1 OVS DPDK ADVANCED INSTALL GUIDE
2 =================================
6 1. [Overview](#overview)
7 2. [Building Shared Library](#build)
8 3. [System configuration](#sysconf)
9 4. [Performance Tuning](#perftune)
10 5. [OVS Testcases](#ovstc)
11 6. [Vhost Walkthrough](#vhost)
13 8. [Rate Limiting](#rl)
14 9. [Flow Control](#fc)
17 ## <a name="overview"></a> 1. Overview
19 The Advanced Install Guide explains how to improve OVS performance using
20 DPDK datapath. This guide also provides information on tuning, system configuration,
21 troubleshooting, static code analysis and testcases.
23 ## <a name="build"></a> 2. Building Shared Library
25 DPDK can be built as static or shared library and shall be linked by applications
26 using DPDK datapath. The section lists steps to build shared library and dynamically
27 link DPDK against OVS.
29 Note: Minor performance loss is seen with OVS when using shared DPDK library as
30 compared to static library.
32 Check section [INSTALL DPDK], [INSTALL OVS] of INSTALL.DPDK on download instructions
35 * Configure the DPDK library
37 Set `CONFIG_RTE_BUILD_SHARED_LIB=y` in `config/common_base`
38 to generate shared DPDK library
41 * Build and install DPDK
43 For Default install (without IVSHMEM), set `export DPDK_TARGET=x86_64-native-linuxapp-gcc`
44 For IVSHMEM case, set `export DPDK_TARGET=x86_64-ivshmem-linuxapp-gcc`
47 export DPDK_DIR=/usr/src/dpdk-16.07
48 export DPDK_BUILD=$DPDK_DIR/$DPDK_TARGET
49 make install T=$DPDK_TARGET DESTDIR=install
52 * Build, Install and Setup OVS.
54 Export the DPDK shared library location and setup OVS as listed in
55 section 3.3 of INSTALL.DPDK.
57 `export LD_LIBRARY_PATH=$DPDK_DIR/x86_64-native-linuxapp-gcc/lib`
59 ## <a name="sysconf"></a> 3. System Configuration
61 To achieve optimal OVS performance, the system can be configured and that includes
62 BIOS tweaks, Grub cmdline additions, better understanding of NUMA nodes and
63 apt selection of PCIe slots for NIC placement.
65 ### 3.1 Recommended BIOS settings
68 | Settings | values | comments
69 |---------------------------|-----------|-----------
70 | C3 power state | Disabled | -
71 | C6 power state | Disabled | -
72 | MLC Streamer | Enabled | -
73 | MLC Spacial prefetcher | Enabled | -
74 | DCU Data prefetcher | Enabled | -
76 | CPU power and performance | Performance -
77 | Memory RAS and perf | | -
78 config-> NUMA optimized | Enabled | -
81 ### 3.2 PCIe Slot Selection
83 The fastpath performance also depends on factors like the NIC placement,
84 Channel speeds between PCIe slot and CPU, proximity of PCIe slot to the CPU
85 cores running DPDK application. Listed below are the steps to identify
88 - Retrieve host details using cmd `dmidecode -t baseboard | grep "Product Name"`
89 - Download the technical specification for Product listed eg: S2600WT2.
90 - Check the Product Architecture Overview on the Riser slot placement,
91 CPU sharing info and also PCIe channel speeds.
93 example: On S2600WT, CPU1 and CPU2 share Riser Slot 1 with Channel speed between
94 CPU1 and Riser Slot1 at 32GB/s, CPU2 and Riser Slot1 at 16GB/s. Running DPDK app
95 on CPU1 cores and NIC inserted in to Riser card Slots will optimize OVS performance
98 - Check the Riser Card #1 - Root Port mapping information, on the available slots
99 and individual bus speeds. In S2600WT slot 1, slot 2 has high bus speeds and are
100 potential slots for NIC placement.
