From owner-freebsd-questions Tue Jul 25 07:45:24 1995 Return-Path: questions-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.11/8.6.6) id HAA26863 for questions-outgoing; Tue, 25 Jul 1995 07:45:24 -0700 Received: from prades.cesca.es (prades.cesca.es [192.94.163.152]) by freefall.cdrom.com (8.6.11/8.6.6) with SMTP id HAA26853 for ; Tue, 25 Jul 1995 07:45:20 -0700 Received: by prades.cesca.es (1.38.193.4/16.2) id AA26865; Tue, 25 Jul 1995 16:41:49 +0200 Date: Tue, 25 Jul 1995 16:34:00 +0200 (METDST) From: Carlos Amengual Reply-To: Carlos Amengual Subject: Re: Headless/keyboardless booting... To: Michael Smith Cc: joe@via.net, questions@freebsd.org In-Reply-To: <199507242336.JAA01071@genesis.atrad.adelaide.edu.au> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII Sender: questions-owner@freebsd.org Precedence: bulk On Tue, 25 Jul 1995, Michael Smith wrote: > This is just so totally confused it's not funny 8( That's because your sense of humour is not very technical. > A20 is an address line, not an interrupt. A20 is an address line, but I was meaning the A20 "gate" or "switch" (calling it "interrupt" was a lapse, admittedly). > No systems (that I'm aware of anyway) use the keyboard to access any > memory anywhere. I never mentioned a system using a keyboard to access memory, however the PCs use the keyboard controller to turn the A20 address line off and on to access memory above 1 MB. This may seem silly to you, but what you are missing is that the PC is barely a computer... > A20 _is_ (obviously, really) used to address memory above the 1M mark, > and is manipulated in an odd fashion in some modes. None have anything > to do with the keyboard. The A20 gate opens when non-XT PCs are powered up, for compatibility with the XT; then must be closed to allow full memory addressing, and all this is done via the keyboard controller. I (really) cannot believe that you have never experienced a memory parity error, for example, because of a keyboard failure (or absence). The keyboard controller is very sensitive to this kind of stuff, at least for the PCs that I know. It should be possible to build a motherboard that gets rid of this, and perhaps it exists, but the point is that, as I stated in my message, in principle it will not. > What you're probably confusing this with is the trick that old versions > of OS/2 used to return from protected mode to real mode on the '286. This is only a particular subset of the general problem. > This has _nothing_ to do with FreeBSD, memory above 1M, A20 or my > (badly wilted) geraniums. Do not give up, geraniums sometimes grow in surprising places... > ]] Mike Smith, Software Engineer msmith@atrad.adelaide.edu.au [[ ------------------------------------------------- Carlos Amengual, amengual@prades.cesca.es Sociedad Astronomica de Espana y America (SADEYA) Av. Diagonal, 377, 2 ; 08008 Barcelona, Spain -------------------------------------------------