From owner-freebsd-hackers Wed Mar 26 13:27:58 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id NAA27674 for hackers-outgoing; Wed, 26 Mar 1997 13:27:58 -0800 (PST) Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.50]) by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id NAA27661 for ; Wed, 26 Mar 1997 13:27:55 -0800 (PST) Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id OAA28636; Wed, 26 Mar 1997 14:13:28 -0700 From: Terry Lambert Message-Id: <199703262113.OAA28636@phaeton.artisoft.com> Subject: Re: how to name fs specific programs To: cjs@portal.ca (Curt Sampson) Date: Wed, 26 Mar 1997 14:13:28 -0700 (MST) Cc: sommerfeld@orchard.east-arlington.ma.us, jonathan@dsg.stanford.edu, terry@lambert.org, perry@piermont.com, hackers@freebsd.org, port-i386@netbsd.org In-Reply-To: from "Curt Sampson" at Mar 26, 97 11:02:28 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 > > ...and he believes that the name > > of the filesystem type should be defined purely by the name used in > > the /sbin/fs directory, and not by anything inherent in the binaries > > found inside /sbin/fs/foo/. > > In other words, one wants /sbin/nfs/mount, and /sbin/nfs/nfsmount > is wrong because the word `nfs' appears in the files under /sbin/nfs? > > I can't see what advantage this offers besides aesthetic. Certainly > the code doesn't care whether or not there's an _ or a / between > the two %s strings in the printf format. It doesn't make a lot of difference, unless you decide to replace the agregation (generic) "mount" command with an FS specific "mount" command at boot time. If you do that, then it's possible that the names won't match like you expect them to, and you won't find "/sbin/nfsmount" when you are looking for "/sbin/mount". > It does seem to me to be a disadvantage in that if the file is ever > encountered outside of the tree (say, you restore a single file > from backup) it's very easy to get confused as to which mount > program it is. Well, I can't speak for a restore of a partial directory tree into a place other than the one written on the tape label... 8-). But the ability to replace the "generic" (agregating) commands with an FS specific command at varios stages of the boot is valuable to replacing existing kernels on non-BS systems with your BSD kernel, and having things "just work". It's an important factor if you ever want to pursue the "competitive upgrade" market with BSD kernels replacing the native kernels for the OS installed on the machine. This is the same issue as a device file system, since it allows you to replace the system device nodes with your own without having to destroy/disable the native devices, and without worrying about device major number differences/conflicts. A devfs also has the side effect of allowing you to NFS mount root from a system that doesn't understand BSD device nodes at all, either because of numbers of bits (like OSF on an Alpha) or because device nodes simply are not supported on the host system's NFS server (NetWare, VMS, etc.). Similar wins can be achieved in the FS module area using similar techniques (see previous posting for some "out there" examples). Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.