(about) 2006 when i replaced my power supply and CPU grease on a machine
which is now ‘retired’…my current machine (a little ACER +nettop
bought in Feb 2011) has never shown any of those problems, not
once…so, i’m not likely to think it is either a kernel or Linux
problem…it MAY be a problem that Linux does not like your hardware,
but that is a different kettle of fish.

–
dd http://tinyurl.com/DD-Caveat http://tinyurl.com/DD-Hardware
http://tinyurl.com/DD-Software

#7

On 2012-07-03 15:06, psaxioti wrote:
>
> Hello and thanks for your response.
>
> I don’t know why you guess that I am running 11.1, but you are wrong. I
> am running 12.2 beta2 32bit. The thing is that I had the same issues
> with 12.1 32bit. Before that I had 11.4 64bit, where I don’t remember
> having any problems, and before that 11.3 32bit that I am almost 100%
> positive that it was working properly.

I typed 11.1 but I meant 12.1, as you mentioned it. I was confused with
what you posted.

> As far as the kernel panic log. I have checked and never got anything
> in /var/log/messages

Pity.

> Usually I find the system frozen with a black screen, or if the thing
> happens while working, it just freezes. I am using nvidia drivers and it
> doesn’t go to the terminal with the kernel panic output. I can try
> uninstalling the nvidia drivers and check the output (I think without
> the drivers it falls back to the terminal).

It would be best to reproduce the problem with the nouveau driver. The
kernel devs will not look at it if not. :frowning:

> For sure I dont know what to do with a serial cable.

Well, it is explained in that link I posted. It is the best alternative.
They also mention a method with USB which is new to me.

For serial port you first try connection from that computer (with a real
serial port in hardware) to another running minicom and dumping to file.
When that runs, you modify the kernel boot parameters so that messages are
logged to the serial port, and reboot. I do not know if the procedure works
with systemd, I think the support is incomplete.

The messages appear in the second computer in the minicom session.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

#8

OK, I guess that the words I used are not the right ones (freeze, kernel panic, instability, …). I think that what I get is always a kernel panic, but depending from the graphics drivers I use and when it happens (working on it or being away) the state that I find the pc (black screen, frozen, …) is different.

My keyboard is wireless with no leds, so no way to check this.

I have also thought of the PSU, so I have checked it. The voltages are fine, even with the pc running, and then I guess I would have problems in windows also.

The memory I have checked it in the past. I will leave it overnight today again, and maybe over the weekend also to see what is happening. But if it was a memory problem I cannot understand why it happens only without the nosmp flag.

Just for the history, when I install a new system I usually format the whole system (this I haven’t done with 12.2, it is a distribution upgrade from 12.1) apart from the /home partition, from which I delete all the hidden files and directories and reconfigure the desktop. I leave only my personal files + .tcsrc or things like this. Ubuntu I installed it recently on a different partition (it uses the same /home partition though) to check if I get the same instabilities.

You asked me for the output of zypper lr -d, so here it is:

#  | Alias                             | Name                                    | Enabled | Refresh | Priority | Type   | URI                                                                                | Service
---+-----------------------------------+-----------------------------------------+---------+---------+----------+--------+------------------------------------------------------------------------------------+--------
 1 | download.opensuse.org-Education   | openSUSE BuildService - Education       | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/Education/openSUSE_Factory/              |        
 2 | download.opensuse.org-Extra       | openSUSE BuildService - KDE:Extra       | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/KDE:/Extra/openSUSE_Factory/             |        
 3 | download.opensuse.org-Wine        | openSUSE BuildService - Wine CVS Builds | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/Emulators:/Wine/openSUSE_Factory/        |        
 4 | download.opensuse.org-filesystems | openSUSE BuildService - filesystems     | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/filesystems/openSUSE_Factory/            |        
 5 | download.opensuse.org-games       | openSUSE BuildService - Games           | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/games/openSUSE_Factory/                  |        
 6 | download.opensuse.org-mozilla     | openSUSE BuildService - Mozilla         | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/mozilla/SUSE_Factory/                    |        
 7 | download.opensuse.org-packages    | openSUSE BuildService - Java:packages   | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/Java:/packages/openSUSE_Factory/         |        
 8 | editors                           | editors                                 | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/editors/openSUSE_Factory/                |        
 9 | home:Lord_LT:drivers              | home:Lord_LT:drivers                    | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/home:/Lord_LT:/drivers/openSUSE_Factory/ |        
10 | home:flowabcd                     | home:flowabcd                           | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/home:/flowabcd/openSUSE_Factory/         |        
11 | network                           | network                                 | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/repositories/network/openSUSE_Factory/                |        
12 | openSUSE-12.2_12.2-1.14           | openSUSE-12.2 12.2-1.14                 | Yes     | No      |   99     | yast2  | iso:///?iso=openSUSE-12.2B1.iso&url=file:///home/axiotis/prosopika/Programs/SuSE/  |        
13 | packman.inode.at-suse             | Packman Repository                      | Yes     | Yes     |   99     | rpm-md | http://packman.inode.at/suse/Factory/                                              |        
14 | repo-debug                        | openSUSE-12.2-Debug                     | No      | Yes     |   99     | NONE   | http://download.opensuse.org/debug/distribution/12.2/repo/oss/                     |        
15 | repo-debug-update                 | openSUSE-12.2-Update-Debug              | No      | Yes     |   99     | NONE   | http://download.opensuse.org/debug/update/12.2/                                    |        
16 | repo-non-oss                      | openSUSE-12.2-Non-Oss                   | Yes     | Yes     |   99     | yast2  | http://download.opensuse.org/distribution/12.2/repo/non-oss/                       |        
17 | repo-oss                          | openSUSE-12.2-Oss                       | Yes     | Yes     |   99     | yast2  | http://download.opensuse.org/distribution/12.2/repo/oss/                           |        
18 | repo-source                       | openSUSE-12.2-Source                    | No      | Yes     |   99     | NONE   | http://download.opensuse.org/source/distribution/12.2/repo/oss/                    |        
19 | repo-update                       | openSUSE-12.2-Update                    | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/update/12.2/                                          |

