Bug#886731: linux-image-4.9.0-4-amd64: DKMS sends system in suspend since linux-image-4.9.0-3-amd64

[Marc-Robin Wendt]

On Fri, 2018-01-12 at 01:12 +0000, Ben Hutchings wrote:
> On Thu, 2018-01-11 at 12:25 +0100, Marc-Robin Wendt wrote:
> > Hello Ben and Michael,
> > 
> > thanks for taking time. Of course its DPMS. Sorry for the typo.
> > 
> > Reproducing the bug turns out, it first appears in linux-image-
> > 4.9.0-4-
> > amd64 and now linux-image-4.9.0-5-amd64.
> > 
> > When I boot the system from grub with old kernel linux-image-4.9.0-
> > 3-
> > amd64 or linux-image-4.9.0-2-amd64, behavior is as expected ok,
> > means
> > switching off the monitor doesnt trigger suspend of the system.
> 
> So are *you* switching off the monitor?  You mentioned DPMS, which is
> the way for the display adapter to control the monitor - it doesn't
> involve any signalling in the other direction.
> 
> And is the monitor attached to a VGA or digital port?  If it's a
> digital port, turning off power to the monitor will look exactly the
> same as unplugging it...
> 

I examined the problem more closely. Lid is closed, because the laptop
is on the dock. Connection via display port. When I connect via HDMI
there are no problems with turning on or off the monitor.

It doesn't matter how the monitor is switched off, manual or
screensaver or even by command from command line.
The suspend is actually triggered when you switch the monitor on(!)
again. I didn't noticed it so far, because it happens instantly, when
you touch anything keyboard or mouse to wake up the monitor from
screensaver.

So, monitor is switched off, system runs, I even could ping it. Wake up
the monitor will send the system into suspend. Even more strange
behavior.

Graphic card, see my first mail. Its a Lenovo T430s, so probably
NVIDIA
> > NVS 5200M inside. Monitor is LG 34UM88C-P.
> 
> ...and if you have a laptop that is already closed, unplugging the
> last
> external display is *supposed* to trigger suspend by default.  (The
> same as when you have the laptop open with no external display, and
> then close the laptop.)
> 
So you imply the old kernel had a bug? When I have my laptop on the
docking station, I want it to behave like a desktop. I least this was
the behavior the last 20 years. So of course the lid is closed and
turning off and on the last monitor shouldn't suspend the system. 


> But if this happens with the laptop still open, then this might be a
> kernel problem with lid switch detection.
> 
> Ben.
> 
> > Attached is the journalctl output once with linux-image-4.9.0-3-
> > amd64
> > booted and once with linux-image-4.9.0-4-amd64.
> 
>