External displays go white after a display sleep

Bug #2052452 reported by Alex Lowe
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
New
Undecided
Unassigned

Bug Description

After letting the displays timeout and go to sleep, external displays have a white layer between the windows and the mouse. It's still possible to interact with the windows below this layer, but it's not possible to see what's happening without moving them to a display that doesn't have this glitch.

This occurs in both x11 and wayland sessions, and it's new on noble (didn't see it on mantic, including with the backports ppa).

ProblemType: Bug
DistroRelease: Ubuntu 24.04
Package: kwin-wayland 4:5.27.10-0ubuntu1
ProcVersionSignature: Ubuntu 6.6.0-14.14-generic 6.6.3
Uname: Linux 6.6.0-14-generic x86_64
ApportVersion: 2.27.0-0ubuntu6
Architecture: amd64
CasperMD5CheckResult: pass
CurrentDesktop: KDE
Date: Mon Feb 5 11:41:37 2024
InstallationDate: Installed on 2023-10-23 (106 days ago)
InstallationMedia: Kubuntu 23.10 "Mantic Minotaur" - Release amd64 (20231010)
SourcePackage: kwin
UpgradeStatus: Upgraded to noble on 2024-01-26 (10 days ago)

Revision history for this message
Alex Lowe (lengau) wrote :
Revision history for this message
Alex Lowe (lengau) wrote :

Hmm this may not be a kwin bug - I tried logging out and back in, and on SDDM the displays worked fine, but logging back into plasma I had the problem again. I've installed sway and will try logging into it the next time this occurs. I'll report back whether sway presents the same issues when logging back in.

Revision history for this message
Alex Lowe (lengau) wrote :

I was also unable to trigger this by running `kscreen-doctor -d off` and waiting for the monitors to properly sleep, so there may be a more complex trigger needed.

Revision history for this message
Alex Lowe (lengau) wrote :

Definitely not a KDE bug since this occurs in other desktops as well. I'm guessing it has to do with the amdgpu driver?

affects: kwin (Ubuntu) → linux (Ubuntu)
Revision history for this message
Alex Lowe (lengau) wrote :

This occurred after a fresh reboot this morning, after letting my displays go to sleep for ~5 minutes. dmesg log also attached. The IO_PAGE_FAULT entries begin shortly after moving my mouse to turn the displays back on.

Revision history for this message
Alex Lowe (lengau) wrote :

Attached: a photograph of the displays after exiting display sleep but before unlocking. If it's relevant, the affected display is attached to the front right port of an AMD Framework laptop (USB 3.2 gen 2) using a device recognised by lsusb as `Bus 001 Device 002: ID 0bda:5442 Realtek Semiconductor Corp. Cable Matters USB-C Video Cable`

This display seems to be more often affected than the one attached to my docking station. However, disconnecting it and moving it to the rear left port (USB 4) does not fix the issue.

Revision history for this message
Alex Lowe (lengau) wrote :

After more research (including getting different symptoms with kernel 6.8-rc3), this appears to be the amdgpu scatter/gather bug: https://community.frame.work/t/responded-screen-flashing-partially-white-on-framework-13-amd-ubuntu-22-04/40626/2

I've applied the `amdgpu.sg_display=0` kernel command line workaround and will report back.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.