Occasionally, I do something (i.e., attempt to run VirtualBox) that provokes a panic on my workstation. When this happens, the system becomes completely unresponsive where not even a shutdown signal from pressing the power button works. It is probably a kernel panic, but there is no dump and no reboot. There is only a hard freeze. Due to running GEOM_RAID, this leads to a nice long synchronization compared to if it did reboot, so I am trying to figure out how to fix that problem. I suspect the Nvidia driver (v440.100 built from ports) is somehow related to this, I am not certain. System details: FreeBSD 12-STABLE (r365263), but this has been happening for awhile nvidia-driver-440.100 GeForce GTX 960 UFS + Intel ICH8+ RAID1 Dump disabled or enabled matters not. To see if it mattered if X was the active console or the type of panic mattered, I forced a panic from the console (ttyv0) while X was running. This is all it gave me before becoming unresponsive. ------------- panic: kdb_sysctl_panic cpuid = 1 time = 1599367793 KDB: stack backtrace #0 0xffffffff80731ec5 at kdb_backtrace+0x65 #1 0xffffffff806e615b at vpanic+0x17b #2 0xffffffff806e5fd3 at panic+0x43 #3 0xffffffff80732891 at kdb_sysctl_panic+0x61 #4 0xffffffff806f503a at sysctl_root_handler_locked+0x8a #5 0xffffffff806f4469 at sysctl_root+0x249 #6 0xffffffff806f4ad8 at userland_sysctl+0x178 #7 0xffffffff806f491f at sys___sysctl+0x5f #8 0xffffffff80a324c7 at amd64_syscall+0x387 #9 0xffffffff80a0995e at fast_syscall_common+0xf8 Uptime: 1m52s ------------- Now, if I do the same without X running, then it spits out a stacktrace and reboots. Should the system at least reboot when this happens? Does it matter what options are used to build the Nvidia driver such as ACPI, which I have enabled? Sean -- scf at FreeBSD.org