This bug has been around since at least FC3, and I figured it'd be fixed by now, but it hasn't (I've just been putting up with it) and it's my #1 annoyance, so I'm submitting this bug to hopefully have some attention drawn to it again (there are no open dupes). So since FC3 I've been using KDE + Fedora, and in that time I've used two different MS Natural keyboards where the mod4 (windows key) would only work *randomly* after X login. To fix it I have to reapply the same keyboard layout over and over until it sticks using either the GUI ("kcmshell keyboard_layout") config launched from the "kde keyboard tool" systray util, or directly from the commandline with: "setxkbmap -model microsoftprooem -layout us". The common keycombo I use to test that my layout is working is Win+O which tells Amarok to show the current song info in an overlay OSD. When everything IS working, "xev" will output the following for Win+O: --------------------------------------------------------------------- KeyPress event, serial 32, synthetic NO, window 0x4e00001, root 0x1a5, subw 0x0, time 2739616650, (333,488), root:(462,766), state 0x10, keycode 115 (keysym 0xffeb, Super_L), same_screen YES, XLookupString gives 0 bytes: XmbLookupString gives 0 bytes: XFilterEvent returns: False FocusOut event, serial 32, synthetic NO, window 0x4e00001, mode NotifyGrab, detail NotifyAncestor FocusIn event, serial 32, synthetic NO, window 0x4e00001, mode NotifyUngrab, detail NotifyAncestor KeymapNotify event, serial 32, synthetic NO, window 0x0, keys: 2 0 0 0 1 0 0 0 0 0 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 KeyRelease event, serial 32, synthetic NO, window 0x4e00001, root 0x1a5, subw 0x0, time 2739617194, (333,488), root:(462,766), state 0x50, keycode 32 (keysym 0x6f, o), same_screen YES, XLookupString gives 1 bytes: (6f) "o" XFilterEvent returns: False KeyRelease event, serial 32, synthetic NO, window 0x4e00001, root 0x1a5, subw 0x0, time 2739617378, (333,488), root:(462,766), state 0x50, keycode 115 (keysym 0xffeb, Super_L), same_screen YES, XLookupString gives 0 bytes: XFilterEvent returns: False --------------------------------------------------------------------- When NOT working, the KeymapNotify event is missing, and there are 4 press,press,release,release events (and Amarok doesn't recieve it - instead the "o" goes to the commandline). Related bug (CLOSED w/ INSUFFICIENT_DATA): https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=214427
From the above-mentioned bug you could know that I am pretty insisting on getting /etc/X11/xorg.conf and /var/log/Xorg.*.log. Please attach them to this bug as uncompressed separate attachments. Also (especially because you seem to have using this system for a long time), could you please try to restart X without any /etc/X11/xorg.conf whatsoever and tell us (with /var/log/Xorg.*.log is appropriate) what happened? Thanks a lot.
Created attachment 151608 [details] my day-to-day xorg.conf
Created attachment 151609 [details] Xorg.0.log with my xorg.conf
Created attachment 151610 [details] Xorg.0.log with no xorg.conf present
I created a testuser to do a little more testing: 1) This bug occurs under both KDE and Gnome. 2) It occurs with my xorg.conf, or with the automagically detected defaults when no xorg.conf is present. Also, I notice the following error *sometimes* being appended to Xorg.0.log each time I attempt to get my keyboard layout to stick (with "setxkbmap -model microsoftprooem -layout us"): BOGUS LENGTH in write keyboard desc, expected 4800, got 4804 BOGUS LENGTH in write keyboard desc, expected 5288, got 5292 BOGUS LENGTH in write keyboard desc, expected 4812, got 4816 ...
Bug is still present in Fedora7, and it isn't dependent on the keyboard model. If I repeat the following a random number of times, suddenly the WinKey works again: $ setxkbmap -model pc104 -layout us Do it one more time, and... the key's ignore again. rinse/repeat. So aggravating.
... and still present in Fedora8
FWIW - I haven't been able to reproduce this since switching from F8 i386 to F8 x86_64 a few weeks ago. WinKey has worked reliably.
Confirming this still occurs under F8 on i686, and only random restarts of X or repetition-unto-success reruns of (setxkbmap -model pc105 -layout us), get Super properly remapped to the Win-keys again. I run Xfce but problem occurs randomly on gnome and fluxbox alike, as well. current kernel is 2.6.24.3-34.fc8 (later ones as of this moment are still causing panics) If there's any other details I can provide, dont' hesitate to ask.
Can you cross-check with Bug 239409 please? Sounds like it's the same.
(In reply to comment #10) > Can you cross-check with Bug 239409 please? Sounds like it's the same. It does look to be similar. I haven't been able to reproduce this problem in F9 x86_64 (or F10a x86_64). It's possible that F9 i386 is still affected.
*** This bug has been marked as a duplicate of bug 239409 ***