I have personal XKB settings. From my .initrc file:

  xkbcomp -w0 -I$HOME/.xkb -R$HOME/.xkb keymap/custom $DISPLAY

They are lost after suspend (hibernate) / resume.

Possible related bugs:

 #541388  xserver-xorg: Xmodmap settings lost across suspend/hibernate
 #568868  key repeat for caps lock goes away after suspend/resume
 #582566  bell settings (xset -b) not preserved after suspend/resume

severity 633849 important
thanks

Increasing severity as after I restore the settings with

  xkbcomp -w0 -I$HOME/.xkb -R$HOME/.xkb keymap/custom $DISPLAY

(the same command I use in my .xinitrc, and "xkbcomp $DISPLAY xkb.dump"
shows that they are completely restored), the behavior is still
incorrect. For instance, I have:

    type "CUSTLEVELTHREE" {
        modifiers= LevelThree;
        map[LevelThree]= Level2;
        level_name[Level1]= "Base";
        level_name[Level2]= "Alt";
    };

    key <LEFT> {
        type= "CUSTLEVELTHREE",
        symbols[Group1]= [            Left,       0x1000001 ]
    };

(see attached file) but Alt+Left doesn't yield a ^A. This means that
after suspend/resume, I need to log out!

retitle 633849 xserver-xorg: XKB settings lost after suspend (hibernate) / resume or USB keyboard plugged in
found 633849 1:7.6+11
thanks
do not take these settings into account (the keyboard of the laptop
is unaffected).

And running

  xkbcomp -w0 -I$HOME/.xkb -R$HOME/.xkb keymap/custom $DISPLAY

after plugging in the keyboard is not really useful because some
X applications (like xterm and fvwm) still partly behave as with
the old settings (for this one, I suspect a bug in libx11-6,
since both xterm and fvwm depend on it). There's also the problem
concerning the user CUSTLEVELTHREE type mentioned in another
message.

I remap modifier keys quite a bit:

partial modifier_keys
xkb_symbols "lisp_machine_modifiers" {
    key <LCTL> { [ Super_L, Super_L ] };
    key <RCTL> { [ Super_R, Super_R ] };
    key <LALT> { [ Control_L, Control_L ] };
    key <RALT> { [ Control_R, Control_R ] };
    key <LWIN> { [ Alt_L, Meta_L ] };
    key <MENU> { [ Alt_R, Meta_R ] };
    modifier_map Control { <RALT>, <LALT> };
    modifier_map Mod1 { <LWIN>, <MENU> };
    modifier_map Mod4 { <LCTL>, <RCTL> };
};

After suspend/hibernate LCTL, RCTL and MENU revert to their original
meaning. Switching to the console and back is also enough to trigger it.
I'm running a fully updated (as of today) Wheezy.

I've also found that the plugging in a usb headset also upsets the
keyboard layout ...

Actually when I suspend and resume the laptop, the USB keyboard
is unaffected. Only when unplug it and plug it in again.

its builtin keyboard, I noticed that the XKB settings were restored
(without the need to run xkbcomp), and everything worked fine.

I looked at the kernel and Xorg logs (see below), and saw that the
builtin keyboard was recognized as a USB HID keyboard instead of an
AT Translated Set 2 keyboard as usual. But I've done other tests and
I can't make it work again.

USB keyboard recognition needs the RF_KILL bit to be set to enable
radio. I don't know whether this is a bug or there is a reason.
But the bug (XKB settings lost after suspend) needs to be fixed
one way or another. Since the problem may come from the kernel in
the first place, this could be a kernel bug, the problem in Xorg
being just a consequence.

