>> And dc driver will have a following structure;
>>
>>       dc0
>>          lkkbd0
>>             wskbd0
>>          vsms0
>>             wsmouse0  
>
> That doesn't work. On the 5100, channel 0 and 1 are wired to serial
> ports. The default console is on channel 0. Always hardiwring these to
> mouse/keyboard is wrong.

My driver will have 'line' locator, and lkkbd/vsms/dctty stuff is not
hardwired on specfic ports.

With explicit designations as follows;

	dc0
	   dctty? line 0
	   dctty? line 1

line 0 and line 1 can be useful for ttys.  Current implementation does
not allow this.

>>          dctty0
>>          dctty1
>
> Needs to be should be dctty2 and dctty3, since that's what they areon
> a 5100. (well, line 3 may not be available).

Digital had bad practice to assign 4.ch serial ports to various
purpose, and kept the bad habit even for z85c30 combo.  What's they
really did was to emulate a pair of z85c30 as if it was an old DEC
chip.  That's the reason why external serial ports of IOASIC
DECstation and TC alpha are numbered for #2 and #3, and so as you
already noticed, that did provide complications on #0/B/A and #1/B/A
device driver.

> 3100s, and /5000/200, there is a hardware dongle to use the
> `mouse'/`keyboard; ports as tty ports. The default numbering should
> take that into account, because some NetBSD users use them.

It's very easy to absorb with 'line' locator by allowing both of wildcard
matching and explicit matching.  Mattias Drochner already made a solution
for zs_ioasic.c in [nisimura-pax-wscons] branch.

Tohru Nishimura