DPMS not working on GTX980 with DisplayPort connection
Hi, I have two displays attached to the GTX980 card: Samsung SA850 is connected via DisplayPort and a Sony TV through HDMI. The problem is with Samsung, it doesn't turn off in power saving mode. After the idle timeout the screen goes black but the backlight remains on and never turns off. I also tried the 'sleep 1; xset dpms force off' command from the console, but result is the same. I have a dual-boot Windows 7 on the same machine, and on Windows the monitor turns off as expected, so the hardware seems to be fine. I also tried connecting the monitor via DVI-I and it seems to work as well. So the problem is with DisplayPort connection. Is there a way to fix it? Thanks. Kubuntu 14.10 x86_64, Xorg 1.16, nvidia driver 346.16
Hi,

I have two displays attached to the GTX980 card: Samsung SA850 is connected via DisplayPort and a Sony TV through HDMI. The problem is with Samsung, it doesn't turn off in power saving mode. After the idle timeout the screen goes black but the backlight remains on and never turns off. I also tried the 'sleep 1; xset dpms force off' command from the console, but result is the same.

I have a dual-boot Windows 7 on the same machine, and on Windows the monitor turns off as expected, so the hardware seems to be fine. I also tried connecting the monitor via DVI-I and it seems to work as well. So the problem is with DisplayPort connection. Is there a way to fix it?

Thanks.

Kubuntu 14.10 x86_64, Xorg 1.16, nvidia driver 346.16

#1
Posted 11/24/2014 11:15 PM   
Is any other drivers version help you to resolve this issue? Please share EDID of your display. Not sure internally we have exact same Samsung SMS27A850 display to test.
Is any other drivers version help you to resolve this issue? Please share EDID of your display. Not sure internally we have exact same Samsung SMS27A850 display to test.

Thanks,
Sandip.

#2
Posted 11/25/2014 08:58 AM   
> Is any other drivers version help you to resolve this issue? I shall try 343.22 later. Is there any version I should try in particular? > Please share EDID of your display. I tried to save EDID in nvidia-settings but it crashed (crashes every time, whether you save to text or binary file). It did save something but I'm not sure the file is complete. I'll attach it anyway.
> Is any other drivers version help you to resolve this issue?

I shall try 343.22 later. Is there any version I should try in particular?

> Please share EDID of your display.

I tried to save EDID in nvidia-settings but it crashed (crashes every time, whether you save to text or binary file). It did save something but I'm not sure the file is complete. I'll attach it anyway.
Attachments

samsung_edid.bin.gz

#3
Posted 11/25/2014 09:45 AM   
Also test with 340.58 . Off the screensaver too and check if this helps https://wiki.archlinux.org/index.php/Display_Power_Management_Signaling http://ptspts.blogspot.be/2009/10/screen-blanking-dpms-screen-saver.html

Thanks,
Sandip.

#4
Posted 11/25/2014 11:21 AM   
The 343.22 driver behaves the same, the problem persists. 340.58 doesn't seem to support Maxwell, so I didn't install it. The screen saver and screen lock are disabled in my case, and 'xset dpms force off' doesn't turn off the display, as I mentioned. I tried adding 'Option "DPMS" "True"' in my xorg.conf.d but that didn't help as well (and 'xset q' says "DPMS is Enabled" regardless of this option). Disabling screen blanking ('xset s off') doesn't help either.
The 343.22 driver behaves the same, the problem persists. 340.58 doesn't seem to support Maxwell, so I didn't install it.

The screen saver and screen lock are disabled in my case, and 'xset dpms force off' doesn't turn off the display, as I mentioned. I tried adding 'Option "DPMS" "True"' in my xorg.conf.d but that didn't help as well (and 'xset q' says "DPMS is Enabled" regardless of this option). Disabling screen blanking ('xset s off') doesn't help either.

#5
Posted 11/25/2014 06:21 PM   
Generally when this sort of thing happens, it's because the monitor doesn't respond properly to the DP power management commands. If you use the xrandr command-line tool's --off command, does the backlight turn off?
Generally when this sort of thing happens, it's because the monitor doesn't respond properly to the DP power management commands. If you use the xrandr command-line tool's --off command, does the backlight turn off?

Aaron Plattner
NVIDIA Linux Graphics

#6
Posted 11/26/2014 03:08 AM   
Yes, 'xrandr --output DP-4 --off' does disable the monitor properly, although I cannot enable it afterwards with activity or from command line. It seems, the command removes the display from X server layout because afterwards I can see that in the multi-monitor configuration in nvidia-settings the displays lost their positions relative to each other. I don't think the problem is in the display because, as I mentioned, it works in Windows and with DVI.
Yes, 'xrandr --output DP-4 --off' does disable the monitor properly, although I cannot enable it afterwards with activity or from command line. It seems, the command removes the display from X server layout because afterwards I can see that in the multi-monitor configuration in nvidia-settings the displays lost their positions relative to each other.