Here are the logs, when the XKB settings were restored:

In the /var/log/syslog file, keyboard information did not appear after
wake up, until the RF_KILL bit was toggled to enable radio:

[...]
Aug 10 21:11:25 xvii kernel: [235310.726023] iwlwifi 0000:0c:00.0: RF_KILL bit toggled to enable radio.
Aug 10 21:11:26 xvii kernel: [235311.208293] usb 3-1: new full-speed USB device number 16 using uhci_hcd
Aug 10 21:11:26 xvii kernel: [235311.385578] usb 3-1: New USB device found, idVendor=0a5c, idProduct=4500
Aug 10 21:11:26 xvii kernel: [235311.385588] usb 3-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Aug 10 21:11:26 xvii kernel: [235311.385595] usb 3-1: Product: BCM2046B1
Aug 10 21:11:26 xvii kernel: [235311.385599] usb 3-1: Manufacturer: Broadcom
Aug 10 21:11:26 xvii kernel: [235311.387503] hub 3-1:1.0: USB hub found
Aug 10 21:11:26 xvii kernel: [235311.389335] hub 3-1:1.0: 3 ports detected
Aug 10 21:11:26 xvii kernel: [235311.670578] usb 3-1.1: new full-speed USB device number 17 using uhci_hcd
Aug 10 21:11:26 xvii kernel: [235311.788586] usb 3-1.1: New USB device found, idVendor=413c, idProduct=8157
Aug 10 21:11:26 xvii kernel: [235311.788596] usb 3-1.1: New USB device strings: Mfr=0, Product=0, SerialNumber=0
Aug 10 21:11:26 xvii kernel: [235311.794006] input: HID 413c:8157 as /devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1.1/3-1.1:1.0/input/input28
Aug 10 21:11:26 xvii kernel: [235311.794332] generic-usb 0003:413C:8157.000A: input,hidraw0: USB HID v1.11 Keyboard [HID 413c:8157] on usb-0000:00:1a.0-1.1/input0
[...]

In the /var/log/Xorg.0.log file after wake up:

[235125.841] (II) Open ACPI successful (/var/run/acpid.socket)
[235125.841] (II) AIGLX: Resuming AIGLX clients after VT switch
[235125.841] (II) NOUVEAU(0): NVEnterVT is called.
[235125.990] (II) NOUVEAU(0): EDID vendor "AUO", prod id 16455
[235125.990] (II) NOUVEAU(0): Printing DDC gathered Modelines:
[235125.990] (II) NOUVEAU(0): Modeline "1440x900"x0.0  108.00  1440 1504 1536 1970  900 903 906 912 +hsync -vsync (54.8 kHz eP)
[235125.990] (II) NOUVEAU(0): Modeline "1440x900"x0.0   72.50  1440 1504 1536 1970  900 903 906 912 +hsync -vsync (36.8 kHz e)
[235126.117] (**) Option "Device" "/dev/input/event10"
[235126.117] (--) synaptics: AlpsPS/2 ALPS DualPoint TouchPad: touchpad found
[235312.020] (II) config/udev: Adding input device HID 413c:8157 (/dev/input/event1)
[235312.020] (**) HID 413c:8157: Applying InputClass "evdev keyboard catchall"
[235312.020] (II) Using input driver 'evdev' for 'HID 413c:8157'
[235312.020]    Option "XkbRules" "evdev"
[235312.020]    Option "xkb_model" "pc105"
[235312.020]    Option "xkb_layout" "gb"
[235312.020]    Option "_source" "server/udev"
[235312.020]    Option "name" "HID 413c:8157"
[235312.020]    Option "path" "/dev/input/event1"
[235312.020]    Option "device" "/dev/input/event1"
[235312.020]    Option "xkb_options" "lv3:ralt_switch"
[235312.020]    Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1.1/3-1.1:1.0/input/input28/event1"
[235312.020]    Option "driver" "evdev"
[235312.020] (**) HID 413c:8157: always reports core events
[235312.020] (**) evdev: HID 413c:8157: Device: "/dev/input/event1"
[235312.020] (--) evdev: HID 413c:8157: Vendor 0x413c Product 0x8157
[235312.020] (--) evdev: HID 413c:8157: Found keys
[235312.020] (II) evdev: HID 413c:8157: Configuring as keyboard
[235312.020] (**) Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1.1/3-1.1:1.0/input/input28/event1"
[235312.020] (II) XINPUT: Adding extended input device "HID 413c:8157" (type: KEYBOARD, id 15)
[235312.020] (**) Option "xkb_rules" "evdev"
[235312.020] (**) Option "xkb_model" "pc105"
[235312.020] (**) Option "xkb_layout" "gb"
[235312.020] (**) Option "xkb_options" "lv3:ralt_switch"
[235312.020] (II) XKB: Reusing cached keymap

