Subject: Re: PCI and subsystem IDs.
To: Michael Richardson <mcr@sandelman.ottawa.on.ca>
From: Chris G. Demetriou <cgd@netbsd.org>
List: tech-kern
Date: 10/06/1998 11:58:13
cgd@netbsd.org (Chris G. Demetriou) writes:
> Michael Richardson <mcr@sandelman.ottawa.on.ca> writes:
> >   Do we currently have any support for probing based on PCI Subsystem Vendor
> > IDs and Subsystem IDs?
> 
> Some devices certainly do check those things, yes.

In case it's not clear: they check them by digging the data out of the
configuration space header, and checking it.  8-)

Right now, we only provide the vendor & product IDs, and the
class/subclass/interface/revision register.  Those are good enough to
support the vast majority of the devices we've run into.

It's not at all clear to me that we should provide support for them in
the common pci device attach args structure, because then the generic
PCI code needs to worry about _finding_ them.  (They don't exist at
all in type-1 headers, and exist at different places in type 0 and
type 2 headers.)

I mean, if we're going to start providing these bits of
variable-header information, we've started down the slippery slope,
and most drivers really don't need that information.




cgd
-- 
Chris Demetriou - cgd@netbsd.org - http://www.netbsd.org/People/Pages/cgd.html
Disclaimer: Not speaking for NetBSD, just expressing my own opinion.