Subject: Mice, KVM, and X, oh my!
To: None <netbsd-help@netbsd.org>
From: Peter Seebach <seebs@plethora.net>
List: netbsd-help
Date: 02/25/2002 20:17:15
I have a mouse.  It's a pmsi, and it works fine.  (It's actually using my
hacks to probe as a 5-button mouse, but I don't think this affects anything.)

I start X.  X works.  Buttons work.  Joy is complete.

I use a KVM switch.  I boot Windows.  Mouse works, everything works.

I use a KVM switch, *and* I use X.  Everything works, but if I ever switch
away, and come back, the mouse is *ALWAYS* screwed up in ways I do not
understand; it looks like a protocol mismatch.

More interestingly:  I have a 2-port KVM switch, and if I use this, it's fine
most of the time, unless I move the mouse right after switching.  If I use my
4-port switch, though, the protocol error is 100% reproducible.

So... It's probably a problem with the switch, *and yet*, it's not a problem
under Windows, and if I leave X, and restart the X server, everything works
fine.  So, it's some kind of a very transient problem, and if the X server
could recover, I'd never see it.

Anyone have any ideas?  I'm using the wsmouse protocol, and I would have
assumed that, since "od -x /dev/wsmouse" produces nice stable results before
and after a switch, so would everything else.

-s