In the last hours I did some tests so here is the outcome. The 3.4.4-2.4-desktop kernel seems somewhat more stable without the nosmp flag than the 3.4.2-1-default (I just installed to compare with the new one). Anyway I managed to get 2 kernel panics with each of the above mentioned kernels. I uninstalled the nvidia drivers so I photographed the outputs. You can find them here. The first two (horrible quality, they are with a mobile phone) is the first kernel panic with the 3.4.2-1-default kernel and the next three from the second kernel panic. Then there are two + two photos from the kernel panics of the 3.4.4-2.4-desktop kernel.
In all of them there is something about SMP and in the newest kernel pics there is something about a bug at the posix-cpu-timers.c:1288 and PREEMPT. The pics from the newest kernel (with bug at the posix-cpu-timers.c) are closer to what I remember getting at opensuse 12.1.

Although I didnt notice anything on /var/log/messages, I have uploaded it here.

If you need any more info I will try to provide it.
Thank you very much for your time.

#9

To answer a question that was asked at the beginning of this thread, when the
kernel panics nothing is logged! Why? By definition, a kernel panic indicates a
condition from which there is no recovery. Attempting to write to a log file
could cause disk corruption and severe data loss.

If you have a serial console, you can tell the system to log data there. If not,
you can always see the last few lines of the crash dump on virtual terminal 10;
however, you must have selected that terminal before the crash. If the crash
can be triggered by a given command, then enter the string “sleep 10 ;
<bad_command>”, then switch to CTRL-ALT-F10. That gives you 10 seconds to get
there. If the panic occurs “at random”, then just switch to that terminal and
wait. when the panic occurs, copy the info from the screen, or take a picture of it.

#10

On 07/03/2012 06:36 PM, psaxioti wrote:
> If you need any more info I will try to provide it.

ok, i’m not good enough at understanding the messages log (even when
none are fuzzy) so i can’t know if the problems you had all along (from
11.3) are the same as you now are having now…pity you don’t have the
kernel panics captured in your log for 11.3/4 and 12.1…

OH: have you checked to see if the manufacturer has an update for the
BIOS? sometimes those make all the difference!

OH 2: are you overclocking? and, the other BIOS settings: are you sure
they are as needed…especially look at the multiple processor
settings…and…well, i don’t know where to point you: have you
changed any settings off of default?

maybe some of this (or similar search strings) will find someone else
with the same problem, and a clue:

https://www.google.dk/search?q=MSI+k8n+neo4-f+“linux+problem”

i just don’t know if you should post to the factory mail list and ask if
they want a bug report…or, just do the bug report…

Larry/Carlos/whoever: what do you think is best (i’m thinking it is as
likely a hardware problem as software…despite the “it works good in
Windows”)

i can’t add much more to this thread: good luck!

–
dd

#11

First: on a 64bit system, I would use a 64bit OS. The …32bit… packages you see are there for 32bit compatibility
Second: with the load of repos you have, I wouldn’t dare guessing or saying anything about a potential cause of the issue.

Does the issue exist when booting from a LiveCD / USB medium?
Did you check BIOS settings for the CPU ?

