i386/67739: smth was broken in keyboard processing since 5.2.1-RELEASE

[Seva Gluschenko]

>Number:         67739
>Category:       i386
>Synopsis:       smth was broken in keyboard processing since 5.2.1-RELEASE
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    freebsd-i386
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Jun 09 07:20:20 GMT 2004
>Closed-Date:
>Last-Modified:
>Originator:     Seva Gluschenko
>Release:        FreeBSD 5.2-CURRENT i386
>Organization:
Yandex LLC
>Environment:
System: FreeBSD road.yandex.ru 5.2-CURRENT FreeBSD 5.2-CURRENT #8: Fri Jun 4 13:37:30 MSD 2004 gvs at road.yandex.ru:/local/obj/local/usr/src/sys/ROAD i386


	
>Description:
	My workstation operates on 5.2-CURRENT and I noticed some misterious
	keyboard behaviour after some point in March, I think. At first, an
	attempt to switch from X to text console and type something there
	resulted to X Server crash. I didn't bothered much considering it
	just XFree86-4.3.0 (from ports) bug.

	Next I noticed occasional "cycling" of keyboard input, only <Space>
	and <Enter> was affected. It means both ignorance for the key pressed
	(lost event?) and occasional repeating (just like Enter or Space was
	pressed many times while it actually was pressed just once) which stops
	right after the next keypress.

	It was hardly to believe it's XFree86 bug again, 'cause 5.2 before March
	and 4.X didn't show that behaviour, but it happens not to often, so I
	didn't care much.

	But, I tried to update my homestation from 5.2.1-RELEASE to 5.2-CURRENT
	last evening, and the result was terrible. The key difference is my
	workstation has PS/2 keyb but my homestation has USB keyb.

	'twas alright after installworld/mergemaster/reboot, but I noticed the
	absence of acpi.ko (complained during boot process), so I did
	cd /sys/modules/acpi ; make install clean ; reboot
	Right after reboot things went really bad: the keypresses were totally
	ignored for about first ten seconds, then the first symbol entered (say, l)
	was repeated ad infinitum or, speaking determiningly, up to reset button
	pressed (because there were no more method eventually).

	I was forced to install 5.2.1 back from CD because there were no more ways
	to return the homestation under my control.

	I apologize for very long and not too technical bug report, but I just don't
	know what should I inspect to provide the needed amount of technical details.
>How-To-Repeat:
	
>Fix:

	


>Release-Note:
>Audit-Trail:
>Unformatted: