From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mailman by lists.gnu.org with archive (Exim 4.43) id 1OEMAI-0000nx-NF for mharc-grub-devel@gnu.org; Tue, 18 May 2010 08:44:54 -0400 Received: from [140.186.70.92] (port=47674 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1OEMAG-0000nk-1G for grub-devel@gnu.org; Tue, 18 May 2010 08:44:53 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.69) (envelope-from ) id 1OEMAE-0006O4-3T for grub-devel@gnu.org; Tue, 18 May 2010 08:44:51 -0400 Received: from mail.gmx.net ([213.165.64.20]:51239) by eggs.gnu.org with smtp (Exim 4.69) (envelope-from ) id 1OEMAD-0006Na-KB for grub-devel@gnu.org; Tue, 18 May 2010 08:44:50 -0400 Received: (qmail invoked by alias); 18 May 2010 12:44:46 -0000 Received: from 165.126.46.212.adsl.ncore.de (HELO 192.168.2.69) [212.46.126.165] by mail.gmx.net (mp008) with SMTP; 18 May 2010 14:44:46 +0200 X-Authenticated: #2145628 X-Provags-ID: V01U2FsdGVkX1/mj9J8NX3ms5aC/j665pvtWtbdlmzE73PJhPN52z bjX0b3RI3121he Date: Tue, 18 May 2010 14:44:09 +0200 From: "Thomas Schmitt" To: grub-devel@gnu.org References: <20100518120311.GL21862@riva.ucam.org> In-Reply-To: <20100518120311.GL21862@riva.ucam.org> Message-Id: <106270736515853@192.168.2.69> X-Y-GMX-Trusted: 0 X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. Subject: Re: Using efi image on USB stick X-BeenThere: grub-devel@gnu.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: The development of GNU GRUB List-Id: The development of GNU GRUB List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 18 May 2010 12:44:53 -0000 Hi, Colin Watson wrote: > I don't know how well > BIOSes handle MBR partition tables on CD-ROMs. According to ECMA-119 (aka ISO 9660) the first 32 kB of an image are "System Area" with arbitrary custom content. El Torito specs mention that this area may contain a bootable disk image. (Figure 1, case "Multiple Boot-Image Configuration"). To my view, a bootable disk image may well begin by an MBR which does not necessarily have to describe the geometry of the entire CD. The BIOS learns the positions of boot images from the El Torito boot catalog. The position of the boot catalog is given by the El Torito record in sector 17 (decimal) of the CD. So it seems natural that the BIOS should not interpret the system area of a CD unless the boot catalog points to it as boot image. Currently the xorriso generated boot catalogs do not point to the system area. But even if they would, then the bytes at the position of a partition table should not be interpreted as partition table of the CD. > http://fedoraproject.org/wiki/Anaconda/Features/UEFI : "Many BIOS implementations, ... do not correctly handle the multiple El-Torito boot blocks" El Torito specs (of 1995) talk of "BIOS with Single Boot-Image capability" in contrast to "BIOS with Multiple Boot-Image capability". I interpret the following statement as a prescription that single-boot BIOSes shall ignore further boot images: "Single-Image INT 19 knows nothing about multiple-images, nor does it know about their possible entries listed in the Booting Catalog." > I guess all we can do is test ... Maybe one should publish a little test image which is alread yknown to work on some systems. (I could test with a 64 bit AMD quadcore and an old 32 bit AMD of 2004.) Have a nice day :) Thomas