Por que o light-locker mantém a tela preta após o desbloqueio?

7

Uma máquina totalmente atualizada do Xubuntu 16.04 com um usuário recém-criado não me deixa desbloquear depois de usar o light-locker para bloquear minha tela (independentemente de o bloqueio ter sido ativado manualmente ou por tempo limite). A tela permanece preta (mas ativada) na entrada de senha correta

Matar o light-locker de tty1 irá revelar a área de trabalho (e consequentemente evitará mais bloqueios).

Passos tentados:

  • ativado / desativado bloqueio nas configurações do defletor de luz & amp; nas configurações de energia (sem alteração)
  • usado xscreensaver em vez de light-locker (funciona)
  • testado se o problema está relacionado a suspender ou fechar a tampa (não é)
    • no entanto, após a suspensão, a tela preta revela a mensagem This session is locked: You will be redirected to the unlock dialog automatically in a few seconds - sem que nenhum redirecionamento ocorra
  • purgado + reinstalado lightdm, light-locker e xubuntu-default-settings
  • usando loginctl unlock-session não tem efeito
  • systemctl status lightdm apenas revela mensagens (inofensivas, esperadas) sobre o módulo pam do kwallet ausente
  • examinou o log do dpkg para atualizações entre "funcionou" e "não funciona mais" (nenhum suspeito) e inicializou o kernel anterior (o problema persiste)
  • Comportamento comparado entre o light-locker iniciado pela sessão e iniciado manualmente um (sem diferença)
  • adicionado usuário ao grupo nopasswdlogin group (tela preta, em seguida, ocorre logo após o bloqueio, não apenas após a digitação da senha)

Quais etapas adicionais podem ser úteis para descobrir como corrigir isso?

Log de depuração:

$ export LC_ALL=C; light-locker --debug & ( sleep 5; echo "## Locking now"; light-locker-command --lock; sleep 10; echo "### 10 seconds have passed - entering password now"; sleep 10; echo "### 10 more seconds have passwd - Killing"; killall light-locker; sleep 5; echo "### Killed now" )  
[1] 23463
[gs_debug_init] gs-debug.c:106 (00:08:32):   Debugging enabled
[main] light-locker.c:142 (00:08:32):    initializing light-locker 1.7.0
[main] light-locker.c:144 (00:08:32):    lock after screensaver 10
[main] light-locker.c:145 (00:08:32):    late locking 0
[main] light-locker.c:146 (00:08:32):    lock on suspend 1
[main] light-locker.c:147 (00:08:32):    lock on lid 0
[main] light-locker.c:148 (00:08:32):    idle hint 1
[query_session_id] gs-listener-dbus.c:2088 (00:08:32):   org.freedesktop.login1.NoSessionForPID raised:
 PID 23463 does not belong to any known session


[init_session_id] gs-listener-dbus.c:2180 (00:08:32):    Got session-id: (null)
[query_sd_session_id] gs-listener-dbus.c:2164 (00:08:32):    Couldn't determine our own sd session id: No such process
[init_session_id] gs-listener-dbus.c:2185 (00:08:32):    Got sd-session-id: (null)
[init_seat_path] gs-listener-dbus.c:2262 (00:08:32):     Got seat: /org/freedesktop/DisplayManager/Seat0
[gs_listener_delay_suspend] gs-listener-dbus.c:449 (00:08:32):   Delay suspend
[gs_listener_x11_acquire] gs-listener-x11.c:172 (00:08:32):  ScreenSaver Registered
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:32):    obj_path=/org/freedesktop/DBus interface=org.freedesktop.DBus method=NameAcquired destination=:1.316
## Locking now
[gs_grab_grab_root] gs-grab-x11.c:647 (00:08:37):    Grabbing the root window
[gs_grab_get_keyboard] gs-grab-x11.c:153 (00:08:37):     Grabbing keyboard widget=E3
[gs_grab_get_mouse] gs-grab-x11.c:213 (00:08:37):    Grabbing mouse widget=E3
[gs_manager_create_windows_for_screen] gs-manager.c:548 (00:08:37):  Creating 1 windows for screen 0
[gs_manager_create_window_for_monitor] gs-manager.c:324 (00:08:37):  Creating window for monitor 0 [0,0] (1600x900)
[update_geometry] gs-window-x11.c:197 (00:08:37):    got geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:210 (00:08:37):    using geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:197 (00:08:37):    got geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:210 (00:08:37):    using geometry for monitor 0: x=0 y=0 w=1600 h=900
[gs_window_move_resize_window] gs-window-x11.c:243 (00:08:37):   Move and/or resize window on monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:197 (00:08:37):    got geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:210 (00:08:37):    using geometry for monitor 0: x=0 y=0 w=1600 h=900
[gs_window_move_resize_window] gs-window-x11.c:243 (00:08:37):   Move and/or resize window on monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:197 (00:08:37):    got geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:210 (00:08:37):    using geometry for monitor 0: x=0 y=0 w=1600 h=900
[gs_window_move_resize_window] gs-window-x11.c:243 (00:08:37):   Move and/or resize window on monitor 0: x=0 y=0 w=1600 h=900
[gs_manager_timed_switch] gs-manager.c:445 (00:08:37):   Start switch to greeter timer
[gs_window_xevent] gs-window-x11.c:369 (00:08:37):   not raising our windows
[window_map_event_cb] gs-manager.c:233 (00:08:37):   Handling window map_event event
[gs_listener_resume_suspend] gs-listener-dbus.c:513 (00:08:37):  Resume suspend: fd=14
[manager_maybe_grab_window] gs-manager.c:204 (00:08:37):     Moving grab to 0x5595f8d5a290
[gs_grab_move_keyboard] gs-grab-x11.c:450 (00:08:37):    Moving keyboard grab from E3 to 3A0000C
[gs_grab_move_keyboard] gs-grab-x11.c:457 (00:08:37):    *** doing X server grab
[gs_grab_release_keyboard] gs-grab-x11.c:279 (00:08:37):     Ungrabbing keyboard
[gs_grab_get_keyboard] gs-grab-x11.c:153 (00:08:37):     Grabbing keyboard widget=3A0000C
[gs_grab_move_keyboard] gs-grab-x11.c:479 (00:08:37):    *** releasing X server grab
[gs_grab_move_mouse] gs-grab-x11.c:395 (00:08:37):   Moving pointer grab from E3 to 3A0000C
[gs_grab_move_mouse] gs-grab-x11.c:402 (00:08:37):   *** doing X server grab
[gs_grab_release_mouse] gs-grab-x11.c:320 (00:08:37):    Ungrabbing pointer
[gs_grab_get_mouse] gs-grab-x11.c:213 (00:08:37):    Grabbing mouse widget=3A0000C
[gs_grab_move_mouse] gs-grab-x11.c:425 (00:08:37):   *** releasing X server grab
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:37):    obj_path=/org/freedesktop/login1 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[update_geometry] gs-window-x11.c:197 (00:08:37):    got geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:210 (00:08:37):    using geometry for monitor 0: x=0 y=0 w=1600 h=900
[gs_window_move_resize_window] gs-window-x11.c:243 (00:08:37):   Move and/or resize window on monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:197 (00:08:37):    got geometry for monitor 0: x=0 y=0 w=1600 h=900
[update_geometry] gs-window-x11.c:210 (00:08:37):    using geometry for monitor 0: x=0 y=0 w=1600 h=900
[gs_window_move_resize_window] gs-window-x11.c:243 (00:08:37):   Move and/or resize window on monitor 0: x=0 y=0 w=1600 h=900
[gs_listener_send_lock_session] gs-listener-dbus.c:180 (00:08:38):   Send lock session
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:38):    obj_path=(null) interface=(null) method=(null) destination=:1.316
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/session/c5 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/session/c12 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/seat/seat0 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
process 23463: arguments to dbus_message_new_method_call() were incorrect, assertion "path != NULL" failed in file ../../dbus/dbus-message.c line 1340.
This is normally a bug in some application using the D-Bus library.
[query_session_active] gs-listener-dbus.c:1124 (00:08:39):   Couldn't allocate the dbus message
[listener_dbus_handle_system_message] gs-listener-dbus.c:1361 (00:08:39):    systemd notified ActiveSession 0
[listener_session_switched_cb] gs-monitor.c:298 (00:08:39):  Session switched: 0
[gs_manager_stop_switch] gs-manager.c:456 (00:08:39):    Stop switch to greeter timer
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/seat/seat0 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/session/c12 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/user/_108 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/seat/seat0 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
process 23463: arguments to dbus_message_new_method_call() were incorrect, assertion "path != NULL" failed in file ../../dbus/dbus-message.c line 1340.
This is normally a bug in some application using the D-Bus library.
[query_session_active] gs-listener-dbus.c:1124 (00:08:39):   Couldn't allocate the dbus message
[listener_dbus_handle_system_message] gs-listener-dbus.c:1361 (00:08:39):    systemd notified ActiveSession 0
[listener_session_switched_cb] gs-monitor.c:298 (00:08:39):  Session switched: 0
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:39):    obj_path=/org/freedesktop/login1/session/c5 interface=org.freedesktop.login1.Session method=Lock destination=(null)
### 10 seconds have passed - entering password now
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:51):    obj_path=/org/freedesktop/login1/session/c5 interface=org.freedesktop.login1.Session method=Unlock destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:51):    obj_path=/org/freedesktop/login1/session/c13 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:51):    obj_path=/org/freedesktop/login1/session/c5 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
[listener_dbus_handle_system_message] gs-listener-dbus.c:1330 (00:08:51):    obj_path=/org/freedesktop/login1/seat/seat0 interface=org.freedesktop.DBus.Properties method=PropertiesChanged destination=(null)
process 23463: arguments to dbus_message_new_method_call() were incorrect, assertion "path != NULL" failed in file ../../dbus/dbus-message.c line 1340.
This is normally a bug in some application using the D-Bus library.
[query_session_active] gs-listener-dbus.c:1124 (00:08:51):   Couldn't allocate the dbus message
[listener_dbus_handle_system_message] gs-listener-dbus.c:1361 (00:08:51):    systemd notified ActiveSession 0
[listener_session_switched_cb] gs-monitor.c:298 (00:08:51):  Session switched: 0
[xroot_filter] gs-listener-x11.c:124 (00:08:52):     ScreenSaver stopped
### 10 more seconds have passwd - Killing
### Killed now

Diffing isso para um sistema bem conhecido produz:

$ comm -3 <(grep -iv grab good | sed 's/[0-9:]\+/[NN]/g;s/[ \t]\+/ /g' | sort) <(grep -iv grab bad | sed 's/[0-9:]\+/[NN]/g;s/[ \t]\+/ /g' | sort) | head -n2
[init_session_id] gs-listener-dbus.c[NN] ([NN])[NN] Got session-id[NN] (null)
[init_session_id] gs-listener-dbus.c[NN] ([NN])[NN] Got session-id[NN] /org/freedesktop/login[NN]/session/c[NN]

E, de fato, algo parece ser muito estranho sobre a configuração da sessão:

$ loginctl session-status
Could not get properties: No such process
    
por anx 02.10.2017 / 01:56

1 resposta

4

TL; DR: o light-locker mantém a tela preta se o sistema não puder ler / proc

  • light-locker depende de dbus fornecendo IPC .
  • dbus depende de systemd-logind fornecendo session informação.
  • systemd-logind depende de /proc fornecendo process informação.

Isso não funcionará muito bem em conjunto:

$ file /sbin/init
/sbin/init: symbolic link to /lib/systemd/systemd
$ mount | grep proc
proc on /proc type proc (rw,nosuid,nodev,noexec,relatime,hidepid=2,gid=1337)

E não se pode esperar que nenhum deles:

  

O hidepid não é realmente compatível com o systemd. desculpa. [..] Enfim, fechando como isso não foi causado pelo systemd. - Lennart Poettering

Solução parcial 1

Parcial solução, porque vem com divulgação de informações.

Permitir que programas sem privilégios (como systemd, depois de eliminar privilégios) acessem informações sobre processos de outros usuários em /proc .

$ sudo mount -o remount,hidepid=0 /proc
 # and fix /etc/fstab accordingly

Solução parcial 2

Parcial solução, porque systemd pode quebrar em outros lugares, não apenas logind.

Adicione systemd-logind ao grupo apropriado, mais convenientemente obtido adicionando um serviço Drop-In.

$ addgroup showpid
$ adduser myuser showpid
$ sudo mount -o remount,hidepid=2,gid=showpid /proc
 # and fix /etc/fstab accordingly
$ echo -e '[Service]\nSupplementaryGroups=showpid' | sudo tee /etc/systemd/system/systemd-logind.service.d/10-showpid.conf
$ sudo systemctl daemon-reload
$ sudo systemctl restart systemd-logind

Após a solução qualquer , a mensagem Got session-id: (null) deve parecer-se mais com Got session-id: /org/freedesktop/login1/c7 e o light-locker poderá se comunicar corretamente por meio de dbus .

O que devo ter feito para descobrir isso mais rapidamente?

  • Se algo mudou, e você não consegue descobrir rapidamente o que mudou, basta pegar o backup já e diff -ruiN todo o sistema.
  • Documente a primeira vez que um problema ocorreu com mais precisão, portanto, a classificação de arquivos de registros / relatórios de IDS por horário revelará rapidamente a causa relevante.
  • Arquive mais relatórios de bugs. Os aplicativos que falham silenciosamente sob condições que posteriormente causam dor de cabeça e / ou bloqueio do sistema são inaceitáveis.
por anx 17.10.2017 / 02:08