#12

Hello and thank you all,
I run memtest and I had some errors. Although I don’t understand why the system changes behavior with and without the nosmp flag I will check which dimm module is bad and try to replace it.
I will also run on a liveusb or make a clean install (with no repositories apart the official ones) and see if anything change.

The pc is not overclocked and the bios is not the latest. I had the same problems with the latest one, so I reverted to the one I remembered that was working fine (probably the problems were not from the bios as it seems).

All this will take some time, as this is my work pc and I need it working, but I will report back when I have something new.

#13

On 07/04/2012 10:06 AM, psaxioti wrote:
> I run memtest and I had some errors

note 1: read the caveat in my sig: if you don’t know how to do anything
i might suggest (either inside the machine or out) ask!

“some” means more than one, and even just one is far too many…because
every time you try to stuff data into that part of the RAM you will
have a big problem…

but, very happy to hear of the progress!! knowing where to begin is a
BIG part of the fixing…

it could also be a bad socket…or just one, very slightely corroded
contact!

note 2: if yours is a laptop do NOT start taking it apart to do any of
the following (these instructions suppose it is a
desktop/tower/blade/rack-mount/etc):

i would begin by removing all RAM (note which slot each is removed from)
and using a soft eraser (like on the end of a wooden lead pencil)
“erase” all the stuff (which may or may not be visible) off the gold
contacts…just treat it like you are erasing penciled words off of
paper–NO need to apply a lot of pressure OR use a hard eraser (like to
remove ink)

then reinsert the RAM into different slots than the one they came out of…

maybe that is all you need…if you continue to have this range of
problems: do you know how to determine which stick (or socket) is bad?

–
dd http://tinyurl.com/DD-Caveat

#14

Hello again. After one week I am coming back to report my progress.

First of all I have taken out my memory modules, cleaned the contacts and reinstalled in different slots. I run memtest through the whole weekend with no errors at all. But I have to report that this is the case with memtest 4.10 coming with the live enlightment usb (based on opensuse 11.4). Trying to run memtest 4.20 from live kde usb build 0451 (opensuse 12.2) I cannot pass test 7, it gives so many errors that I believe that it is a problem with memtest, otherwise I think I couldn’t even boot. I also would like to remind you that with the nosmp flag I have no problems, so I wouldn’t expect really a memory problem.

Secondly I removed almost all the repositories that I had. Now I am left only with the following:

# | Alias                 | Name                       | Enabled | Refresh | Priority | Type   | URI                                                             | Service
--+-----------------------+----------------------------+---------+---------+----------+--------+-----------------------------------------------------------------+--------
1 | packman.inode.at-suse | Packman Repository         | Yes     | Yes     |   99     | rpm-md | http://packman.inode.at/suse/Factory/                           |        
2 | repo-debug            | openSUSE-12.2-Debug        | No      | Yes     |   99     | NONE   | http://download.opensuse.org/debug/distribution/12.2/repo/oss/  |        
3 | repo-debug-update     | openSUSE-12.2-Update-Debug | No      | Yes     |   99     | NONE   | http://download.opensuse.org/debug/update/12.2/                 |        
4 | repo-non-oss          | openSUSE-12.2-Non-Oss      | Yes     | Yes     |   99     | yast2  | http://download.opensuse.org/distribution/12.2/repo/non-oss/    |        
5 | repo-oss              | openSUSE-12.2-Oss          | Yes     | Yes     |   99     | yast2  | http://download.opensuse.org/distribution/12.2/repo/oss/        |        
6 | repo-source           | openSUSE-12.2-Source       | No      | Yes     |   99     | NONE   | http://download.opensuse.org/source/distribution/12.2/repo/oss/ |        
7 | repo-update           | openSUSE-12.2-Update       | Yes     | Yes     |   99     | rpm-md | http://download.opensuse.org/update/12.2/                       |

I have made “zypper dup” with these repositories and manually uninstalled the remaining packages that were left. So now the system has packages only from the official repositories + some from packman repository + dropbox (from network repository) + nedit (from editors repository).

Even like this I get kernel panics. I made two photos of them (here and here). As far as I can see they both look the same and the only thing I can understand, although I am not a programmer, is that the system doesn’t like something that has to do with posix-cpu-timers.c file. I have tried to find something about this on the web but with no success.

I guess that I will have to live with only the one cpu, except if you can suggest me to write to some other place/forum where some kernel developers might look at it and might find a problem, either with my system or the kernel.

