No visible display in non gnome sessions when using nvidia drivers via nvidia-prime until screen goes to sleep, then waked up

[Doug McMahon]

** this appears to be that the display is blanked/asleep at greeter & login.
Forcing or waiting for screen to blank/sleep then waking up restores display
See comments 8 & 10 for further info, demonstrations **

Please note that this bug is being filed on a recent 15.10 Ubuntu image install but from a gnome session

In a nutshell -
Once nividia drivers are installed & used from nvidia-prime there is no longer a visible display in unity-greeter nor in an ubuntu session once logged in. (a blind log in successfully loads an ubuntu session.
Everything in the ubuntu session works fine, there are no errors & all the related logs look ok, there is just no visible display.

I can open apps, run commands, copy/create logs, play music, play videos, just no display so has to be done blindly from keyboard.

On the other hand if I install gnome-shell & log in (blindly) to a gnome session from greeter then all is well & I get a visible display.

The same behaviour is seen on an Xubuntu image install, once nvidia is installed & switched to, no visible display. Again there installing & logging into a gnome session works fine with visible display

nvidia driver version doesn't matter, exact same is seen with current 15.10 versions & the one I'm currently using (355) from drivers ppa.

** this appears to be that the display is blanked/asleep at greeter & login.
Forcing or waiting for screen to blank/sleep then waking up restores display
See comment 8 for further info **

ProblemType: Bug
DistroRelease: Ubuntu 15.10
Package: xserver-xorg-core 2:1.17.2-1ubuntu7
ProcVersionSignature: Ubuntu 4.2.0-11.13-generic 4.2.1
Uname: Linux 4.2.0-11-generic x86_64
.tmp.unity.support.test.0:

ApportVersion: 2.19-0ubuntu1
Architecture: amd64
CompizPlugins: No value set for `/apps/compiz-1/general/screen0/options/active_plugins'
CompositorRunning: None
CurrentDesktop: GNOME
Date: Tue Sep 29 16:15:21 2015
DistUpgraded: Fresh install
DistroCodename: wily
DistroVariant: ubuntu
DkmsStatus:
 bbswitch, 0.7, 4.2.0-11-generic, x86_64: installed
 nvidia-355, 355.11, 4.2.0-11-generic, x86_64: installed
ExtraDebuggingInterest: Yes
GraphicsCard:
 Intel Corporation 4th Gen Core Processor Integrated Graphics Controller [8086:0416] (rev 06) (prog-if 00 [VGA controller])
   Subsystem: Lenovo Device [17aa:3801]
 NVIDIA Corporation GK107M [GeForce GT 755M] [10de:0fcd] (rev ff) (prog-if ff)
InstallationDate: Installed on 2015-09-28 (0 days ago)
InstallationMedia: Ubuntu 15.10 "Wily Werewolf" - Alpha amd64 (20150928)
MachineType: LENOVO 20217
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.2.0-11-generic.efi.signed root=UUID=3712249e-a4c8-4b29-9ffb-f9f9e7d5259a ro quiet splash vt.handoff=7
SourcePackage: xorg-server
UdevLog: Error: [Errno 2] No such file or directory: '/var/log/udev'
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 09/18/2013
dmi.bios.vendor: LENOVO
dmi.bios.version: 74CN44WW(V3.05)
dmi.board.asset.tag: No Asset Tag
dmi.board.name: VIQY0Y1
dmi.board.vendor: LENOVO
dmi.board.version: 31900058STD
dmi.chassis.asset.tag: No Asset Tag
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: Lenovo IdeaPad Y510P
dmi.modalias: dmi:bvnLENOVO:bvr74CN44WW(V3.05):bd09/18/2013:svnLENOVO:pn20217:pvrLenovoIdeaPadY510P:rvnLENOVO:rnVIQY0Y1:rvr31900058STD:cvnLENOVO:ct10:cvrLenovoIdeaPadY510P:
dmi.product.name: 20217
dmi.product.version: Lenovo IdeaPad Y510P
dmi.sys.vendor: LENOVO
version.compiz: compiz 1:0.9.12.2+15.10.20150908-0ubuntu1
version.ia32-libs: ia32-libs N/A
version.libdrm2: libdrm2 2.4.64-1
version.libgl1-mesa-dri: libgl1-mesa-dri 11.0.0-1ubuntu1
version.libgl1-mesa-dri-experimental: libgl1-mesa-dri-experimental N/A
version.libgl1-mesa-glx: libgl1-mesa-glx 11.0.0-1ubuntu1
version.xserver-xorg-core: xserver-xorg-core 2:1.17.2-1ubuntu7
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev 1:2.9.2-1ubuntu1
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:7.5.0+git20150819-0ubuntu1
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.99.917+git20150808-0ubuntu2
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.11-1ubuntu3
xserver.bootTime: Tue Sep 29 16:05:24 2015
xserver.configfile: default
xserver.errors:

xserver.logfile: /var/log/Xorg.0.log
xserver.outputs:
 product id 729
 vendor LGD
xserver.version: 2:1.17.2-1ubuntu7

