Bhyve driver
Bhyve is a FreeBSD hypervisor. It first appeared in FreeBSD 10.0. However, it's
recommended to keep tracking FreeBSD 10-STABLE to make sure all new features
of bhyve are supported.
In order to enable bhyve on your FreeBSD host, you'll need to load the vmm
kernel module. Additionally, if_tap
and if_bridge
modules
should be loaded for networking support. Also, since 3.2.0 the
virt-host-validate(1)
supports the bhyve host validation and could be
used like this:
$ virt-host-validate bhyve BHYVE: Checking for vmm module : PASS BHYVE: Checking for if_tap module : PASS BHYVE: Checking for if_bridge module : PASS BHYVE: Checking for nmdm module : PASS $
Additional information on bhyve could be obtained on bhyve.org.
Connections to the Bhyve driver ¶
The libvirt bhyve driver is a single-instance privileged driver. Some sample connection URIs are:
bhyve:///system (local access) bhyve+unix:///system (local access) bhyve+ssh://root@example.com/system (remote access, SSH tunnelled)
Example guest domain XML configurations ¶
Example config
The bhyve driver in libvirt is in its early stage and under active development. So it supports only limited number of features bhyve provides.
Note: in older libvirt versions, only a single network device and a single disk device were supported per-domain. However, since 1.2.6 the libvirt bhyve driver supports up to 31 PCI devices.
Note: the Bhyve driver in libvirt will boot whichever device is first. If you want to install from CD, put the CD device first. If not, put the root HDD first.
Note: Only the SATA bus is supported. Only cdrom
- and
disk
-type disks are supported.
<domain type='bhyve'> <name>bhyve</name> <uuid>df3be7e7-a104-11e3-aeb0-50e5492bd3dc</uuid> <memory>219136</memory> <currentMemory>219136</currentMemory> <vcpu>1</vcpu> <os> <type>hvm</type> </os> <features> <apic/> <acpi/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <disk type='file'> <driver name='file' type='raw'/> <source file='/path/to/bhyve_freebsd.img'/> <target dev='hda' bus='sata'/> </disk> <disk type='file' device='cdrom'> <driver name='file' type='raw'/> <source file='/path/to/cdrom.iso'/> <target dev='hdc' bus='sata'/> <readonly/> </disk> <interface type='bridge'> <model type='virtio'/> <source bridge="virbr0"/> </interface> </devices> </domain>
(The <disk> sections may be swapped in order to install from cdrom.iso.)
Example config (Linux guest)
Note the addition of <bootloader>.
<domain type='bhyve'> <name>linux_guest</name> <uuid>df3be7e7-a104-11e3-aeb0-50e5492bd3dc</uuid> <memory>131072</memory> <currentMemory>131072</currentMemory> <vcpu>1</vcpu> <bootloader>/usr/local/sbin/grub-bhyve</bootloader> <os> <type>hvm</type> </os> <features> <apic/> <acpi/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <disk type='file' device='disk'> <driver name='file' type='raw'/> <source file='/path/to/guest_hdd.img'/> <target dev='hda' bus='sata'/> </disk> <disk type='file' device='cdrom'> <driver name='file' type='raw'/> <source file='/path/to/cdrom.iso'/> <target dev='hdc' bus='sata'/> <readonly/> </disk> <interface type='bridge'> <model type='virtio'/> <source bridge="virbr0"/> </interface> </devices> </domain>
Example config (Linux UEFI guest, VNC, tablet)
This is an example to boot into Fedora 25 installation:
<domain type='bhyve'> <name>fedora_uefi_vnc_tablet</name> <memory unit='G'>4</memory> <vcpu>2</vcpu> <os> <type>hvm</type> <loader readonly="yes" type="pflash">/usr/local/share/uefi-firmware/BHYVE_UEFI.fd</loader> </os> <features> <apic/> <acpi/> </features> <clock offset='utc'/> <on_poweroff>destroy</on_poweroff> <on_reboot>restart</on_reboot> <on_crash>destroy</on_crash> <devices> <disk type='file' device='cdrom'> <driver name='file' type='raw'/> <source file='/path/to/Fedora-Workstation-Live-x86_64-25-1.3.iso'/> <target dev='hdc' bus='sata'/> <readonly/> </disk> <disk type='file' device='disk'> <driver name='file' type='raw'/> <source file='/path/to/linux_uefi.img'/> <target dev='hda' bus='sata'/> </disk> <interface type='bridge'> <model type='virtio'/> <source bridge="virbr0"/> </interface> <serial type="nmdm"> <source master="/dev/nmdm0A" slave="/dev/nmdm0B"/> </serial> <graphics type='vnc' port='5904'> <listen type='address' address='127.0.0.1'/> </graphics> <controller type='usb' model='nec-xhci'/> <input type='tablet' bus='usb'/> </devices> </domain>
Please refer to the UEFI section for a more detailed explanation.
