Date: Mon, 28 Sep 1998 19:55:48 -0400 From: David Holland <dholland@cs.toronto.edu> To: eivind@yes.no (Eivind Eklund) Cc: reg@shale.csir.co.za, current@FreeBSD.ORG Subject: Re: Current is Really Broken(tm) Message-ID: <98Sep28.195549edt.37768-5346@qew.cs.toronto.edu> In-Reply-To: <19980928182931.41626@follo.net> from "Eivind Eklund" at Sep 28, 98 12:29:31 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > [root ramdisk]
>
> I believe the main reason it doesn't do that is that too many people
> believe it too radical. I know a number of developers (hi phk!) would
> like to have the actual device probes controlled from a very small
> userland running out of a ramdisk - personally, I think this might be
> a good architecture, but as I haven't seen an implmentation yet, I'm
> not quite sure.
Linux has something like this for loading modules before mounting
root, and it's gorss.
If you were really going to try to have a ramdisk root, it should just
be a framework for mounting other things and not have any data on it.
IMO. Something like the AmigaDOS "assign", or a cross between a union
mount and a symlink would then be nice so you can have /bin actually
reference /usr/bin, /usr/local/bin, /disk1/stuff/bin, and whatnot, all
transparently.
But this rapidly stops being Unix.
--
- David A. Holland | (please continue to send non-list mail to
dholland@cs.utoronto.ca | dholland@hcs.harvard.edu. yes, I moved.)
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?98Sep28.195549edt.37768-5346>