Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 9 May 1998 15:21:37 +0200
From:      Stefan Esser <se@FreeBSD.ORG>
To:        Stefan Bethke <stefan@promo.de>
Cc:        FreeBSD Hackers <freebsd-hackers@FreeBSD.ORG>, Stefan Esser <se@FreeBSD.ORG>
Subject:   Re: ISA PnP / snd PnP developments?
Message-ID:  <19980509152137.30068@mi.uni-koeln.de>
In-Reply-To: <326400.3103442037@d254.promo.de>; from Stefan Bethke on Wed, May 06, 1998 at 11:13:57AM %2B0200
References:  <326400.3103442037@d254.promo.de>

next in thread | previous in thread | raw e-mail | index | archive | help
On 1998-05-06 11:13 +0200, Stefan Bethke <stefan@promo.de> wrote:
> --On Die, 5. Mai 1998 23:47 Uhr +0200 "Stefan Esser" <se@freebsd.org> wrote:
> 
> 
> > On 1998-05-05 11:37 +0200, Stefan Bethke <stefan@promo.de> wrote:
> >> Btw., you might want to adjust the comments in sys/i386/isa/pcibus.c,
> they
> >> might prove to be somewhat misleading.
> > 
> > Hmmm, what's wrong with the comments in that file ?
> > Under which version of FreeBSD, BTW ?
> 
> 1.27.2.1 (RELENG_2_2_6_RELEASE, RELENG_2_2) says:
> ..
> /*-----------------------------------------------------------------
> **
> **      The following functions are provided by the pci bios.
> **      They are used only by the pci configuration.
> **
> **      pcibus_setup():
> ..
> 
> And from what I currently can see in that file, it doesn't do any BIOS
> calls.

I see ...

Well, the comment is a little misleading, but not wrong,
nonetheless ... :)

In fact, the functions listed in that comment are really 
primitives provided by the PCI BIOS. The PCI code in 2.2
originates from a time, when the PCI BIOS could not be
trusted, and since the PCI BIOS is a PC only thing (not
defined for processors other than Intel x86, anyway), we
just implemented those functions based on register accesses
instead of BIOS calls.

The PCI BIOS uses terrible parameter passing conventions,
and although we have BIOS32 support in the kernel, I did
not rewrite the low-level PCI code to take advantage of
the BIOS, yet. Only two primitives should rely on the BIOS,
anyway, IMHO: Testing for the PCI configuration mechanism
to use, and the function that returns the highest PCI bus
number that has to be probed.

Regards, STefan

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19980509152137.30068>