#15

On 07/10/2012 09:46 AM, psaxioti wrote:
> if you
> can suggest me to write to some other place/forum where some kernel
> developers might look at it and might find a problem, either with my
> system or the kernel.

i can’t help more (maybe others here can–so check back)…

but, if you want to try the more technically capable you could join an
openSUSE mail list…give a smooth run down of your symptoms and what
you have tried, so far to solve the problem…and, reference this thread
http://forums.opensuse.org/showthread.php?t=476488

i’d suggest you begin here:
http://en.opensuse.org/openSUSE:Communication_channels#Mailing_lists

how to join, post, and etc is all explained in those few wiki pages…

see the link on that page and read about “Mailing list netiquette” and
then join the opensuse@opensuse.org list…

who knows, maybe the first thing they do is tell you to log a bug…

when you finally get it resolved, please come back here and report the
found problem and solution (that way we all learn something AND those
folks who google in see what is what…)

good luck.

–
dd

#16

On 2012-07-10 09:46, psaxioti wrote:
> Even like this I get kernel panics. I made two photos of them (‘here’
> (https://www.dropbox.com/sh/nq31j0cmrsa7hxl/PxUNnxtfYM/IMG_5570.JPG) and
> ‘here’
> (https://www.dropbox.com/sh/nq31j0cmrsa7hxl/PxUNnxtfYM/IMG_5571.JPG)).

The screen shows that the kernel says “kernel bug”, thus report directly
that to Bugzilla.

Your second photo gives me error 404.

–
Cheers / Saludos,

Carlos E. R.
(from 11.4 x86_64 “Celadon” at Telcontar)

#17

OK, sorry wrong link for the second photo. Since I cannot edit the post I have here the correct link, although they are very similar, if not identical.

So, I will keep monitoring this post for any suggestions and I will subscribe to the mailing list/bugzilla.
In case the problem is resolved I will post here the solution.

Thanks everybody for their help.

#18

I’m confused - you had memory errors from memtest, so you moved the ram sticks around and then you got more errors from memtest, but you think the problem is with memtest and not your ram??? I would strongly suggest that you get another stick of ram.

I, coincidentally, have the exact same M/B and CPU as you do and its been nothing but stable for me. I’m running 12.1 64-bit.

One other thing you can try, is in that BIOS you can underclock the CPU and RAM - try that and see if it becomes more stable. My guess is that yes, it will.

#19

On 07/10/2012 09:26 AM, quantamm wrote:
>
> I’m confused - you had memory errors from memtest, so you moved the ram
> sticks around and then you got more errors from memtest, but you think
> the problem is with memtest and not your ram??? I would strongly suggest
> that you get another stick of ram.
>
> I, coincidentally, have the exact same M/B and CPU as you do and its
> been nothing but stable for me. I’m running 12.1 64-bit.
>
> One other thing you can try, is in that BIOS you can underclock the CPU
> and RAM - try that and see if it becomes more stable. My guess is that
> yes, it will.

The kernel panic happens because interrupts are disabled at line 1200 of
kernel/posix-cpu-timers.c and they must be enabled there. It is hard to see how
a memory error could cause this.

I think we can rule out most kernel bugs in this section as there have been no
changes since 2010; however, a good test would be to boot a 12.1 Live CD and see
if the system works OK with a 3.1 kernel.

As it works with only 1 CPU enabled, trying underclocking is a good suggestion
as there may be a subtle timing error in the CPU or support chips.

#20

The story with memtest, to clear the thing, is the following. At the beginning I run memtest 4.20 overnight and I found it giving a lot of errors at test 7, it was as the whole memory was bad. I reported this and till the weekend I cleaned the contacts and changed the slots. During the weekend I run a new memtest run, this time 4.10 (I had changed the live usb I had), and on monday morning it was still running with no errors reported (after 70 passes if I am not wrong). Then just for “fun” I run with memtest 4.20 and it was giving again a lot of errors at test 7.
As I cannot believe that the system would even run with all the errors that memtest 4.20 was reporting, I guess that something changed from memtest 4.10 to 4.20. Then on my favor I guess stands also the fact that with the nosmp flag I run stable. I would really love to have memory problems and sort the problem out, but it doesn’t seem it is so.

I will check underclocking the CPU, if the bios supports it as I haven’t overclocked it (I think everything is at auto, I haven’t make any changes from the default). I would appreciate if quantamm could send me his bios settings to see if I have anything wrong there.

The kernel panic happens because interrupts are disabled at line 1200 of kernel/posix-cpu-timers.c and they must be enabled there. It is hard to see how a memory error could cause this.

So can I do something for the above? do I have to compile the kernel with some changes? some bios setting that I can check?

I will also try to boot 12.1 from live usb and see what happens.

Thank you very much.