I just ordered my Duo Elite set last night, and was double-checking for compatibility problems ( it hasn't even shipped yet ), and I found this thread -> http://www.linuxquestions.org/questi...tech+duo+elite

One person said they called Logitech and got a replacement set which did not have any problems ( such as inserting characters ), while the old set continued to have them ( logitech did not require him to send in the faulty gear ). Unplugging and replugging the keyboard connector seems to be the fix du jour.

I do use a USB KVM switch, so I will be on the lookout for any problems after switching between Linux and Windows boxen.

~Fortezza

I have noticed that when I have my KVM switched to my windows machine throughout the boot process, then switch to the linux machine and watch the full boot process it will switch back and forth with no problems of the 'w' occuring, when I started both machines at the same time, it screwed up.

I've just purchased 3 Logitech Cordless MX Duo keyboard/mouse sets and was having similar problems to what is described in this thread.

Basically you can see the problem by starting in run level 3 (Non gui) and running showkey

[root@posidon root] showkey -s

0x0e
0x8e
0xe1 0x5c 0xe1 0x39 0xe5 0x91


Every time I pressed the backspace. Which resulted in a 'w' appearing in X. This was VERY annoying. I have a suspicion that windows may be tinkering/programming the keyboard controller chip because after putting up with this problem for a week and writing a patch for XFree86 to eat the crap scancodes generated after the backspace. It's now entirely dissapeared and the only reason I can think of is I restarted one of my Linux systems last & the KVM switch was set to it. Unfortunately I cannot reboot my XP computer for the next few hours. (I'm rebuilding the Redhat 9 XFree86 rpms in a terminal window and it takes about 3 hours on the 333mhz that I use for that stuff)

Anyway the following patches solved my problems
---------------------------- cut here-----------------------------------------

--- XFree86-4.3.0/xc/programs/Xserver/hw/xfree86/common/xf86Events.c 2004-07-28 14:51:53.000000000 +0930
+++ XFree86-4.3.0.logitech/xc/programs/Xserver/hw/xfree86/common/xf86Events.c 2004-07-28 15:00:23.000000000 +0930
@@ -446,6 +446,8 @@
switch (scanCode) {
case KEY_Prefix0:
case KEY_Prefix1:
+ case KEY_Prefix2:
+ case KEY_Prefix5:
#if defined(PCCONS_SUPPORT) || defined(SYSCONS_SUPPORT) || defined(PCVT_SUPPORT) || defined(WSCONS_SUPPORT)
if (xf86Info.consType == PCCONS || xf86Info.consType == SYSCONS
|| xf86Info.consType == PCVT
@@ -534,6 +536,22 @@
scanCode = KEY_Pause; /* pause */
}

+ else if (xf86Info.scanPrefix == KEY_Prefix2)
+ {
+ /* Eat it for now */
+ xf86Info.scanPrefix = 0;
+ return;
+ /* return; */
+ }
+
+ else if (xf86Info.scanPrefix == KEY_Prefix5)
+ {
+ /* Eat Logitech BackSpace CRAP */
+ xf86Info.scanPrefix = 0;
+ return;
+ }
+
+
/*
* PC keyboards generate separate key codes for
* Alt+Print and Control+Pause but in the X keyboard model

-------------------------------- end here -------------------------------------------------------
[root@sigma BUILD]# cat ../SOURCES/XFree86-4.3.0.logi-atKeynames.patch
---------------------------------cut here ------------------------------------------------------------
--- XFree86-4.3.0/xc/programs/Xserver/hw/xfree86/common/atKeynames.h 2001-09-30 06:10:30.000000000 +0930
+++ XFree86-4.3.0.logitech/xc/programs/Xserver/hw/xfree86/common/atKeynames.h 2004-07-28 14:43:14.000000000 +0930
@@ -160,6 +160,8 @@

