NetBSD Problem Report #3541

Received: (qmail-queue invoked from smtpd); 26 Apr 1997 00:18:16 -0000
Message-Id: <199704260017.RAA23479@adagio.fugue.com>
Date: Fri, 25 Apr 1997 17:17:44 -0700 (PDT)
From: Ted Lemon <mellon@adagio.fugue.com>
Reply-To: mellon@adagio.fugue.com
To: gnats-bugs@gnats.netbsd.org
Subject: le0 fails to probe on DS5000/133
X-Send-Pr-Version: 3.95

>Number:         3541
>Category:       port-pmax
>Synopsis:       le0 fails to probe on DS5000/133
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    gnats-admin
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Apr 26 00:20:01 +0000 1997
>Closed-Date:    Thu May 15 08:21:29 +0000 1997
>Last-Modified:  Wed Aug 07 22:54:42 +0000 2002
>Originator:     Ted Lemon
>Release:        4/25/97
>Organization:
Barefoot Consulting
>Environment:
System: NetBSD adagio.fugue.com 1.2D NetBSD 1.2D (ADAGIO) #8: Sun Apr 20 21:51:46 PDT 1997 mellon@toccata.fugue.com:/NetBSD/current/current/sys/arch/pmax/compile/ADAGIO pmax


>Description:
	During autoconfig on a DS5000/133, it fails to match the on-board
	lance ethernet chip.   This doesn't appear to happen with kernels
	that have root on nfs - I've no idea why.   The code in ioasic_submatch 
	seems destined never to match anything, since it's matching against
	something that appears to be constant.   I don't really understand
	what this code is supposed to be doing, though, so perhaps I'm
	getting it wrong.   In any case, other ioasic drivers don't call
	ioasic_submatch, so I just took that out, and everything worked.
>How-To-Repeat:
	Configure a kernel without root on NFS.   Boot it.   It'll come
	up just fine (assuming all else is well) but will not configure
	the lance.   It does print an error message indicating that the
	lance is not configured, though.
>Fix:
	No guarantees:

Index: if_le_ioasic.c
===================================================================
RCS file: /cvsroot/src/sys/dev/tc/if_le_ioasic.c,v
retrieving revision 1.6
diff -c -2 -r1.6 if_le_ioasic.c
*** if_le_ioasic.c	1997/03/17 03:19:13	1.6
--- if_le_ioasic.c	1997/04/26 00:06:44
***************
*** 91,96 ****
--- 91,98 ----
  	struct ioasicdev_attach_args *d = aux;

+ #if 0
  	if (!ioasic_submatch(match, aux))
  		return (0);
+ #endif
  	if (strncmp("lance", d->iada_modname, TC_ROM_LLEN))
  		return (0);
>Release-Note:
>Audit-Trail:

From: Jonathan Stone <jonathan@netbsd1.cygnus.com>
To: GNATS administrator <gnats-admin@NetBSD.ORG>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-pmax/3541: le0 fails to probe on DS5000/133
Date: 15 May 1997 01:06:34 -0000

 	`Jonathan Stone' made changes to this PR.

 --- /tmp/gnats012445	Wed May 14 17:59:32 1997
 +++ /tmp/gnatsa12445	Wed May 14 17:59:43 1997
 @@ -19,7 +19,7 @@
  >Severity:       critical
  >Priority:       high
  >Responsible:    gnats-admin (GNATS administrator)
 ->State:          open
 +>State:          feedback
  >Class:          sw-bug
  >Submitter-Id:   net
  >Arrival-Date:   Fri Apr 25 17:20:01 1997

 The cause of the problem is that the pmax and alpha ports share a
 driveer for ioasic-attached lance devices.  The Alpha configuration
 includes an `offset' locator for ioasic-attached devices, but the pmax
 port did not.

 The ioasic lance front-end is derived from the Alpha port and includes
 an ioasic_submatch() which checks that the `offset' locator is either
 wildcarded or matches the non-wildcard value in the config file.

 I've added a wildcarded `offset' locator to all the pmax config entries.
 That will cause config(8) to  output wildcarded `offset'  locators.
 The ioasic_submatch() call should now work on any  5000/133 kernel.

 Note that the `offset' locator is contentious; the pmax documentation
 describes the ioasic space as composed of fixed-sized `subslots',
 rather than arbitrary offsets. The alpha port maintainer says teh the
 correct design is a flat, linear address space with an offset locator.
 I think configuring devices via a subslot, (possibly with a
 device-specific offset) is a better design.

 Either way,  auto-configuration of le0 at ioasic?  should now work.

From: Jonathan Stone <jonathan@netbsd1.cygnus.com>
To: GNATS administrator <gnats-admin@NetBSD.ORG>
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-pmax/3541: le0 fails to probe on DS5000/133
Date: 15 May 1997 08:22:30 -0000

 	`Jonathan Stone' made changes to this PR.

 --- /tmp/gnats015725	Thu May 15 01:21:13 1997
 +++ /tmp/gnatsa15725	Thu May 15 01:21:30 1997
 @@ -19,7 +19,7 @@
  >Severity:       critical
  >Priority:       high
  >Responsible:    gnats-admin
 ->State:          feedback
 +>State:          closed
  >Class:          sw-bug
  >Submitter-Id:   net
  >Arrival-Date:   Fri Apr 25 17:20:01 1997

 Fixed by adding an offset locator to  ioasic-attached devices on pmaxes.
>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.