NetBSD Problem Report #10129

Received: (qmail 15322 invoked from network); 16 May 2000 18:09:50 -0000
Message-Id: <200005161809.OAA00497@zorkmid.mit.edu>
Date: Tue, 16 May 2000 14:09:27 -0400 (EDT)
From: John Hawkinson <jhawk@mit.edu>
Reply-To: jhawk@mit.edu
To: gnats-bugs@gnats.netbsd.org
Subject: wscons loses when console is serial
X-Send-Pr-Version: 3.95

>Number:         10129
>Category:       kern
>Synopsis:       wscons loses when console is serial
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    drochner
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue May 16 18:10:00 +0000 2000
>Closed-Date:    Fri Oct 06 22:33:04 +0000 2006
>Last-Modified:  Fri Oct 06 22:33:04 +0000 2006
>Originator:     John Hawkinson
>Release:        -current of 15 May 2000
>Organization:
MIT

>Environment:

System: NetBSD zorkmid.mit.edu 1.4Y NetBSD 1.4Y (ZORKMID-$Revision: 1.5 $) #88: Mon May 15 19:31:34 EDT 2000 jhawk@zorkmid.mit.edu:/usr/local/current-src/sys/arch/i386/compile/ZORKMID i386


>Description:
	After booting on a serial console, wscons-based displays fail
catastrophically. Output to /dev/ttyE* is displayed, however no input
can be read. Keystrokes can be read from /dev/wskbd0, however.

Consoles cannot be switched via ctrl-alt-fN, nor by calling
wsdisplay_switch() (fails because sc->sc_displaydv has never been
initialized).

Most poigently, observe a friendly "login: " prompt that accepts
no input and echos no characters.


This didn't occur under 1.4.2.

>How-To-Repeat:
	Boot a serial console. Try to use the physical console. Be stymied.

>Fix:
	it appears from preliminary code inspection that much of the critical
wscons setup only happens if the console is actually on wscons.
I don't really understand what the theory is here, and the documentation
is sparse-to-nonexistant...

--jhawk
>Release-Note:
>Audit-Trail:

From: Matthias Drochner <M.Drochner@fz-juelich.de>
To: jhawk@mit.edu
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: kern/10129: wscons loses when console is serial 
Date: Thu, 18 May 2000 16:14:19 +0200

 jhawk@mit.edu said:
 >  Output to /dev/ttyE* is displayed, however no input can be read.
 > Keystrokes can be read from /dev/wskbd0, however. 

 The assignment of keyboards to displays is not done automatically
 anymore unless the device is used as system console.
 It has to be done later, by "wsconscfg -k", or -easier-
 by uncommenting the "keyboard" line in /etc/wscons.conf.

 best regards
 Matthias



From: John Hawkinson <jhawk@MIT.EDU>
To: M.Drochner@fz-juelich.de
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/10129: wscons loses when console is serial 
Date: Thu, 18 May 2000 13:22:39 -0400

 In message <200005181414.QAA0000011175@zel459.zel.kfa-juelich.de>, Matthias Dro
 chner writes:
 >
 >jhawk@mit.edu said:
 >>  Output to /dev/ttyE* is displayed, however no input can be read.
 >> Keystrokes can be read from /dev/wskbd0, however. 
 >
 >The assignment of keyboards to displays is not done automatically
 >anymore unless the device is used as system console.
 >It has to be done later, by "wsconscfg -k", or -easier-
 >by uncommenting the "keyboard" line in /etc/wscons.conf.

 Indeed, this does the trick.

 Can you speak to why this is this way? Is there some reason that
 "keyboard" is not uncommented by default? It seems to me that the
 handing of output and input devices should be handled similarly, and
 the current state violates the principal of least confusion.

 I would say, at a bare minimum, this needs a doc/CHANGES
 entry, since it was not this way in 1.4.

