From owner-freebsd-hackers  Tue Jun 27 14:10:45 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id OAA03258
          for hackers-outgoing; Tue, 27 Jun 1995 14:10:45 -0700
Received: from gndrsh.aac.dev.com (gndrsh.aac.dev.com [198.145.92.241])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id OAA03250
          for <hackers@freebsd.org>; Tue, 27 Jun 1995 14:10:42 -0700
Received: (from rgrimes@localhost) by gndrsh.aac.dev.com (8.6.11/8.6.9) id OAA07820; Tue, 27 Jun 1995 14:10:31 -0700
From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
Message-Id: <199506272110.OAA07820@gndrsh.aac.dev.com>
Subject: Re: 2.05R reboot hangs
To: bde@zeta.org.au (Bruce Evans)
Date: Tue, 27 Jun 1995 14:10:31 -0700 (PDT)
Cc: bde@zeta.org.au, hackers@freebsd.org, wpaul@skynet.ctr.columbia.edu
In-Reply-To: <199506272033.GAA22671@godzilla.zeta.org.au> from "Bruce Evans" at Jun 28, 95 06:33:46 am
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1601      
Sender: hackers-owner@freebsd.org
Precedence: bulk

> 
> >> You initialize it before jumping to the reset vector.
> 
> >One that is very very hard to get right for most [3-4]86's is the
> >CPUID that is stored id DX:AX on power up reset since that is the
> >only way to get this value.
> 
> "Genuine Intel NOT!" :-).
> 
> >I don't want to call cold boot, I guess that was not clear from
> >above.  Also jumping to 0xFFFF:0000 (note your address above is
> >not the reset vector the value here is) is not a documented interface
> 
> There is no value there, only code that normally has a value embedded
> in it.
> 
> >to the BIOS.  Calling cold boot is about as reliable as triple faulting
> >or hitting the keyboard reset.  The one and only proper way to
> >make this work reliable on all machines as far as I can tell would
> >be a good old int 0x19, and that means I had better leave the BDA's
> >intact.
> 
> As I explained about 10 messages ago, int 0x19 isn't even useful for
> rebooting from DOS because it doesn't reinitialize hooked vectors.
> 
> I now think the keyboard reset is best.  It apparently requires disabling
> gateA20.  The code that disables it needs to run from an address < 1MB.
> Page 0 is a good place.

Okay, I can agree with that one, let me see about hacking some code
into the 0x500->0x1000 range that does the drop out of protected mode,
gate A20 off, then slam the keyboard controller.  I think you are right
in that this is going to be the most reliable mechanism.

-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                 Reliable computers for FreeBSD