102 ### 3.3 Advanced Hugepage setup
104 Allocate and mount 1G Huge pages:
106 - For persistent allocation of huge pages, add the following options to the kernel bootline
108 Add `default_hugepagesz=1GB hugepagesz=1G hugepages=N`
110 For platforms supporting multiple huge page sizes, Add options
112 `default_hugepagesz=<size> hugepagesz=<size> hugepages=N`
113 where 'N' = Number of huge pages requested, 'size' = huge page size,
114 optional suffix [kKmMgG]
116 - For run-time allocation of huge pages
118 `echo N > /sys/devices/system/node/nodeX/hugepages/hugepages-1048576kB/nr_hugepages`
119 where 'N' = Number of huge pages requested, 'X' = NUMA Node
121 Note: For run-time allocation of 1G huge pages, Contiguous Memory Allocator(CONFIG_CMA)
122 has to be supported by kernel, check your Linux distro.
126 `mount -t hugetlbfs -o pagesize=1G none /dev/hugepages`
128 Note: Mount hugepages if not already mounted by default.
130 ### 3.4 Enable Hyperthreading
132 Requires BIOS changes
134 With HT/SMT enabled, A Physical core appears as two logical cores.
135 SMT can be utilized to spawn worker threads on logical cores of the same
136 physical core there by saving additional cores.
138 With DPDK, When pinning pmd threads to logical cores, care must be taken
139 to set the correct bits in the pmd-cpu-mask to ensure that the pmd threads are
140 pinned to SMT siblings.
142 Example System configuration:
143 Dual socket Machine, 2x 10 core processors, HT enabled, 40 logical cores
145 To use two logical cores which share the same physical core for pmd threads,
146 the following command can be used to identify a pair of logical cores.
148 `cat /sys/devices/system/cpu/cpuN/topology/thread_siblings_list`, where N is the
151 In this example, it would show that cores 1 and 21 share the same physical core.
152 The pmd-cpu-mask to enable two pmd threads running on these two logical cores
153 (one physical core) is.
155 `ovs-vsctl set Open_vSwitch . other_config:pmd-cpu-mask=100002`
157 ### 3.5 Isolate cores
159 'isolcpus' option can be used to isolate cores from the linux scheduler.
160 The isolated cores can then be used to dedicatedly run HPC applications/threads.
161 This helps in better application performance due to zero context switching and
162 minimal cache thrashing. To run platform logic on core 0 and isolate cores
163 between 1 and 19 from scheduler, Add `isolcpus=1-19` to GRUB cmdline.
165 Note: It has been verified that core isolation has minimal advantage due to
166 mature Linux scheduler in some circumstances.
168 ### 3.6 NUMA/Cluster on Die
170 Ideally inter NUMA datapaths should be avoided where possible as packets
171 will go across QPI and there may be a slight performance penalty when
172 compared with intra NUMA datapaths. On Intel Xeon Processor E5 v3,
173 Cluster On Die is introduced on models that have 10 cores or more.
174 This makes it possible to logically split a socket into two NUMA regions
175 and again it is preferred where possible to keep critical datapaths
176 within the one cluster.
178 It is good practice to ensure that threads that are in the datapath are
179 pinned to cores in the same NUMA area. e.g. pmd threads and QEMU vCPUs
180 responsible for forwarding. If DPDK is built with
181 CONFIG_RTE_LIBRTE_VHOST_NUMA=y, vHost User ports automatically
182 detect the NUMA socket of the QEMU vCPUs and will be serviced by a PMD
183 from the same node provided a core on this node is enabled in the
184 pmd-cpu-mask. libnuma packages are required for this feature.
186 ### 3.7 Compiler Optimizations
188 The default compiler optimization level is '-O2'. Changing this to
189 more aggressive compiler optimization such as '-O3 -march=native'
190 with gcc(verified on 5.3.1) can produce performance gains though not
191 siginificant. '-march=native' will produce optimized code on local machine
192 and should be used when SW compilation is done on Testbed.
194 ## <a name="perftune"></a> 4. Performance Tuning
198 For superior performance, DPDK pmd threads and Qemu vCPU threads
199 needs to be affinitized accordingly.
201 * PMD thread Affinity
203 A poll mode driver (pmd) thread handles the I/O of all DPDK
204 interfaces assigned to it. A pmd thread shall poll the ports
205 for incoming packets, switch the packets and send to tx port.
206 pmd thread is CPU bound, and needs to be affinitized to isolated
207 cores for optimum performance.
209 By setting a bit in the mask, a pmd thread is created and pinned
210 to the corresponding CPU core. e.g. to run a pmd thread on core 2
212 `ovs-vsctl set Open_vSwitch . other_config:pmd-cpu-mask=4`
214 Note: pmd thread on a NUMA node is only created if there is
215 at least one DPDK interface from that NUMA node added to OVS.
