Date: Wed, 3 Jan 1996 10:16:11 +0100 (MET) From: J Wunsch <j@uriah.heep.sax.de> To: peterb@telerama.lm.com (Peter Berger) Cc: freebsd-hackers@freebsd.org Subject: Re: PS/2 mouse driver & keyboard lockups. Message-ID: <199601030916.KAA19702@uriah.heep.sax.de> In-Reply-To: <Pine.BSI.3.91.960103010235.21096A-100000@ivory.lm.com> from "Peter Berger" at Jan 3, 96 01:04:25 am
next in thread | previous in thread | raw e-mail | index | archive | help
As Peter Berger wrote: > > 1) Is anyone currently responsible for the PS/2 mouse driver? Not that i know of. The only actions that have been taken on it lately are general cleanups. > 2) Is the core team aware of the keyboard lockups that occur if you don't > bounce on (for example) the numlock key during autoconfig? You are expecting way too much from the core team. We aren't gods, nor do we have any contract with the devil that would provide us with more than 48 hours/day, say in exchange for our souls. :-) The ``must bounce on numlock'' problem came up every now and then, for some minority of the keyboards. > 3) Anyone have an idea why this is happening? Especially if the answer > to question (1) is "no". I'm getting sick of this problem, and would > like to fix it, especially since it doesn't seem to occur in NetBSD or > BSD/OS. Two reasons: · The archtitectural solution for the psm and sc or vt drivers isnt't very clean yet. The ideal solution would consist of an independent kbd driver, and psm plus either sc or vt layered on top of this. Nobody volunteered to really do that job. . These lockup problems arise out of an architectural misdesign of the PC/AT keyboard itself. It's got a three-wire interface (actually four, but the fourth is for power supply only) with some strange serial protocol. It has been designed to operate one-way only: from the keyboard to the computer. Starting with the PC/AT, Little Blue decided to add the reverse direction, too. It's most obvious when it comes to control the keyboard LEDs, and that's also the correlation to ``I have to bounce on the NumLock key!'': this bouncing causes several reverse operations where data is being fed from the computer to the keyboard. Alternatively, some machines lock up during VT switch for the very same reason: the keyboard LEDs are updated, while you're holding one or more keys pressed (that's what has just triggered the VT switch, after all). -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601030916.KAA19702>