#define KEY_Prefix0 /* special 0x60 */ 96
#define KEY_Prefix1 /* specail 0x61 */ 97
+#define KEY_Prefix2 /* specail 0x62 */ 98
+#define KEY_Prefix5 /* specail 0x65 */ 101

/*
* The 'scancodes' below are generated by the server, because the MF101/102
@@ -199,6 +201,9 @@
#define KEY_NFER /* No Kanji Transfer 0x7b */ 123
#define KEY_Yen /* Yen 0x7d */ 125

+#define KEY_ScrollUp /* Logitech scrollwheel 0x99 */ 153
+#define KEY_ScrollDown /* Logitech scrollwheel 0x9a */ 154
+
/* These are for "notused" and "unknown" entries in translation maps. */
#define KEY_NOTUSED 0
#define KEY_UNKNOWN 255

------------------------------------------- end here-----------------------------------

Please note these diffs were done against the patched Redhat 9.0 src RPM so your milage may vary if you try to apply them against different distributions, but they *should* work.

Well I still do not know if XP is reprogramming the Keyboard controller but the Logitech software certianly does!. When I booted through Linux the little scroll wheel on the left of the keyboard was returning 0xe0 0x48 the same as the up arrow. I switched my KVM to XP and changed the scroll wheels function from scroll to switch apps and it now has a new scancode of 0xe2 0x81. The funny part is no matter what I do it keeps the new scancode. The good news (for me) is this will make it easier to use in X

   Rebuild X to fix a Winbloze issue. Why didn't I guess that?? (smirk)
   None-the-less, many, many bows and thanks to jadx for the patch. Until I know enough karate to rebuild X, the keyboard upluggaroo routine appears to work just dandy. Haven't tried the "boot to the linux box" trick ... yet! (wide grin) That would certainly simplify things.
   FYI, here's the setup:

IOGear MiniView 4 port KVM (not a micro, *no* keyboard shortcuts, just one little ol' button on front to cycle the ports):
- (shiny new) Firewall running RH 9 on port A ... this post is its first foray into being a functional part of the LAN.
- Personal computer running XP Pro on port B
- Development system (ultimately ... old Notes server & CD farm) running RH 9 on port C
- Old Firewall (future undecided) running RH 7.2 on port D ... old reliable has been booting from a floppy forever and just refuses to quit, but it's too slow for content filtering and traffic shaping, plus whatever else I may decide to toss in there. (snicker)

Logitech Cordless Optical Trackman on a separate receiver
Logitech Cordless Access (early multimedia keys version) Keyboard on a separate receiver

   As soon as I got the Cordless Access (no duo, single band receiver) and plugged it into the KVM ... POOF ... broken Linux all over the place. I was in the process of setting up the new firewall (using it now thanks to the replugaroo workaround), so the spurrious "w" characters were, shall we say, problematic. Fortunately (knocks on head) the TrackMan has been exemplary in its performance.
   FYI, this IOGear KVM has been a trooper. Maybe it's the fact that it's just a button, but I've had it for years and years, and (knock on head) it's never even hiccoughed, except that time I shorted out the keyboard that was hooked to it. Port A wasn't the same for a while, but then it got better. (wipes forehead) (bemused grin)
   Now to set up that development platform so I can figure out how to build things. (wrings hands expectantly)

Thanks everyone!!

(            )
 Ragnorok

I've posted the src rpm & patches at http://atlas.netyp.com/patches/XFree86

This rpm incorperates the patches so if you grab the rpm you do not need to grab the patches seperately.

If you run the command rpm -i XFree86-4.3.0-2.90.55.jad.8.src.rpm you will find the patches in /usr/src/redhat/SOURCES/XFree86-4.3.0.logi* along with all the rest of the source code.

The command
rpmbuild --rebuild -ba --clean XFree86-4.3.0-2.90.55.jad.8.src.rpm
will install the src rpm & build binaries or you can do

rpm -i XFree86-4.3.0-2.90.55.jad.8.src.rpm
cd /usr/src/redhat
rpmbuild -ba --clean SPECS/XFree86.spec