vhost.rst revision f7a9461e
1
2..  BSD LICENSE
3    Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
4    All rights reserved.
5
6    Redistribution and use in source and binary forms, with or without
7    modification, are permitted provided that the following conditions
8    are met:
9
10    * Redistributions of source code must retain the above copyright
11    notice, this list of conditions and the following disclaimer.
12    * Redistributions in binary form must reproduce the above copyright
13    notice, this list of conditions and the following disclaimer in
14    the documentation and/or other materials provided with the
15    distribution.
16    * Neither the name of Intel Corporation nor the names of its
17    contributors may be used to endorse or promote products derived
18    from this software without specific prior written permission.
19
20    THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
21    "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
22    LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
23    A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
24    OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
25    SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
26    LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
27    DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
28    THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
29    (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
30    OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
31
32
33Vhost Sample Application
34========================
35
36The vhost sample application demonstrates integration of the Data Plane
37Development Kit (DPDK) with the Linux* KVM hypervisor by implementing the
38vhost-net offload API. The sample application performs simple packet
39switching between virtual machines based on Media Access Control (MAC)
40address or Virtual Local Area Network (VLAN) tag. The splitting of Ethernet
41traffic from an external switch is performed in hardware by the Virtual
42Machine Device Queues (VMDQ) and Data Center Bridging (DCB) features of
43the Intel® 82599 10 Gigabit Ethernet Controller.
44
45Testing steps
46-------------
47
48This section shows the steps how to test a typical PVP case with this
49vhost-switch sample, whereas packets are received from the physical NIC
50port first and enqueued to the VM's Rx queue. Through the guest testpmd's
51default forwarding mode (io forward), those packets will be put into
52the Tx queue. The vhost-switch example, in turn, gets the packets and
53puts back to the same physical NIC port.
54
55Build
56~~~~~
57
58Follow the *Getting Started Guide for Linux* on generic info about
59environment setup and building DPDK from source.
60
61In this example, you need build DPDK both on the host and inside guest.
62Also, you need build this example.
63
64.. code-block:: console
65
66    export RTE_SDK=/path/to/dpdk_source
67    export RTE_TARGET=x86_64-native-linuxapp-gcc
68
69    cd ${RTE_SDK}/examples/vhost
70    make
71
72
73Start the vswitch example
74~~~~~~~~~~~~~~~~~~~~~~~~~
75
76.. code-block:: console
77
78        ./vhost-switch -c f -n 4 --socket-mem 1024  \
79             -- --socket-file /tmp/sock0 --client \
80             ...
81
82Check the `Parameters`_ section for the explanations on what do those
83parameters mean.
84
85.. _vhost_app_run_vm:
86
87Start the VM
88~~~~~~~~~~~~
89
90.. code-block:: console
91
92    qemu-system-x86_64 -machine accel=kvm -cpu host \
93        -m $mem -object memory-backend-file,id=mem,size=$mem,mem-path=/dev/hugepages,share=on \
94                -mem-prealloc -numa node,memdev=mem \
95        \
96        -chardev socket,id=char1,path=/tmp/sock0,server \
97        -netdev type=vhost-user,id=hostnet1,chardev=char1  \
98        -device virtio-net-pci,netdev=hostnet1,id=net1,mac=52:54:00:00:00:14 \
99        ...
100
101.. note::
102    For basic vhost-user support, QEMU 2.2 (or above) is required. For
103    some specific features, a higher version might be need. Such as
104    QEMU 2.7 (or above) for the reconnect feature.
105
106.. _vhost_app_run_dpdk_inside_guest:
107
108Run testpmd inside guest
109~~~~~~~~~~~~~~~~~~~~~~~~
110
111Make sure you have DPDK built inside the guest. Also make sure the
112corresponding virtio-net PCI device is bond to a uio driver, which
113could be done by:
114
115.. code-block:: console
116
117   modprobe uio_pci_generic
118   $RTE_SDK/tools/dpdk-devbind.py -b uio_pci_generic 0000:00:04.0
119
120Then start testpmd for packet forwarding testing.
121
122.. code-block:: console
123
124    ./x86_64-native-gcc/app/testpmd -c 0x3 -- -i
125    > start tx_first
126
127Inject packets
128--------------
129
130While a virtio-net is connected to vhost-switch, a VLAN tag starts with
1311000 is assigned to it. So make sure configure your packet generator
132with the right MAC and VLAN tag, you should be able to see following
133log from the vhost-switch console. It means you get it work::
134
135    VHOST_DATA: (0) mac 52:54:00:00:00:14 and vlan 1000 registered
136
137
138.. _vhost_app_parameters:
139
140Parameters
141----------
142
143**--socket-file path**
144Specifies the vhost-user socket file path.
145
146**--client**
147DPDK vhost-user will act as the client mode when such option is given.
148In the client mode, QEMU will create the socket file. Otherwise, DPDK
149will create it. Put simply, it's the server to create the socket file.
150
151
152**--vm2vm mode**
153The vm2vm parameter sets the mode of packet switching between guests in
154the host.
155
156- 0 disables vm2vm, impling that VM's packets will always go to the NIC port.
157- 1 means a normal mac lookup packet routing.
158- 2 means hardware mode packet forwarding between guests, it allows packets
159  go to the NIC port, hardware L2 switch will determine which guest the
160  packet should forward to or need send to external, which bases on the
161  packet destination MAC address and VLAN tag.
162
163**--mergeable 0|1**
164Set 0/1 to disable/enable the mergeable Rx feature. It's disabled by default.
165
166**--stats interval**
167The stats parameter controls the printing of virtio-net device statistics.
168The parameter specifies an interval (in unit of seconds) to print statistics,
169with an interval of 0 seconds disabling statistics.
170
171**--rx-retry 0|1**
172The rx-retry option enables/disables enqueue retries when the guests Rx queue
173is full. This feature resolves a packet loss that is observed at high data
174rates, by allowing it to delay and retry in the receive path. This option is
175enabled by default.
176
177**--rx-retry-num num**
178The rx-retry-num option specifies the number of retries on an Rx burst, it
179takes effect only when rx retry is enabled.  The default value is 4.
180
181**--rx-retry-delay msec**
182The rx-retry-delay option specifies the timeout (in micro seconds) between
183retries on an RX burst, it takes effect only when rx retry is enabled. The
184default value is 15.
185
186**--dequeue-zero-copy**
187Dequeue zero copy will be enabled when this option is given.
188
189**--vlan-strip 0|1**
190VLAN strip option is removed, because different NICs have different behaviors
191when disabling VLAN strip. Such feature, which heavily depends on hardware,
192should be removed from this example to reduce confusion. Now, VLAN strip is
193enabled and cannot be disabled.
194
195Common Issues
196-------------
197
198* QEMU fails to allocate memory on hugetlbfs, with an error like the
199  following::
200
201      file_ram_alloc: can't mmap RAM pages: Cannot allocate memory
202
203  When running QEMU the above error indicates that it has failed to allocate
204  memory for the Virtual Machine on the hugetlbfs. This is typically due to
205  insufficient hugepages being free to support the allocation request. The
206  number of free hugepages can be checked as follows:
207
208  .. code-block:: console
209
210      cat /sys/kernel/mm/hugepages/hugepages-<pagesize>/nr_hugepages
211
212  The command above indicates how many hugepages are free to support QEMU's
213  allocation request.
214
215* vhost-user will not work with QEMU without the ``-mem-prealloc`` option
216
217  The current implementation works properly only when the guest memory is
218  pre-allocated.
219
220* vhost-user will not work with a QEMU version without shared memory mapping:
221
222  Make sure ``share=on`` QEMU option is given.
223
224* Failed to build DPDK in VM
225
226  Make sure "-cpu host" QEMU option is given.
227