https://bugs.kde.org/show_bug.cgi?id=526072
[email protected] changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #2 from [email protected] --- I can reproduce a very similar DisplayPort wake failure with an AMD GPU on KDE Plasma 6.8.1. System: OS: EndeavourOS Session: Wayland Plasma / KWin / PowerDevil: 6.8.1 Kernel: 7.2.11-arch1-1 GPU: AMD Radeon RX 7900 XTX (amdgpu) Mesa: 26.2.5 Monitor: Titan Army P32A6V, DisplayPort (DP-1) Resolution: 3840x2160, 144 Hz Steps to reproduce: 1. Run: kscreen-doctor --dpms off 2. Wait 30 seconds. 3. Run: kscreen-doctor --dpms on Actual result: The monitor wakes briefly, displays "No Signal", and returns to standby. KWin reports DPMS as "on", and kscreen-doctor reports the display as "connected" and "enabled", despite no image being displayed. The issue also occurs after automatic screen-off in Plasma. Additional observations: - Reproduced at both 4K 144 Hz and 4K 60 Hz. - Reproduced with HDR, VRR and DDC/CI disabled and 8-bit color depth selected. - The problem does not appear to depend exclusively on DSC or HDR. - No amdgpu/DRM messages were found by the filtered kernel journal query. Recovery attempts: - Running "kscreen-doctor --dpms on" alone does not help. - Running "kscreen-doctor output.DP-1.enable" alone does not help. - Changing the display mode from 144 Hz to 60 Hz and back successfully restored the image in one test. - Running "kscreen-doctor output.DP-1.disable" followed by "kscreen-doctor output.DP-1.enable" restored the image in repeated tests. Interestingly, KWin rejects the disable operation with: "applying config failed! Disabling all outputs through configuration changes is not allowed" This is expected because DP-1 is my only connected display. However, the disable/enable sequence still appears to trigger successful recovery. Power-cycling the monitor also restores the image. This may be related to DisplayPort link training or video stream initialization after DPMS wake. I cannot yet determine whether the cause is KWin, amdgpu, or monitor firmware. I can provide additional logs and run further tests if needed. -- You are receiving this mail because: You are watching all bug changes.