I don't think the problem is in the display because, as I mentioned, it works in Windows and with DVI.

#7
Posted 11/26/2014 09:52 AM   
It probably is in the display, actually, since most other displays work. In Windows, the operating system does the equivalent of an xrandr --off right after it triggers DPMS, which explains how the monitor manufacturer may have missed that it doesn't respond to DP DPMS commands correctly.
It probably is in the display, actually, since most other displays work. In Windows, the operating system does the equivalent of an xrandr --off right after it triggers DPMS, which explains how the monitor manufacturer may have missed that it doesn't respond to DP DPMS commands correctly.

Aaron Plattner
NVIDIA Linux Graphics

#8
Posted 11/26/2014 03:12 PM   
Is there a way to confirm that the problem is indeed in the display? Any additional logs? I'm not an expert in DPMS and DisplayPort protocol, so I don't quite understand the difference between 'xrandr --output DP-4 --off' and 'xset dpms force off' on the hardware side. But if there are displays which, as you say, don't fully support DPMS commands, can you provide an option for the driver to behave similarly to how it works on Windows? Or just always behave that way?
Is there a way to confirm that the problem is indeed in the display? Any additional logs?

I'm not an expert in DPMS and DisplayPort protocol, so I don't quite understand the difference between 'xrandr --output DP-4 --off' and 'xset dpms force off' on the hardware side. But if there are displays which, as you say, don't fully support DPMS commands, can you provide an option for the driver to behave similarly to how it works on Windows? Or just always behave that way?

#9
Posted 11/26/2014 04:13 PM   
I have the same issue with GTX 970 using driver 346.16 on Ubuntu 14.10. I have 4 monitors setup 3 on DP and one on HDMI. The issue is that the monitors will not go to sleep unless I force screen lock. "xset dpms force off" shuts only one monitor. This monitor cannot be awaken with mouse movement. I would like to add that My monitors are FHD, QHD, 4k and QHD. When the computer wake-up it is in confused overlapping configuration.
I have the same issue with GTX 970 using driver 346.16 on Ubuntu 14.10.
I have 4 monitors setup 3 on DP and one on HDMI. The issue is that the monitors will not go to sleep unless I force screen lock. "xset dpms force off" shuts only one monitor. This monitor cannot be awaken with mouse movement. I would like to add that My monitors are FHD, QHD, 4k and QHD. When the computer wake-up it is in confused overlapping configuration.

#10
Posted 12/06/2014 09:17 PM   
It might help if you could attach the output of nvidia-bug-report.sh.
It might help if you could attach the output of nvidia-bug-report.sh.

#11
Posted 12/07/2014 11:03 AM   
nvidia-bug-report.sh: [url]https://drive.google.com/file/d/0ByeY7uHIPmjMZGFiWDEwd3J6ajQ/view?usp=sharing[/url] I you notice there are strage mode lines in Xorg.0.log file [code][ 67731.976] (II) NVIDIA(0): Setting mode "DP-0: nvidia-auto-select @2560x1440 +1920+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}, HDMI-0: nvidia-auto-select @1920x1080 +0+0 {ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DP-4: nvidia-auto-select @2560x1440 +8320+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}" [ 67732.340] (II) NVIDIA(0): Setting mode "DP-0: nvidia-auto-select @2560x1440 +1920+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}, HDMI-0: nvidia-auto-select @1920x1080 +0+0 {ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DP-4: nvidia-auto-select @2560x1440 +8320+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}" [ 67732.647] (II) NVIDIA(0): Setting mode "DP-0: nvidia-auto-select @2560x1440 +1920+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}, DP-2: nvidia-auto-select @3840x2160 +4480+0 {ViewPortIn=3840x2160, ViewPortOut=3840x2160+0+0}, HDMI-0: nvidia-auto-select @1920x1080 +0+0 {ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DP-4: nvidia-auto-select @2560x1440 +8320+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}" [/code] These are not the ones that are specified in xorg.conf [code] Option "metamodes" "HDMI-0: nvidia-auto-select +0+0, DP-0: nvidia-auto-select +1920+0, DP-2: nvidia-auto-select +4480+0, DP-4: nvidia-auto-select +8320+0" [/code]
nvidia-bug-report.sh: https://drive.google.com/file/d/0ByeY7uHIPmjMZGFiWDEwd3J6ajQ/view?usp=sharing

I you notice there are strage mode lines in Xorg.0.log file
[ 67731.976] (II) NVIDIA(0): Setting mode "DP-0: nvidia-auto-select @2560x1440 +1920+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}, HDMI-0: nvidia-auto-select @1920x1080 +0+0 {ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DP-4: nvidia-auto-select @2560x1440 +8320+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}"
[ 67732.340] (II) NVIDIA(0): Setting mode "DP-0: nvidia-auto-select @2560x1440 +1920+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}, HDMI-0: nvidia-auto-select @1920x1080 +0+0 {ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DP-4: nvidia-auto-select @2560x1440 +8320+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}"
[ 67732.647] (II) NVIDIA(0): Setting mode "DP-0: nvidia-auto-select @2560x1440 +1920+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}, DP-2: nvidia-auto-select @3840x2160 +4480+0 {ViewPortIn=3840x2160, ViewPortOut=3840x2160+0+0}, HDMI-0: nvidia-auto-select @1920x1080 +0+0 {ViewPortIn=1920x1080, ViewPortOut=1920x1080+0+0}, DP-4: nvidia-auto-select @2560x1440 +8320+0 {ViewPortIn=2560x1440, ViewPortOut=2560x1440+0+0}"


These are not the ones that are specified in xorg.conf
Option         "metamodes" "HDMI-0: nvidia-auto-select +0+0, DP-0: nvidia-auto-select +1920+0, DP-2: nvidia-auto-select +4480+0, DP-4: nvidia-auto-select +8320+0"

#12
Posted 12/07/2014 04:47 PM   
[s]I have the same problem[/s] Just notice that it's because I switched off the dpms feature of my monitor
I have the same problem

Just notice that it's because I switched off the dpms feature of my monitor

#13
Posted 03/18/2015 11:00 AM   
[quote="aplattner"]It probably is in the display, actually, since most other displays work. In Windows, the operating system does the equivalent of an xrandr --off right after it triggers DPMS, which explains how the monitor manufacturer may have missed that it doesn't respond to DP DPMS commands correctly.[/quote] That seems possible. But given that xrandr can turn off the display, there must be a good way to hook up xrandr to do the work in linux. On the other hand, I wonder if there is a way to convince the display to cooperate. I have two systems, both running Fedora 19, both show that DPMS is enabled (via 'xset q'): [olist] [.]one has a G98 Quadro NVS 295 with one Acer K272HUL and one ASUS (2560x1440), driver version 331.67[/.] [.]one has a gtx660 with two Acer K272HUL (2560x1440), driver version 346.47[/.] [/olist] The first turns off the displays just fine. The second cannot turn off the display. 'xset dpms force off' just blanks the screens, but the backlight remains on. That's what happens when the screen saver turns off the display. I have tried it with and without a locker and the displays never turn off. I have spent a bunch of time with this system trying to convince it to turn off with dpms, but no luck. xrandr works, but that's an incomplete description. One of the comments was whether the display itself had dmps enabled, but I don't see a way to do that with the ACERs. Any suggestions?
aplattner said:It probably is in the display, actually, since most other displays work. In Windows, the operating system does the equivalent of an xrandr --off right after it triggers DPMS, which explains how the monitor manufacturer may have missed that it doesn't respond to DP DPMS commands correctly.

That seems possible. But given that xrandr can turn off the display, there must be a good way to hook up xrandr to do the work in linux. On the other hand, I wonder if there is a way to convince the display to cooperate.

I have two systems, both running Fedora 19, both show that DPMS is enabled (via 'xset q'):
  1. one has a G98 Quadro NVS 295 with one Acer K272HUL and one ASUS (2560x1440), driver version 331.67
  2. one has a gtx660 with two Acer K272HUL (2560x1440), driver version 346.47

The first turns off the displays just fine.

The second cannot turn off the display. 'xset dpms force off' just blanks the screens, but the backlight remains on. That's what happens when the screen saver turns off the display. I have tried it with and without a locker and the displays never turn off. I have spent a bunch of time with this system trying to convince it to turn off with dpms, but no luck. xrandr works, but that's an incomplete description. One of the comments was whether the display itself had dmps enabled, but I don't see a way to do that with the ACERs.

Any suggestions?

#14
Posted 04/06/2015 03:50 PM   
I am having this issue just recently with my 980 Ti and DisplayPort monitor. It actually worked correctly prior to the big update (version 355). It works just fine in Windows with the latest drivers as well.
I am having this issue just recently with my 980 Ti and DisplayPort monitor. It actually worked correctly prior to the big update (version 355). It works just fine in Windows with the latest drivers as well.

#15
Posted 08/25/2015 03:19 AM   
Scroll To Top

Add Reply