Bug 112983

Summary: AMD64 reboots spontaneously
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Klaus Singvogel <kssingvo>
Component: KernelAssignee: Andreas Kleen <ak>
Status: RESOLVED INVALID QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P5 - None    
Version: Preview 2   
Target Milestone: ---   
Hardware: 64bit   
OS: All   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: hwinfo of e162.suse.de

Description Klaus Singvogel 2005-08-25 13:49:19 UTC 
After reboot, I login into my AMD64, wait that all my KDE tools got started 
and running, and then I just need to start konqueror, try to login at 
bugzilla.novell.com and see how my machine stops running... 
and a second later the BIOS (reboot) is visible. 
   
Before question comes up: I installed latest BIOS (v2.30 2005/07/29) for my 
ASROCK K8 Combo-Z motherboard this afternoon and it still happens.  :-) 
 
The hostname is: e162.suse.de
Comment 1 Olaf Kirch 2005-08-25 13:57:58 UTC 
Andi, this will be of interest to you. 
 
Klaus, does this happen reliably when starting konqueror, i.e. could this 
be network related? 
 
Please append the hwinfo output. Thanks!
Comment 2 Klaus Singvogel 2005-08-25 14:04:37 UTC 
I'm not able to guarantee its reliability, as you know the 
"Vorfuehreffekt". :-) But it happened now in 3 of 3 tries.
Comment 3 Klaus Singvogel 2005-08-25 14:05:24 UTC 
Created attachment 47558 [details]
hwinfo of e162.suse.de
Comment 4 Klaus Singvogel 2005-08-25 14:28:08 UTC 
Regarding comment#1 "network related": I doubt that this is the case, as other 
network related tools are already started: kopete to login iat the IM of 
groupwise.novell.com, and no crash happens. 
 
My view is this: I already contacted bugzilla.suse.de and it doesn't crash 
neither. But after the login is done and the buzilla page is getting displayed 
(with all its bloats and whistles, like javascript), the machine crashes. 
 
My assumption is (but very unsure regarding this), the kernel crashes as soon 
as real computing is necessary. Maybe its related to Cool'n'Quiet, maybe CnQ 
and network together? But I really don't know. 
 
Before I forgot: the builtin ethernet connection wasn't running under SuLi 
9.3. Therefore the machine is having a second ethernet card, with which the 
only network is done. Don't stumble on this if you read the hwinfo file. :-)
Comment 5 Klaus Singvogel 2005-08-25 15:16:14 UTC 
This is the console output: 
CPU 0: Machine Check Exception:                4 Bank 4: b200000000070f0f 
TSC 79b93e4fbe  
Kernel panic - not syncing: Machine check
Comment 6 Olaf Kirch 2005-08-25 15:21:17 UTC 
Indeed. I was expecting an MCE.  
 
Most likely a HW/driver problem. Try using a different network card.
Comment 7 Klaus Singvogel 2005-08-25 16:05:08 UTC 
Ok. This was a good advice. 
I exchanged the DLINK DFE-530TX netword card against a Intel EtherExpress 
(exact model unknown) and the bug is no longer reproduceable. 
 
So I suggest to close this problem.
Comment 8 Andreas Kleen 2005-08-25 16:16:40 UTC 
The machine check decoded is:  
  
CPU 0 4 northbridge TSC 79b93e4fbe   
  Northbridge Watchdog error  
       bit57 = processor context corrupt  
       bit61 = error uncorrected  
  bus error 'generic participation, request timed out  
      generic error mem transaction  
      generic access, level generic'  
STATUS b200000000070f0f MCGSTATUS 4  
Kernel panic - not syncing: Machine check   
  
This happens when a page table entry points to not existant memory  
or in most cases it's some kind of hardware bug: the hardware doesn't  
respond to an MMIO access when it should. 
  
One workaround is nomce   
 
I'll close it as suggested.