[ ... kernel boot stucks when no mouse attached on video console boot ... ] 

>> The Problem was very easy to fix: I just attached a mouse to the system.
>> Oh man... Why can't NetBSD just say: "No mouse attached - I'm going to
>> hang now" :-).
>
> This problem also occurs on my 5000/125.
>
> The mouse is initialized in arch/pmax/tc/scc.c:scc_mouse_init().
> If keyboard/framebuffer are used as console, it calls lk_mouseinit()
> in arch/pmax/dev/lk201.c. But it uses sccPutc() to send command
> to the mouse, and sccPutc() will never return if transmitter
> is not ready.

I decided my mind to fix the issue by not probing mouse existence
during autoconfiguration process.  Instead, open(2) system call would
fail with ENXIO when no VSXXX mouse is attached.  DEC3000 TCWSCONS
now behaves in this way when open("/dev/wsmouse0", ) is attempted.

Slightly brighter news on NetBSD/pmax WSCONS transition.  A guy of MIT
created VS3100 WSCONS keyboard/mouse driver and Xserver for it (!).

       http://mail-index.netbsd.org/port-vax/2000/10/03/0004.html

I will try to make the driver compatible DS3100 and 3MAX.

Tohru Nishimura