Chipset autodetect -- pro and con

[Niklas Ekström]

On Mon, 30 Nov 1998, Andrew Kohlsmith wrote:

> > I think it should DEFINETLTY need to be seperated.
> 
> heh, I think the opposite..
> 
> > Why?
> > 
> > Too Big. Each chipset got it's own data and specs (don't forget we're
> > talking about 386, 486, LX, BX, VX, FX chipset and they got lots of
> > differences) - combining them all together will give a binary file that
> > will probably won't fit into the flash (not mentioning other features we
> > want to have).
> 
> I kind of doubt it will be too big.  The amount of code to initialize
> the subsystems is NOT that large in comparison to something like the
> generic display and generic disk and various other generic systems.  I
> can't see more than about 32-64k of code to support all of the above
> mentioned chipsets.
> 
> I'm not a BIOS designer by trade but I *am* a designer of embedded
> systems.  The code to handle the hardware is invariably much smaller
> than the code to handle the user.

After alot of thinking about this I've come to the conclusion that I agree
that there should be autodetection of chipsets. I have been reading some
pciset documents from Intel, and as Andrew says there is really only
smaller differences (between the PCI chipsets that is). But all this
depends on one things; if it is possible to, with 100% accuracy, determine
what chipset the BIOS is running on. If this is possible, is see no reason
not to include autodetection. A wild guess is that the code for the
various chipsets (all together) is not more then a couple of kB.

> 
> Why do I think it should be autodetect?
> 
> simple: chipsets aren't THAT different.  There is software out now
> which can differentiate between them fairly accurately.  There are
> subtle differences which can increase performance and extra features
> of the individual chipsets but deep down they're all the same.
> 
> Perhaps we will need to seperate 386, 486 and Pentium chipsets or
> perhaps a 386/486 and a Pentium* class BIOS but really that's about it
> I think.  Autodetection is a really great idea in my opinion. 
> Besides, you *don't* want the end user to look up our site because
> they think they're technical minded, download the wrong flash and
> boom, be dead in the water.  We need bare minimum "generic" BIOS
> support to protect against this because no matter how hard we try, it
> *will* happen and OpenBIOS will be a black name, no matter how good it
> is.
> 
> Andrew
>