[CentOS] VIA C3 Nehemiah Locks Up

Thu Mar 22 05:37:33 UTC 2007
Jed Reynolds <lists at benrey.is-a-geek.net>

Chris Miller wrote:
>
> Hello,
>     I've got an EPIA-M series motherboard in a system that has locked 
> up twice under random conditions. Nothing is logged to the console or 
> logs, the system just hangs.
>
> The system is installed with CentOS 4.4 and running the 
> 2.6.9-42.0.8.EL i686 kernel. I am aware that older C3 processors had 
> the CMOV problem, but that is resolved in the later boards (procinfo 
> below). I've seen a few posts about similar problems, but it's not 
> clear whether this is a bad board, or if there's an issue with the 
> kernel.
>
> cpuspeed was running, I've since disabled and stopped the service as 
> it's been rumored to cause problems. This was mentioned in this bug 
> report that is still open :
>
> http://bugs.centos.org/view.php?id=1270
>
> Can anyone confirm that there are no known problems with the VIA C3 
> Nehemiah Model 9 and CentOS, or if some (kernel) tweaks are required 
> to make these systems stable?
>

Try running a large rsync to system with "hdparm -d" DMA turned on. I've 
read a thread on Via's forums that some of their southbridge chipsets do 
not handle DMA contention and freak out. Turn DMA off using
 hdparm and try the rsync again. I was able to get around this by 
turning dma off for the hard drive.

http://forums.viaarena.com/messageview.aspx?catid=28&threadid=60131&STARTPAGE=3&FTVAR_FORUMVIEWTMP=Linear

I've had a sales engineer that works a lot with Via systems say 
something to me of this topic:
> Are you using any additional peripherals (wireless LAN, USB devices,
> etc...)?  I've found that the DMA issues are usually triggered when a
> DMA-enabled PCI device enters the picture, however, I have seen reports
> of these problems arising in situations where considerable network data
> transfer occurs simultaneously with heavy disk activity.


Jed