217 * Qemu vCPU thread Affinity
219 A VM performing simple packet forwarding or running complex packet
220 pipelines has to ensure that the vCPU threads performing the work has
221 as much CPU occupancy as possible.
223 Example: On a multicore VM, multiple QEMU vCPU threads shall be spawned.
224 when the DPDK 'testpmd' application that does packet forwarding
225 is invoked, 'taskset' cmd should be used to affinitize the vCPU threads
226 to the dedicated isolated cores on the host system.
228 ### 4.2 Multiple poll mode driver threads
230 With pmd multi-threading support, OVS creates one pmd thread
231 for each NUMA node by default. However, it can be seen that in cases
232 where there are multiple ports/rxq's producing traffic, performance
233 can be improved by creating multiple pmd threads running on separate
234 cores. These pmd threads can then share the workload by each being
235 responsible for different ports/rxq's. Assignment of ports/rxq's to
236 pmd threads is done automatically.
238 A set bit in the mask means a pmd thread is created and pinned
239 to the corresponding CPU core. e.g. to run pmd threads on core 1 and 2
241 `ovs-vsctl set Open_vSwitch . other_config:pmd-cpu-mask=6`
243 For example, when using dpdk and dpdkvhostuser ports in a bi-directional
244 VM loopback as shown below, spreading the workload over 2 or 4 pmd
245 threads shows significant improvements as there will be more total CPU
248 NIC port0 <-> OVS <-> VM <-> OVS <-> NIC port 1
250 ### 4.3 DPDK physical port Rx Queues
252 `ovs-vsctl set Interface <DPDK interface> options:n_rxq=<integer>`
254 The command above sets the number of rx queues for DPDK physical interface.
255 The rx queues are assigned to pmd threads on the same NUMA node in a
258 ### 4.4 Exact Match Cache
260 Each pmd thread contains one EMC. After initial flow setup in the
261 datapath, the EMC contains a single table and provides the lowest level
262 (fastest) switching for DPDK ports. If there is a miss in the EMC then
263 the next level where switching will occur is the datapath classifier.
264 Missing in the EMC and looking up in the datapath classifier incurs a
265 significant performance penalty. If lookup misses occur in the EMC
266 because it is too small to handle the number of flows, its size can
267 be increased. The EMC size can be modified by editing the define
268 EM_FLOW_HASH_SHIFT in lib/dpif-netdev.c.
270 As mentioned above an EMC is per pmd thread. So an alternative way of
271 increasing the aggregate amount of possible flow entries in EMC and
272 avoiding datapath classifier lookups is to have multiple pmd threads
273 running. This can be done as described in section 4.2.
275 ### 4.5 Rx Mergeable buffers
277 Rx Mergeable buffers is a virtio feature that allows chaining of multiple
278 virtio descriptors to handle large packet sizes. As such, large packets
279 are handled by reserving and chaining multiple free descriptors
280 together. Mergeable buffer support is negotiated between the virtio
281 driver and virtio device and is supported by the DPDK vhost library.
282 This behavior is typically supported and enabled by default, however
283 in the case where the user knows that rx mergeable buffers are not needed
284 i.e. jumbo frames are not needed, it can be forced off by adding
285 mrg_rxbuf=off to the QEMU command line options. By not reserving multiple
286 chains of descriptors it will make more individual virtio descriptors
287 available for rx to the guest using dpdkvhost ports and this can improve
290 ## <a name="ovstc"></a> 5. OVS Testcases
291 ### 5.1 PHY-VM-PHY [VHOST LOOPBACK]
293 The section 5.2 in INSTALL.DPDK guide lists steps for PVP loopback testcase
294 and packet forwarding using DPDK testpmd application in the Guest VM.
295 For users wanting to do packet forwarding using kernel stack below are the steps.
298 ifconfig eth1 1.1.1.2/24
299 ifconfig eth2 1.1.2.2/24
300 systemctl stop firewalld.service
301 systemctl stop iptables.service
302 sysctl -w net.ipv4.ip_forward=1
303 sysctl -w net.ipv4.conf.all.rp_filter=0
304 sysctl -w net.ipv4.conf.eth1.rp_filter=0
305 sysctl -w net.ipv4.conf.eth2.rp_filter=0
306 route add -net 1.1.2.0/24 eth2
307 route add -net 1.1.1.0/24 eth1
308 arp -s 1.1.2.99 DE:AD:BE:EF:CA:FE
309 arp -s 1.1.1.99 DE:AD:BE:EF:CA:EE
312 ### 5.2 PHY-VM-PHY [IVSHMEM]
314 The steps (1-5) in 3.3 section of INSTALL.DPDK guide will create & initialize DB,
315 start vswitchd and add dpdk devices to bridge br0.
