Qubes became extremely sluggish after a dom0 update

User Support
1

Qubes has become incredibly sluggish since I updated dom0 today. What happened? Is there a solution?

2

:police_car_light: Strange things observed today.

3

What does this sluggishness manifest as for you? Did it happen immediately or only after a reboot? Are you on stable, security-testing or testing repos?

4

It’s extremely slow—it’s become sluggish, like an old computer. It looks like dom0 is using an excessive amount of CPU. The slowdown started right after I updated dom0 and rebooted. I believe I’m using the stable version; I haven’t made any changes since installation, so it’s still on the default settings.

5

Could you check which process is using this much CPU? htop or btop are good commands to do that.

6

it’s Xorg , qubes-guid and xfwm4

7

I noticed this too.

It appears that when fcitx5 is launched on a VM with gui-agent-linux v4.3.19 installed, it hangs. So, when I reverted it to the previous version, it now works properly:

sudo dnf install qubes-gui-agent-4.3.18

It seems the debian repositories don’t have packages for previous versions…

8

@kzlz

I believe you should make an issue on the issue tracker for this.

9

I think you’re right; the VM doesn’t become incredibly slow if I use a template without fcitx. However, since I only use Kicksecure and Whonix templates, I’m not really sure how to revert to the previous version.

10

Do you see the high cpu usage in dom0 or in the appvms / qubes?
Also as you mentioned xfwm4. Is everyone who experiences this running xfce? Anyone running KDE or something else?

11

I pushed a fix here: Limit XEmbed size normalization to 1x1 clients by besendorf · Pull Request #265 · QubesOS/qubes-gui-agent-linux · GitHub

12

For some reason after the dom0 update none of my qubes are starting anymore nothing happens when i click on any applications or on the start qube button

13

Does anyone know how to fix this

14

thank you

15

Did someone test this?

16

Thank you for working on this issue so promptly.

So I’ve built and tested a package that included your PR on fedora-44-xfce, but it looks like Xorg won’t start…

journalctl -xe -u qubes-gui-agent -l --no-pager 
Aug 20 04:30:13 fedora-44-xfce qubes-gui[630]: open status file: Permission denied
Aug 20 04:30:13 fedora-44-xfce qubes-gui[630]: Waiting on /var/run/xf86-qubes-socket socket...
Aug 20 04:30:14 fedora-44-xfce qubes-gui[630]: Xorg exited in the meantime, aborting
Aug 20 04:30:14 fedora-44-xfce systemd[1]: qubes-gui-agent.service: Main process exited, code=exited, status=1/FAILURE
cat ~/.xsession-errors 
X.Org X Server 1.21.1.24
X Protocol Version 11, Revision 0
Build ID: xorg-x11-server 21.1.24-1.fc44 
Current version of pixman: 0.46.2
	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/home/user/.local/share/xorg/Xorg.0.log", Time: Thu Aug 20 04:30:14 2026
(++) Using config file: "/etc/X11/xorg-qubes.conf"
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
Parse error on line 26 of section Monitor in file /etc/X11/xorg-qubes.conf
	ModeLine VSyncEnd expected
(EE) 
Fatal server error:
(EE) no screens found(EE) 
(EE) 
Please consult the The X.Org Foundation support 
	 at http://wiki.x.org
 for help. 
(EE) Please also check the log file at "/home/user/.local/share/xorg/Xorg.0.log" for additional information.
(EE) 
(EE) Server terminated with error (1). Closing log file.
/usr/bin/xinit: giving up
/usr/bin/xinit: unable to connect to X server: Connection refused
/usr/bin/xinit: server error
nl -ba /etc/X11/xorg-qubes.conf 
     1	Section "Module"
     2	        Load "fb"
     3	        Load "glamoregl"
     4	EndSection
     5	
     6	Section "ServerLayout"   
     7	        Identifier     "Default Layout"
     8	        Screen      0  "Screen0" 0 0  
     9	        InputDevice "qubesdev pointer"
    10	EndSection
    11	
    12	Section "Device"
    13	        Identifier  "Videocard0"
    14	        Driver      "dummyqbs"
    15	        VideoRam 8100
    16	        Option "GUIDomID" "0"
    17	        # Render option requires to override default Intel device
    18	        # Option "Render" "/dev/dri/renderD128"
    19	EndSection
    20	
    21	Section "Monitor"
    22	        Identifier "Monitor0"
    23	        HorizSync 0-1
    24		VertRefresh 0-1
    25		Modeline "QBx" 0  1 2 3  1 2 3 
    26	EndSection
    27	
    28	Section "Screen"
    29	        Identifier "Screen0"
    30	        Device     "Videocard0"
    31		Monitor    "Monitor0"
    32	        DefaultDepth     
    33	        SubSection "Display"
    34	                Viewport   0 0
    35	                Depth      
    36			Modes "QBx" 
    37	        EndSubSection
    38	EndSection
    39	
    40	
    41	Section "InputDevice"
    42	        Identifier  "qubesdev pointer"
    43	        Driver      "qubes"
    44	EndSection

I noticed that the Modeline in the “Monitor” section of xorg-qubes.conf looks strange.
On a properly working qube, that section shows as
Modeline "QB1920x1080" 96 1920 1921 1922 1923 1080 1081 1082 1083

It might be PEBKAC. Let’s see the results of ongoing OpenQA tests…

17

OpenQA is happy about this patch. It doesn’t change anything related to starting of the gui-agent, so it’s likely unrelated. The open status file: Permission denied message suggests some issues with installing it - maybe wrong SELinux labels (if the binary was copied manually)? You can try running restorecon on the gui agent binary.