[Doug McMahon (mc3man)]

journalctl from an ubuntu session

[Doug McMahon (mc3man)]

journalctl from a gnome session

[Doug McMahon (mc3man)]

gpu-manager log from an ubuntu session

[Doug McMahon (mc3man)]

glxinfo log from an ubuntu session

[Doug McMahon (mc3man)]

Ex. of stdout from playing a vid in cli player in an ubuntu session with nvidia(vdpau), no errors
(obviously I only here sound as there is no visible display

Playing: 1.mp4
 (+) Video --vid=1 (*) (h264)
 (+) Audio --aid=1 --alang=und (*) (aac)
AO: [pulse] 44100Hz stereo 2ch float
Using hardware decoding (vdpau).
VO: [opengl-hq] 1920x1080 vdpau

Exiting... (Quit)

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Doug McMahon (mc3man)]

What happens in an ubuntu session is -
After logging in, no display. If the session is left inactive until display is put to sleep, then awakining it thru keyboard or mouse brings back a visible display. (default is 5 min. of inactivity

A simple startup script via startup applications demonstrates -
script in ~/bin

#!/bin/bash
xset dpms force off
xset dpms force on

~/.config/autostart/unblank.desktop

[Desktop Entry]
Type=Application
Exec=unblank1
Hidden=false
NoDisplay=false
X-GNOME-Autostart-enabled=true
Name=unblank
X-GNOME-Autostart-Delay=3

So with above in place after the blind login > 3 secs > visible display is returned

It's been suggested that xorg.conf needs some serverflags, ie.

Section "ServerFlags"
        Option "BlankTime" "0"
        Option "OffTime" "0"
EndSection

Though not effective here as a standalone section & if there is an xorg.conf solution then it needs to be patched into ubuntu-drivers-common

[Doug McMahon (mc3man)]

Also affects Xubuntu

[Doug McMahon (mc3man)]

Demonstration 2
If the autostart script was in place then disable

Alter the sleep time of unity-greeter from 300 to 6 (sec) in /usr/share/glib-2.0/schemas/com.canonical.unity-greeter.gschema.xml
(at bottom of file, assumes 6 sec is enough time to enter password, if not increase 6 to something suitable

sudo -H gedit /usr/share/glib-2.0/schemas/com.canonical.unity-greeter.gschema.xml

recompile schemas
sudo glib-compile-schemas /usr/share/glib-2.0/schemas

Log out or restart, at black greeter screen wait at least 6 sec., then press a key. (or wait for value set above in the .xml
The greeter screen will become visible - Quickly enter password & press enter while the screen is still visible

Login session will have a visible display as expected

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Alexander (sality)]

same problem in Ubuntu 15.10 (Unity)

[Tim Fisher (k2trf)]

I'm having a similar issue, but not quite the same. The display works fine upon booting up (Kubuntu 15.10 beta 2), but if I let it go to sleep (or turn off the TV I use as a monitor) and return, then I get no display, and have to reset the computer to get a display back.

Just yesterday something changed (been installing upgrades like crazy, being a brand new PC on beta 2), and the display now works when I wake the computer, but I cannot get back into my established session (although I can start a new one).

I'm using SDDM to login, KWin as my WM (with Plasma 5.4 as my DE); I'm honestly not sure what logs I could provide to aid this bug, but if somebody else knows I'll gladly spew any output from my system that can help!

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Ubuntu QA Website (ubuntuqa)]

This bug has been reported on the Ubuntu Package testing tracker.

A list of all reports related to this bug can be found here:
http://packages.qa.ubuntu.com/qatracker/reports/bugs/1501041

[Doug McMahon (mc3man)]

Not really fixed so opening again

[sebastien.caille (sebastien-caille)]

Hello,

I had the same issue and managed to get back to a sane situation

To summarize (more details below), the pattern I saw on my laptop is:
- it works fine as long as the default screen configuration is used (that is, default screen resolution, default screen position)
 (actually, no log "switch to mode" in Xorg.0.log)
- to work around the issue: I switched to vt1, waited for a while (maybe 30 seconds, I'm not 100% sure), switched back to vt7. Then the login screen was displayed and I was able to log in.
- It works better after applying the default configuration using the System Settings (the login screen appears after a reboot)

Installed packages:
xserver-xorg-core 2:1.17.2-1ubuntu9
xserver-xorg-video-intel 2:2.99.917+git20150808-0ubuntu4
nvidia-355 355.11-0ubuntu0~gpu15.10.1

Hardware is Intel HD Graphics 4600 + NVIDIA GT740M
[ 2.612] (--) PCI:*(0:0:2:0) 8086:0416:1043:122d rev 6, Mem @ 0xf7400000/4194304, 0xd0000000/268435456, I/O @ 0x0000f000/64
[ 2.612] (--) PCI: (0:1:0:0) 10de:1292:1043:122d rev 161, Mem @ 0xf6000000/16777216, 0xe0000000/268435456, 0xf0000000/33554432, I/O @ 0x0000e000/128, BIOS @ 0x????????/524288

Outputs are eDP1 (laptop's screen, 1600x900) and HDMI1 (1920x1080) (now they are named eDP-1-0 and HDMI-1-0... could be an update I applied)

Xorg correctly loads the drivers, prime-offload correctly associates the cards.
Then blank screen

After some vt switches, the login screen appeared on vt7.
The display was corrupted (only for a while), the screen resolution was unusual and I had the same display on both screens.

The last vt switch likely matches the log at 67.771 (it's definitely an unusual resolution, with mirror mode)

[ 3.460] (WW) config/udev: device ETPS/2 Elantech Touchpad already added. Ignoring.
[ 3.808] (II) intel(G0): switch to mode 1600x900@60.0 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
[ 3.855] (II) intel(G0): switch to mode 1600x900@60.0 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
[ 4.646] (II) intel(G0): switch to mode 1920x1080@60.0 on HDMI1 using pipe 1, position (1600, 0), rotation normal, reflection none
[ 34.168] (II) NVIDIA(0): Setting mode "NULL"
[ 34.169] (II) intel(G0): switch to mode 1600x900@60.0 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
[ 34.169] (II) intel(G0): switch to mode 1920x1080@60.0 on HDMI1 using pipe 1, position (1600, 0), rotation normal, reflection none
[ 34.256] (--) synaptics: ETPS/2 Elantech Touchpad: touchpad found
[ 36.296] (II) NVIDIA(0): Setting mode "NULL"
[ 36.297] (II) intel(G0): switch to mode 1600x900@60.0 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
[ 36.297] (II) intel(G0): switch to mode 1920x1080@60.0 on HDMI1 using pipe 1, position (1600, 0), rotation normal, reflection none
[ 36.356] (--) synaptics: ETPS/2 Elantech Touchpad: touchpad found
[ 64.086] (II) NVIDIA(0): Setting mode "NULL"
[ 64.086] (II) intel(G0): switch to mode 1600x900@60.0 on eDP1 using pipe 0, position (0, 0), rotation normal, reflection none
[ 64.086] (II) intel(G0): switch to mode 1920x10...

[sebastien.caille (sebastien-caille)]

The Xorg.0.log after the switch that made the login appear on the screens

[sebastien.caille (sebastien-caille)]

The Xorg.0.log with a working setup (reboot + login, no workaround needed)

[Alberto Salvia Novella (es20490446e)]

Please:

1. Report to https://bugs.freedesktop.org/
2. Paste the new report URL here.
3. Set this bug status back to confirmed for "xorg-server (Ubuntu)".

Thank you.

[pdf (pdffs)]

I believe this may be a kernel regression. I can repro on 15.04 with sddm by installing any 4.x series kernel, haven't done any bisecting, and can't find anything interesting in dmesg (when I can switch to a vc sucessfully after X starts but fails to output anything), but with a single change from 3.19 to any of the mainline PPA 4.x kernels, the problem is immediately evident, and occasinally CPU locks up (SKL), but only when both GPUs are enabled (I have a BIOS hardware mux that I can use to force Nvidia only, which is what I'm doing currently to have a usable system and avoid recompiling iwl 8260 drivers for 3.19 every kernel update).

[Brian Murray (brian-murray)]

ubuntu-drivers-common only detects and helps to install driver packages, it doesn't actually affect them.

[Dmitry (grayscale)]

Could anyone ask me in which package the main reason for this issue?

Regards,
/Dmitriy

[Launchpad Janitor (janitor)]

[Expired for xorg-server (Ubuntu) because there has been no activity for 60 days.]