xref: /freebsd/usr.sbin/bhyve/bhyve.8 (revision 4f52dfbb)
1.\" Copyright (c) 2013 Peter Grehan
2.\" All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\"    notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\"    notice, this list of conditions and the following disclaimer in the
11.\"    documentation and/or other materials provided with the distribution.
12.\"
13.\" THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
14.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
16.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
17.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
18.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
19.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
20.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
21.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
22.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
23.\" SUCH DAMAGE.
24.\"
25.\" $FreeBSD$
26.\"
27.Dd May 1, 2018
28.Dt BHYVE 8
29.Os
30.Sh NAME
31.Nm bhyve
32.Nd "run a guest operating system inside a virtual machine"
33.Sh SYNOPSIS
34.Nm
35.Op Fl abehuwxACHPSWY
36.Oo
37.Fl c\~ Ns
38.Oo
39.Op Ar cpus= Ns
40.Ar numcpus Ns
41.Oc Ns
42.Op Ar ,sockets=n Ns
43.Op Ar ,cores=n Ns
44.Op Ar ,threads=n
45.Oc
46.Op Fl g Ar gdbport
47.Op Fl l Ar lpcdev Ns Op , Ns Ar conf
48.Op Fl m Ar memsize Ns Op Ar K|k|M|m|G|g|T|t
49.Op Fl p Ar vcpu:hostcpu
50.Op Fl s Ar slot,emulation Ns Op , Ns Ar conf
51.Op Fl G Ar port
52.Op Fl U Ar uuid
53.Ar vmname
54.Sh DESCRIPTION
55.Nm
56is a hypervisor that runs guest operating systems inside a
57virtual machine.
58.Pp
59Parameters such as the number of virtual CPUs, amount of guest memory, and
60I/O connectivity can be specified with command-line parameters.
61.Pp
62If not using a boot ROM, the guest operating system must be loaded with
63.Xr bhyveload 8
64or a similar boot loader before running
65.Nm ,
66otherwise, it is enough to run
67.Nm
68with a boot ROM of choice.
69.Pp
70.Nm
71runs until the guest operating system reboots or an unhandled hypervisor
72exit is detected.
73.Sh OPTIONS
74.Bl -tag -width 10n
75.It Fl a
76The guest's local APIC is configured in xAPIC mode.
77The xAPIC mode is the default setting so this option is redundant.
78It will be deprecated in a future version.
79.It Fl A
80Generate ACPI tables.
81Required for
82.Fx Ns /amd64
83guests.
84.It Fl b
85Enable a low-level console device supported by
86.Fx
87kernels compiled with
88.Cd "device bvmconsole" .
89This option will be deprecated in a future version.
90.It Fl c Op Ar setting ...
91Number of guest virtual CPUs
92and/or the CPU topology.
93The default value for each of
94.Ar numcpus ,
95.Ar sockets ,
96.Ar cores ,
97and
98.Ar threads
99is 1.
100The current maximum number of guest virtual CPUs is 16.
101If
102.Ar numcpus
103is not specified then it will be calculated from the other arguments.
104The topology must be consistent in that the
105.Ar numcpus
106must equal the product of
107.Ar sockets ,
108.Ar cores ,
109and
110.Ar threads .
111If a
112.Ar setting
113is specified more than once the last one has precedence.
114.It Fl C
115Include guest memory in core file.
116.It Fl e
117Force
118.Nm
119to exit when a guest issues an access to an I/O port that is not emulated.
120This is intended for debug purposes.
121.It Fl g Ar gdbport
122For
123.Fx
124kernels compiled with
125.Cd "device bvmdebug" ,
126allow a remote kernel kgdb to be relayed to the guest kernel gdb stub
127via a local IPv4 address and this port.
128This option will be deprecated in a future version.
129.It Fl G Ar port
130Start a debug server that uses the GDB protocol to export guest state to a
131debugger.
132An IPv4 TCP socket will be bound to the supplied
133.Ar port
134to listen for debugger connections.
135Only a single debugger may be attached to the debug server at a time.
136If
137.Ar port
138begins with
139.Sq w ,
140.Nm
141will pause execution at the first instruction waiting for a debugger to attach.
142.It Fl h
143Print help message and exit.
144.It Fl H
145Yield the virtual CPU thread when a HLT instruction is detected.
146If this option is not specified, virtual CPUs will use 100% of a host CPU.
147.It Fl l Ar lpcdev Ns Op , Ns Ar conf
148Allow devices behind the LPC PCI-ISA bridge to be configured.
149The only supported devices are the TTY-class devices
150.Ar com1
151and
152.Ar com2
153and the boot ROM device
154.Ar bootrom .
155.It Fl m Ar memsize Ns Op Ar K|k|M|m|G|g|T|t
156Guest physical memory size in bytes.
157This must be the same size that was given to
158.Xr bhyveload 8 .
159.Pp
160The size argument may be suffixed with one of K, M, G or T (either upper
161or lower case) to indicate a multiple of kilobytes, megabytes, gigabytes,
162or terabytes.
163If no suffix is given, the value is assumed to be in megabytes.
164.Pp
165.Ar memsize
166defaults to 256M.
167.It Fl p Ar vcpu:hostcpu
168Pin guest's virtual CPU
169.Em vcpu
170to
171.Em hostcpu .
172.It Fl P
173Force the guest virtual CPU to exit when a PAUSE instruction is detected.
174.It Fl s Ar slot,emulation Ns Op , Ns Ar conf
175Configure a virtual PCI slot and function.
176.Pp
177.Nm
178provides PCI bus emulation and virtual devices that can be attached to
179slots on the bus.
180There are 32 available slots, with the option of providing up to 8 functions
181per slot.
182.Bl -tag -width 10n
183.It Ar slot
184.Ar pcislot[:function]
185.Ar bus:pcislot:function
186.Pp
187The
188.Ar pcislot
189value is 0 to 31.
190The optional
191.Ar function
192value is 0 to 7.
193The optional
194.Ar bus
195value is 0 to 255.
196If not specified, the
197.Ar function
198value defaults to 0.
199If not specified, the
200.Ar bus
201value defaults to 0.
202.It Ar emulation
203.Bl -tag -width 10n
204.It Li hostbridge | Li amd_hostbridge
205.Pp
206Provide a simple host bridge.
207This is usually configured at slot 0, and is required by most guest
208operating systems.
209The
210.Li amd_hostbridge
211emulation is identical but uses a PCI vendor ID of
212.Li AMD .
213.It Li passthru
214PCI pass-through device.
215.It Li virtio-net
216Virtio network interface.
217.It Li virtio-blk
218Virtio block storage interface.
219.It Li virtio-rnd
220Virtio RNG interface.
221.It Li virtio-console
222Virtio console interface, which exposes multiple ports
223to the guest in the form of simple char devices for simple IO
224between the guest and host userspaces.
225.It Li ahci
226AHCI controller attached to arbitrary devices.
227.It Li ahci-cd
228AHCI controller attached to an ATAPI CD/DVD.
229.It Li ahci-hd
230AHCI controller attached to a SATA hard-drive.
231.It Li e1000
232Intel e82545 network interface.
233.It Li uart
234PCI 16550 serial device.
235.It Li lpc
236LPC PCI-ISA bridge with COM1 and COM2 16550 serial ports and a boot ROM.
237The LPC bridge emulation can only be configured on bus 0.
238.It Li fbuf
239Raw framebuffer device attached to VNC server.
240.It Li xhci
241eXtensible Host Controller Interface (xHCI) USB controller.
242.El
243.It Op Ar conf
244This optional parameter describes the backend for device emulations.
245If
246.Ar conf
247is not specified, the device emulation has no backend and can be
248considered unconnected.
249.Pp
250Network devices:
251.Bl -tag -width 10n
252.It Ar tapN Ns Op , Ns Ar mac=xx:xx:xx:xx:xx:xx
253.It Ar vmnetN Ns Op , Ns Ar mac=xx:xx:xx:xx:xx:xx
254.Pp
255If
256.Ar mac
257is not specified, the MAC address is derived from a fixed OUI and the
258remaining bytes from an MD5 hash of the slot and function numbers and
259the device name.
260.Pp
261The MAC address is an ASCII string in
262.Xr ethers 5
263format.
264.El
265.Pp
266Block storage devices:
267.Bl -tag -width 10n
268.It Pa /filename Ns Oo , Ns Ar block-device-options Oc
269.It Pa /dev/xxx Ns Oo , Ns Ar block-device-options Oc
270.El
271.Pp
272The
273.Ar block-device-options
274are:
275.Bl -tag -width 8n
276.It Li nocache
277Open the file with
278.Dv O_DIRECT .
279.It Li direct
280Open the file using
281.Dv O_SYNC .
282.It Li ro
283Force the file to be opened read-only.
284.It Li sectorsize= Ns Ar logical Ns Oo / Ns Ar physical Oc
285Specify the logical and physical sector sizes of the emulated disk.
286The physical sector size is optional and is equal to the logical sector size
287if not explicitly specified.
288.El
289.Pp
290TTY devices:
291.Bl -tag -width 10n
292.It Li stdio
293Connect the serial port to the standard input and output of
294the
295.Nm
296process.
297.It Pa /dev/xxx
298Use the host TTY device for serial port I/O.
299.El
300.Pp
301Boot ROM device:
302.Bl -tag -width 10n
303.It Pa romfile
304Map
305.Ar romfile
306in the guest address space reserved for boot firmware.
307.El
308.Pp
309Pass-through devices:
310.Bl -tag -width 10n
311.It Ns Ar slot Ns / Ns Ar bus Ns / Ns Ar function
312Connect to a PCI device on the host at the selector described by
313.Ar slot ,
314.Ar bus ,
315and
316.Ar function
317numbers.
318.El
319.Pp
320Guest memory must be wired using the
321.Fl S
322option when a pass-through device is configured.
323.Pp
324The host device must have been reserved at boot-time using the
325.Va pptdev
326loader variable as described in
327.Xr vmm 4 .
328.Pp
329Virtio console devices:
330.Bl -tag -width 10n
331.It Li port1= Ns Pa /path/to/port1.sock Ns ,anotherport= Ns Pa ...
332A maximum of 16 ports per device can be created.
333Every port is named and corresponds to a Unix domain socket created by
334.Nm .
335.Nm
336accepts at most one connection per port at a time.
337.Pp
338Limitations:
339.Bl -bullet -offset 2n
340.It
341Due to lack of destructors in
342.Nm ,
343sockets on the filesystem must be cleaned up manually after
344.Nm
345exits.
346.It
347There is no way to use the "console port" feature, nor the console port
348resize at present.
349.It
350Emergency write is advertised, but no-op at present.
351.El
352.El
353.Pp
354Framebuffer devices:
355.Bl -tag -width 10n
356.It Oo rfb= Ns Oo Ar IP: Oc Ns Ar port Oc Ns Oo ,w= Ns Ar width Oc Ns Oo ,h= Ns Ar height Oc Ns Oo ,vga= Ns Ar vgaconf Oc Ns Oo Ns ,wait Oc Ns Oo ,password= Ns Ar password Oc
357.Bl -tag -width 8n
358.It Ar IP:port
359An
360.Ar IP
361address and a
362.Ar port
363VNC should listen on.
364The default is to listen on localhost IPv4 address and default VNC port 5900.
365Listening on an IPv6 address is not supported.
366.It Ar width No and Ar height
367A display resolution, width and height, respectively.
368If not specified, a default resolution of 1024x768 pixels will be used.
369Minimal supported resolution is 640x480 pixels,
370and maximum is 1920x1200 pixels.
371.It Ar vgaconf
372Possible values for this option are
373.Dq io
374(default),
375.Dq on
376, and
377.Dq off .
378PCI graphics cards have a dual personality in that they are
379standard PCI devices with BAR addressing, but may also
380implicitly decode legacy VGA I/O space
381.Pq Ad 0x3c0-3df
382and memory space
383.Pq 64KB at Ad 0xA0000 .
384The default
385.Dq io
386option should be used for guests that attempt to issue BIOS
387calls which result in I/O port queries, and fail to boot if I/O decode is disabled.
388.Pp
389The
390.Dq on
391option should be used along with the CSM BIOS capability in UEFI
392to boot traditional BIOS guests that require the legacy VGA I/O and
393memory regions to be available.
394.Pp
395The
396.Dq off
397option should be used for the UEFI guests that assume that
398VGA adapter is present if they detect the I/O ports.
399An example of such a guest is
400.Ox
401in UEFI mode.
402.Pp
403Please refer to the
404.Nm
405.Fx
406wiki page
407.Pq Lk https://wiki.freebsd.org/bhyve
408for configuration notes of particular guests.
409.It wait
410Instruct
411.Nm
412to only boot upon the initiation of a VNC connection, simplifying the installation
413of operating systems that require immediate keyboard input.
414This can be removed for post-installation use.
415.It password
416This type of authentication is known to be cryptographically weak and is not
417intended for use on untrusted networks.
418Many implementations will want to use stronger security, such as running
419the session over an encrypted channel provided by IPsec or SSH.
420.El
421.El
422.Pp
423xHCI USB devices:
424.Bl -tag -width 10n
425.It Li tablet
426A USB tablet device which provides precise cursor synchronization
427when using VNC.
428.El
429.El
430.It Fl S
431Wire guest memory.
432.It Fl u
433RTC keeps UTC time.
434.It Fl U Ar uuid
435Set the universally unique identifier
436.Pq UUID
437in the guest's System Management BIOS System Information structure.
438By default a UUID is generated from the host's hostname and
439.Ar vmname .
440.It Fl w
441Ignore accesses to unimplemented Model Specific Registers (MSRs).
442This is intended for debug purposes.
443.It Fl W
444Force virtio PCI device emulations to use MSI interrupts instead of MSI-X
445interrupts.
446.It Fl x
447The guest's local APIC is configured in x2APIC mode.
448.It Fl Y
449Disable MPtable generation.
450.It Ar vmname
451Alphanumeric name of the guest.
