[TEST/REVIEW]: rootfs mounting code rework

[Garance A Drosihn]

At 10:12 AM -0800 12/5/04, Tim Kientzle wrote:
>
>Here's a situation that may be worth thinking about:
>
>   * Working FreeBSD system that uses a
>     separate controller.  / is /dev/da4.
>
>   * Controller blows up.
>
>   * Connect hard disk to the on-board controller.
>
>Note: kernel is now being loaded from /dev/da0, but /etc/fstab
>still says / is /dev/da4.  With the old rootfs code, this system
>is unbootable.

This is the kind of situation I was thinking of, where you get
into a mess due to some hardware shuffling.  It does not come
up often, but it can be frustrating when you do run into it.  I
was just wondering if Poul's work might help in that situation.

In a later message, Poul-Henning Kamp wrote:
>First: I have _not_ changed the logic that decides _which_ rootfs
>to mount, only the mechanics of how it is mounted.
>
>Second: I belive you could simply type something like:
>
>	set vfs.root.mountfrom="ufs:/dev/da0"
>
>in the loader and your system would do what you expect it to.

Ah, very good.  I'll save that away for future reference.

So with your changes, should we just check to see that it works
the same as it used to work?  Or is there any kind of new
behavior that we should try to exercise?

-- 
Garance Alistair Drosehn            =   gad at gilead.netel.rpi.edu
Senior Systems Programmer           or  gad at freebsd.org
Rensselaer Polytechnic Institute    or  drosih at rpi.edu