https://bugs.kde.org/show_bug.cgi?id=525886
Sergey <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- Resolution|WAITINGFORINFO |--- Status|NEEDSINFO |REPORTED --- Comment #6 from Sergey <[email protected]> --- Checked against the recommendation page you linked. This machine is on the 595+ branch (NVIDIA UNIX Open Kernel Module 615.78.08, the report was filed on 615.71.09), so the relevant check is UseKernelSuspendNotifiers and not PreserveVideoMemoryAllocations. Results here: * cat /sys/module/nvidia_drm/parameters/modeset -> Y * grep UseKernelSuspendNotifiers /proc/driver/nvidia/params -> UseKernelSuspendNotifiers: 1 * nvidia-suspend.service and nvidia-resume.service are enabled (nvidia-hibernate.service is disabled, hibernate is not used on this machine) * driver is the open kernel module, not the proprietary one So the configuration already follows the recommendation and there is nothing to change on my side. Two notes that may be relevant: 1. The freeze has not come back since the 615.78.08 update, but I have only had a couple of resume cycles since then, so I am not calling that a fix. The one resume that did break was a bad one at the driver level overall: kwin logged "atomic commit failed: Permission denied" and nvidia_drm logged a sync FD semaphore registration error in the same boot. 2. PreserveVideoMemoryAllocations is 2 here, which is what the 615 series sets on its own. If you expected 1 even on 595+, tell me and I will override it and retry, since it is one modprobe.d line plus an initramfs rebuild. I have a journal hook that captures drm_info plus both journals on every resume, so if the freeze shows up again I will attach a dump from the actual failure. -- You are receiving this mail because: You are watching all bug changes.
