amd64/67745: boot fails on compaq presario r3000z

[Andrew Lankford]

 >Because it removes code that does something on other machines.  No one
 >has done a good analysis of why this fixes booting on a few models of
 >laptops.  Thus we don't understand the consequences of this change.
 >-- David  (obrien at FreeBSD.org)

On account of the fact that my machine was no longer bootable shortly 
after the long awaited kernel module loading improvements (due to 
acpi/keyboard driver/whoknowswhat), I've been reluctant to fiddle around 
further with Kim's patch.  I just noticed though that I never tried 
compiling the kernel on any machine with KBDIO_DEBUG set to some high 
value (I take it you can do this from buildkernel?).  Maybe that would 
be useful, provided that the tester doesn't set it to a value high 
enough to bring the machine to its knees.

In the end, I suppose some major FreeBSDer will do a keyboard code 
overhaul eventually as "atkbd0: [GIANT-LOCKED]" still shows up on my p3 
machine boot log.  Then everything will be peachy or peachy/much 
worse/peachy/much much worse/peachy...

Andrew Lankford