I tested the requested 595.84 package with a different laptop configuration: Kubuntu 26.04, kernel 7.0.0-29-generic, KDE Plasma Wayland, GTX 1650 Mobile (10de:1f91), and nvidia-driver-595-open 595.84-0ubuntu0.26.04.1.
With the default package integration, UseKernelSuspendNotifiers: 0 left the /proc/driver/nvidia/suspend interface and nvidia-suspend.service enabled. That caused nvidia-sleep.sh to switch to VT 63. During that transition KWin logged an atomic DRM "Permission denied" and temporary renderD128 loss; earlier cycles could end in a black screen with cursor after resume. I changed only the suspend integration for the Open module: options nvidia NVreg_UseKernelSuspendNotifiers=1 and disabled the packaged NVIDIA /proc suspend/resume/hibernate units. VRAM preservation remained enabled. An instrumented S3 deep cycle on 595.84 used nv_pm_notifier, did not invoke nvidia-sleep.sh or a VT switch, did not log the atomic DRM denial or the prior DRM semaphore warning, and resumed Plasma correctly. This does not prove the same root cause as the Xwayland freeze described here, but it demonstrates a package-level configuration difference that avoids a separate Open/Wayland suspend failure on 595.84. Please consider making the Open package use the kernel-notifier path by default rather than the shared /proc/VT path. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2158993 Title: System hard-freezes after failed suspend (Xwayland freeze timeout) with nvidia-driver-595-open — deadlock in nvidia_modeset/console To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2158993/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