With a second try...

In the /var/log/syslog file:

[...]
Aug 10 22:03:07 xvii kernel: [238341.135708] input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input32
[...]

And after toggling the RF_KILL bit to enable radio:

Aug 10 22:10:17 xvii kernel: [238774.107477] iwlwifi 0000:0c:00.0: RF_KILL bit toggled to enable radio.
Aug 10 22:10:18 xvii kernel: [238774.588232] usb 3-1: new full-speed USB device number 22 using uhci_hcd
Aug 10 22:10:18 xvii kernel: [238774.765505] usb 3-1: New USB device found, idVendor=0a5c, idProduct=4500
Aug 10 22:10:18 xvii kernel: [238774.765515] usb 3-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
Aug 10 22:10:18 xvii kernel: [238774.765522] usb 3-1: Product: BCM2046B1
Aug 10 22:10:18 xvii kernel: [238774.765527] usb 3-1: Manufacturer: Broadcom
Aug 10 22:10:18 xvii kernel: [238774.767445] hub 3-1:1.0: USB hub found
Aug 10 22:10:18 xvii kernel: [238774.769264] hub 3-1:1.0: 3 ports detected
Aug 10 22:10:18 xvii kernel: [238775.050498] usb 3-1.1: new full-speed USB device number 23 using uhci_hcd
Aug 10 22:10:19 xvii kernel: [238775.169504] usb 3-1.1: New USB device found, idVendor=413c, idProduct=8157
Aug 10 22:10:19 xvii kernel: [238775.169514] usb 3-1.1: New USB device strings: Mfr=0, Product=0, SerialNumber=0
Aug 10 22:10:19 xvii kernel: [238775.175898] input: HID 413c:8157 as /devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1.1/3-1.1:1.0/input/input33
Aug 10 22:10:19 xvii kernel: [238775.176371] generic-usb 0003:413C:8157.000C: input,hidraw0: USB HID v1.11 Keyboard [HID 413c:8157] on usb-0000:00:1a.0-1.1/input0
[...]

In the /var/log/Xorg.0.log file:

[238344.158] (II) Open ACPI successful (/var/run/acpid.socket)
[238344.158] (II) AIGLX: Resuming AIGLX clients after VT switch
[238344.158] (II) NOUVEAU(0): NVEnterVT is called.
[238344.297] (II) NOUVEAU(0): EDID vendor "AUO", prod id 16455
[238344.297] (II) NOUVEAU(0): Printing DDC gathered Modelines:
[238344.297] (II) NOUVEAU(0): Modeline "1440x900"x0.0  108.00  1440 1504 1536 1970  900 903 906 912 +hsync -vsync (54.8 kHz eP)
[238344.297] (II) NOUVEAU(0): Modeline "1440x900"x0.0   72.50  1440 1504 1536 1970  900 903 906 912 +hsync -vsync (36.8 kHz e)
[238344.420] (**) Option "Device" "/dev/input/event10"
[238344.420] (--) synaptics: AlpsPS/2 ALPS DualPoint TouchPad: touchpad found
[238344.421] (II) config/udev: Adding input device AT Translated Set 2 keyboard (/dev/input/event0)
[238344.421] (**) AT Translated Set 2 keyboard: Applying InputClass "evdev keyboard catchall"
[238344.421] (II) Using input driver 'evdev' for 'AT Translated Set 2 keyboard'
[238344.421]    Option "XkbRules" "evdev"
[238344.421]    Option "xkb_model" "pc105"
[238344.421]    Option "xkb_layout" "gb"
[238344.421]    Option "_source" "server/udev"
[238344.421]    Option "name" "AT Translated Set 2 keyboard"
[238344.421]    Option "path" "/dev/input/event0"
[238344.421]    Option "device" "/dev/input/event0"
[238344.421]    Option "xkb_options" "lv3:ralt_switch"
[238344.421]    Option "config_info" "udev:/sys/devices/platform/i8042/serio0/input/input32/event0"
[238344.421]    Option "driver" "evdev"
[238344.421] (**) AT Translated Set 2 keyboard: always reports core events
[238344.421] (**) evdev: AT Translated Set 2 keyboard: Device: "/dev/input/event0"
[238344.421] (--) evdev: AT Translated Set 2 keyboard: Vendor 0x1 Product 0x1
[238344.421] (--) evdev: AT Translated Set 2 keyboard: Found keys
[238344.421] (II) evdev: AT Translated Set 2 keyboard: Configuring as keyboard
[238344.421] (**) Option "config_info" "udev:/sys/devices/platform/i8042/serio0/input/input32/event0"
[238344.421] (II) XINPUT: Adding extended input device "AT Translated Set 2 keyboard" (type: KEYBOARD, id 9)
[238344.421] (**) Option "xkb_rules" "evdev"
[238344.421] (**) Option "xkb_model" "pc105"
[238344.421] (**) Option "xkb_layout" "gb"
[238344.421] (**) Option "xkb_options" "lv3:ralt_switch"
[238344.421] (II) XKB: Reusing cached keymap
[238775.401] (II) config/udev: Adding input device HID 413c:8157 (/dev/input/event1)
[238775.401] (**) HID 413c:8157: Applying InputClass "evdev keyboard catchall"
[238775.401] (II) Using input driver 'evdev' for 'HID 413c:8157'
[238775.401]    Option "XkbRules" "evdev"
[238775.401]    Option "xkb_model" "pc105"
[238775.401]    Option "xkb_layout" "gb"
[238775.401]    Option "_source" "server/udev"
[238775.401]    Option "name" "HID 413c:8157"
[238775.401]    Option "path" "/dev/input/event1"
[238775.401]    Option "device" "/dev/input/event1"
[238775.401]    Option "xkb_options" "lv3:ralt_switch"
[238775.401]    Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1.1/3-1.1:1.0/input/input33/event1"
[238775.401]    Option "driver" "evdev"
[238775.401] (**) HID 413c:8157: always reports core events
[238775.401] (**) evdev: HID 413c:8157: Device: "/dev/input/event1"
[238775.401] (--) evdev: HID 413c:8157: Vendor 0x413c Product 0x8157
[238775.401] (--) evdev: HID 413c:8157: Found keys
[238775.401] (II) evdev: HID 413c:8157: Configuring as keyboard
[238775.401] (**) Option "config_info" "udev:/sys/devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1.1/3-1.1:1.0/input/input33/event1"
[238775.401] (II) XINPUT: Adding extended input device "HID 413c:8157" (type: KEYBOARD, id 11)
[238775.402] (**) Option "xkb_rules" "evdev"
[238775.402] (**) Option "xkb_model" "pc105"
[238775.402] (**) Option "xkb_layout" "gb"
[238775.402] (**) Option "xkb_options" "lv3:ralt_switch"
[238775.402] (II) XKB: Reusing cached keymap

But the XKB settings were not restored (whether before or after
the RF_KILL bit was toggled).

A third try, letting the RF_KILL bit always on to enable radio,
i.e. before suspend and after wake up (resume).