Guest usage / management ¶
Connecting to a guest console ¶
Guest console connection is supported through the nmdm
device. It could be enabled by adding
the following to the domain XML (Since 1.2.4):
... <devices> <serial type="nmdm"> <source master="/dev/nmdm0A" slave="/dev/nmdm0B"/> </serial> </devices> ...
Make sure to load the nmdm
kernel module if you plan to use that.
Then virsh console
command can be used to connect to the text console
of a guest.
NB: Some versions of bhyve have a bug that prevents guests from booting
until the console is opened by a client. This bug was fixed in FreeBSD
r262884. If
an older version is used, one either has to open a console manually with virsh console
to let a guest boot or start a guest using:
start --console domname
NB: A bootloader configured to require user interaction will prevent
the domain from starting (and thus virsh console
or start
--console
from functioning) until the user interacts with it manually on
the VM host. Because users typically do not have access to the VM host,
interactive bootloaders are unsupported by libvirt. However, if you happen to
run into this scenario and also happen to have access to the Bhyve host
machine, you may select a boot option and allow the domain to finish starting
by using an alternative terminal client on the VM host to connect to the
domain-configured null modem device. One example (assuming
/dev/nmdm0B
is configured as the slave end of the domain serial
device) is:
cu -l /dev/nmdm0B
Converting from domain XML to Bhyve args ¶
The virsh domxml-to-native
command can preview the actual
bhyve
commands that will be executed for a given domain.
It outputs two lines, the first line is a bhyveload
command and
the second is a bhyve
command.
Please note that the virsh domxml-to-native
doesn't do any
real actions other than printing the command, for example, it doesn't try to
find a proper TAP interface and create it, like what is done when starting
a domain; and always returns tap0
for the network interface. So
if you're going to run these commands manually, most likely you might want to
tweak them.
# virsh -c "bhyve:///system" domxml-to-native --format bhyve-argv --xml /path/to/bhyve.xml /usr/sbin/bhyveload -m 214 -d /home/user/vm1.img vm1 /usr/sbin/bhyve -c 2 -m 214 -A -I -H -P -s 0:0,hostbridge -s 3:0,virtio-net,tap0,mac=52:54:00:5d:74:e3 -s 2:0,virtio-blk,/home/user/vm1.img -s 1,lpc -l com1,/dev/nmdm0A vm1
Using ZFS volumes ¶
It's possible to use ZFS volumes as disk devices since 1.2.8. An example of domain XML device entry for that will look like:
... <disk type='volume' device='disk'> <source pool='zfspool' volume='vol1'/> <target dev='vdb' bus='virtio'/> </disk> ...
Please refer to the Storage documentation for more details on storage management.
Using grub2-bhyve or Alternative Bootloaders ¶
It's possible to boot non-FreeBSD guests by specifying an explicit
bootloader, e.g. grub-bhyve(1)
. Arguments to the bootloader may be
specified as well. If the bootloader is grub-bhyve
and arguments
are omitted, libvirt will try and infer boot ordering from user-supplied
<boot order='N'> configuration in the domain. Failing that, it will boot
the first disk in the domain (either cdrom
- or
disk
-type devices). If the disk type is disk
, it will
attempt to boot from the first partition in the disk image.
... <bootloader>/usr/local/sbin/grub-bhyve</bootloader> <bootloader_args>...</bootloader_args> ...
Caveat: bootloader_args
does not support any quoting.
Filenames, etc, must not have spaces or they will be tokenized incorrectly.
