In exploration of "jail/bhyve everything for security," I have successfully jailed the NIC (soon to bhyve with PCI passthrough), and jailed a VPN tunnel. I'm now considering 2 options for GUI jails:
1. `ssh -Y` between host/jail (definitely secure)
2. `mount_nullfs /tmp/.X11-host` to my jails, with xhost +
Since I jailed the NIC, blocked all traffic to the host, and don't intend to X Forward outside of the physical machine/host, I have this naive hope that ssh is unnecessary and no more secure than option #2.
Is this correct? Or are multiple GUI jails simultaneously sharing /tmp/.X11-unix a recipe for security vulnerabilities?
EDIT: I'm not able to start Falkon, OpenOffice, and getting crashes on many KDE applications. I am thinking that I need to install Qt and KDE to the host, as I'm guessing that many of these libraries require direct access to kernel resources? I'm not sure. I was hoping to avoid installing anything more than X11 to host.
1. `ssh -Y` between host/jail (definitely secure)
2. `mount_nullfs /tmp/.X11-host` to my jails, with xhost +
Since I jailed the NIC, blocked all traffic to the host, and don't intend to X Forward outside of the physical machine/host, I have this naive hope that ssh is unnecessary and no more secure than option #2.
Is this correct? Or are multiple GUI jails simultaneously sharing /tmp/.X11-unix a recipe for security vulnerabilities?
EDIT: I'm not able to start Falkon, OpenOffice, and getting crashes on many KDE applications. I am thinking that I need to install Qt and KDE to the host, as I'm guessing that many of these libraries require direct access to kernel resources? I'm not sure. I was hoping to avoid installing anything more than X11 to host.