I ran into this problem as well.  Doing searching of the kernel source,
I found the major device number to be 93.  I guessed at a minor number
of 0 and it worked.
bash-2.04$ ls -l cfs0
crw-r--r--  1 root  wheel   93,   0 Nov 14 20:23 cfs0

Later, I found the answer in the docs confirming my guess:
http://www.coda.cs.cmu.edu/doc/html/manual/x1050.html

I have submitted a pr to fix MAKEDEV.  Since I have confirmed my guess,
I should actually send in a patch for MAKEDEV.

Coda does run fine on FreeBSD; I have it running here (although only for
a week or so so far).

Kenneth

On Fri, Nov 17, 2000 at 07:50:25PM +0100, Gabriel Ambuehl wrote:
> Hello,
> I've been trying to get CODA 5.3.9 working in the last days and I
> guess I was successful doing the vice-setup tasks but I'm obviously
> failing when it comes to venus-setup as this will complain about
> missing device cfs0 which really doesn't exists (but also isn't listed
> in makedev). I've rebuilt the kernel using
> OPTIONS      CODA
> to compile coda support in but this won't solve the problem. Has
> anyone got any good ideas to solve this? Or, if this is allowed here,
> might also point me to other solutions which allow you to replicate
> servers in close to realtime, which means to us that the
> backup server shouldn't lag more than 1min behind in normal operation.
> 
> Any comments would be greatly appreciated.
> 
>   
> 
> 
> Best regards,
>  Gabriel
> 
>