vhost_lib.rst revision 6b3e017e
1..  BSD LICENSE
2    Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
3    All rights reserved.
4
5    Redistribution and use in source and binary forms, with or without
6    modification, are permitted provided that the following conditions
7    are met:
8
9    * Redistributions of source code must retain the above copyright
10    notice, this list of conditions and the following disclaimer.
11    * Redistributions in binary form must reproduce the above copyright
12    notice, this list of conditions and the following disclaimer in
13    the documentation and/or other materials provided with the
14    distribution.
15    * Neither the name of Intel Corporation nor the names of its
16    contributors may be used to endorse or promote products derived
17    from this software without specific prior written permission.
18
19    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20    "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21    LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22    A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23    OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24    SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25    LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26    DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27    THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28    (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29    OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
30
31Vhost Library
32=============
33
34The vhost library implements a user space virtio net server allowing the user
35to manipulate the virtio ring directly. In another words, it allows the user
36to fetch/put packets from/to the VM virtio net device. To achieve this, a
37vhost library should be able to:
38
39* Access the guest memory:
40
41  For QEMU, this is done by using the ``-object memory-backend-file,share=on,...``
42  option. Which means QEMU will create a file to serve as the guest RAM.
43  The ``share=on`` option allows another process to map that file, which
44  means it can access the guest RAM.
45
46* Know all the necessary information about the vring:
47
48  Information such as where the available ring is stored. Vhost defines some
49  messages (passed through a Unix domain socket file) to tell the backend all
50  the information it needs to know how to manipulate the vring.
51
52
53Vhost API Overview
54------------------
55
56The following is an overview of the Vhost API functions:
57
58* ``rte_vhost_driver_register(path, flags)``
59
60  This function registers a vhost driver into the system. ``path`` specifies
61  the Unix domain socket file path.
62
63  Currently supported flags are:
64
65  - ``RTE_VHOST_USER_CLIENT``
66
67    DPDK vhost-user will act as the client when this flag is given. See below
68    for an explanation.
69
70  - ``RTE_VHOST_USER_NO_RECONNECT``
71
72    When DPDK vhost-user acts as the client it will keep trying to reconnect
73    to the server (QEMU) until it succeeds. This is useful in two cases:
74
75    * When QEMU is not started yet.
76    * When QEMU restarts (for example due to a guest OS reboot).
77
78    This reconnect option is enabled by default. However, it can be turned off
79    by setting this flag.
80
81  - ``RTE_VHOST_USER_DEQUEUE_ZERO_COPY``
82
83    Dequeue zero copy will be enabled when this flag is set. It is disabled by
84    default.
85
86    There are some truths (including limitations) you might want to know while
87    setting this flag:
88
89    * zero copy is not good for small packets (typically for packet size below
90      512).
91
92    * zero copy is really good for VM2VM case. For iperf between two VMs, the
93      boost could be above 70% (when TSO is enableld).
94
95    * for VM2NIC case, the ``nb_tx_desc`` has to be small enough: <= 64 if virtio
96      indirect feature is not enabled and <= 128 if it is enabled.
97
98      The is because when dequeue zero copy is enabled, guest Tx used vring will
99      be updated only when corresponding mbuf is freed. Thus, the nb_tx_desc
100      has to be small enough so that the PMD driver will run out of available
101      Tx descriptors and free mbufs timely. Otherwise, guest Tx vring would be
102      starved.
103
104    * Guest memory should be backended with huge pages to achieve better
105      performance. Using 1G page size is the best.
106
107      When dequeue zero copy is enabled, the guest phys address and host phys
108      address mapping has to be established. Using non-huge pages means far
109      more page segments. To make it simple, DPDK vhost does a linear search
110      of those segments, thus the fewer the segments, the quicker we will get
111      the mapping. NOTE: we may speed it by using tree searching in future.
112
113* ``rte_vhost_driver_session_start()``
114
115  This function starts the vhost session loop to handle vhost messages. It
116  starts an infinite loop, therefore it should be called in a dedicated
117  thread.
118
119* ``rte_vhost_driver_callback_register(virtio_net_device_ops)``
120
121  This function registers a set of callbacks, to let DPDK applications take
122  the appropriate action when some events happen. The following events are
123  currently supported:
124
125  * ``new_device(int vid)``
126
127    This callback is invoked when a virtio net device becomes ready. ``vid``
128    is the virtio net device ID.
129
130  * ``destroy_device(int vid)``
131
132    This callback is invoked when a virtio net device shuts down (or when the
133    vhost connection is broken).
134
135  * ``vring_state_changed(int vid, uint16_t queue_id, int enable)``
136
137    This callback is invoked when a specific queue's state is changed, for
138    example to enabled or disabled.
139
140* ``rte_vhost_enqueue_burst(vid, queue_id, pkts, count)``
141
142  Transmits (enqueues) ``count`` packets from host to guest.
143
144* ``rte_vhost_dequeue_burst(vid, queue_id, mbuf_pool, pkts, count)``
145
146  Receives (dequeues) ``count`` packets from guest, and stored them at ``pkts``.
147
148* ``rte_vhost_feature_disable/rte_vhost_feature_enable(feature_mask)``
149
150  This function disables/enables some features. For example, it can be used to
151  disable mergeable buffers and TSO features, which both are enabled by
152  default.
153
154
155Vhost-user Implementations
156--------------------------
157
158Vhost-user uses Unix domain sockets for passing messages. This means the DPDK
159vhost-user implementation has two options:
160
161* DPDK vhost-user acts as the server.
162
163  DPDK will create a Unix domain socket server file and listen for
164  connections from the frontend.
165
166  Note, this is the default mode, and the only mode before DPDK v16.07.
167
168
169* DPDK vhost-user acts as the client.
170
171  Unlike the server mode, this mode doesn't create the socket file;
172  it just tries to connect to the server (which responses to create the
173  file instead).
174
175  When the DPDK vhost-user application restarts, DPDK vhost-user will try to
176  connect to the server again. This is how the "reconnect" feature works.
177
178  .. Note::
179     * The "reconnect" feature requires **QEMU v2.7** (or above).
180
181     * The vhost supported features must be exactly the same before and
182       after the restart. For example, if TSO is disabled and then enabled,
183       nothing will work and issues undefined might happen.
184
185No matter which mode is used, once a connection is established, DPDK
186vhost-user will start receiving and processing vhost messages from QEMU.
187
188For messages with a file descriptor, the file descriptor can be used directly
189in the vhost process as it is already installed by the Unix domain socket.
190
191The supported vhost messages are:
192
193* ``VHOST_SET_MEM_TABLE``
194* ``VHOST_SET_VRING_KICK``
195* ``VHOST_SET_VRING_CALL``
196* ``VHOST_SET_LOG_FD``
197* ``VHOST_SET_VRING_ERR``
198
199For ``VHOST_SET_MEM_TABLE`` message, QEMU will send information for each
200memory region and its file descriptor in the ancillary data of the message.
201The file descriptor is used to map that region.
202
203``VHOST_SET_VRING_KICK`` is used as the signal to put the vhost device into
204the data plane, and ``VHOST_GET_VRING_BASE`` is used as the signal to remove
205the vhost device from the data plane.
206
207When the socket connection is closed, vhost will destroy the device.
208
209Vhost supported vSwitch reference
210---------------------------------
211
212For more vhost details and how to support vhost in vSwitch, please refer to
213the vhost example in the DPDK Sample Applications Guide.
214