From owner-freebsd-fs@FreeBSD.ORG Tue Apr 26 18:46:42 2011 Return-Path: Delivered-To: freebsd-fs@FreeBSD.org Received: from mx2.freebsd.org (mx2.freebsd.org [IPv6:2001:4f8:fff6::35]) by hub.freebsd.org (Postfix) with ESMTP id B3CCB10656E7; Tue, 26 Apr 2011 18:46:42 +0000 (UTC) (envelope-from dougb@FreeBSD.org) Received: from 65-241-43-5.globalsuite.net (hub.freebsd.org [IPv6:2001:4f8:fff6::36]) by mx2.freebsd.org (Postfix) with ESMTP id 0A025153BCA; Tue, 26 Apr 2011 18:46:41 +0000 (UTC) Message-ID: <4DB71311.7080203@FreeBSD.org> Date: Tue, 26 Apr 2011 11:46:41 -0700 From: Doug Barton Organization: http://SupersetSolutions.com/ User-Agent: Mozilla/5.0 (X11; U; FreeBSD amd64; en-US; rv:1.9.2.15) Gecko/20110319 Thunderbird/3.1.9 MIME-Version: 1.0 To: Alexander Motin References: <4DB70949.6090104@FreeBSD.org> <4DB70F97.1070809@FreeBSD.org> In-Reply-To: <4DB70F97.1070809@FreeBSD.org> X-Enigmail-Version: 1.1.2 OpenPGP: id=1A1ABC84 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: freebsd-fs@FreeBSD.org Subject: Re: Why not just name the cam-ata devices the same as the old names? X-BeenThere: freebsd-fs@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Filesystems List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 26 Apr 2011 18:46:42 -0000 On 04/26/2011 11:31, Alexander Motin wrote: > Hi. > > Doug Barton wrote: >> I'm not nearly as smart as you are, so please explain to me like I'm >> dense. :) Why can we not simply give the devices created by ata-cam the >> same names they have under the old ata system? > > Don't underestimate yourself, or don't make me blush. ;) > > There are two problems: names and unit numbers. > > We can't use same names because old and new stacks coexisting last 18 > months (and they will forever in 8.x), and it was possible to just load > single ahci module of new stack for SATA devices, while PATA were > working via old stack. Using same name would cause collisions. Now when > we are going to switch to the new stack completely, coexistence could be > a bit less important, but there is already number of migrated systems > and they would suffer from the second renaming. > > Even if we do something with names, there is a problem with device unit > numbers. Previous numbering scheme of ATA_STATIC_ID reserved only two > device numbers per ATA channel. It was working fine for PATA times, but > it is not now. When SATA port multipliers are used (and it is not so > rare now), there could be up to 15 devices per channel, plus multiplier > itself. They just won't fit. Thanks for breaking this down so that even I can understand it. :) Given what you've explained here it sounds like the dev symlink stuff that you just committed is the right way to go. If that plan proves successful what I'm thinking is that I can put some code in mergemaster on 9.0-RELEASE that checks for the old device names in /etc/fstab and suggests that the user change them to the new names. I don't generally like to give mergemaster specific knowledge of certain files, but there is precedent for doing so for changes like this, and as long as we are not relying on mergemaster to _fix_ the problem, I'm Ok with doing it that way. So I will stay tuned to see how this develops. Doug -- Nothin' ever doesn't change, but nothin' changes much. -- OK Go Breadth of IT experience, and depth of knowledge in the DNS. Yours for the right price. :) http://SupersetSolutions.com/