Bug#961132: lightdm: with the nvidia driver, lightdm suspends the system when locking or logging out
On 2020-05-21 14:01:04 +0200, Yves-Alexis Perez wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA256
> control: tag -1 unreproducible
Not sure the conditions to reproduce the bug, but to be sure, with
1. Set up a systemd-inhibit lock for handle-lid-switch.
2. Close the lid.
3. Lock the screen with light-locker or log out.
Just guessing, because the systemd log does not mention lid closing
at all for the reason of suspend.
Note that this was in my bug report against systemd:
(merged with this bug).
> On Wed, 2020-05-20 at 19:35 +0200, Vincent Lefevre wrote:
> > Both locking and logging out trigger the lightdm login prompt,
> > so that I suppose that this is the same issue, and it was not
> > light-locker that was suspending the system, but lightdm.
> this is weird indeed, but I'm not going to investigate an nvidia
> binary issue. For what it's worth, I do have a system with an nvidia
> card and the binary driver, running lightdm, light-locker and Xfce,
> and I don't have that behavior at all (running Buster).
Well, Michael Biebl says that's light-locker / lightdm that suspends
(note: I don't use XFCE, just lightdm + light-locker at that time).
There is no reason to do that under any condition, and the code that
does that needs to be removed.
Now, if there isn't any such code, I would tend to think that
Michael Biebl is wrong, because I also get a suspend issue
without lightdm / light-locker:
and it seems to be related to systemd-inhibit lock vs which
login session is associated with the display, in particular
when switching from X to a VT:
Since the use of lightdm does something similar, this could explain
Vincent Lefèvre <firstname.lastname@example.org> - Web: <https://www.vinc17.net/>
100% accessible validated (X)HTML - Blog: <https://www.vinc17.net/blog/>
Work: CR INRIA - computer arithmetic / AriC project (LIP, ENS-Lyon)