Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 22 Mar 1996 11:20:31 -0700 (MST)
From:      Terry Lambert <terry@lambert.org>
To:        bde@zeta.org.au (Bruce Evans)
Cc:        bde@zeta.org.au, phk@critter.tfs.com, current@FreeBSD.ORG, julian@ref.tfs.com, scrappy@ki.net
Subject:   Re: PATCH: small, syntax changes for devfs
Message-ID:  <199603221820.LAA02981@phaeton.artisoft.com>
In-Reply-To: <199603220716.SAA24449@godzilla.zeta.org.au> from "Bruce Evans" at Mar 22, 96 06:16:51 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> wd0s1 is there, and opening it would reveal wd0s1[a-h], but this isn't
> much use.  Perhaps wd0s1 should be a directory containing [a-h].  I
> prefer a flat namespace.

Why?

It's possible, of course, to "chunk up" names, 2 characters at a time,
to build a flat name space, but is this necessary or even desirable?

At the very least, each device should be treated as a directory for
the purposes of getdirents and the ISDIR attributes, anyway.  It
lets you have "status" and other information in the "directory".

For the purposes of read/write, you treat it as a file.

Consider: a device reference is a vnode reference in devfs: all
lookup instances are predicated on the path to the devfs.  That
means I can access devices remotely without modifications (assuming
no layer 5 differentiation between systems -- if B9600 isn't the
same on both boxes, well, you're pretty much screwed).


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199603221820.LAA02981>