Subject: Re: port-i386/13118: pc keyboard does not work when console is on the
To: None <netbsd-bugs@NetBSD.ORG>
From: Todd Vierling <tv@pobox.com>
List: netbsd-bugs
Date: 05/06/2003 18:21:41
On Tue, 6 May 2003, Greg A. Woods wrote:

: Note that anything which requires some command to be run after the
: physical connection is made in order to configure the keyboard device is
: broken by design.  The attachment _MUST_ be automatic.
:
: I don't know yet if this can be done by wiring down the attachments in
: the kernel config or not, but such an option is not ideal.

IMHO, on any system where pckbc0 is configured, it should always be attached
(and just disable itself, polling for validity every now and then, if
invalid spew on the port is seen).  There's precious little that abuses the
IO port and IRQ of the keyboard controller, so I don't see attaching all the
time as a problem.  It'll go onto the wsmux and will propagate no events,
thus doing nothing harmful.

I think the only oddity mentioned about this before is that probing is a
little strange if there's no keyboard attached.  Perhaps only real testing
can show if it'd be a problem to attach the device always.

-- 
-- Todd Vierling <tv@pobox.com>