0.Update the binutils to the latest one from ports. We need to do this because
  the binutils that is shipped with base does not understand how to assemble
  the newly added VT-x instructions.

  pkg_add -r binutils

  Also make sure that your processor supports VT-x with nested page tables.
  Anything after 2010 should work just fine.

1.Check out /projects/bhyve and install world and kernel on the host.

2.Compile the bhyve guest kernel
  make -DNO_MODULES KERNCONF=BHYVE buildkernel

  Note that if you plan to use the pre-built guest kernel then you don't
  need to do this. See step (9)

3.Set the 'hw.physmem' variable in /boot/loader.conf to restrict the memory
  seen by the host kernel. The remaining memory is used to allocate to VMs.

  For e.g. the following line in /boot/loader.conf restricts the host to
  seeing only the first 4GB of physical memory.

  hw.physmem="0x100000000"

4.The vmm.ko module causes the kernel to spit out many, many witness warnings
  so for the time being turn off witness warnings by setting the following
  tunables in /boot/loader.conf.

  debug.witness.watch="0"

5.Reboot so that the newly installed kernel and the hw.physmem tunable
  take effect.

6.Load the following kernel modules:
  
  kldload vmm
  kldload if_tap

7.Create the tap0 interface

  ifconfig tap0 create

8.Check out /projects/virtio and make sys/modules/virtio

9.Create a VM boot area (for e.g. /usr/share/vm1) and populate it with the
  following:

A pre-built tarball of the VM boot area is available at:
http://people.freebsd.org/~neel/bhyve/vm1.tar.gz

It was created as follows:

mkdir /usr/share/vm1
mkdir /usr/share/vm1/boot
mkdir /usr/share/vm1/boot/kernel

cd /usr/share/vm1
cp ${OBJDIR}/sys/boot/userboot/userboot/userboot.so .

# create the virtio backing disk device (32MB in size in this example)
dd if=/dev/zero of=diskdev count=32768 bs=1024

cd /usr/share/vm1/boot
cp /boot/*.4th boot
cp -a /boot/defaults .
cp /boot/loader.help .
cp /boot/loader.rc .
cp /boot/menu.rc .

cat > loader.conf << EOF
kernel="/kernel"

virtio_load="YES"
if_vtnet_load="YES"
virtio_pci_load="YES"
virtio_blk_load="YES"

kern.hz="100"
hw.pci.enable_msix="0"
hw.pci.honor_msi_blacklist="0"
bootverbose="1"

mfsroot_load="YES"
mfsroot_type="mfs_root"
mfsroot_name="mdroot"
EOF

# Guest state
cd /usr/share/vm1/boot/kernel
# Copy guest kernel created in step (2)
cp kernel.bhyve kernel
# Copy virtio modules created in step (8)
find /projects/virtio/sys/modules/virtio -name '*.ko' -exec cp {} ./ \;
# Root filesystem for the guest
cp /guest/memory/filesystem mdroot

10. start the virtual machine
    ./vmrun.sh vm1

11. Interacting with the virtual machine

    The virtual machine has two virtio devices connected to it

    vtnet0: virtual network interface that is connected to the host's
    tap0 interface

    vtbd0: virtual block device that uses the backing file
    /usr/share/vm1/diskdev on the host

    The pre-build root filesystem creates a VM with a "root" user without
    any password. It also configures the vtnet0 interface with an IP address
    of 192.168.1.1

12. stopping and exiting the virtual machine

    Type "quit" in the loader cli.