Simon Burge writes:

>On Tue, 03 Feb 1998 19:22:42 -0800  Jonathan Stone wrote:
>
>> 
>> Is there anyone out there using an X server on a 5000/200
>> with NetBSD 1.3_BETA or newer?  If so, does the mouse work?
>> 
>> If it does, that would help enormously wiht limiting the
>> possible reasons why mouse movement seems to not work on 3100s.
>
>My 3100 at work was running 1.3_ALPHA - I just started an install before
>I left for home last night.  I'll let you know later today about this.
>I also have a 5000/200 idle - I can test that too...
>
>One thing - when I first turned it on after I got back from holidays
>the mouse _didn't_ work, but it must have been hardware - another mouse
>_did_ work...

Thanks, simon My suspicion was that the DELAY() calibration based on
CPU speed is slighly too low for the 3100 and 2100, and this tends to
confirm it.

There's a code fragment in sys/arch/pmax/dev/lk201.c which
does DELAY(100) between the mouse self-test and setting the mouse
into auto-report mode.  Michael Hitch was going to try changing
that to DELAY(1000).  

I think an even safer test is to try changing the  two
	cpuspeed=8

lines in sys/arch/mips/mips/mips_mcclock.c (nominally for 2100  and 3100) 
to be
	cpuspeed=13

which is the same DELAY() multipiler that the 5000/200 is using.
If that doesn't work, we need another theory.

(Apologies to those with 3100s, I cant easily get mine booted right now.)