[SOLVED] Wrong screen resolution after nvidia driver update

piepo · 2021-03-07 12:46:08

I have an Nvidia GK106GL (Quadro K4000) graphic card and a 34'' (21:9) LG monitor with a resolution of 2560x1080. After the last drivers update I can't get the right screen resolution: the maximum I can set is a 1920x1080 (16:9). I have done what is said in the Nvidia wiki for their drivers installation: added the kernel parameters, kernel module and pacman hook (I didn't do this the first time I installed them).
Anyone knows what is the problem/what I'm doing wrong?
This is the output of xrandr:

$ xrandr
Screen 0: minimum 8 x 8, current 1920 x 1080, maximum 16384 x 16384
DVI-I-0 disconnected (normal left inverted right x axis y axis)
DVI-I-1 disconnected (normal left inverted right x axis y axis)
DP-0 disconnected (normal left inverted right x axis y axis)
DP-1 connected primary 1920x1080+0+0 (normal left inverted right x axis y axis) 798mm x 334mm
   1920x1080     60.00*+  59.94    50.00  
   1680x1050     59.95  
   1600x900      60.00  
   1280x1024     75.02    60.02  
   1280x800      59.81  
   1280x720      60.00    59.94    50.00  
   1152x864      60.00  
   1024x768      75.03    60.00  
   800x600       75.00    60.32  
   720x576       50.00  
   720x480       59.94  
   640x480       75.00    59.94    59.93  
DP-2 disconnected (normal left inverted right x axis y axis)
DP-3 disconnected (normal left inverted right x axis y axis)

Last edited by piepo (2021-03-09 18:15:25)

seth · 2021-03-07 13:27:47

Please post the xorg log.
You can in dobut https://wiki.archlinux.org/index.php/Xr … esolutions

piepo · 2021-03-07 14:19:18

this is the the Xorg.0.log
I also have another log (named Xorg.1.log) though I have only one gnome session opened, is it normal?
I tried to add the undetected resolution but I got a "BadMatch" as a response.

$ xrandr --addmode DP-1 2560x1080_60.00
X Error of failed request:  BadMatch (invalid parameter attributes)
  Major opcode of failed request:  140 (RANDR)
  Minor opcode of failed request:  18 (RRAddOutputMode)
  Serial number of failed request:  39
  Current serial number in output stream:  40

seth · 2021-03-07 14:34:12

GDM uses a second server for the session, that's normal expectable.

1. Did you first "xrandr --newmode …"?
2. Is this actually a display port connection or HDMI in disguise?
3. Can you read the edit w/ get-edid from https://archlinux.org/packages/communit … read-edid/ ?

piepo · 2021-03-07 15:11:41

1. Yes:

$cvt 2560 1080
# 2560x1080 59.98 Hz (CVT) hsync: 67.17 kHz; pclk: 230.00 MHz
Modeline "2560x1080_60.00"  230.00  2560 2720 2992 3424  1080 1083 1093 1120 -hsync +vsync
$xrandr --newmode "2560x1080_60.00"  230.00  2560 2720 2992 3424  1080 1083 1093 1120 -hsync +vsync4

2. It's a display port on the graphic card side and an HDMI on the monitor side with an adapter in the middle. Maybe it's not the best solution but until Friday everything worked fine. Could it create some problems?
3. This is the output of

$ sudo get-edid

and this

Partial Read... Try again

is the output of parse-edid. (if I understood well how this package has to be used)
It seems that something is wrong...

seth · 2021-03-07 15:20:51

display port on the graphic card side and an HDMI on the monitor side with an adapter in the middle … Could … create some problems

If you check your pacman log, does the update to 460.56 (10 days ago on the mirrors) coincide with the problem emerging?
Also

I didn't do this the first time I installed them

did you interim use nouveau?

piepo · 2021-03-07 16:08:14

Looking at the pacman log, In February I used it until the 22nd (the last driver update is on 27) and I did a pacman -Syu on the march 5th at 10.29 (from 22 to last Friday I don't remember to have used this machine). In this general update, nvidia drivers weren't updated.
I remember that something strange happened Friday: at a certain point in the evening (21.27 according to the logs) I was installing VirtualBox and I had problems with the USB (I forgot to install the usb-utils...). So I installed them and rebooted: after the reboot my mouse pointer wasn't an arrow but a sort of "rectangle cloud". While I was trying to figure out what happend I noticed that nvidia-smi was missing: at that point I realised that maybe the drivers were "gone" and I did a pacman -S nvidia at 22.19 according to the logs. After that I added the kernel parameters, modules and hook (which solved me some start up issues I started to notice in the last three weeks): the mouse pointer was an arrow again but the screen resolution was gone.

seth wrote:

did you interim use nouveau?

