Date: Sat, 30 Sep 1995 15:14:39 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: dennis@etinc.com (dennis) Cc: gibbs@freefall.freebsd.org, jkh@time.cdrom.com, hackers@FreeBSD.ORG Subject: Re: FreeBSD 2.1 will require a minimum of 8MB for installation. Message-ID: <199509302214.PAA18839@phaeton.artisoft.com> In-Reply-To: <199509301945.PAA06042@etinc.com> from "dennis" at Sep 30, 95 03:45:04 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> but that every > driver (almost) be implemented as a loadable module and that the modules be > extremely clean and non-destrucitve. In theory, with the recent SYSINT changes are fully propagated, there will be no code differences between a set of driver objects "ld -r"'ed into a single object for later static linking in a kernel and a set of driver objects "ld -r"'ed into a single object for later dynamic loading. The issue in this regard is the devfs registration mechanism and whether or not the need for a backing store can be completely removed from the devfs. Part of the support for this is the support for multiple "root" device mounts (a relatively trivial set of changes, though somewhat broadly strewn) and loopback mounting -- required to have the ability to mount the devfs first thing, but not as the file system root, an then mount the file system root using the devfs name space, then loopback mount the devfs onto /dev. An alternate approach might be to support a "frame" file system that is implicitly mounted, and then loopback/union mount / on its /, leaving the /dev intact in the process and falling through to the devfs as a covered vnode of /dev in the frame. 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?199509302214.PAA18839>