 --jhawk

From: Matthias Drochner <M.Drochner@fz-juelich.de>
To: John Hawkinson <jhawk@MIT.EDU>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/10129: wscons loses when console is serial 
Date: Fri, 19 May 2000 19:03:34 +0200

 jhawk@MIT.EDU said:
 > Can you speak to why this is this way? Is there some reason that
 > "keyboard" is not uncommented by default?

 It was enabled per default, but got commented out when the
 wsmux code came in. Perhaps there are nasty warnings if
 there is no keyboard to connect (ie in the graphics console
 case). Lennart?

 > I would say, at a bare minimum, this needs a doc/CHANGES
 > entry

 Or in the default /etc/wscons.conf. (There is already the
 "screen 0" line which must be touched if a serial console
 is used.)
 I'll have a look at it next week.

 best regards
 Matthias



From: woods@weird.com (Greg A. Woods)
To: M.Drochner@fz-juelich.de, 
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/10129: wscons loses when console is serial 
Date: Tue,  3 Jul 2001 17:01:15 -0400 (EDT)

 I'm just at a point here were I've brought up a new machine running near
 -current and wanted to have it on serial console, but with wscons too,
 and since there's not been any activity on this PR for over a year I
 thought I'd give it an update....

 [ On Thursday, May 18, 2000 at 13:22:39 (-0400), John Hawkinson wrote: ]
 > Subject: Re: kern/10129: wscons loses when console is serial 
 >
 > In message <200005181414.QAA0000011175@zel459.zel.kfa-juelich.de>, Matthias Dro
 > chner writes:
 > >
 > >jhawk@mit.edu said:
 > >>  Output to /dev/ttyE* is displayed, however no input can be read.
 > >> Keystrokes can be read from /dev/wskbd0, however. 
 > >
 > >The assignment of keyboards to displays is not done automatically
 > >anymore unless the device is used as system console.
 > >It has to be done later, by "wsconscfg -k", or -easier-
 > >by uncommenting the "keyboard" line in /etc/wscons.conf.
 > 
 > Indeed, this does the trick.
 > 
 > Can you speak to why this is this way? Is there some reason that
 > "keyboard" is not uncommented by default? It seems to me that the
 > handing of output and input devices should be handled similarly, and
 > the current state violates the principal of least confusion.
 > 
 > I would say, at a bare minimum, this needs a doc/CHANGES
 > entry, since it was not this way in 1.4.


 Not only that but it also causes anyone with any PC experience to
 believe that the system is totally hung!  :-)


 Personally I'd like to see the pckbd configured automatically all the
 time, regardless of where the console is.  For the same reasons it makes
 sense to auto-configure screen 0 all the time too.

 -- 
 							Greg A. Woods

 +1 416 218-0098      VE3TCP      <gwoods@acm.org>     <woods@robohack.ca>
 Planix, Inc. <woods@planix.com>;   Secrets of the Weird <woods@weird.com>
State-Changed-From-To: open->feedback 
State-Changed-By: martin 
State-Changed-When: Fri Aug 16 01:33:01 PDT 2002 
State-Changed-Why:  

I think this has been fixed in -current and the 1.6 branch, 
can you confirm this? 

Responsible-Changed-From-To: kern-bug-people->drochner 
Responsible-Changed-By: perry 
Responsible-Changed-When: Fri Apr 4 13:55:12 PST 2003 
Responsible-Changed-Why:  

on Fri, 19 May 2000, Matthias said: 
I'll have a look at it next week. 
I take that as his accepting responsibility for this. :) 


From: "Erik E. Fair" <fair@NetBSD.org>
To: NetBSD GNATS Bug Tracking System <gnats-bugs@NetBSD.org>
Cc:  
Subject: Re: kern/10129
Date: Mon, 29 Mar 2004 12:05:23 -0800

 Has there been any progress on this issue in a year?

 	Erik <fair@netbsd.org>

State-Changed-From-To: feedback->closed
State-Changed-By: elad@netbsd.org
State-Changed-When: Fri, 06 Oct 2006 22:33:04 +0000
State-Changed-Why:
should be fixed, feedback timeout (2+ years)


>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.