Date: Thu, 21 Mar 1996 03:02:53 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: George.Scott@cc.monash.edu.au (George Scott) Cc: current@FreeBSD.org Subject: Re: DEVFS vs "regular /dev" Message-ID: <199603211002.DAA29477@phaeton.artisoft.com> In-Reply-To: <199603210607.QAA12180@moa.cc.monash.edu.au> from "George Scott" at Mar 21, 96 04:07:07 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> I'm not sure I understand all this, but never mind. > > What would be wrong with doing something like this.. > > At boot time the kernel creates a memory-filesystem (mfs or other) with a /dev > directory. During the device probes this /dev gets populated with appropriate > device entries. At the end of the kernel initialisations the kernel union > mounts the root disk over the memory /. This is exactly what I was suggesting. My suggestion differs only in that I think it is only necessary to skeletonize / and /dev; an FS that only supports directories and doesn't suppor file creation or anything else could be built into the devfs itself as static data. This would be a lot less work, and you would not have to drag in a memfs image into kernel static data. > For those detractors in a previous thread this would also allow those who > are so inclined to have their own custom modifications to /dev or even to > remount their root without the union to keep using the old method. Yes. Though I would rip out specfs so the old method would fail. 8-). The vnops code for specfs sucks uot something fierce. This is what I was describing when I said the Heidemann framework had been pounded into 4.4BSD without much care... 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?199603211002.DAA29477>