On Tue, Jan 30, 2007 at 03:47:57AM -0600, David Young wrote:
> I am seeing some hard-to-believe traps on a Soekris net4521 (AMD Elan
> SC520 processor).  I have included two traps below.

This *seems* to be related to my adjusting the latency timer on my
rtw at cardbus.  LT=0xf8 leads to a panic within a couple of hours.
LT=0x50 leads to a panic, too, but not for about a day.  I believe
LT=0x10 was usual, before.  I have noticed that 0x50 and 0xf8 are both
greater than the LT on the PCI-Cardbus bridge, 0x3f.  I don't think
that matters.  (Well, I am sure it matters, but I don't know how it
would lead to a trap.)  I am going to try LT=0x3f on the rtw, just to
see if it makes any difference.

I am interested to hear interpretations and suggestions concerning these
strange traps.

Dave

-- 
David Young             OJC Technologies
dyoung@ojctech.com      Urbana, IL * (217) 278-3933