[Pkg-xfce-devel] Bug#738124: Bug#738124: xfce4: Screen does not always lock during suspend/hibernate despite that option being enabled

[David da Medic]

So, unfortunately, I've had to give up the computer that was producing
this bug. But I'll try to continue the conversation as best as I can,
referencing a nearly identical install (fairly stock debian wheezy w/
xfce4.8).

> That's how it's supposed to happen, actually. Can you try to suspend
> from the xfce4-power-manager menu and see if it works?

I can't seem to find an option to do that... perhaps that's from a newer
version?

> Do you use systemd (or maybe upstart)?

I'm really not sure... It seems that debian does not install plain
"systemd," but libsystemd-login0 and libsystemd-daemon0 are present, and
I imagine either one might be relevant.

Upstart is not installed.

On Sat, 2014-02-08 at 11:32 +0100, Yves-Alexis Perez wrote:
> On Fri, Feb 07, 2014 at 04:54:17PM -0500, David Z wrote:
> > Package: xfce4
> > Version: 4.8.0.3
> > Severity: important
> > Tags: upstream
> > 
> > Dear Maintainer,
> > 
> >    * What led up to the situation?
> > 
> > In the Xfce Power Manager dialog, under "Extended," the checkbox for "Lock
> > screen when going for suspend/hibernate" is checked. (An aside - that phrasing
> > is a little clunky. Might I suggest "Lock screen during suspend/hibernate"?)
> > 
> > I find that SOME times when I suspend for a short time (<5 minutes or so), I
> > can wake the computer and return to my session without the screen being locked
> > or authenticating myself. That has some concerning security implications, of
> > course.
> 
> At first sight, it looks like it's not xfpm which is actually
> suspending, but something else.
> 
> The “short time suspend = no lock” looks simply that the screensaver
> did not activate at all.
> > 
> > I imagine that the cause is that the "lock screen" command is not completed
> > before the ACPI suspend occurs. Could this be fixed with a simple wait for the
> > lock screen portion of the sequence to return a good exit code before signaling
> > an ACPI suspend? This does not explain, though, why it has failed to lock only
> > when the suspend is for less than 5 minutes or so - I suspect that that may
> > even just be a fluke, since the length of the suspend really shouldn't matter
> > at all, as the way I understand Suspend to work, it does not allow any programs
> > to do anything while suspended. Perhaps there is a time-checking component to
> > how XFCE resumes from suspend which is interfering, wherein if the time
> > difference between suspend and wake is less than the configured "put the
> > computer to sleep" time (or its default, since I have that setting set to
> > "never"), it will not bother to lock the screen?
> > 
> > Further, on some resumes, I am able to see my desktop session for a split-
> > second before the lock screen draws over it - this suggests that the locking
> > happens upon resume, and not on suspend.
> 
> Definitely, and it activates because of the timing, not because it was
> instructed so, I think.
> 
> > I feel that locking the screen PRIOR
> > to engaging the suspend would be much better, for reliability of operation (as
> > in the main topic of this bug), security, and privacy (with regard to the
> > split-second view of the desktop prior to being locked - imagine if some
> > "unprofessional" image were being displayed for a split second when another
> > person happened to open the lid of my laptop!).
> 
> That's how it's supposed to happen, actually. Can you try to suspend
> from the xfce4-power-manager menu and see if it works?
> 
> Do you use systemd (or maybe upstart)?
> 
> Regards,