Using UEFI bootrom, VNC, and USB tablet ¶
Since 3.2.0, in addition to grub-bhyve,
non-FreeBSD guests could be also booted using an UEFI boot ROM, provided both guest OS and
installed bhyve(1)
version support UEFI. To use that, loader
should be specified in the os
section:
<domain type='bhyve'> ... <os> <type>hvm</type> <loader readonly="yes" type="pflash">/usr/local/share/uefi-firmware/BHYVE_UEFI.fd</loader> </os> ...
This uses the UEFI firmware provided by the sysutils/bhyve-firmware FreeBSD port.
VNC and the tablet input device could be configured this way:
<domain type='bhyve'> <devices> ... <graphics type='vnc' port='5904'> <listen type='address' address='127.0.0.1'/> </graphics> <controller type='usb' model='nec-xhci'/> <input type='tablet' bus='usb'/> </devices> ... </domain>
This way, VNC will be accessible on 127.0.0.1:5904
.
Please note that the tablet device requires to have a USB controller
of the nec-xhci
model. Currently, only a single controller of this
type and a single tablet are supported per domain.
Since 3.5.0, it's possible to configure how the video device is exposed
to the guest using the vgaconf
attribute:
<domain type='bhyve'> <devices> ... <graphics type='vnc' port='5904'> <listen type='address' address='127.0.0.1'/> </graphics> <video> <driver vgaconf='on'/> <model type='gop' heads='1' primary='yes'/> </video> ... </devices> ... </domain>
If not specified, bhyve's default mode for vgaconf
will be used. Please refer to the
bhyve(8)
manual page and the bhyve wiki for more details on using
the vgaconf
option.
Since 3.7.0, it's possible to use autoport
to let libvirt allocate VNC port automatically (instead of explicitly specifying
it with the port
attribute):
<graphics type='vnc' autoport='yes'>
Clock configuration ¶
Originally bhyve supported only localtime for RTC. Support for UTC time was introduced in r284894 for 10-STABLE and in r279225 for -CURRENT. It's possible to use this in libvirt since 1.2.18, just place the following to domain XML:
<domain type="bhyve"> ... <clock offset='utc'/> ... </domain>
Please note that if you run the older bhyve version that doesn't support UTC time, you'll fail to start a domain. As UTC is used as a default when you do not specify clock settings, you'll need to explicitly specify 'localtime' in this case:
<domain type="bhyve"> ... <clock offset='localtime'/> ... </domain>
e1000 NIC ¶
As of r302504 bhyve supports Intel e1000 network adapter emulation. It's supported in libvirt since 3.1.0 and could be used as follows:
... <interface type='bridge'> <source bridge='virbr0'/> <model type='e1000'/> </interface> ...
Wiring guest memory ¶
Since 4.4.0, it's possible to specify that guest memory should be wired and cannot be swapped out as follows:
<domain type="bhyve"> ... <memoryBacking> <locked/> </memoryBacking> ... </domain>
CPU topology ¶
Since 4.5.0, it's possible to specify guest CPU topology, if bhyve supports that. Support for specifying guest CPU topology was added to bhyve in r332298 for -CURRENT. Example:
<domain type="bhyve"> ... <cpu> <topology sockets='1' cores='2' threads='1'/> </cpu> ... </domain>
Pass-through of arbitrary bhyve commands ¶
Since 5.1.0, it's possible to pass additional command-line
arguments to the bhyve process when starting the domain using the
<bhyve:commandline>
element under domain
.
To supply an argument, use the element <bhyve:arg>
with
the attribute value
set to additional argument to be added.
The arg element may be repeated multiple times. To use this XML addition, it is necessary
to issue an XML namespace request (the special xmlns:name
attribute)
that pulls in http://libvirt.org/schemas/domain/bhyve/1.0
;
typically, the namespace is given the name of bhyve
.
Example:
<domain type="bhyve" xmlns:bhyve="http://libvirt.org/schemas/domain/bhyve/1.0"> ... <bhyve:commandline> <bhyve:arg value='-somebhyvearg'/> </bhyve:commandline> </domain>
Note that these extensions are for testing and development purposes only. They are unsupported, using them may result in inconsistent state, and upgrading either bhyve or libvirtd maybe break behavior of a domain that was relying on a specific commands pass-through.