Subject: Keyboard Difficulties revisited
To: None <tech-kern@netbsd.org>
From: Matthew Fincham <matthewf@cat.co.za>
List: tech-kern
Date: 12/14/2001 17:15:34
Hi

A while ago I posted about having difficulties when a keyboard is plugged in
after boot-up. Despite some investigating, I still have not managed a fix. I
have found some strange (explained a bit more later) behaviour when using a
particular motherboard and a switchbox. As with my last e-mail, I am going
to break this down into separate scenarios:

Firstly, when the mouse and keyboard are plugged in, but the switchbox is
switched to a different PC at boot time, the keyboard fails to work. If I
look at what is happening, it seems that firstly, the keyboard is probed.
Part of this probe is to reset the keyboard. This fails, and (following the
comments in the driver) it is assumed that the keyboard is to be connected
later (this seems logical). The next step is to probe the ps2 mouse. For
some reason, this seems to fail, even thought the mouse IS plugged in (a
series of polled commands are executed, with all of them asking for a
resend).... When the switchbox is then swithched to the PC, the keyboard
fails to work.

The next scenario is when the keyboard is plugged in, but the mouse isn't,
and the switchbox is switched to a different PC at boot time. In this case
the keyboard works when the switchbox is switched to the PC, BUT when the
mouse is plugged in, the keyboard stops working....  The keyboard, again, is
assumed to be connected later (the reset fails). When the ps2 mouse is
probed and reset, the polled commands fail ( no data as opposed to  a series
of unseccessful resends), and the reset again fails.

This is quite confusing to me (as my writing most probably is to anyone who
has managed it this far!). I have probably put too much detail here already.
I have kernel printouts of the commands that are issued and the responses as
well as traces through the keyboard and mouse drivers if they would help.

I think part of the problem is  a lack of understanding on my side of how
the keyboard and mouse drivers fit into wscons. I anyone has somethiing to
read on this it would be greatly appreciated. I anyone has any pointers to
follow, I would appreciate that too.

Thank you for reading the essay!

Matthew Fincham