NetBSD Problem Report #32575

From www@netbsd.org  Thu Jan 19 20:06:11 2006
Return-Path: <www@netbsd.org>
Received: by narn.netbsd.org (Postfix, from userid 31301)
	id 6ADCD63BA0A; Thu, 19 Jan 2006 20:06:11 +0000 (UTC)
Message-Id: <20060119200611.6ADCD63BA0A@narn.netbsd.org>
Date: Thu, 19 Jan 2006 20:06:11 +0000 (UTC)
From: Ingolf Steinbach <v04vmaj02@sneakemail.com>
Reply-To: Ingolf Steinbach <v04vmaj02@sneakemail.com>
To: gnats-bugs@netbsd.org
Subject: GENERIC crashes with non-root vi on Tadpole 3GX
X-Send-Pr-Version: www-1.0

>Number:         32575
>Category:       port-sparc
>Synopsis:       GENERIC hangs with vi on console of Tadpole 3GX
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    chs
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Jan 19 20:10:00 +0000 2006
>Closed-Date:    
>Last-Modified:  Sat Mar 08 09:14:28 +0000 2008
>Originator:     Ingolf Steinbach
>Release:        NetBSD 3.0
>Organization:
none
>Environment:
System: NetBSD ares.steinba.ch 3.0 NetBSD 3.0 (GENERIC.DDB) #0: Thu Jan 19 00:36:31 CET 2006 ingolf@ares.steinba.ch:/usr/src/sys/arch/sparc/compile/GENERIC.DDB sparc
Architecture: sparc
Machine: sparc
Model: Tadpole 3GX
>Description:
On a Tadpole 3GX, the GENERIC kernel can easily be crashed
by non-root users. For instance, when editing a file in vi
(on the console) and moving the cursor around a bit, the
console will eventually freeze. After a while, the watchdog
resets the SPARCbook.

In the hope of finding more information, I have re-compiled
the GENERIC kernel but with
 - options DDB
 - options DDB_HISTORY_SIZE=100
 - options DDB_ONPANIC=1
 - makeoptions DEBUG="-g"
enabled. After the freeze, the machine still drops to the
PROM ("cont" is not possible).

I have verified that I can enter ddb when everything is still
ok. However, as soon as the machine freezes, I can also no
longer enter the debugger.

Maybe the console code cannot handle the rather limited
capabilities of the 3GX when using the default fonts
selected in the GENERIC kernel. When booting, the kernel
messages are first displayed in a rather small font which
allows for sufficiently long lines. Then during the boot,
the font is switched to a larger one. Only 66x27 characters
can be displayed with this font. (This also makes several
messages unreadable during system installation...)
>How-To-Repeat:
Boot a plain 3.0 system on a Tadpole 3GX.
Login (either as root or real user)
vi /etc/ssh/sshd_config
In vi, go to the last line (G), then move the cursor up
line by line.
After reaching the top line which is visible and moving
further up, the cursor will disappear and the machine
>Fix:
No idea.

>Release-Note:

>Audit-Trail:

Responsible-Changed-From-To: port-sparc-maintainer->chs
Responsible-Changed-By: chs@netbsd.org
Responsible-Changed-When: Tue, 24 Jan 2006 01:02:43 +0000
Responsible-Changed-Why:
to me


State-Changed-From-To: open->feedback
State-Changed-By: chs@netbsd.org
State-Changed-When: Tue, 24 Jan 2006 01:02:43 +0000
State-Changed-Why:
I may have fixed this recently, could you try a -current kernel
and see if it still happens there?


From: "Ingolf Steinbach" <v04vmaj02@sneakemail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: port-sparc/32575 (GENERIC hangs with vi on console of Tadpole 3GX)
Date: Fri, 27 Jan 2006 21:33:05 +0100

 > I may have fixed this recently, could you try a -current kernel
 > and see if it still happens there?

 Unfortunately not -- at least with 3.99.15, the behaviour is
 still the same.

From: Michael Lorenz <macallan@netbsd.org>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: PR/32575
Date: Sat, 25 Feb 2006 12:14:28 -0500

 Can you reproduce the problem with the TADPOLE3GX kernel?

From: "Ingolf Steinbach" <v04vmaj02@sneakemail.com>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: PR/32575
Date: Mon, 6 Mar 2006 21:50:54 +0100

 Hi Michael.

 you wrote:
 >  Can you reproduce the problem with the TADPOLE3GX kernel?

 Is a TADPOLE3GX kernel (NetBSD-3.0) available for download
 somewhere (I'd like to avoid getting the source tree and
 compiling one myself).

 BTW, Chuck Silvers asked me to try a current kernel from
 end of January. He seems to have checked in a fix for the
 problem:

 At least, the situation has improved: the machine no longer
 hangs (when moving the cursor around in vi as described in
 the original PR).

 However, when moving further up when the cursor is already
 in the top row of the display, vi does not scroll the display
 correctly. Only at the 7th "move-up" of the cursor, the
 display is scrolled. But it seems that vi thinks that the
 cursor is at a completely other line... Could now be a
 curses (or whatever is responsible for the display) problem
 with the small number of rows/columns.

 Cheers
     Ingolf

State-Changed-From-To: feedback->open
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Sat, 08 Mar 2008 09:14:28 +0000
State-Changed-Why:
feedback has been received.


>Unformatted:

Home	PR Database Search