From owner-freebsd-current  Fri Feb 23 08:17:27 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id IAA18403
          for current-outgoing; Fri, 23 Feb 1996 08:17:27 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id IAA18379
          Fri, 23 Feb 1996 08:17:05 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id DAA19724; Sat, 24 Feb 1996 03:13:53 +1100
Date: Sat, 24 Feb 1996 03:13:53 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199602231613.DAA19724@godzilla.zeta.org.au>
To: joerg_wunsch@uriah.heep.sax.de, olah@cs.utwente.nl
Subject: Re: Keyboard lockout on 2.x.x
Cc: freebsd-current@FreeBSD.org, sos@FreeBSD.org
Sender: owner-current@FreeBSD.org
Precedence: bulk

>> > > Just FYI, specifying ASYNCH apparently fixed the lockups for me.
>> > 
>> > So far me too.
>> 
>> Søren, time to make it the default?

Time to rewrite the keyboard driver?

>I'm not sure that everyone would like it, because a virtual console
>switch takes now a few 10th's of a second during which the system
>seems to be halted (at least the HD stops grinding).  I'm still

The delay is because ASYNCH is less asynchronous than !ASYNCH.
kbd_cmd() is always called at spltty(), so keyboard interrupts are
blocked, so `kbd_reply' is as nonvolatile as it looks in the ASYNCH
spinloop in kbd_cmd(), so the spinloop always times out.  The
keyboard reply is received and discarded some time later.  Perhaps
this sort of works by waiting for a few thousand times as long as
necessary instead of a few microseconds shorter than necessary.

Bruce