From gibbs@estienne.cs.berkeley.edu  Sat Apr  1 15:15:29 1995
Received: from narnia.hip.berkeley.edu (narnia.HIP.Berkeley.EDU [136.152.64.57]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id PAA09265 for <FreeBSD-gnats-submit@freebsd.org>; Sat, 1 Apr 1995 15:15:28 -0800
Received: (from gibbs@localhost) by narnia.hip.berkeley.edu (8.6.11/8.6.9) id PAA25490; Sat, 1 Apr 1995 15:16:14 -0800
Message-Id: <199504012316.PAA25490@narnia.hip.berkeley.edu>
Date: Sat, 1 Apr 1995 15:16:14 -0800
From: gibbs@estienne.cs.berkeley.edu
Reply-To: gibbs@estienne.cs.berkeley.edu
To: FreeBSD-gnats-submit@freebsd.org
Subject: PCI devices still probe/attach after being disabled via usercfg
X-Send-Pr-Version: 3.2

>Number:         291
>Category:       kern
>Synopsis:       PCI devices still probe/attach after being disabled via usercfg
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    se
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sat Apr  1 15:20:00 1995
>Closed-Date:    Sun Apr 12 10:05:29 PDT 1998
>Last-Modified:  Sun Apr 12 10:05:41 PDT 1998
>Originator:     gibbs
>Release:        FreeBSD 2.1.0-Development i386
>Organization:
>Environment:

	

>Description:

	PCI devices don't follow the same rules as all other types of
	devices in the way they are probed.  PCI devices will steal
	unit numbers from non PCI devices as well as probe devices
	that was disabled.  This is probably because most drivers only
	look to see if the unit number they are passed is < NFOO and
	userconfig does not update NFOO.  Even updating NFOO would
	not be enough since (as in the case of the ahc driver) other
	units may be on other types of buses and a PCI device will just
	take the place of an EISA or VL device.

	Both of these behaviors will make userconfig less powerful and
	more confusing to users.

>How-To-Repeat:

	Configure a kernel with both eisa and PCI ahc devices.  Attempt
	to disable the PCI device.

	For stealing, configure one EISA and one PCI ahc device.  Put
	two aic7870 based controllers in the machine.

>Fix:
	
	Haven't looked yet.

>Release-Note:
>Audit-Trail:
Responsible-Changed-From-To: freebsd-bugs->se 
Responsible-Changed-By: wollman 
Responsible-Changed-When: Thu Feb 8 08:44:08 PST 1996 
Responsible-Changed-Why:  
se is Mr. PCI. 
State-Changed-From-To: open->suspended 
State-Changed-By: se 
State-Changed-When: Sat Dec 14 06:11:36 PST 1996 
State-Changed-Why:  
Recent "userconfig" changes point out, that PCI devices can not be 
re-configured or disabled. It would in fact be possible to disable 
a PCI device by resetting one bit in its configuration space command 
register, but the current PCI probe would assume a failure of the  
BIOS to enable that device, and would re-activate it ... 
(The PCI BIOS may have enabled only the boot devices, leaving all  
PCI cards it doesn't know about in a disabled state.) 

From: Studded <Studded@dal.net>
To: freebsd-gnats-submit@freebsd.org, gibbs@estienne.cs.berkeley.edu
Cc:  Subject: Re: kern/291: PCI devices still probe/attach after being disabled via usercfg
Date: Sun, 22 Mar 1998 15:30:37 -0800

 Mail to originator fails with "user unknown."
 
 Doug

From: Studded <Studded@san.rr.com>
To: freebsd-gnats-submit@freebsd.org, gibbs@estienne.cs.berkeley.edu
Cc:  Subject: Re: kern/291: PCI devices still probe/attach after being disabled via usercfg
Date: Wed, 08 Apr 1998 23:26:27 -0700

 Mail ping fails with user unknown.
State-Changed-From-To: suspended->closed 
State-Changed-By: phk 
State-Changed-When: Sun Apr 12 10:05:29 PDT 1998 
State-Changed-Why:  
timed out 
>Unformatted:


