Re: USB keyboard problems

On Tue, Aug 31, 2004 at 04:55:26PM -0400, John Baldwin wrote:
> On Friday 27 August 2004 11:29 pm, M. Warner Losh wrote:
> > In message: <20040828025829.GA51618_at_prophecy.dyndns.org>
> >
> >             Christopher Nehren <apeiron_at_comcast.net> writes:
> > : On Fri, Aug 27, 2004 at 22:41:32 EDT, M. Warner Losh scribbled these
> > :
> > : curious markings:
> > : > I guess it all depends on what you mean by recently.  I'd thought that
> > : > some of my recent changes had broken it, but the breakage goes back
> > : > further than that.  5.2 release isn't 'recent', and I fully believe
> > : > that things may have changed since then...  I see a major uhid upgrade
> > : > in that time frame, which may be the time of breakage...
> > :
> > : I understand your point, but that doesn't sit well with the fact that my
> > : joypad broke *after* I updated from 5.2-CURRENT of ~ August 15 to
> > : 6.0-CURRENT of ~ August 19. Logically, if it had been busted by the uhid
> > : upgrade in March, then it could not possibly have worked with 5.2-CURRENT
> > : (which it did, just as well as it did with 5.2.1-RELEASE[1]), correct?
> > :
> > : [1] Better, in fact, as the device wasn't detected at boot time in 5.2.1
> > : but was in 5.2-CURRENT.
> >
> > Then I'm very confused...  Time to dig deeper...
> 
> Alfred turned the atkbd probing off a while back:
> 
> alfred      2004/04/01 13:48:31 PST
> 
>   FreeBSD src repository
> 
>   Modified files:
>     sys/i386/conf        GENERIC.hints
>   Log:
>   Fix booting with ps2 keyboards.
> 
>   Revision  Changes    Path
>   1.13      +0 -1      src/sys/i386/conf/GENERIC.hints
> 
> This didn't actually affect booting with a PS/2 keyboard, but only meant that 
> you could boot up without a PS/2 keyboard and then plug it in, at the cost of 
> breaking all uses of USB keyboards unless you use explicit kbdcontrol 
> commands.  The instant-MFC was backed out of RELENG_4 at the request of re_at_ 
> due to POLA.

The following entries in my devd.conf make things much happier on my
laptop.  Maybe we should commit something like this to the default file:

# When a keyboard arrives, attach it as the console keyboard
attach 100 {
	device-name "ukbd0";
	action "test -c /dev/kbd1 && kbdcontrol -k /dev/kbd1 < /dev/console";
}
detach 100 {
	device-name "ukbd0";
	action "kbdcontrol -k /dev/kbd0 < /dev/console";
}

It has the slightly weird effect that the built in keyboard stops
working for console when you plug a USB keyboard in, but IMO that's less
lame then the current setup.

-- Brooks

-- 
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4