> On Unix root can break out of the chrooted environment by the following: > > # mknod hda1 b 3 1; # assume hda1 is your root filesystem > # mkdir mnt > # mount -n hda1 mnt > # chroot mnt > # echo "I'm now in the original root directory!" > > Of course, you can still use system calls to achieve the same result. > > Is there such an escape possible too?
The Hurd equivalent of this would not put in the original root directory, but in the root directory of a second filesystem using the same disk device. There will always be many ways for a process that is root to get access to the system-wide state.
