From owner-freebsd-current Wed Mar 15 19:33:27 1995 Return-Path: current-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id TAA06043 for current-outgoing; Wed, 15 Mar 1995 19:33:27 -0800 Received: from ref.tfs.com (ref.tfs.com [140.145.254.251]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id TAA06037 for ; Wed, 15 Mar 1995 19:33:24 -0800 Received: (from phk@localhost) by ref.tfs.com (8.6.8/8.6.6) id TAA09321; Wed, 15 Mar 1995 19:33:15 -0800 From: Poul-Henning Kamp Message-Id: <199503160333.TAA09321@ref.tfs.com> Subject: Re: newfs: sectors per cylinder (4096) disagrees with disk label (36) To: terry@cs.weber.edu (Terry Lambert) Date: Wed, 15 Mar 1995 19:33:13 -0800 (PST) Cc: rgrimes@gndrsh.aac.dev.com, current@FreeBSD.org In-Reply-To: <9503160321.AA16377@cs.weber.edu> from "Terry Lambert" at Mar 15, 95 08:21:21 pm Content-Type: text Content-Length: 1558 Sender: current-owner@FreeBSD.org Precedence: bulk > > > 1. Physical volume spanning. Like AIX, it is useful to have a > > > logical partition on which you actually build your file system > > > capable of spanning one or more phical partitions on one or > > > more physical drives. > > Is this an argument for or against ? > > An argument for seperation of geometry information from /dev accessable > devices, but against seperation of them from device drivers. > > Basically, there need to be devices that access the raw disk (machine > specific) which are used to access the DOS partitions from user space > and to mount (using a clone) do partitions without a device being > dedicated to the DOS partition. No. The slice code will do this the right way. On device in /dev for each logical area of the disk. > ... > There need to be geometry "bleed through" "hints" to the upper > layers. In the case of a non-translated MFM or ESDI device, ROTdelay > still has meaning. Yes, that's why I wanted that bit of logic in newfs. > The "perfection" of media (bad144, etc.) is handled at the middle > layer, or in a layer between the middle layer and the first layer, > depending on how machine specific this might get. bad144 is a relic. Forget it, it won't bother us for long. I think that most of what you mention, is right as such. I'm just not going to worry about it for now. When I see the code: yes, now: no. -- Poul-Henning Kamp -- TRW Financial Systems, Inc. 'All relevant people are pertinent' && 'All rude people are impertinent' => 'no rude people are relevant'