https://bugs.kde.org/show_bug.cgi?id=505139
Bug ID: 505139
Summary: Wayland: Display output severely corrupted (horizontal
green/white lines) after resume from prolonged
standby; persistent ACPI errors & SDDM/IBus
instability (AMD HD 5870)
Classification: Plasma
Product: kwin
Version First 6.3.5
Reported In:
Platform: Manjaro
OS: Linux
Status: REPORTED
Severity: normal
Priority: NOR
Component: wayland-generic
Assignee: kwin-bugs-n...@kde.org
Reporter: stefan_...@yahoo.de
Target Milestone: ---
SUMMARY
Wayland: Display output severely corrupted (horizontal green/white lines) after
resume from prolonged standby; persistent ACPI errors & SDDM/IBus instability
(AMD HD 5870)
STEPS TO REPRODUCE
Scenario 1: Standby from SDDM (HIGHLY FREQUENT FAILURE - almost 100% after
overnight standby)
1. Boot system to the SDDM login screen (Wayland session pre-selected or
active).
2. Do NOT log in.
3. Activate standby mode using the option available on the SDDM screen.
4. Leave the system in standby for several hours (e.g., overnight).
5. Wake up the system (e.g., by pressing a key or mouse button).
Scenario 2: Standby from active Plasma session (RARE FAILURE - but occurs)
1. Boot system and log into a KDE Plasma Wayland session.
2. Activate standby mode via the application launcher.
3. Leave the system in standby for several hours (e.g., overnight).
4. Wake up the system.
OBSERVED RESULT
The display output is completely corrupted and unusable. The screen shows
horizontal, fragmented green, white, and discolored lines, rendering it
completely illegible. (See attached `screenshot_corrupted_display.jpg`). [HIER
BITTE BESCHREIBEN, WAS SIE SEHEN - z.B. statisches Rauschen, zerrissene Farben,
Fragmente des letzten Bildschirminhalts, wie ein alter analoger Fernseher ohne
Empfang etc. Eventuell erwähnen, ob der Mauszeiger sichtbar/beweglich ist,
falls zutreffend.] The system requires a hard reboot to become usable again.
The kernel log for the boot session *preceding* the failure (`journalctl -b
-1`, see attached `kernel_meldungen_vorheriger_boot_STANDBY_FAILURE.txt`) shows
a seemingly normal suspend entry (`PM: suspend entry (deep)` at `Jun 02
01:04:41`) and a seemingly normal resume sequence approximately 18 hours later
(starting `Jun 02 19:47:26`). The kernel reports `PM: suspend exit` at `Jun 02
19:47:27`. Radeon DRM also reports successful re-initialization (ring tests,
UVD init, IB tests) during this resume.
However, immediately following `PM: suspend exit` (at `Jun 02 19:47:27` in
`kernel_meldungen_vorheriger_boot_STANDBY_FAILURE.txt`), `ibus-ui-gtk3` crashes
with a segfault (details also in `standby_fehler_vorheriger_boot_ERRORS.txt`).
This might indicate an issue with the graphical userspace failing to restore
correctly.
KScreenLocker PAM errors (`pam_unix(kde:auth): conversation failed`) were
observed shortly before the system entered suspend (`Jun 02 01:04:41` in
`standby_fehler_vorheriger_boot_ERRORS.txt`).
EXPECTED RESULT
The login screen (if resumed from SDDM standby) or the desktop/lock screen (if
resumed from session standby) should be displayed correctly, and the system
should be usable after prolonged standby without graphical corruption.
SOFTWARE/OS VERSIONS
(Information primarily from KInfoCenter and `fastfetch`)
Linux/KDE Plasma: Manjaro Linux
KDE Plasma Version: 6.3.5
KDE Frameworks Version: 6.14.0
Qt Version: 6.9.0
Kernel Version: 6.14.6-2-MANJARO (64-bit)
Graphics Platform: Wayland
(Additional details from `fastfetch` output, which will be attached as
`fastfetch_output.txt`)
OS: Manjaro Linux x86_64
Host: G73Jh (1.0) (Manufacturer: ASUSTeK Computer Inc., Product Name: G73Jh,
System Version: 1.0)
DE: KDE Plasma 6.3.5 (Wayland)
WM: KWin (Wayland)
Display Manager (DM): sddm-autologin 0.21.0 (Wayland)
SDDM Package Version: 0.21.0-6
Graphics Processor (GPU): AMD JUNIPER (AMD Mobility Radeon HD 5870 [Discrete])
Mesa Package Version: 1:25.0.5-1
(glxinfo reports: OpenGL core profile version string: 4.5 (Core Profile)
Mesa 25.0.5-arch1.1, LLVM 19.1.7)
Processors (CPU): 8 × Intel® Core™ i7 CPU Q 840 @ 1.87GHz
Memory: 31.3 GiB (fastfetch reported: 3.36 GiB / 31.28 GiB (11%) used at the
time of a `fastfetch` run)
ADDITIONAL INFORMATION
- The issue is specific to Wayland sessions. X11 sessions resume correctly from
prolonged standby.
- Short standby periods (e.g., up to an hour) do not trigger the issue under
Wayland.
- Persistent ACPI errors occur during every boot sequence (see attached logs,
e.g., `standby_fehler_vorheriger_boot_ERRORS.txt` and
`dmesg_nach_neustart_aktueller_boot.txt`):
`ACPI Error: AE_ERROR, Returned by Handler for [PCI_Config]
(20240827/evregion-301)`
`ACPI Error: Aborting method \_SB.PCI0.PDRC._CRS due to previous error
(AE_ERROR) (20240827/psparse-529)`
These might indicate an underlying firmware/ACPI issue relevant to power
management.
- During the problematic boot session logged in
`standby_fehler_vorheriger_boot_ERRORS.txt`, SDDM crashed twice (coredumps at
`Mai 31 12:49:46` and `Mai 31 13:09:09`) when `sddm --version` was attempted by
the user. `sudo sddm --version` causes a black screen and system hang.
- The `ibus-ui-gtk3` segfault immediately after kernel resume (`Jun 02
19:47:27`) might be a symptom of the graphical session failing to restore.
- PipeWire "broken pipe" errors related to rtsp-client/raop-sink were observed
during the long-running session prior to suspend (see
`standby_fehler_vorheriger_boot_ERRORS.txt`). While likely unrelated to the
primary graphics corruption, they indicate other potential instabilities during
long sessions.
ATTACHMENTS:
1. `fastfetch_output.txt` (Full output of `fastfetch` from current boot)
2. `standby_fehler_vorheriger_boot_ERRORS.txt` (`journalctl -b -1 -p err
--no-pager` from the boot *preceding* the standby failure)
3. `kernel_meldungen_vorheriger_boot_STANDBY_FAILURE.txt` (`journalctl -b -1
-k -T --no-pager` from the boot *preceding* the standby failure, including
suspend/resume)
4. `dmesg_nach_neustart_aktueller_boot.txt` (`sudo dmesg -T` from a typical
boot *after* a standby failure and reboot)
5. `screenshot_corrupted_display.jpg` (Photo of the screen corruption after
resume)
6. [Optional, if requested by developers:
`standby_komplett_vorheriger_boot.txt` (`journalctl -b -1 --no-pager` - very
large)]
--
You are receiving this mail because:
You are watching all bug changes.
