NetBSD Problem Report #7478
Received: (qmail 22218 invoked from network); 26 Apr 1999 18:19:13 -0000
Message-Id: <199904261819.UAA14353@panay.sis.co.at>
Date: Mon, 26 Apr 1999 20:19:11 +0200 (MEST)
From: Charlie Root <root@panay.sis.co.at>
Reply-To: pm@sis.co.at
To: gnats-bugs@gnats.netbsd.org
Subject: kernel panic
X-Send-Pr-Version: 3.95
>Number: 7478
>Category: kern
>Synopsis: kernel panic on i386 in gdt_compact() during fork bomb
>Confidential: no
>Severity: critical
>Priority: high
>Responsible: chs
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Mon Apr 26 11:20:01 +0000 1999
>Closed-Date: Sat Dec 14 01:16:20 +0000 2002
>Last-Modified: Sat Dec 14 01:16:20 +0000 2002
>Originator: Manfred Petz
>Release: 1.4_ALPHA
>Organization:
SIS Datenverarbeitung GmbH, Vienna
>Environment:
System: NetBSD bsd 1.4_BETA NetBSD 1.4_BETA (POROS) #0: Mon Apr 26 16:49:21 CEST 1999 root@bsd:/usr/src/sys/arch/i386/compile/POROS i386
>Description:
System can be crashed with a fork bomb.
>How-To-Repeat:
How I did:
while (1) {
sleep (1);
fork ();
}
The sleep() appears to be important, but maybe that's just a
coincidence.
- I started top(1) on a VC.
- Telnet into the box and started the fork bomb from the telnet session.
- After a few seconds the system couldn't fork anymore. I pressed
^C. All processes disappeared.
- Then I started the fork bomb again, waited again until no more
processes could be started.
- Pressed ^C -> panic in the same moment.
I repeated that several times. Always the same situation.
kernel: page fault trap, code=0
Stopped at _gdt_compact+0x40 movl 0x68(%ecx),%esi
Sorry - no dump since this should be nicely reproducible and the
system panics again inside ddb when trying to dump. :-)
If you cannot reproduce then please drop me an email so I can tell you
more. My config is available at
ftp://ftp.sis.co.at/pub/system/NetBSD/POROS.conf
But I don't think this has to do with a particular configuration.
>Fix:
>Release-Note:
>Audit-Trail:
See also kern/7714.
Responsible-Changed-From-To: kern-bug-people->chs
Responsible-Changed-By: fair
Responsible-Changed-When: Thu Mar 22 11:51:42 PST 2001
Responsible-Changed-Why:
Our UVM expert is Chuck Silvers.
State-Changed-From-To: open->closed
State-Changed-By: junyoung
State-Changed-When: Thu Dec 12 19:23:26 PST 2002
State-Changed-Why:
Fork bomb no longer causes the system to panic or die.
Responsible-Changed-From-To: chs->junyoung
Responsible-Changed-By: junyoung
Responsible-Changed-When: Thu Dec 12 19:23:26 PST 2002
Responsible-Changed-Why:
I'm closing it.
State-Changed-From-To: closed->open
State-Changed-By: chs
State-Changed-When: Thu Dec 12 21:40:53 PST 2002
State-Changed-Why:
this is not the usual fork bomb hang, there's a bug in the GDT code.
I've seen it a couple times myself since this PR was opened.
Responsible-Changed-From-To: junyoung->chs
Responsible-Changed-By: chs
Responsible-Changed-When: Thu Dec 12 21:40:53 PST 2002
Responsible-Changed-Why:
back to me
State-Changed-From-To: open->closed
State-Changed-By: chs
State-Changed-When: Fri Dec 13 17:14:51 PST 2002
State-Changed-Why:
bang pointed out that the GDT bug must be fixed since
we never shrink or compact the GDT anymore after the x86 MP merge.
>Unformatted:
(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.