[Mageia-dev] ANN: For the brave. systemd v185 in cauldron updates_testing

Fri Jun 29 02:48:02 CEST 2012

'Twas brillig, and Colin Guthrie at 29/06/12 00:35 did gyre and gimble:
+> 'Twas brillig, and Guillaume Rousse at 26/06/12 12:41 did gyre and gimble:
+>> Le 26/06/2012 11:27, Colin Guthrie a écrit :
+>>> Therefore, I'd hope this is just a product of me not rebuilding
+>>> everything under the sun for new udev...
+>>>
+>>> That's my current guess... :)
+>> ...and the winner is....(heavy rolling drum sound)... invalid keyboard
+>> mappings. I'd say some udev rule is just missing, but what is puzzling
+>> me is that only X display is affected, whereas TTY display is OK.
+>>
+>> How are keyboards mapping managed nowadays ?
+> 
+> Right I think I sussed it.
+> 
+> It seems it's done via mageia-setup-keyboard in x11-server-keyboard.
+> 
+> It detects when it's not run via udev, and calls udevadm trigger to
+> import the keyboard settings into the udev database which are
+> subsequently read by evdev etc. when loading.
+> 
+> 
+> In the new systemd the path to udevadm has changed. In order to fix
+> things up for now, I've added a compat symlink to it's old path such
+> that external calls using the full path still succeed.
+> 
+> I've just submitted a new systemd build to updates_testing.
+> 
+> 
+> Can users please try deleting the /etc/X11/xorg.conf.d/00-keyboard.conf
+> file and rebooting to make sure the keymap is set correctly (especially
+> in the DM if you are using Gnome as user-specific settings may override
+> the initial defaults).
+> 
+> Assuming I get confirmation that this is OK, I'll submit this tomorrow.
+
+Very Important!!!!! Also update x11-server too!
+
+It's actually really lucky the path to udevadm changed as I'd have
+otherwise been super confused! It turns out correcting this problem
+highlighted another - an infinite trigger loop in udev.
+
+The rules file:
+  /etc/udev/rules.d/61-x11-input.rules
+calls /sbin/mageia-setup-keyboard.
+
+
+mageia-setup-keyboard in turn uses the variable UDEV_LOG to work out if
+we're called within udev or not. If not it assumes it's being run
+"standalone" rather than via udev.
+
+
+I've fixed this looping now and it should be backwards compatible with
+the current udev.
+
+So make sure you update the x11-server at the same time otherwise you
+might have a hard time until you reboot.
+
+If you do get stuck, just rm /sbin/udevadm symlink in rescue.target (aka
+runlevel 1). udevd will be happily going quite nuts until you reboot :)
+
+Col
+
+
+-- 
+
+Colin Guthrie
+colin(at)mageia.org
+http://colin.guthr.ie/
+
+Day Job:
+  Tribalogic Limited http://www.tribalogic.net/
+Open Source:
+  Mageia Contributor http://www.mageia.org/
+  PulseAudio Hacker http://www.pulseaudio.org/
+  Trac Hacker http://trac.edgewall.org/
+
+
+