In the /var/log/syslog file:

[...]
Aug 10 22:21:34 xvii kernel: [239409.521324] input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input35
Aug 10 22:21:34 xvii kernel: [239409.564618] ata2.00: configured for UDMA/100
Aug 10 22:21:34 xvii kernel: [239409.572067] usb 3-1: reset full-speed USB device number 22 using uhci_hcd
Aug 10 22:21:34 xvii kernel: [239410.009423] dell_wmi: Received unknown WMI event (0x11)
Aug 10 22:21:34 xvii kernel: [239410.013989] usb 3-1.1: reset full-speed USB device number 23 using uhci_hcd
Aug 10 22:21:34 xvii kernel: [239410.185946] usb 3-1.2: reset full-speed USB device number 24 using uhci_hcd
[...]

In the /var/log/Xorg.0.log file:

[239412.722] (II) config/udev: Adding input device AT Translated Set 2 keyboard (/dev/input/event0)
[239412.722] (**) AT Translated Set 2 keyboard: Applying InputClass "evdev keyboard catchall"
[239412.722] (II) Using input driver 'evdev' for 'AT Translated Set 2 keyboard'
[239412.722]    Option "XkbRules" "evdev"
[239412.722]    Option "xkb_model" "pc105"
[239412.722]    Option "xkb_layout" "gb"
[239412.722]    Option "_source" "server/udev"
[239412.722]    Option "name" "AT Translated Set 2 keyboard"
[239412.722]    Option "path" "/dev/input/event0"
[239412.722]    Option "device" "/dev/input/event0"
[239412.722]    Option "xkb_options" "lv3:ralt_switch"
[239412.722]    Option "config_info" "udev:/sys/devices/platform/i8042/serio0/input/input35/event0"
[239412.722]    Option "driver" "evdev"
[239412.722] (**) AT Translated Set 2 keyboard: always reports core events
[239412.722] (**) evdev: AT Translated Set 2 keyboard: Device: "/dev/input/event0"
[239412.722] (--) evdev: AT Translated Set 2 keyboard: Vendor 0x1 Product 0x1
[239412.722] (--) evdev: AT Translated Set 2 keyboard: Found keys
[239412.722] (II) evdev: AT Translated Set 2 keyboard: Configuring as keyboard
[239412.722] (**) Option "config_info" "udev:/sys/devices/platform/i8042/serio0/input/input35/event0"
[239412.722] (II) XINPUT: Adding extended input device "AT Translated Set 2 keyboard" (type: KEYBOARD, id 11)
[239412.722] (**) Option "xkb_rules" "evdev"
[239412.722] (**) Option "xkb_model" "pc105"
[239412.722] (**) Option "xkb_layout" "gb"
[239412.722] (**) Option "xkb_options" "lv3:ralt_switch"
[239412.722] (II) XKB: Reusing cached keymap
[239412.755] (II) Open ACPI successful (/var/run/acpid.socket)
[239412.755] (II) AIGLX: Resuming AIGLX clients after VT switch
[239412.755] (II) NOUVEAU(0): NVEnterVT is called.
[239412.896] (II) NOUVEAU(0): EDID vendor "AUO", prod id 16455
[239412.896] (II) NOUVEAU(0): Printing DDC gathered Modelines:
[239412.896] (II) NOUVEAU(0): Modeline "1440x900"x0.0  108.00  1440 1504 1536 1970  900 903 906 912 +hsync -vsync (54.8 kHz eP)
[239412.896] (II) NOUVEAU(0): Modeline "1440x900"x0.0   72.50  1440 1504 1536 1970  900 903 906 912 +hsync -vsync (36.8 kHz e)
[239413.023] (**) Option "Device" "/dev/input/event10"
[239413.023] (--) synaptics: AlpsPS/2 ALPS DualPoint TouchPad: touchpad found

But the XKB settings were not restored.

I have the same bug and I would really love to have it fixed.  Since
about 2 years (before, it worked) I have been forced to execute xmodmap
~/.xmodmap after each resume, since the key bindings I use in that file
are lost at suspend.

xmodmap simply adds French and Romanian accentuated characters éèôîșț...
to my English keyboard.

