https://bugs.kde.org/show_bug.cgi?id=522564
--- Comment #6 from Michael <[email protected]> --- I found several KDE reports/discussions describing related output/placeholder-screen behaviour, which may help put this report in context. In particular, KDE Bug 511757 concerns libkscreen/KWayland output handling and includes the same diagnostic: `There are no outputs - creating placeholder screen` and: `requesting unexisting screen available rect -1` There are also KDE Discuss reports describing similar behaviour with an external monitor on Wayland/AMD hardware, including the same placeholder-screen messages. My case appears related, but there are some important differences: * I am running Plasma/KWin/libkscreen 6.7.5. * The external HDMI-A-1 display is the only enabled display; the internal eDP-1 panel is disabled. * The failure occurs while the machine is at the lock/login screen after being idle. * `kscreenlocker_greet` is one of the processes that simultaneously reports no outputs. * Many other independent KDE processes report the same condition at essentially the same timestamp. * KWin does not crash or restart. * PowerDevil does not restart. * There is no kernel/DRM/AMDGPU message in the relevant interval. * There is no logged HDMI hotplug event. * There is no logged suspend/resume or DPMS event. * The kernel journal is completely silent from monotonic 6500 through 6650 seconds, covering the failure. * The display becomes available again after logging in. The relevant sequence is approximately: 6586.634 — many KDE processes simultaneously report: `There are no outputs - creating placeholder screen` 6586.637 — PowerDevil reports the same condition. 6587.681 — another widespread wave of the same placeholder-screen condition. The DDC/I2C permission errors from PowerDevil occur after the initial zero-output transition, so they appear to be a consequence rather than the initiating event. Because there is no corresponding DRM/kernel event, I am wondering whether this represents a userspace KWin/libkscreen output-state transition in which the set of available outputs temporarily becomes empty. I would be interested in knowing whether this should be investigated as part of the same output-device lifecycle/configuration code involved in Bug 511757, or whether a new bug would be more appropriate. -- You are receiving this mail because: You are watching all bug changes.
