RE: Subsystem Vendor and Device IDs

[<mek@sco.com>]

	From pci-sig-request@znyx.com Tue Jul 30 06:58 EDT 1996
	From: ingvar_berg@x400.icl.co.uk
	Subject: RE: Subsystem Vendor and Device IDs

	It makes a big difference if the controller lives on the mb, because 
	that defines the implementation to the OS, by combining the mb ID and 
	the controller ID. One possible way to get a "full" ID would be that 
	the enumerator SW uses the mb ID as subvendor ID if the subvendor ID 
	field is empty in HW.

But, this is in no way any part of the PCI standard. Motherboard ID? Where
does that live? Maybe you can intuit it from, say, Host Bridge vendor and
device
id or some such, but all you'd be doing is providing a 'convention,' not
a standard.

An interesting idea though; use the vendor and device id for the host
bridge
as the 'motherboard' id. It'd probably work for all the simpler PCI
motherboards
out there.

	As for the early, buggy BIOSes: we're discussing how systems delivered 
	after January 1, 1997 has to be implemented. Of course one must require 
	that they have a BIOS that behaves well...

Is that the date Subsystem Vendor and Device ID become mandatory for Win
'95?

	/Ingvar

Matt


§Œy