Re: Linux keyboard support, Logitech Access Keyboard, Multimedia keys

Hi!

Thank you for your reply.

Dances With Crows schrieb:

On Mon, 10 Jul 2006 23:12:10 +0200, Andreas Schallenberg staggered into
the Black Sun and said:
I got a [Logitech Access] wireless keyboard and found some keys not
working. The keyboard has 10 extra keys with just one mapping.

You mean all 10 extra keys produce the same keycode according to xev
and/or showkey? That's weird.

These ten keys have one individual scancode each. The other "special keys"
are on the same (physical) keys than the F-Keys. It depends on the
"F Lock" status (which is invisible to the computer) wether they
produce the normal scancodes for the F-keys (F1 to F12) or the
scancodes for their special meaning. This is what I meant with
a second mapping. I hope it is more understandable now :)

Two of the ten extra keys are "Messenger/SMS" and "Webcam". These
produce scancodes which are known to the kernel and should be "e0 11"
(M/SMS) and "e0 12" (Webcam) according to
http://www.johanneshuebner.com/ge/linux.shtml#tastatur
However, they don't work as keys! "xev" gives the following output:

Key "Webcam"
ButtonPress event, serial 31, synthetic NO, window 0x4200001,
state 0x10, button 4, same_screen YES

Key "Messenger/SMS"
ButtonPress event, serial 31, synthetic NO, window 0x4200001,
state 0x10, button 5, same_screen YES

I am not sure, but this looks like these are [being] interpreted as
mouse events! This confuses me.

Sounds like a problem with your X configuration somewhere, or possibly
the scancode->keycode map you're using isn't right.

Are the keycodes used for mouse events, too? I expected
a device to produce scancodes (translated into keycodes)
if detected as a keyboard, OR mouse events if detected
as a mouse.

What's the
keyboard's InputDevice section of your xorg.conf look like? Reproduce
that section in a followup. Are you using XKB? Which keyboard layout
are you using?

I am using Xorg 6.8.2. The input section is this:

Section "InputDevice"
Driver "kbd"
Identifier "Keyboard[0]"
Option "Protocol" "Standard"
Option "XkbLayout" "de"
Option "XkbModel" "pc105"
Option "XkbRules" "xfree86"
Option "XkbVariant" "nodeadkeys"
EndSection

Section "InputDevice"
Driver "mouse"
Identifier "Mouse[1]"
Option "Buttons" "7"
Option "Device" "/dev/input/mice"
Option "Name" "ImExPS/2 Logitech Explorer Mouse"
Option "Protocol" "explorerps/2"
Option "Vendor" "Sysp"
Option "ZAxisMapping" "4 5"
EndSection

Is there any way this [problem has anything to do with] the kernel?

Impossible to tell without more info. Is this thing's interface PS/2 or
USB? Which kernel version are you using?

It depends on the interface? Both are connected to the computer
via PS/2. For the keyboard I use an USB to PS/2 adaptor which
came with the keyboard. I'll omit it this evening and connect the
keyboard via USB.

The twelve F keys which have a second mapping produce scancodes which
are not known by the kernel. atkbd.c prints a warning to the kernel
log.

The device is probably too new for your kernel to grok. I still don't
understand what you mean by "second mapping" here. When you press F1,
do you get kernel messages, or do you only get them when you press
{Shift, Ctrl, CapsLock, NumLock}-F1?

I hope it is clear now. See the first paragraph of this posting.
The kernel is "2.6.11.4-21.12-default". SUSE 9.2 here.
I'll download a linux live-cd (maybe dapper drake or knoppix)
and see what newer kernels do.

What is the usual way of dealing with this situation?

Once the thing's reporting sane scancodes, use xbindkeys (or
gtk-xbindkeys) to make the extra keys do useful things. Making the
keyboard report sane scancodes may require a bit of screwing around with
loadkeys.

Who assigns the static mapping between the scancode and keycode
tables? That is, who [do I] send the scancode table [to]?

"man loadkeys showkey" to do it in console. In X, I have no idea.
There's a table somewhere, but XKB is the world's most awesome mess and
I don't have any idea where that table is.

Thank your for these pointers!

I'll post again when the experiments (PS/2->USB, newer kernel) are done.

RU,
Andreas
.

Quantcast