Note that xset settings are lost too.  For ex., I execute xset b off, or
xset r rate 200 30, but they go back to default values after suspend/resume.

¡Hola Eugen!

El 2013-10-31 a las 05:02 +0100, Eugen Dedu escribió:

How are you suspending your computer?
Under which desktop environment?

I also have a custom ~/.Xmodmap, and using pm-suspend or powerdevil (kde)
won't loss those changes.

¡Hola Eugen!

El 2013-10-31 a las 05:02 +0100, Eugen Dedu escribió:

How are you suspending your computer?
Under which desktop environment?

I also have a custom ~/.Xmodmap, and using pm-suspend or powerdevil (kde)
won't loss those changes.

¡Hola

The bug appears when I close the lid, and also when executing pm-suspend
as root.

Under gnome until one week ago, and under awesome since then.

xset r rate setting was not visible on gnome, since it is itself which
changes it when resuming, I suppose.

This is strange then.  How can I see what happens?  Where to look at?

Could that be an issue of a kbd-related package?

¡Hola

The bug appears when I close the lid, and also when executing pm-suspend
as root.

Under gnome until one week ago, and under awesome since then.

xset r rate setting was not visible on gnome, since it is itself which
changes it when resuming, I suppose.

This is strange then.  How can I see what happens?  Where to look at?

Could that be an issue of a kbd-related package?

Yes, there are several related bugs about lost settings after
suspend/resume:

  #541388  xserver-xorg: Xmodmap settings lost across suspend/hibernate
  #568868  key repeat for caps lock goes away after suspend/resume
  #582566  bell settings (xset -b) not preserved after suspend/resume
  #633849  xserver-xorg: XKB settings lost after suspend (hibernate) / resume
           or USB keyboard plugged in

In my case, I suspend with Fn-F1, after which

  acpid: received input layer event "button/sleep SBTN 00000080 00000000"
  acpid: rule from 23102[0:0] matched
  acpid: notifying client 23102[0:0]
  acpid: rule from /etc/acpi/events/sleepbtn matched
  acpid: executing action "/etc/acpi/sleep_suspend.sh sleep"
  [...]

appears in the logs.

I don't use a desktop environment, just the fvwm window manager.

Note that once, my keyboard settings where automatically restored,
under my usual environment. I couldn't reproduce this. But this
makes me think that this may be related to how the kernel detects
the keyboard (the logs were different, but I don't know why, and
I don't know whether this is related).

Do you know if the upstream is informed about all these bugs?

Control: forwarded -1 https://bugs.freedesktop.org/show_bug.cgi?id=71168

I've just reported this bug upstream.

pm-utils. I wrote the attached /etc/pm/sleep.d/40xkb-save-restore script
(I use XKB instead of the obsolete xmodmap solution, but this script
might also work if you have changed the settings with xmodmap, I have
not tried). The following command is needed in ~/.xsession

  xhost +si:localuser:root

since the script is run by root, and root doesn't have the user's
xauth information (for MIT-MAGIC-COOKIE-1) in his environment.

Most often this works perfectly. This solution is not even affected by

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=661295

However the settings are not always restored:

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=756268

Thank you for the help.

Just to say that since a few days (I track unstable) this bug appears
very rarely (once in say 6 resumes).

[I should have Cc'd this to the bug. Sending a copy now.]

tags 633849 wontfix - fixed-upstream
thanks

The upstream bug was closed because "there isn't enough maintainer
time (or motivation) to ever fix this", not because it was fixed.

IMHO, this limitation should be documented (probably in the man pages
of the concerned commands).

I'm also wondering whether the suspend/resume system allows the user
to run scripts to save the current config and restore it.

On 2023-06-11 19:43:03 +0000, Debian Bug Tracking System wrote:
[...]
[...]