[Primary] Screen stays black after waking up from energy saving

[Kansi]

Hi all,

I have a Dell M4800 notebook with an Nvidia card + driver (NVIDIA Corporation GK106GLM [Quadro K2100M] according to lspci) running 14.04.2 LTS.

When "Screen Energy Saving" kicks in on my machine, the screen turns off/black, as expected. However, when waking up the machine (e.g., by moving the mouse), the screen will stay black. As a workaround, I can switch to a virtual console with CTRL+ATL+F1 and back with CTRL+ALT+F7 to make the screen wake up. In the long run, this is frustrating.

A few notes:

• This problem does not occur when the machine goes to Sleep (configured in "Energy saving -> Suspend session"); when waking the machine from sleep, the screen comes back normal.
• When I have a second monitor attached via VGA, that display comes back normal from energy saving. Only the notebook's own, primary display stays black.
• When switching to the virtual console for the workaround, I do not actually get a virtual console - the screen will just stay black (or, in the case of the second monitor, also turn black); this is always the case: I also do not get a visible virtual console when I switch over while my screen is displaying stuff. (I am not overly concerned about this behaviour, just mentioning it in case somebody sees a connection.)

Most posts I have found seem to be about not returning from Suspend/Sleep/Hibernate. I tried out turning up the brightness when the screen stayed black, and I tried adding an automatic virtual console switch in /etc/pm/sleep.d, but neither changed anything. Does anybody have any other ideas how to go about this?

Thanks!
Karla

[SteveRiley]

Similar bug reports with your hardware.

https://bugs.launchpad.net/ubuntu/+s...g/+bug/1268151 -- workarounds in comments #32, #42.
https://bugs.launchpad.net/ubuntu/+s...x/+bug/1400235

What version of the NVIDIA driver and what release of the kernel are you running?

[Kansi]

Hi Steve,

thanks for your quick reply and those links - I think 1400235 might be different (problem there occurs on Resume, and my workaround with switching to the invisible VT and back does not work there); but 1268151 sounds very much like my issue. I can reproduce the behaviour from the initial comment (xset dpms force off; ...). Unfortunately, the approaches in comments #32 and #42 did not change anything.

The versions you asked about: The kernel is 3.13.0-48, and the driver is nvidia-331-updates.

A few things I have tried:

• Updated the Bios to A13 (was mentioned in one the bugs you linked) - no effect.
• Adding a line with "acpiphp" to /etc/modules as per a Dell script from a driver bundle found at http://www.dell.com/support/home/us/...driverId=V9HNV (also mentioned in one of those bugs) - no effect.
• Switched to nvidia-331 (instead of nvidia-331-updates). Interestingly, this solves the issue! Sadly, however, it really only solves that one issue :-) Nothing else works anymore with that driver ... The resolution of the primary screen is fixed at something around the 1600x1200 (not the actual 3200x1800), and the secondary screen, when attached, is not even recognized. And the Nvidia configuration settings are simply void of any options.

I found another lengthy thread on https://devtalk.nvidia.com/default/t...on-qhd-k2100m/ - it's primarily about the invisible VTs, but somebody incidentally mentions having my problem. Seems this issue persists even with current Nvidia drivers.

I will dig through that thread again - there might be some more options hidden in there somewhere. Then again, I should perhaps try updating my Kubuntu to 14.10?

Cheers
Karla

[SteveRiley]

Then again, I should perhaps try updating my Kubuntu to 14.10?

Hard to know for sure. Generally, hardware support improves with newer software. But this isn't universally the case.

[Kansi]

Quick update for the sake of potential reference: After having postponed doing anything about this for a while, I have now updated my nvidia driver to version 346.82 (from nvidia-346). And just like that, the problem has disappeared.

(I should note that, although I had observed the problem regularly over the last months, I cannot rule out that the change has come about - unnoticed - by other updates in the last few weeks or so.)

- K.