452This should be the same as that created by
453.Xr bhyveload 8 .
454.El
455.Sh DEBUG SERVER
456The current debug server provides limited support for debuggers.
457.Ss Registers
458Each virtual CPU is exposed to the debugger as a thread.
459.Pp
460General purpose registers can be queried for each virtual CPU, but other
461registers such as floating-point and system registers cannot be queried.
462.Ss Memory
463Memory (including memory mapped I/O regions) can be read by the debugger,
464but not written.  Memory operations use virtual addresses that are resolved
465to physical addresses via the current virtual CPU's active address translation.
466.Ss Control
467The running guest can be interrupted by the debugger at any time
468.Pq for example, by pressing Ctrl-C in the debugger .
469.Pp
470Single stepping is only supported on Intel CPUs supporting the MTRAP VM exit.
471.Pp
472Breakpoints are not supported.
473.Sh SIGNAL HANDLING
474.Nm
475deals with the following signals:
476.Pp
477.Bl -tag -width indent -compact
478.It SIGTERM
479Trigger ACPI poweroff for a VM
480.El
481.Sh EXIT STATUS
482Exit status indicates how the VM was terminated:
483.Pp
484.Bl -tag -width indent -compact
485.It 0
486rebooted
487.It 1
488powered off
489.It 2
490halted
491.It 3
492triple fault
493.El
494.Sh EXAMPLES
495If not using a boot ROM, the guest operating system must have been loaded with
496.Xr bhyveload 8
497or a similar boot loader before
498.Xr bhyve 4
499can be run.
500Otherwise, the boot loader is not needed.
501.Pp
502To run a virtual machine with 1GB of memory, two virtual CPUs, a virtio
503block device backed by the
504.Pa /my/image
505filesystem image, and a serial port for the console:
506.Bd -literal -offset indent
507bhyve -c 2 -s 0,hostbridge -s 1,lpc -s 2,virtio-blk,/my/image \\
508  -l com1,stdio -A -H -P -m 1G vm1
509.Ed
510.Pp
511Run a 24GB single-CPU virtual machine with three network ports, one of which
512has a MAC address specified:
513.Bd -literal -offset indent
514bhyve -s 0,hostbridge -s 1,lpc -s 2:0,virtio-net,tap0 \\
515  -s 2:1,virtio-net,tap1 \\
516  -s 2:2,virtio-net,tap2,mac=00:be:fa:76:45:00 \\
517  -s 3,virtio-blk,/my/image -l com1,stdio \\
518  -A -H -P -m 24G bigvm
519.Ed
520.Pp
521Run an 8GB quad-CPU virtual machine with 8 AHCI SATA disks, an AHCI ATAPI
522CD-ROM, a single virtio network port, an AMD hostbridge, and the console
523port connected to an
524.Xr nmdm 4
525null-modem device.
526.Bd -literal -offset indent
527bhyve -c 4 \\
528  -s 0,amd_hostbridge -s 1,lpc \\
529  -s 1:0,ahci,hd:/images/disk.1,hd:/images/disk.2,\\
530hd:/images/disk.3,hd:/images/disk.4,\\
531hd:/images/disk.5,hd:/images/disk.6,\\
532hd:/images/disk.7,hd:/images/disk.8,\\
533cd:/images/install.iso \\
534  -s 3,virtio-net,tap0 \\
535  -l com1,/dev/nmdm0A \\
536  -A -H -P -m 8G
537.Ed
538.Pp
539Run a UEFI virtual machine with a display resolution of 800 by 600 pixels
540that can be accessed via VNC at: 0.0.0.0:5900.
541.Bd -literal -offset indent
542bhyve -c 2 -m 4G -w -H \\
543  -s 0,hostbridge \\
544  -s 3,ahci-cd,/path/to/uefi-OS-install.iso \\
545  -s 4,ahci-hd,disk.img \\
546  -s 5,virtio-net,tap0 \\
547  -s 29,fbuf,tcp=0.0.0.0:5900,w=800,h=600,wait \\
548  -s 30,xhci,tablet \\
549  -s 31,lpc -l com1,stdio \\
550  -l bootrom,/usr/local/share/uefi-firmware/BHYVE_UEFI.fd \\
551   uefivm
552.Ed
553.Sh SEE ALSO
554.Xr bhyve 4 ,
555.Xr nmdm 4 ,
556.Xr vmm 4 ,
557.Xr ethers 5 ,
558.Xr bhyvectl 8 ,
559.Xr bhyveload 8
560.Sh HISTORY
561.Nm
562first appeared in
563.Fx 10.0 .
564.Sh AUTHORS
565.An Neel Natu Aq Mt neel@freebsd.org
566.An Peter Grehan Aq Mt grehan@freebsd.org
567