Hm...  Yes space is a requirement.  I'm trying to make an embedded
system with a really small footprint.  I chose to get the integrated
processor because it was much cheaper to do it that way and I knew i'd
never be upgrading it.... i'd just toss the whole thing if it gets out
of date.

Specifially my requirements are a microatx board with 3 or 4 pci slots,
an integrated LAN on the motherboard, and it must run either Via C3 or
celeron.  Something powerful but conserves power at the same time..
this is an embedded system.  I'm wondering if it would be better if i
ordered this $79 motherboard with SiS chipsets on it instead:
http://www.ecs.com.tw/products/p6vem3_3x.htm   
SiS seems to be pretty good from what i've heard.  I just don't know
about NetBSD support.

Another option I was thinking of was to try to get FreeBSD running on
this system, then start replacing some of the memory code that netbsd
has with freebsd code..  (yes I know this has to be very carefully
done).

Any other suggestions are welcome and if you know any kernel hackers
that could help explain more to me whats going on would be much
appreciated..  I'm not giving up very easily..  I'll go through the
kernel memory parts line by line if i have to.

Greg


On Thu, Nov 14, 2002 at 10:57:10AM -0500, Keith Mastin wrote:

> ...nothing off the top of my head, but I do know from experience that just 
> about any board with integrated devices have turned into bigger headaches 
> and cost more time and money in configuring them for anything but windoh$ 
> than they are worth. Best to check the supported hardware list, and if you 
> are going to the store with a fistful of $$, try a bare board and add on 
> the devices you need through expansion cards, as you have a better chance 
> of getting true hardware (as opposed to "assisted") than with an 
> integrated board.
> 
> My understanding is you have a baby-atx board... is physical space a 
> limitation? A 1U or 2U case, for example?
> 
> IMHO, getting a kernel hack for this chipset is not a real option, unless 
> the coders have nothing else really pressing to do at the moment.
>