Hello,

I am using KDE on Arch Linux and since I got a new Laptop, I cannot really use external Monitors. After some time (ranges from a few minutes to a few hours) my desktops start to freeze, that is the taskbar and the desktop icons and background just freeze and cannot be clicked anymore. The applications run fine and I can also change them by ALT+TAB or start new ones with ALT+F2 although the start menu doesn't come up if I press SUPER. All desktops freeze individually and not all at the same time, which can easily be seen as the clock in each monitors taskbar is stuck at a different time. I can't find a solution for my problem and I don't know how to start debugging it. I haven't noticed that problem when only using the laptop without a external monitor.

When this happens, my workaround is a

Code: Select all

killall plasmashell && plasmashell

in a Konsole window, which then restarts the desktops/taskbars, making them usable again for a few minutes/hours.

When it happens while the screens are locked, I have to change the TTY, kill kscreenlocker_greet, which fortunately gets restarted automatically.

Can someone help me debug this?

I am running the following:

plasmashell 5.17.5
xorg-server 1.20.6-3
linux 5.4.10-arch1-1

Laptop Dell Latitude 7480 (i5-7300U CPU @ 2.60GHz, no discrete graphics)
Dock Dell WD15


Thanks in advance for any help or tips

I don't have a solution. I just want to give this a bump because I'm also seeing this issue on NixOS. I've looked at what plasmashell prints out to console when this happens, and nothing jumps out to me as an obvious error, but maybe I'm missing something.

I'm running:

Code: Select all

plasmashell 5.17.8
X.Org X Server 1.20.8
Linux kernel version 5.4.48


If there's anything else I can provide to help get to the bottom of it, don't hesitate to ask.

Hoe about updating first? Especially on Arch, 5.17.x is pretty old now, the most recent Plasma release is 5.19.2

I experience similar issue.

Here my versions that I'm using:

Code: Select all

Linux lenovo-pc 5.4.58-1-MANJARO #1 SMP PREEMPT Tue Aug 11 15:46:30 UTC 2020 x86_64 GNU/Linux
plasmashell 5.19.4
xorg-server                   [Installed] 1.20.8-4     


During normal use of laptop with external monitor attached plasmashell will freeze with random intervals: can't use any panels/start menu but ALT+F2 respond slowly. Also during that time I can switch between virtual desktops, run new programs, use Yakuake and switch language.

Usual solution is to restart plasmashell with script similar to one proposed above.

journalctl -f shows nothing on time of freeze. Freeze happened on 13:24:51, but as we can see there is no suspicious logs:

Code: Select all

Aug 27 13:23:31 lenovo-pc kwin_x11[1859]: Qt Quick Layouts: Detected recursive rearrange. Aborting after two iterations.
Aug 27 13:23:31 lenovo-pc kwin_x11[1859]: Qt Quick Layouts: Detected recursive rearrange. Aborting after two iterations.
Aug 27 13:23:31 lenovo-pc kwin_x11[1859]: file:///usr/lib/qt/qml/org/kde/plasma/components/Highlight.qml:47:5: QML Connections: Implicitly defined onFoo properties in Connections are deprecated. Use this syntax instead: function onFoo(<arguments>) { ... }
Aug 27 13:27:45 lenovo-pc org_kde_powerdevil[2270]: powerdevil: Screen brightness value:  1365
Aug 27 13:27:45 lenovo-pc org_kde_powerdevil[2270]: powerdevil: Screen brightness value:  1365


System doesn't have high load or memory usage on time of freeze and plasmashell process doesn't use CPU endlessly as well.
Any help would be appreciated, I'm ready to provide any logs that would be helpful to resolve this problem.

A person I am helping has been experiencing a similar issue on a Lenovo T570 for several months now. The laptop has an Nvidia graphics device, but no special configuration has been made to make use of it. External monitor does not factor into the problem. The problem was first observed on Manjaro, but now it continues on a freshly installed Arch (installed via EndeavourOS). The desktop freezes, while the mouse works. Switching to another virtual console with Ctrl-Alt-F2 and back with Ctrl-Alt-F1 usually heals the issue, but not every time.

Here is a snippet of the journal leading up to the virtual console switch at 18:52:13:

Code: Select all

elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1289, resource id: 39846624, major code: 19 (DeleteProperty), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1292, resource id: 39846624, major code: 19 (DeleteProperty), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1293, resource id: 39846624, major code: 18 (ChangeProperty), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1294, resource id: 39846624, major code: 19 (DeleteProperty), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1295, resource id: 39846624, major code: 19 (DeleteProperty), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1296, resource id: 39846624, major code: 19 (DeleteProperty), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1297, resource id: 39846624, major code: 7 (ReparentWindow), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1298, resource id: 39846624, major code: 6 (ChangeSaveSet), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1299, resource id: 39846624, major code: 2 (ChangeWindowAttributes), minor code: 0
elo 26 18:44:06 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 1300, resource id: 39846624, major code: 10 (UnmapWindow), minor code: 0
elo 26 18:44:09 computer audit[347]: USYS_CONFIG pid=347 uid=0 auid=4294967295 ses=4294967295 msg='op=statistics arg="refresh-rate-ms" pid=1102 uid=1000 result=success exe="/usr/bin/NetworkManager" hostname=? addr=? terminal=? res=success'
elo 26 18:44:09 computer NetworkManager[347]: <info>  [1598456649.4236] audit: op="statistics" arg="refresh-rate-ms" pid=1102 uid=1000 result="success"
elo 26 18:44:09 computer kernel: audit: type=1111 audit(1598456649.418:91): pid=347 uid=0 auid=4294967295 ses=4294967295 msg='op=statistics arg="refresh-rate-ms" pid=1102 uid=1000 result=success exe="/usr/bin/NetworkManager" hostname=? addr=? terminal=? res=success'
elo 26 18:44:09 computer plasmashell[1102]: file:///usr/lib/qt/qml/org/kde/plasma/extras/ExpandableListItem.qml:327: TypeError: Cannot read property 'width' of null
elo 26 18:44:11 computer systemd[1]: NetworkManager-dispatcher.service: Succeeded.
elo 26 18:44:11 computer audit[1]: SERVICE_STOP pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=NetworkManager-dispatcher comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
elo 26 18:44:11 computer kernel: audit: type=1131 audit(1598456651.235:92): pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=NetworkManager-dispatcher comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
elo 26 18:45:08 computer kdeinit5[12237]: Qt: Session management error: networkIdsList argument is NULL
elo 26 18:48:09 computer kwin_x11[1043]: qt.qpa.xcb: QXcbConnection: XCB error: 3 (BadWindow), sequence: 32816, resource id: 94371851, major code: 18 (ChangeProperty), minor code: 0
elo 26 18:52:13 computer kernel: nouveau 0000:02:00.0: Enabling HDA controller
elo 26 18:52:13 computer kernel: nouveau 0000:02:00.0: bus: MMIO read of 00000000 FAULT at 6013d4 [ IBUS ]
elo 26 18:52:13 computer org_kde_powerdevil[1153]: powerdevil: ACTIVE SESSION PATH CHANGED: "/"
elo 26 18:52:13 computer kernel: audit: type=1130 audit(1598457133.855:93): pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=getty@tty2 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
elo 26 18:52:13 computer audit[1]: SERVICE_START pid=1 uid=0 auid=4294967295 ses=4294967295 msg='unit=getty@tty2 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'


Like others in this thread, I am wondering what sort of debug switch I might flip to get more stuff in the journal. I know how to do it for individual units, but here I am at a loss.

I asked about this in Reddit and one idea was to observe, if there is a memory leak going on. It doesn't really feel like it, but let's find out next time it happens.

Debugging advice: https://community.kde.org/Plasma/Debugging

Just got mem info from another freeze. No leaking, total mem use was 3.4GB/16GB, so completely normal.

This could be due to some desktop effect running wild, selectively disabling them to find the culprit is probably the way to go.
Another (not infrequent) issue could be the available space on the disk...

I was on Arch for a long time and around kernel 5.7 was a game change for some laptops - could this be related?

robgriff444 wrote:I was on Arch for a long time and around kernel 5.7 was a game change for some laptops - could this be related?

5.7 was released in 31 May 2020, so it can't be related. Notice that this topic was created in January 2020.

And Neon / Ubuntu is still on 5.4. And I didn't say 5.7 SOLVED I said 5.7 was a game changer suggesting you all could do some research just in case. There are instances of people reverting kernels to solve freezing as well as upgrading to solve freezing. Solus recently had a long discussion about which kernels they should be on and one of the problems was about how long they could hold on to an old (and good for laptops) kernel.

I was only offering help, I read the title, I know when this post is from, maybe it's kernel related maybe not.

robgriff444 wrote:And Neon / Ubuntu is still on 5.4. And I didn't say 5.7 SOLVED I said 5.7 was a game changer suggesting you all could do some research just in case. There are instances of people reverting kernels to solve freezing as well as upgrading to solve freezing. Solus recently had a long discussion about which kernels they should be on and one of the problems was about how long they could hold on to an old (and good for laptops) kernel.

I thought you meant 5.7 CAUSED the problem. It definitely didn't solve it as it keeps happening with 5.8 kernel.

I will do a sweeping desktop effects deactivation on the Lenovo (noting the original settings beforehand). Disk space is not the problem, should be hundreds of gigs free.

Can you try the LTS kernel to rule it out?

robgriff444 wrote:Can you try the LTS kernel to rule it out?

You mean 4.19 or 5.4? Original poster was using linux 5.4.10.

beluga wrote:I asked about this in Reddit and one idea was to observe, if there is a memory leak going on. It doesn't really feel like it, but let's find out next time it happens.

During several freezes I cannot see any strange memory increases. Memory usage is way below available RAM and no swap usage.

Mamarok wrote:This could be due to some desktop effect running wild, selectively disabling them to find the culprit is probably the way to go.
Another (not infrequent) issue could be the available space on the disk...

Have more than 200GB free space.

I tried to disable Compositor completely but problem still persist.

robgriff444 wrote:Can you try the LTS kernel to rule it out?

I'm on Manjaro and problem was present with LTS 5.4 series kernel.
Currently trying to run latest 5.8 but no luck.

Will try to return to the 4.19 (previous LTS kernel) and will report how it will go.