From owner-freebsd-current Tue Nov 26 14:28:54 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id OAA01880 for current-outgoing; Tue, 26 Nov 1996 14:28:54 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id OAA01859 for ; Tue, 26 Nov 1996 14:28:48 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id PAA25659; Tue, 26 Nov 1996 15:13:05 -0700 From: Terry Lambert Message-Id: <199611262213.PAA25659@phaeton.artisoft.com> Subject: Re: 2.2-ALPHA install failure To: bde@zeta.org.au (Bruce Evans) Date: Tue, 26 Nov 1996 15:13:05 -0700 (MST) Cc: bde@zeta.org.au, terry@lambert.org, current@freefall.freebsd.org, darrylo@sr.hp.com, jkh@time.cdrom.com, rgrimes@GndRsh.aac.dev.com In-Reply-To: <199611262202.JAA21885@godzilla.zeta.org.au> from "Bruce Evans" at Nov 27, 96 09:02:22 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk > >If you are going to present fake geometry numbers, they should percoalte > >up from where the real geometry numbers are coming from. > > They do percolate up, but there is no "real" geometry. There lots of > different geometries: > > default for newfs: independent, X/1/4096 > values in the label: independent, usually unused > values in the partition table: should agree with BIOS > BIOS indepependent, important only for booting and > in other OS's > physical: independent, usually an average for sectors/ > track, usually unused > > These geometries should NOT be visible in all layers. Oh, I agree. Anything other than the "DOS PRIMARY PARTITIONING" or "DOS EXTENDED PARTITIONING" layers could care less: they deal strictly with sectors. The layer itself would make the geometry-to-sector-start+length translation. All subsequent layers wouldn't care, unless they were also geometry sensitive (say Partition #1 in the example I posted was a "DOS EXTENDED PARTITIONING" instead of "none"; it would care). The point is that as long as it can get an absolute sector start and length (for a linear run) or an absolute sector start and a "length" that represents the total number of sectors in the run, and a function to take a linear offset within the run and translate it to a physical sector, then the upper layer only cares about some sector number 0-N. All the FS cares about is that it has N sectors, and heres a device where they can be linearly accessed. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.