[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: etch to lenny upgrade - X apps no longer see keystrokes?



Thanks again Florian.....

Received Sat 18 Apr 2009  3:34am +1000 from Florian Kulzer:
> On Fri, Apr 17, 2009 at 14:46:58 +1000, Graham Williams wrote:
> > Received Wed 15 Apr 2009  5:32am +1000 from Florian Kulzer:
> > > On Tue, Apr 14, 2009 at 09:21:11 +1000, Graham Williams wrote:
> > > > Received Fri 10 Apr 2009  6:31am +1000 from Florian Kulzer:
> > > > > On Thu, Apr 09, 2009 at 18:10:41 +1000, Graham Williams wrote:
> > > > > > Have just upgraded
[...]
> > > > >From etch to lenny, as per Subject.
[...]
> > > > 07:00.0 VGA compatible controller [0300]: nVidia Corporation NV43GL [Quadro FX 550] [10de:014d] (rev a2)
[...]
> > > dpkg -l udev {,lib}hal\* {,lib}dbus\* xserver-xorg\* libx11\* xkb\* | awk '/ii/{print$2,$3}'
> 
> [ output edited ]
> 
> > udev 0.125-7
> 
> You should upgrade udev to version 0.125-7+lenny1 (security.debian.org).
> 
> > xserver-xorg-core 2:1.4.2-10
> 
> Rmadison tells me that the current version of this package for Lenny is
> 2:1.4.2-10.lenny1. I would try to upgrade to that. 
> 
> Other than those two packages, I did not see anything unusual your list.

I have upgrade both. No change in the X/kbd behaviour

> > > Another thing to check is which processes are using files in
> > > /dev/input/.  Ideally, this check should be done after X has started.
> > > Using CTRL-ALT-Fn does not work for you, but you could use a simple
> > > ~/.xinitrc that runs "sudo chvt 1" in an xterm, which would return you
> > > to the text terminal. (You have to configure your system to allow your
> > > user to run sudo with this command without password.) Then I would like
> > > to the output of:
> > > 
> > > lsof /dev/input/*

I am getting:

COMMAND    PID USER   FD   TYPE DEVICE SIZE NODE NAME
hald-addo 4357 root    4r   CHR  13,69      5819 /dev/input/event5
hald-addo 4357 root    5w   CHR  13,68      5772 /dev/input/event4
hald-addo 4357 root    6r   CHR  13,67      5654 /dev/input/event3
hald-addo 4357 root    7r   CHR  13,66      5620 /dev/input/event2

> Another thing to check is if certain processes are running:
> 
> ps -ef | grep -E 'X|hal|dbus|udev'

root      1370     1  0 13:38 ?        00:00:00 udevd --daemon
103       2862     1  0 13:38 ?        00:00:00 /usr/bin/dbus-daemon --system
root      3775  3638  0 13:39 ?        00:00:00 /sbin/dhclient -1 -lf /var/lib/dhcp3/dhclient.eth1.leases -pf /var/run/dhclient.eth1.pid -q -e dhc_dbus=31 -d eth1
105       4336     1  0 13:44 ?        00:00:00 /usr/sbin/hald
root      4337  4336  0 13:44 ?        00:00:00 hald-runner
root      4357  4337  0 13:44 ?        00:00:00 hald-addon-input: Listening on /dev/input/event5 /dev/input/event4 /dev/input/event3 /dev/input/event2
105       4362  4337  0 13:44 ?        00:00:00 hald-addon-acpi: listening on acpid socket /var/run/acpid.socket
root      4368  4337  0 13:44 ?        00:00:00 hald-addon-storage: polling /dev/hda (every 2 sec)
anet      4421  4404  0 13:44 tty2     00:00:00 xinit /home/anet/.xinitrc -- /etc/X11/xinit/xserverrc :0 -auth /tmp/serverauth.sVdbIjaCrQ
root      4422  4421  0 13:44 tty7     00:00:00 /usr/bin/X11/X -nolisten tcp
root      4439  3960  0 13:45 tty1     00:00:00 grep -E X|hal|dbus|udev

> > Note that I am booting single user mode to do this, so it is logging
> > in as root and running startx as root. I modified the .xinitrc to chvt
> > 1 and to then xterm. I ran lsof on the console.
> 
> I think it would be better to do further tests in the normal runlevel.
> I would temporarily uninstall or at least disable [xkg]dm to allow you
> to log in at the normal tty prompt.

Yes - now in user land.

> > I've been playing with xev, looks like it is getting KeyRelease events
> > but not the KeyPress events for the keys that actually result in the
> > screen resolution being reset. Without knowing how hal and the kbd
> > device works, it is almost as if X is capturing these KeyPress events
> > and not passing them on, instead treating them as a screen resolution
> > change shortcut.
> 
> I think the capturing would be normal for real resolution-change key
> combinations, but we have to figure out why your system misidentifies
> other key press events as this combination.

Regards,
Graham


Reply to: