https://bugs.kde.org/show_bug.cgi?id=527480
Bug ID: 527480
Summary: KDE Plasma Desktop intermittently freezes for a second
after upgrade to 6.8
Classification: Plasma
Product: plasmashell
Version First 6.8.0
Reported In:
Platform: openSUSE
OS: Linux
Status: REPORTED
Severity: normal
Priority: NOR
Component: general
Assignee: [email protected]
Reporter: [email protected]
CC: [email protected]
Target Milestone: 1.0
After upgrading openSUSE Tumbleweed from Plasma 6.7 to 6.8.0, the entire
desktop intermittently freezes for approximately 1–2 seconds. I was already
using Wayland under Plasma 6.7 without this problem.
During a freeze, the pointer stops responding and desktop widgets, including a
CPU monitor, stop updating. The freezes also occur when scrolling web pages
using keyboard keys without moving the mouse.
I have not identified a reliable activity that triggers them. Rebooting does
not resolve the issue, and it also occurs with the external monitor
disconnected.
**Steps to reproduce**
1. Start a Plasma 6.8.0 Wayland session on the system described below, with
normal NVIDIA runtime power management enabled.
2. Use the desktop normally, including browsing and scrolling web pages.
3. Intermittently, the desktop freezes for approximately 1–2 seconds.
The issue is intermittent; I cannot trigger it consistently on demand.
**Observed result**
The desktop temporarily stops updating and responding.
Each observed freeze coincides with the following message in KWin’s journal,
viewed using:
journalctl --user -b -u plasma-kwin_wayland.service --no-pager
kwin_wayland: The main thread was hanging temporarily!
**Expected result**
The desktop remains responsive with NVIDIA runtime power management enabled, as
it did before the update.
**System information**
- Distribution: openSUSE Tumbleweed
- Laptop: HP ZBook Studio 15.6-inch G8 Mobile Workstation
- KWin / Plasma: 6.8.0
- Qt: 6.11.2
- Mesa: 26.2.4
- Kernel: 7.2.11-1-default
- Session: Wayland
- Intel GPU: UHD Graphics, Tiger Lake, using `i915`
- NVIDIA GPU: RTX A2000 Laptop GPU / GA107, using `nouveau`
- NVIDIA’s proprietary driver is not installed
KWin reports:
```text
Compositing Type: OpenGL ES
OpenGL vendor string: Intel
OpenGL renderer string: Mesa Intel(R) UHD Graphics (TGL GT1)
OpenGL platform interface: EGL
```
**GPU and display configuration**
```text
0000:00:02.0 → /dev/dri/card1 → i915
0000:01:00.0 → /dev/dri/card0 → nouveau
```
Both connected displays are driven by Intel:
```text
card1-eDP-1: connected
card1-DP-1: connected
```
All NVIDIA display connectors report disconnected.
The laptop display runs at 120 Hz with 185% scaling; the external display runs
at 60 Hz with 110% scaling. The freezes also occur with only the laptop display
connected.
**Additional diagnostics**
- No DRM events appeared in `udevadm monitor` during the freezes.
- The kernel journal did not show GPU hangs, resets, or graphics timeouts
corresponding to the freezes.
- KWin logs these messages at session startup, although I do not know whether
they are related:
```text
MESA: error: ZINK: failed to choose pdev
drmSetClientCap for Atomic Mode Setting failed. Using legacy mode on GPU
"/dev/dri/card0"
```
**Workaround**
Preventing the NVIDIA GPU from entering runtime power saving stops the freezes
in my testing:
```bash
echo on | sudo tee /sys/bus/pci/devices/0000:01:00.0/power/control
```
Automatic runtime power management can be restored with:
```bash
echo auto | sudo tee /sys/bus/pci/devices/0000:01:00.0/power/control
```
**Possible explanation**
The successful workaround suggests that NVIDIA/Nouveau runtime power management
is involved in blocking KWin’s main thread, despite Intel handling compositing
and both connected displays.
This is an inference from the workaround, not a confirmed root cause. The issue
began after the update to Plasma 6.8, but I have not isolated whether the
regression is in KWin, Mesa, Wayland, or the kernel/Nouveau driver.
--
You are receiving this mail because:
You are watching all bug changes.
