Unable to boot Debian : "PANIC : Double fault"
 

Hello everyone,

I recently decided to build up a Linux computer with the following configuration
- Motherboard : MSI KT4 Ultra
- CPU : Athlon XP 2000+
- RAM : Kingston ValueRAM (2 DIMMs, I put them in all possible arrangements)
I doubt other compoment details are relevant, but if you do think so, feel free to ask.

I installed Debian 5.05 on that machine ; it worked all fine.
However, given that my mobo would not support wake on lan (and that I needed it given that my goal was to build some sort of server), I searched for a new one that did, I then installed an ASRock K7S41GX.

Since doing so, I have not been able to boot Linux ; after doing some fixing on my part, I have come to the point where I have the very same error repeating times and times again every time I try to boot ; here's the error message I get and how I get to it.

GRUB loads like a charm, then Linux starts booting and stuff happens :

Checks filestystem, initrd and stuff (I actually don't understand that part, but it seems to work OK).

Decompressing Linux... Parsing ELF... done.
Booting the kernel.
[ 0.160010] PANIC: double fault, gdt at ######## [255 bytes]
[ 0.160010] double fault, tss at ########
[ 0.160010] eip = ########, esp = ########
[ 0.160010] eax = ########, ebx = ########, ecx = ########, edx = ########
[ 0.160010] esi = ########, edi = ########

Where "########" stands for a series of eight hexa chars (I assme these might be memory adresses ?).


Is there a way I can understand this ? Is this a known bug and is there a solution ? (Well I doubt it's known given that I found nothing on the internet.)
Any help would be most appreciated.


A few more details :
- The ########'s are NOT the same on each attempt to boot.
- When I try to reinstall Linux, I also got various error messages (Mosty "PANIC: attempted to kill init" or "PANIC: attempted to kill the idle task")
- Rescue mode fails the same way re-install does.

- For some reason I just can't explain I ONCE happened to manage to start the installer process, which unfortunately ended up hanging on "Detecting Network hardware". (I attempted to disable the onboard network device to no avail).


Thanks in advance.

Try running memtest on the ram. Also, if you are overclocking, set every thing back to the default settings.

I am not as knowledgeable as I would like to be about the kernel, but I'd suggest this possibility:
When you first installed Debian, you had one motherboard. The installation wrote a system map file (located in /boot). Then you replaced the motherboard, so the hardware has changed. When you try to boot, the hardware the kernel finds doesn't match the system map. So, what to do? PANIC

Can you boot a liveCD? If so, from the opening menu (before the liveCD actually boots) select memtest (to test RAM). If you get no errors, then boot the liveCD. If all goes well, then you should consider deleting the current installation, and re-installing so that the new installation can map the current hardware.

Note: see the first sentence of my reply. I may be completely off the mark, but it makes sense to me, and it's what I would do if I were having the same problem.

First off, thanks for answering,

I'm currently performing the memtest, I'll keep you updated on whether it finds something or not.

Now, I've indeed considered the possibility that the system won't boot because so many things have changed, but I've unfortunately been unable to even start the install program (I actually have been unable, except on a single occasion, to go farther than the first menu. As soon as the memtest's performed, I'll try to find the exact error codes given by the installer.

Now, maybe I need to perform some sort of "clean" install ? I mean, starting with a "blank", formatted (sorry if it's not the exact word, English's not my native language) hdd ? Do you think it could change something ?

So here's a heads up :

I've ran memtest which, to my susrprise, did crash (although not reporting any errors or anything, just hanging), it did so during its 7th test (Random number sequence), so this calls for two questions :

- Could memory crashe be motherboard-related ?
- Why did it work before ?
- Is there a way I can circumvent this problem ?(ie. slowing the ram's frequency down or whatever)

Thanks in advance

I'm wondering if that ram is compatible with your motherboard. You might want to double check that. It really sounds like a hardware problem. Double check your connections to the motherboard. Make sure the motherboard standoff posts are in the right places for your motherboard. If they are not, they can short out the motherboard leading to all kinds of erratic problems. Also, double check your settings in the bios setup and set ram timings and voltages to automatic. You could also try running memtest with one stick of ram at a time.

I did memtest with each stick of RAM separately, to no avail ; the system crashed at the exact same time (test 7).
I'm going to try to find some other DIMMs I might have hanging around somewhere at my place.
I'm going to have a look for the standoffs.

Thanks a lot.

Total shot in the dark, but something along these lines happened to me once.

Is there a motherboard jumper that relates to the RAM frequency or anything like that? If it is set to the highest you could try lowering the jumper setting. I'm not sure if this exists on your board, but you could check the user manual for this information.

Well there is indeed a jumper to set FSB. I set in on 266Mhz (correct frequency for the CPU), do you think I should try lowering that ? Would slow down the pc for sure, but I don't need it to do anything intensive, so...

I just noticed that your having problems with an ASRock board. ASRock is a budget line of boards marketed by Asus and manufactured for Asus by ECS. Just in case you didn't know ECS=PCChips; they are one and the same company. I would be hard pressed to name a motherboard manufacturer with a worse reputation than PCChips. Bottom line, the motherboard may be bad out of the box; a not uncommon occurrence with PCChips. Just something to keep in mind.

If you can't get the board to work with other ram and your Kingston ram checks out OK with memtest on your other MSI motherboard, you can try doing an out of the box build as described here:

http://forum.pcmech.com/showthread.php?t=132409

It's a good way to simplify things to better zero in on what's causing the problem. If you can't get the out of box build to work, rma the motherboard; it's probably defective.

Well, I'll try the board with another DIMM if I manage to find one. Maybe it's just a compatibility issue, although I doubt it.

Anyway, thank you very much for your assistance. My system doesn't work yet, but I've learnt a few things.

I can not speak about your motherboard but I totally agree with kilgoretrout about the pcchips mother boards. I have had quite a bit of experience with them, if you can get a refund or trade out of it do so. If you ever get it going it will only be about 25 to 30 percent as fast as a fair quality mother board. If you run windows xp it will normally take from 3 to 4 minutes just to boot. With linux it is better but very little.