From owner-freebsd-bugs Sun Dec 1 06:40:03 1996 Return-Path: owner-bugs Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id GAA16564 for bugs-outgoing; Sun, 1 Dec 1996 06:40:03 -0800 (PST) Received: (from gnats@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id GAA16558; Sun, 1 Dec 1996 06:40:02 -0800 (PST) Date: Sun, 1 Dec 1996 06:40:02 -0800 (PST) Message-Id: <199612011440.GAA16558@freefall.freebsd.org> To: freebsd-bugs Cc: From: se@freebsd.org (Stefan Esser) Subject: Re: conf/2130: Installation bug Reply-To: se@freebsd.org (Stefan Esser) Sender: owner-bugs@freebsd.org X-Loop: FreeBSD.org Precedence: bulk The following reply was made to PR conf/2130; it has been noted by GNATS. From: se@freebsd.org (Stefan Esser) To: pgiffuni@fps.biblos.unal.edu.co Cc: freebsd-gnats-submit@freebsd.org Subject: Re: conf/2130: Installation bug Date: Sun, 1 Dec 1996 15:27:14 +0100 On Nov 30, pgiffuni@fps.biblos.unal.edu.co wrote: > > >Number: 2130 > >Category: conf > >Synopsis: Installation bug > >Description: > No PCI support can be removed before installing the OS! This is the way it is uspposed to be ... We could add a DISABLE function to the config code, which would directly write some bit in the PCI device's config space command register. This would completely deactivate the device (except for config space accesses) and the PCI code could respect this bit (while it currently tries hard to gat all devices activated that the PCI BIOS might have forgotten, for example behind a PCI to PCI bridge ...) I'm not sure there are really good applications for a ONE time dactivate of a PCI device. And the PCI code assumes that all devices are auto- configuring. And for that reason, a device might be found at a different address (and even in a different 'slot') on each reboot without violating the specs. (This will happen, if you add cards with PCI to PCI bridges and bus. Then, even the bus number of previously installed devices may come out differently ...) Regards, STefan