🚀 go-pugleaf

I have a 4/280 here, with which I run a cg9 frame buffer, and I run X11R6
(built from source) over SunOS 4.1.4. Other cards installed in the chassis
are a -1217 SCSI card, and -1203 ALM-2, and two -1451 32 meg RAM boards.

In the few months I've had the 4/200 board, it has sporadically suffered
serious failures, giving BAD TRAP and DATA FAULT exceptions, sometimes
several within a short period, sometimes only one in a period of several
weeks. I have suspected many causes for these strange failures, and so far
I haven't really deduced what the problem might be. 

However, when thinking about it today (since it's given three exceptions
in the space of about an hour), I've started to wonder whether the problem
might actually be software-related. Why I say this is that the software
support in X11R6 for cg9's is incorporated via a patch written by an
individual, and this cg9 support code is very minimalistic in how much
functionality it gives the 24-bit Xserver when working with the cg9. 

It's just possible that somehow, because I'm running X natively here on 
the cg9, the minimalistic cg9 support that I've patched into the Xsun24 
code in X11R6 is causing things to break.

What do you think? Is this a possibility? I have another 4/200 on the way
to try out, but if it is a problem with X, the cg9, and SunOS fighting
together and breaking the system, it's doubtful that the replacement 4/200
board will help. 

I will soon by upgrading (hopefully!) to a 4/380, but in the meantime it's
worrying that the machine is keeling over in no obvious way. Since this is
my main server, I can't have it failing for no apparent reason. The other
machines I run here (an SS1 as my gateway, and a 4/110 in a 3/60 case as a
special-interest BBS machine) almost never fail. 

I'm still opting towards thinking either there really is an intermittent
problem with the 4/200 board I have now, or something in the SunOS 4.1.4
kernel is not quite right, and because of the load the 24-bit Xserver puts
on it, it's tripping up on this... 

Any thoughts?

Regards,

Craig.
-- 
             Craig Dewick. Send email to 'cdewick@jedi.apana.org.au'
                 The Core Matrix Foundation - Sydney, Australia
                        http://jedi.apana.org.au/core.html
 Always striving for a secure long-term future in an insecure short-term world