xref: /freebsd/share/man/man4/xen.4 (revision 4c8945a06b01a5c8122cdeb402af36bb46a06acc)
1.\" Copyright (c) 2010 Robert N. M. Watson
2.\" All rights reserved.
3.\"
4.\" This software was developed by SRI International and the University of
5.\" Cambridge Computer Laboratory under DARPA/AFRL contract FA8750-10-C-0237
6.\" ("CTSRD"), as part of the DARPA CRASH research program.
7.\"
8.\" Redistribution and use in source and binary forms, with or without
9.\" modification, are permitted provided that the following conditions
10.\" are met:
11.\" 1. Redistributions of source code must retain the above copyright
12.\"    notice, this list of conditions and the following disclaimer.
13.\" 2. Redistributions in binary form must reproduce the above copyright
14.\"    notice, this list of conditions and the following disclaimer in the
15.\"    documentation and/or other materials provided with the distribution.
16.\"
17.\" THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
18.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
21.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27.\" SUCH DAMAGE.
28.\"
29.\" $FreeBSD$
30.\"
31.Dd December 17, 2010
32.Dt XEN 4
33.Os
34.Sh NAME
35.Nm xen
36.Nd Xen Hypervisor Guest (DomU) Support
37.Sh SYNOPSIS
38To compile para-virtualized (PV) Xen guest support into an i386 kernel, place
39the following lines in your kernel configuration file:
40.Bd -ragged -offset indent
41.Cd "options PAE"
42.Cd "options XEN"
43.Cd "nooptions NATIVE"
44.Ed
45.Pp
46To compile hardware-assisted virtualization (HVM) Xen guest support with
47para-virtualized drivers into an amd64 kernel, place the following lines in
48your kernel configuration file:
49.Bd -ragged -offset indent
50.Cd "options XENHVM"
51.Cd "device xenpci"
52.Ed
53.Sh DESCRIPTION
54The Xen Hypervisor allows multiple virtual machines to be run on a single
55computer system.
56When first released, Xen required that i386 kernels be compiled
57"para-virtualized" as the x86 instruction set was not fully virtualizable.
58Primarily, para-virtualization modifies the virtual memory system to use
59hypervisor calls (hypercalls) rather than direct hardware instructions to
60modify the TLB, although para-virtualized device drivers were also required
61to access resources such as virtual network interfaces and disk devices.
62.Pp
63With later instruction set extensions from AMD and Intel to support fully
64virtualizable instructions, unmodified virtual memory systems can also be
65supported; this is referred to as hardware-assisted virtualization (HVM).
66HVM configurations may either rely on transparently emulated hardware
67peripherals, or para-virtualized drivers, which are aware of virtualization,
68and hence able to optimize certain behaviors to improve performance or
69semantics.
70.Pp
71.Fx
72supports a fully para-virtualized (PV) kernel on the i386 architecture using
73.Cd "options XEN"
74and
75.Cd "nooptions NATIVE" ;
76currently, this requires use of a PAE kernel, enabled via
77.Cd "options PAE" .
78.Pp
79.Fx
80supports hardware-assisted virtualization (HVM) on both the i386 and amd64
81kernels; however, PV device drivers with an HVM kernel are only supported on
82the amd64 architecture, and require
83.Cd "options XENHVM"
84and
85.Cd "device xenpci" .
86.Pp
87Para-virtualized device drivers are required in order to support certain
88functionality, such as processing management requests, returning idle
89physical memory pages to the hypervisor, etc.
90.Ss Xen DomU device drivers
91Xen para-virtualized drivers are automatically added to the kernel if a PV
92kernel is compiled using
93.Cd "options XEN" ;
94for HVM environments,
95.Cd "options XENHVM"
96and
97.Cd "device xenpci"
98are required.
99The follow drivers are supported:
100.Bl -hang -offset indent -width blkfront
101.It Nm balloon
102Allow physical memory pages to be returned to the hypervisor as a result of
103manual tuning or automatic policy.
104.It Nm blkback
105Exports local block devices or files to other Xen domains where they can
106then be imported via
107.Nm blkfront .
108.It Nm blkfront
109Import block devices from other Xen domains as local block devices, to be
110used for file systems, swap, etc.
111.It Nm console
112Export the low-level system console via the Xen console service.
113.It Nm control
114Process management operations from Domain 0, including power off, reboot,
115suspend, crash, and halt requests.
116.It Nm evtchn
117Expose Xen events via the
118.Pa /dev/xen/evtchn
119special device.
120.It Nm netback
121Export local network interfaces to other Xen domains where they can be
122imported via
123.Nm netfront .
124.It Nm netfront
125Import network interfaces from other Xen domains as local network interfaces,
126which may be used for IPv4, IPv6, etc.
127.It Nm pcifront
128Allow physical PCI devices to be passed through into a PV domain.
129.It Nm xenpci
130Represents the Xen PCI device, an emulated PCI device that is exposed to
131HVM domains.
132This device allows detection of the Xen hypervisor, and provides interrupt
133and shared memory services required to interact with the hypervisor.
134.El
135.Ss Performance considerations
136In general, PV drivers will perform better than emulated hardware, and are
137the recommended configuration for HVM installations.
138.Pp
139Using a hypervisor introduces a second layer of scheduling that may limit the
140effectiveness of certain
141.Fx
142scheduling optimisations.
143Among these is adaptive locking, which is no longer able to determine whether
144a thread holding a lock is in execution.
145It is recommended that adaptive locking be disabled when using Xen:
146.Bd -unfilled -offset indent
147.Cd "options NO_ADAPTIVE_MUTEXES"
148.Cd "options NO_ADAPTIVE_RWLOCKS"
149.Cd "options NO_ADAPTIVE_SX"
150.Ed
151.Sh SEE ALSO
152.Xr pae 4
153.Sh HISTORY
154Support for
155.Nm
156first appeared in
157.Fx 8.1 .
158.Sh AUTHORS
159.An -nosplit
160.Fx
161support for Xen was first added by
162.An Kip Macy Aq kmacy@FreeBSD.org
163and
164.An Doug Rabson Aq dfr@FreeBSD.org .
165Further refinements were made by
166.An Justin Gibbs Aq gibbs@FreeBSD.org ,
167.An Adrian Chadd Aq adrian@FreeBSD.org ,
168and
169.An Colin Percival Aq cperciva@FreeBSD.org .
170This manual page was written by
171.An Robert Watson Aq rwatson@FreeBSD.org .
172.Sh BUGS
173.Fx
174is only able to run as a Xen guest (DomU) and not as a Xen host (Dom0).
175.Pp
176A fully para-virtualized (PV) kernel is only supported on i386, and not
177amd64.
178.Pp
179Para-virtualized drivers under hardware-assisted virtualization (HVM) kernel
180are only supported on amd64, not i386.
181.Pp
182As of this release, Xen PV DomU support is not heavily tested; instability
183has been reported during VM migration of PV kernels.
184.Pp
185Certain PV driver features, such as the balloon driver, are under-exercised.
186