From owner-freebsd-hackers Sun Jan 12 11:36:47 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.4/8.8.4) id LAA14184 for hackers-outgoing; Sun, 12 Jan 1997 11:36:47 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.211]) by freefall.freebsd.org (8.8.4/8.8.4) with SMTP id LAA14178 for ; Sun, 12 Jan 1997 11:36:45 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id MAA25976; Sun, 12 Jan 1997 12:22:56 -0700 From: Terry Lambert Message-Id: <199701121922.MAA25976@phaeton.artisoft.com> Subject: Re: DEVFS permissions &c. To: bde@zeta.org.au (Bruce Evans) Date: Sun, 12 Jan 1997 12:22:56 -0700 (MST) Cc: davidn@unique.usn.blaze.net.au, joerg_wunsch@uriah.heep.sax.de, hackers@freebsd.org In-Reply-To: <199701121444.BAA14336@godzilla.zeta.org.au> from "Bruce Evans" at Jan 13, 97 01:44:33 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-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk > >Seriously, I've used sysv for many years, and grew to quickly despise > >the sysv approach. It does have some good sides, but, for example, > >Sun's tree of symlinks to init/shutdown scripts is definitely an > >overkill. > > I expect a tree of devices would be overkill too. You would need evil > symlinks to reduce /dev/disks/raw/scsi/bus0/id0/lun0/slice2/partitionh > to something like /dev/rsd0h :-). Why? For all intents and purposes, you have described: o A raw device designator, which need not be a seperate semantic component o The fact that the controller is SCSI, which no one should care about at this level, and which shouln't be under "disks" anyway. Besides which, the information is otherwise retrievable via an ioctl() to the device to retrieve underlying physica device info o The bus on the controller, which *also* no one cares about for the same reasons as above o The SCSI id, which is irrelevant, since it was reflected in the arrival order for the physical devices. So far, we have reduces to /dev/dsk/dsk0 o The SCSI lun, which for devices that fan out by lun, is a physical to logical partitioning scheme, and for devices which do not fan out, are seperate devices anyway. So we are /dev/dsk/dsk0 ... /dev/dsk/dskN or, we are /dev/dsk/dsk0/p0 ... /dev/dsk/dsk0/pN o The slice, which is truly partitioning (invokes physical to logical layering). o The partition, which is truly partitioning (invokes physical to logical layering). The point of the device hierarchy in my examples was to give entry points for a putative "universal fdisk" utility, which could handle all partitioning tasks, regardless of the partitioning schema used to implement them. It was *not* to provide some name space incursion mechanism for all possible fan outs (though one has to agree that any series of 1->N fanouts is probably better handled by hierarch than by changing the exposed device name). Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.