On Fri, 17 Dec 1999, David McConnell wrote:

> I now have "md's" working. There was a "control" node for the "md" devices
> missing from /dev. (As I was suspecting and was pointed out by Izumi.)
> 
> Somebody should perhaps correct this?
> 
	If someone wants to mail me the diff I'll commit it :)
	(They could alway send-pr :)

> Now that I have both md's and mfs working properly, which should I choose
> for runtime ramdisk mounting.
> 
> The md.h source refers to mfs as a "hack" but mfs seems to work fine. Some
> of the advantages I perceived to using mfs have turned out to be not so,
> e.g. leaving MFS out of the kernel only reduces its size by a few hundred
> bytes and whichever way I go I need the newfs (or mount_mfs) utility so I
> see no obvious advantage/disadvantages to distinguish the 2 approaches.
> 
> So, is there any compelling reason to use one rather than the other? (For
> the moment I'm using mfs's).

	I believe mfs is easier to setup on demand, and does not use
	'wired' memory (can be swapped out).


		David/absolute