Kernel crashed, but no oops information!

[wolaile]

I have tried for several days, and I have googled hard, but I can't find the results. So now let me describe my problem as follows:

I have installed an Ubuntu 9.10 on a VMWare machine (system A), and I have installed a Redhat 9 on a VMWare (system B, and yet I have installed an Ubuntu 9.10 on a physical machine (system C).
And I use painic(...) function and/or use die_nmi(...) function in a driver module or in the kernel, also I have tried using the command "cp /dev/zero /dev/mem".

In my testings, all my above methods can crash Linux kernel (keyboard and mouse are dead), but in all the cases, there is no oops information at all.
For system A and system B, the screen remains just as before crash, I mean, the screen is still the desktop and etc.
For system C, system reboots quickly.

So I am strange that how can I see oops information.

[wolaile]

Is it possible that Ubuntu has no oops mechanism?

[robertdaleweir]

Is it possible that Ubuntu has no oops mechanism?

Hi
I assume you mean error messages. Try <dmesg | more> as root (#dmesg | more). It should show any info which should get logged. Most distros have the dmesg function.
Cheers...
Robert

[wolaile]

Hi, compared with Windows, when Windows kernel crashes, there will often (not always) occur a BSOD. The BSOD is "Blue Screen".
And as for Linux, there is a corresponding mechanism, and it is called "Black Screen". So my problem is that I can't see any "Black Screen" while my Linux kernel crashes.
I think dmesg may be helpless, because printk information may be unluckily in the buffer but not into the file while Linux kernel crashes.

So let me ask my question in another way:
If you run "sudo cp /dev/zero /dev/mem" in your terminal, will your machine reboot or go into "Black Screen"? And what is the version of your Linux?

[Kloschüssel]

If you run "sudo cp /dev/zero /dev/mem" in your terminal, will your machine reboot or go into "Black Screen"? And what is the version of your Linux?

what the heck are you trying to accomplish? I don't think cp is the right kind of command to do copy zeros to /dev/mem! Furthermore, if it works (which I quite don't believe it would ever), it would clear your memory and reset it to zeros. Obviously also parts of your apps and OS will go down.

[wolaile]

Hi, I just use "cp /dev/zero /dev/mem" to crash Linux kernel easily.



Now, after kernel crashes and then I reboot system, I have found there will occur a red button on the top-right corner of the screen. And it is obvious that this red button is for reporting "panic information" to developers.
So I can say that I have made some progress in this problem.

[wolaile]

I have found this directory: /var/crash
And this is the key.

[robertdaleweir]

Hi, compared with Windows, when Windows kernel crashes, there will often (not always) occur a BSOD. The BSOD is "Blue Screen".
And as for Linux, there is a corresponding mechanism, and it is called "Black Screen". So my problem is that I can't see any "Black Screen" while my Linux kernel crashes.
I think dmesg may be helpless, because printk information may be unluckily in the buffer but not into the file while Linux kernel crashes.

So let me ask my question in another way:
If you run "sudo cp /dev/zero /dev/mem" in your terminal, will your machine reboot or go into "Black Screen"? And what is the version of your Linux?

Hi woliale
When I issue the command you mentioned I initially get a Black Screen and then it immediately Reboot. My version currently is 2.6.32.11-99.fc12.x86_64. Cheers...
Robert

[tinytux]

what the heck are you trying to accomplish? I don't think cp is the right kind of command to do copy zeros to /dev/mem! Furthermore, if it works (which I quite don't believe it would ever), it would clear your memory and reset it to zeros. Obviously also parts of your apps and OS will go down.

I guess you spotted it

The guy is not trying to accomplish anything productive
He is just f... around. I want for proof

1. He did not explain the purpose of its experimentation
2. He don't talk about the logs
3. He does not talk about tail and dmesg

Which is for me the first step to debunk a kernel panic.

I think he is just trying to "kill the kernel" in a sick purpose

[wolaile]

Thanks all! Yes, I just want to study Linux debugging.