A quick reference to...

In theory, you should never see one of these using the kernel your distro provides. If you are using a hand-rolled kernel, it usually means you've forgotten to build in support for the disk drive controller and filesystem used by the root filesystem, meaning it cannot be mounted

You'd be forgiven for thinking that once this happens, all is lost, but the kernel can still manage one last act of dignity, even if you kill it like this. The kernel can be told to reboot after a panic. Adding kernel. panic=10 to /etc/sysctl.conf will reboot it after 10 seconds, but requires that the system gets as far as mounting the root filesystem to read this file. More robustly, you can add panic=10 to the kernel parameters in your Grub menu.

You may wonder about the point of forcing a reboot if the kernel will only panic again and reboot again, ad infinitum. The answer lies in Grub's ability to set up a fallback boot. This is a godsend when trying a new kernel on a remote computer – far less embarrassing than having to phone and ask someone to reboot because you messed up a kernel config.