From owner-freebsd-stable Tue Nov 21 14:24:54 2000 Delivered-To: freebsd-stable@freebsd.org Received: from spammie.svbug.com (mg128-177.ricochet.net [204.179.128.177]) by hub.freebsd.org (Postfix) with ESMTP id BEE7C37B479 for ; Tue, 21 Nov 2000 14:24:48 -0800 (PST) Received: from spammie.svbug.com (localhost.mozie.org [127.0.0.1]) by spammie.svbug.com (8.9.3/8.9.3) with ESMTP id OAA00538; Tue, 21 Nov 2000 14:23:58 -0800 (PST) (envelope-from jessem@spammie.svbug.com) Message-Id: <200011212223.OAA00538@spammie.svbug.com> Date: Tue, 21 Nov 2000 14:23:55 -0800 (PST) From: opentrax@email.com Reply-To: opentrax@email.com Subject: Re: Dangerously Dedicated To: chad@DCFinc.com Cc: stable@FreeBSD.ORG In-Reply-To: <200011210517.WAA08133@freeway.dcfinc.com> MIME-Version: 1.0 Content-Type: TEXT/plain; charset=us-ascii Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On 20 Nov, Chad R. Larson wrote: > As I recall, Rodney W. Grimes wrote: >> The original IBM AT spec could give a rats ass about a partition >> table, all that it cares about is the boot block signature (magic >> 0xAA55). It is the MBR that knows what a partition table is and how >> to deal with it. The original spec says if there is a valid >> signature, load the code and jump to it passing the drive number in >> reg dl so that the boot code knows where it was loaded from. It was >> up to the MBR to decide what to do from then on. > > I think earlier in this thread was a reference to a document > somewhat later than the BIOS code shipped with an AT. > > Do we want to start a new thread on what exactly =is= the > authoritative documentation for PC architecture? > > The real issue is FreeBSD has to be able to boot on the hardware > that's in the stores. Who's wrong, though interesting, doesn't > matter. > I trimmed the CC line. It was getting a bit long. As per the spec relating to this. There is definitive information out there, That is how HD Mfgs are able to draw up new ATA specs. Most of the issue regarding this, in your outline are correct. In addition, the partition table, as has been, suggested, was not important to the IBM spec. However, many aftermarket vendors, especially the people that wrote drivers for Seagate, required the table to be in place. So, the table is important for that reason. In addition, many WinX and DOS vendors expect that some sort of table will be there. The other missing item is the magic number and a correct cksum for the block. The checksum is part of the spec and is usually written in to many install tools. However, even on this point many BIOS Mfgs ignore the cksums and just boot the system. That is, the load the code in the MBR and pass the instruction pointer to that code. Hence, even in some (now) rare cases the cksum is important. If I've mis-stated some parts, someone will correct me. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message