No when I got this graphic card (like two months ago) I installed the proprietary drivers but I forgot to do the pacman hook. I specified it because maybe this caused some problems and this weird situation?

seth · 2021-03-07 20:25:26

sort of "rectangle cloud"

At this point you were running on the nouveau driver (recently known problem, https://bbs.archlinux.org/viewtopic.php … 9#p1959569 ) - for whatever reason.
So this might have been the "regression" - installing nvidia. You can inspect your older system journals ("sudo journalctl -b -5" for 5 boots ago) and see whether nvidia (nvrm) or nouveau is used.

You could increase the log level (depends on how X11 is started, startx/xinit?) to at least 7 to get nvidias mode validation logged and then disable the blocking validation step.
See /usr/share/doc/nvidia/README, ModeValidation option.

piepo · 2021-03-07 22:17:43

Yes you are right: I was definitely running on nouveau driver since at least the 31st of january!
Maybe it's a stupid question but I read that with Wayland the X server wasn't used, is it wrong?
E.g.:
grepping wayland

$ sudo journalctl -b -15|grep wayland
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Activating service name='org.a11y.Bus' requested by ':1.4' (uid=120 pid=917 comm="/usr/bin/gnome-shell ")
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Successfully activated service 'org.a11y.Bus'
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Activating service name='org.freedesktop.impl.portal.PermissionStore' requested by ':1.3' (uid=120 pid=917 comm="/usr/bin/gnome-shell ")
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Successfully activated service 'org.freedesktop.impl.portal.PermissionStore'
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Activating service name='org.gnome.Shell.Notifications' requested by ':1.3' (uid=120 pid=917 comm="/usr/bin/gnome-shell ")
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[942]: dbus-daemon[942]: Activating service name='org.a11y.atspi.Registry' requested by ':1.0' (uid=120 pid=917 comm="/usr/bin/gnome-shell ")
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[942]: dbus-daemon[942]: Successfully activated service 'org.a11y.atspi.Registry'
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[991]: SpiRegistry daemon is running with well-known name - org.a11y.atspi.Registry
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Successfully activated service 'org.gnome.Shell.Notifications'
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Activating service name='org.freedesktop.systemd1' requested by ':1.9' (uid=120 pid=995 comm="/usr/lib/gsd-sharing ")
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Activated service 'org.freedesktop.systemd1' failed: Process org.freedesktop.systemd1 exited with status 1
Mar 05 22:02:22 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Activating service name='org.gnome.Shell.Screencast' requested by ':1.24' (uid=120 pid=1019 comm="/usr/lib/gsd-media-keys ")
Mar 05 22:02:23 mind-the-swap /usr/lib/gdm-wayland-session[908]: dbus-daemon[908]: [session uid=120 pid=908] Successfully activated service 'org.gnome.Shell.Screencast'
Mar 05 22:20:04 mind-the-swap systemd[1157]: org.gnome.Shell@wayland.service: Succeeded.

grepping X11

$ sudo journalctl -b -15|grep X11
Mar 05 22:02:07 mind-the-swap /usr/lib/gdm-x-session[822]: (==) Using system config directory "/usr/share/X11/xorg.conf.d"
Mar 05 22:02:24 mind-the-swap org.gnome.Shell.desktop[1150]: >                   X11 cannot support keycodes above 255.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Starting GNOME Shell on X11...
Mar 05 22:02:29 mind-the-swap systemd[1157]: Condition check resulted in GNOME Shell on X11 being skipped.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Stopped GNOME Shell on X11.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Starting GNOME Shell on X11...
Mar 05 22:02:29 mind-the-swap systemd[1157]: Condition check resulted in GNOME Shell on X11 being skipped.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Stopped GNOME Shell on X11.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Starting GNOME Shell on X11...
Mar 05 22:02:29 mind-the-swap systemd[1157]: Condition check resulted in GNOME Shell on X11 being skipped.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Stopped GNOME Shell on X11.
Mar 05 22:02:29 mind-the-swap systemd[1157]: Condition check resulted in GNOME Shell on X11 being skipped.
Mar 05 22:02:31 mind-the-swap systemd[1157]: GNOME session X11 services is not active.
Mar 05 22:02:32 mind-the-swap systemd[1157]: Reached target GNOME session X11 services.
Mar 05 22:02:32 mind-the-swap systemd[1157]: Reached target GNOME session X11 services.
Mar 05 22:02:35 mind-the-swap gnome-shell[1728]: >                   X11 cannot support keycodes above 255.
Mar 05 22:20:03 mind-the-swap systemd[1157]: Stopped target GNOME session X11 services.
Mar 05 22:20:03 mind-the-swap systemd[1157]: Stopped target GNOME session X11 services.

and grepping startx or xinit returns nothing... nonetheless I have log in the Xorg.0.log and so I imagine it is used.
I'm sorry if I'm saying something stupid/wrong.

boban_dj · 2021-03-08 05:43:08

whats your

cat /etc/X11/xorg.conf

I always run nvidia-settings as well to check the xorg.conf is set

seth · 2021-03-08 08:08:31

GDM defaults to wayland, https://wiki.archlinux.org/index.php/GD … rg_backend but you can select the gnome display server indepently (so you can run gnome on xorg from a wayland GDM)

About the original issue

seth wrote:

You could increase the log level (depends on how X11 is started, startx/xinit?) to at least 7 to get nvidias mode validation logged and then disable the blocking validation step.
See /usr/share/doc/nvidia/README, ModeValidation option.

piepo · 2021-03-08 09:02:18

This is my xorg.conf

$ cat /etc/X11/xorg.conf
# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig:  version 460.56

Section "ServerLayout"
    Identifier     "Layout0"
    Screen      0  "Screen0"
    InputDevice    "Keyboard0" "CoreKeyboard"
    InputDevice    "Mouse0" "CorePointer"
EndSection

Section "Files"
EndSection

Section "InputDevice"
    # generated from default
    Identifier     "Mouse0"
    Driver         "mouse"
    Option         "Protocol" "auto"
    Option         "Device" "/dev/psaux"
    Option         "Emulate3Buttons" "no"
    Option         "ZAxisMapping" "4 5"
EndSection

Section "InputDevice"
    # generated from default
    Identifier     "Keyboard0"
    Driver         "kbd"
EndSection

Section "Monitor"
    Identifier     "Monitor0"
    VendorName     "Unknown"
    ModelName      "Unknown"
    Option         "DPMS"
EndSection

Section "Device"
    Identifier     "Device0"
    Driver         "nvidia"
    VendorName     "NVIDIA Corporation"
EndSection

Section "Screen"
    Identifier     "Screen0"
    Device         "Device0"
    Monitor        "Monitor0"
    DefaultDepth    24
    SubSection     "Display"
        Depth       24
    EndSubSection
EndSection

boban_dj wrote:

I always run nvidia-settings as well to check the xorg.conf is set

With nvidia-settings I'm able to change "ViewPortIn" and "Panning" under "X Server Display Configuration". "ViewPortOut" remains the same and the new resolution is not seen by gnome (and I don't think it really changed because is not very defined)

seth wrote:

About the original issue

I was asking about X vs Wayland because I can't find how the X server is started and in the documentation is written that I can start a session with

% startx -- -verbose 5 -logverbose 5

I installed xorg-xinit package and started a session with logverbosity 7. I found that this can be the problem (as you said)

[  1672.142] (WW) NVIDIA(GPU-0):   Validating Mode "2560x1080_60":
[  1672.142] (WW) NVIDIA(GPU-0):     Mode Source: EDID
[  1672.142] (WW) NVIDIA(GPU-0):     2560 x 1080 @ 60 Hz
[  1672.142] (WW) NVIDIA(GPU-0):       Pixel Clock      : 181.25 MHz
[  1672.142] (WW) NVIDIA(GPU-0):       HRes, HSyncStart : 2560, 2608
[  1672.142] (WW) NVIDIA(GPU-0):       HSyncEnd, HTotal : 2640, 2720
[  1672.142] (WW) NVIDIA(GPU-0):       VRes, VSyncStart : 1080, 1083
[  1672.142] (WW) NVIDIA(GPU-0):       VSyncEnd, VTotal : 1093, 1111
[  1672.142] (WW) NVIDIA(GPU-0):       Sync Polarity    : +H -V 
[  1672.142] (WW) NVIDIA(GPU-0):     Mode is rejected: PixelClock (181.2 MHz) too high for
[  1672.142] (WW) NVIDIA(GPU-0):     Display Device (Max: 165.0 MHz).
[  1672.142] (WW) NVIDIA(GPU-0):     GPU extended capability check failed.
[  1672.142] (WW) NVIDIA(GPU-0):     Mode "2560x1080_60" is invalid.

Do you know what does it means? Meanwhile I'm trying to understand it

seth · 2021-03-08 09:11:46

The signal is too large for what is announced is possible (likely by the adapter) - and 181.25 is already reduced blanking.

Option "ModeValidation" "NoEdidMaxPClkCheck"

or maybe necessary (but try the other firs)

Option "ModeValidation" "NoMaxPClkCheck"

will allow you to ignore that, but the signal may fail (and in this case neither option will help you)

Nouveau might have been driving the output at 30Hz or 24Hz.

The lowest you can get @60Hz is

cvt12 2560 1080 60 -b
# 2560x1080 @ 60.000 Hz Reduced Blank (CVT) field rate 60.000 Hz; hsync: 66.660 kHz; pclk: 175.98 MHz
Modeline "2560x1080_60.00_rb2"  175.98  2560 2568 2600 2640  1080 1097 1105 1111 +hsync -vsync

which is still too high (175.98 > 165.0)

piepo · 2021-03-08 09:20:39

I see...
I found this on the nvidia forum https://forums.developer.nvidia.com/t/d … gh/71807/3
Maybe the best solution is to buy an active adapter? Thank you very much for the help seth

seth · 2021-03-08 13:38:19

I'd first try to ignore the restriction. If the signal fails, the restriction of the adapter is legit and your next best chance is to lower the pixelclock just a tiny bit more.
If it still fails you need to use a lower refresh rate or a better adapter.

piepo · 2021-03-08 19:35:34

I tried the first two options and seems that the adapter restriction is legit:

[  1010.169] (**) NVIDIA(GPU-0): Mode Validation Overrides for LG Electronics LG HDR WFHD
[  1010.169] (**) NVIDIA(GPU-0):     (DFP-1):
[  1010.169] (**) NVIDIA(GPU-0):     NoMaxPClkCheck
[  1010.169] (WW) NVIDIA(GPU-0):   Validating Mode "2560x1080_60":
[  1010.169] (WW) NVIDIA(GPU-0):     Mode Source: EDID
[  1010.169] (WW) NVIDIA(GPU-0):     2560 x 1080 @ 60 Hz
[  1010.169] (WW) NVIDIA(GPU-0):       Pixel Clock      : 181.25 MHz
[  1010.169] (WW) NVIDIA(GPU-0):       HRes, HSyncStart : 2560, 2608
[  1010.169] (WW) NVIDIA(GPU-0):       HSyncEnd, HTotal : 2640, 2720
[  1010.169] (WW) NVIDIA(GPU-0):       VRes, VSyncStart : 1080, 1083
[  1010.169] (WW) NVIDIA(GPU-0):       VSyncEnd, VTotal : 1093, 1111
[  1010.169] (WW) NVIDIA(GPU-0):       Sync Polarity    : +H -V 
[  1010.170] (WW) NVIDIA(GPU-0):     Mode is rejected: Unable to construct hardware-specific
[  1010.170] (WW) NVIDIA(GPU-0):     mode timings.
[  1010.170] (WW) NVIDIA(GPU-0):     GPU extended capability check failed.
[  1010.170] (WW) NVIDIA(GPU-0):     Mode "2560x1080_60" is invalid.

I tried to lower the pixel check with cvt specifying a lower refresh rate (is it right? I couldn't find something better)

$ cvt 2560 1080 30.00
# 2560x1080 29.97 Hz (CVT) hsync: 33.03 kHz; pclk: 106.75 MHz

but still I get BadMatch.
At this point I think I will get another adapter.

piepo · 2021-03-09 18:12:02

I bought an active adapter and now I get the right resolution without problems.
Thank you seth, I will mark as solved this thread.

Arch Linux

#1 2021-03-07 12:46:08

[SOLVED] Wrong screen resolution after nvidia driver update

#2 2021-03-07 13:27:47

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#3 2021-03-07 14:19:18

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#4 2021-03-07 14:34:12

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#5 2021-03-07 15:11:41

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#6 2021-03-07 15:20:51

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#7 2021-03-07 16:08:14

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#8 2021-03-07 20:25:26

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#9 2021-03-07 22:17:43

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#10 2021-03-08 05:43:08

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#11 2021-03-08 08:08:31

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#12 2021-03-08 09:02:18

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#13 2021-03-08 09:11:46

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#14 2021-03-08 09:20:39

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#15 2021-03-08 13:38:19

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#16 2021-03-08 19:35:34

Re: [SOLVED] Wrong screen resolution after nvidia driver update

#17 2021-03-09 18:12:02

Re: [SOLVED] Wrong screen resolution after nvidia driver update

Board footer