From ats@freebsd.first.gmd.de  Mon Feb 23 14:05:27 1998
Received: from freebsd.first.gmd.de (freebsd.first.gmd.de [194.95.170.200])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id OAA22829
          for <FreeBSD-gnats-submit@freebsd.org>; Mon, 23 Feb 1998 14:05:09 -0800 (PST)
          (envelope-from ats@freebsd.first.gmd.de)
Received: (from ats@localhost) by freebsd.first.gmd.de (8.8.7/8.6.12) id WAA09864; Mon, 23 Feb 1998 22:56:20 +0100 (MET)
Message-Id: <199802232156.WAA09864@freebsd.first.gmd.de>
Date: Mon, 23 Feb 1998 22:56:20 +0100 (MET)
From: Andreas Schulz <ats@freebsd.first.gmd.de>
Reply-To: ats@freebsd.first.gmd.de
To: FreeBSD-gnats-submit@freebsd.org
Subject: Kernel panics in current (3.0)
X-Send-Pr-Version: 3.2

>Number:         5827
>Category:       kern
>Synopsis:       kernel panics in current (3.0)
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Mon Feb 23 14:10:07 PST 1998
>Closed-Date:    Thu Jul 15 00:14:27 PDT 1999
>Last-Modified:  Thu Jul 15 00:18:18 PDT 1999
>Originator:     Andreas Schulz
>Release:        FreeBSD 3.0-971022-SNAP i386
>Organization:
GMD-FIRST
>Environment:

Kernel sources as of last weekend.

>Description:

kernel panics with a "Fatal trap 9: general protection fault while
in kernel mode"  at address 0x8:0xf0196f9f which translates to the
"stosl" in "generic_bzero". 

>How-To-Repeat:

Boot a current kernel on an AMD K6 :-). The panics happen after the
"npx0: INT 16 interface" line.

>Fix:
	
I have fixed it for the moment on my box with a save of %es, move from
%ds to %es before the "stosl" and a restore afterwards. 
It looks like %es segment gets somewhere overwritten in the kernel.
It is not sufficient to explicitly set it in locore.s, also tried
after snooping around in the bootblock code and locore.s. 
I suspect you will only see this problem with the AMD K6 or a Cyrix
chip. The Pentium and 486 use another bzero routine in the kernel.

>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->closed 
State-Changed-By: hoek 
State-Changed-When: Thu Jul 15 00:14:27 PDT 1999 
State-Changed-Why:  
We have developers running the AMD k6, so this has obviously been fixed.  :-) 

(dated Feb. 23, 1998, fwiw...) 
>Unformatted:
