[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#972417: xfce4-power-manager: System left idle => 'display power management' auto-locks session => no X session or lightdm greeter



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

control: forcemerge 870641 -1

On Sun, 2020-10-18 at 10:23 +0200, franck wrote:
> Package: xfce4-power-manager
> Version: 1.6.1-1
> Severity: grave
> 
> Dear Maintainer,
> 
> [ Bug description ]
> Bug happens when the screen auto-locks due to user inactivity. Then the
> screen switches off.
> To easily test that, set following to 1 min or so
> Xfce menu Applications -> Settings -> Power Manager -> Display -> 'Switch
> off after'
> 
> After such auto-lock, it appears not possible to get back the X session
> anymore.
> Because
> - by moving mouse or typing on keyboard the screen does not wake up. It
> stays off, black.
> - with Ctrl+Alt-F7 one gets
>     This session is locked
>     You’ll be redirected to the unlock
>     dialog automatically in a few seconds
>   but after waiting a few seconds the screen switches off again.
> => no way to get to the lightdm greeter.
> In both cases: not possible to get back the running X session anymore.

The lightdm greeter is on tty8, so try with Ctrl-Alt-F8.
> 
> => I filled this bug against xfce4-power-manager
> But I have no clue whether xfce4-power-manager
> - processes the locking completely by itself ?
> - just prepares locking, like deactivating keyboard/mouse? and then trigger
> like light-locker-command --lock...
> - only asks another package to do the full locking procedure.
> 
> 
> 
> Bug occurred under (at least) both kernels:
> linux-image-4.19.0-11-amd64-unsigned     4.19.146-1
> linux-image-5.8.0-0.bpo.2-amd64          5.8.10-1~bpo10+1
> 
> 

> 
> Also this bug gives a bad image to the user:
> - 'system crashed'
> - wondering what is the cause:
>     - is my graphic card not properly waking up ?
>     - or a bug in kernel ?
>     - or Xorg ?
>     - or light-locker ?
>     ...
> Thus they are probably numerous of open bugs that could be closed by fixing
> this bug.

Actually it's (or should be) already fixed in stable.
> 
> 
> [ Bugs reported against other packages, & that seem linked to this bug ]
> Bugs in source package light-locker
>   #906902 System left idle makes system freeze
>   #870641 light-locker, lightdm: screen stays off after resume
>       and all the merged one:
>       805711, 846278, 868087, 908329, 922095, 929461, 929834, 931555
>   #835461 light-locker breaks suspend/resume with nvidia legacy 340 drivers
> Bugs in source package lightdm
>   #867620 lightdm unlock screen randomly doesn't appear

With all those bugs already opened and the mess we're already in, why opening
*yet another one*? Honestly it adds more confusion than anything at that
point.

As indicated in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=870641#218
the bug *should* be fixed in 10.5 with kernel 4.19.0-10. You indicated that
you tested with linux-image-4.19.0-11-amd64-unsigned     4.19.146-1

Can you retry with that kernel and check you have the “atomic” message in the
kernel log:

broken atomic modeset userspace detected, disabling atomic

Regards,
- -- 
Yves-Alexis
-----BEGIN PGP SIGNATURE-----

iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAl+gRRsACgkQ3rYcyPpX
RFvT2Qf+K/oere1FxlQi2jg+f/iL6ejQLL/jowc5R3pg34Jp3jJMPvWe6HYz7c2L
+88W5GqwhI88Qk5oiGMdEf+3s8eKwoJi2kBp92g6HFyk6Xjbe7vzs7FOhUeeHadE
9ezgXfKzrDlhM4Knm3JtTEgKbgnEeOSop78hfzkaTrH94sbR75taK8yMo7Yuec3i
gIezlzkkOTuT37oDzrG2hyD9A/mEnQTWw3pw97kqHZo3meEtTjwZ9/vutd2MuzQM
zJG5qpCLH7G8fywBWzW6SZSDhChAo44Phz2fHDmlPLKtHUTAn03nHPwkdgUgg0Kw
3iaNBwgSZaPfZrPRh4yIhhN9rUodhQ==
=SWS6
-----END PGP SIGNATURE-----


Reply to: