From owner-cvs-src@FreeBSD.ORG Thu Dec 6 18:06:09 2007 Return-Path: Delivered-To: cvs-src@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 9C78516A417; Thu, 6 Dec 2007 18:06:09 +0000 (UTC) (envelope-from xcllnt@mac.com) Received: from smtpoutm.mac.com (smtpoutm.mac.com [17.148.16.73]) by mx1.freebsd.org (Postfix) with ESMTP id 82B3513C4CC; Thu, 6 Dec 2007 18:06:09 +0000 (UTC) (envelope-from xcllnt@mac.com) Received: from mac.com (asmtp001-s [10.150.69.64]) by smtpoutm.mac.com (Xserve/smtpout010/MantshX 4.0) with ESMTP id lB6HoPm1008283; Thu, 6 Dec 2007 09:50:25 -0800 (PST) Received: from mini-g4.jnpr.net (natint3.juniper.net [66.129.224.36]) (authenticated bits=0) by mac.com (Xserve/asmtp001/MantshX 4.0) with ESMTP id lB6HoNS4001796 (version=TLSv1/SSLv3 cipher=AES128-SHA bits=128 verify=NO); Thu, 6 Dec 2007 09:50:24 -0800 (PST) Message-Id: From: Marcel Moolenaar To: John Baldwin In-Reply-To: <200712060421.18367.jhb@freebsd.org> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v915) Date: Thu, 6 Dec 2007 09:50:22 -0800 References: <200712060232.lB62WhFL011174@repoman.freebsd.org> <200712060421.18367.jhb@freebsd.org> X-Mailer: Apple Mail (2.915) Cc: cvs-src@freebsd.org, src-committers@freebsd.org, cvs-all@freebsd.org Subject: Re: cvs commit: src/sys/conf NOTES files options src/sys/geom/part g_part.c g_part.h g_part_bsd.c src/sys/ia64/conf DEFAULTS X-BeenThere: cvs-src@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: CVS commit messages for the src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Dec 2007 18:06:09 -0000 On Dec 6, 2007, at 1:21 AM, John Baldwin wrote: > So one issue with gpart is it doesn't handle attribute requests yet, > so it > doesn't have the feature where geom_bsd only attaches when a parent > partition > has the right type. It supports attributes, but not the ones used by other partitioning classes, because those are not needed in the end. > I ran into this for GPT testing where due to the way we > align disks at work for RAID striping, the beignning of my GPT slice > for / > had a BSD label at the start, so I would end up with bogus /dev/ad0p1a > devices and scary warnings in dmesg about a busted BSD label because > offsets > didn't match up. I fixed this in 6.x by making GEOM_GPT export its > UUID via > an attr and having geom_bsd check that like it does now for > MBR::type and > PC98::type. You will need to handle this in gpart as well. It's all part of the plan. I first go wide to make sure I cover all the partitioning schemes and work out the quirks with them, then I'll add boot code support, flags (i.e. for the active MBR partition) and partition type enforcement. At this time gpart already checks the partition type for kernel dumps, because that was easy :-) -- Marcel Moolenaar xcllnt@mac.com