317 1. Add DPDK ring port to the bridge
320 ovs-vsctl add-port br0 dpdkr0 -- set Interface dpdkr0 type=dpdkr
323 2. Build modified Qemu (Qemu-2.2.1 + ivshmem-qemu-2.2.1.patch)
327 wget http://wiki.qemu.org/download/qemu-2.2.1.tar.bz2
328 tar -jxvf qemu-2.2.1.tar.bz2
329 cd /usr/src/qemu-2.2.1
330 wget https://raw.githubusercontent.com/netgroup-polito/un-orchestrator/master/orchestrator/compute_controller/plugins/kvm-libvirt/patches/ivshmem-qemu-2.2.1.patch
331 patch -p1 < ivshmem-qemu-2.2.1.patch
332 ./configure --target-list=x86_64-softmmu --enable-debug --extra-cflags='-g'
336 3. Generate Qemu commandline
339 mkdir -p /usr/src/cmdline_generator
340 cd /usr/src/cmdline_generator
341 wget https://raw.githubusercontent.com/netgroup-polito/un-orchestrator/master/orchestrator/compute_controller/plugins/kvm-libvirt/cmdline_generator/cmdline_generator.c
342 wget https://raw.githubusercontent.com/netgroup-polito/un-orchestrator/master/orchestrator/compute_controller/plugins/kvm-libvirt/cmdline_generator/Makefile
343 export RTE_SDK=/usr/src/dpdk-16.07
344 export RTE_TARGET=x86_64-ivshmem-linuxapp-gcc
346 ./build/cmdline_generator -m -p dpdkr0 XXX
347 cmdline=`cat OVSMEMPOOL`
353 export VM_NAME=ivshmem-vm
354 export QCOW2_IMAGE=/root/CentOS7_x86_64.qcow2
355 export QEMU_BIN=/usr/src/qemu-2.2.1/x86_64-softmmu/qemu-system-x86_64
357 taskset 0x20 $QEMU_BIN -cpu host -smp 2,cores=2 -hda $QCOW2_IMAGE -m 4096 --enable-kvm -name $VM_NAME -nographic -vnc :2 -pidfile /tmp/vm1.pid $cmdline
360 5. Running sample "dpdk ring" app in VM
363 echo 1024 > /proc/sys/vm/nr_hugepages
364 mount -t hugetlbfs nodev /dev/hugepages (if not already mounted)
366 # Build the DPDK ring application in the VM
367 export RTE_SDK=/root/dpdk-16.07
368 export RTE_TARGET=x86_64-ivshmem-linuxapp-gcc
371 # Run dpdkring application
372 ./build/dpdkr -c 1 -n 4 -- -n 0
373 where "-n 0" refers to ring '0' i.e dpdkr0
376 ## <a name="vhost"></a> 6. Vhost Walkthrough
378 DPDK 16.07 supports two types of vhost:
380 1. vhost-user - enabled default
382 2. vhost-cuse - Legacy, disabled by default
390 - Adding vhost-user ports to Switch
392 Unlike DPDK ring ports, DPDK vhost-user ports can have arbitrary names,
393 except that forward and backward slashes are prohibited in the names.
395 For vhost-user, the name of the port type is `dpdkvhostuser`
398 ovs-vsctl add-port br0 vhost-user-1 -- set Interface vhost-user-1
402 This action creates a socket located at
403 `/usr/local/var/run/openvswitch/vhost-user-1`, which you must provide
404 to your VM on the QEMU command line. More instructions on this can be
405 found in the next section "Adding vhost-user ports to VM"
407 Note: If you wish for the vhost-user sockets to be created in a
408 sub-directory of `/usr/local/var/run/openvswitch`, you may specify
409 this directory in the ovsdb like so:
411 `./utilities/ovs-vsctl --no-wait \
412 set Open_vSwitch . other_config:vhost-sock-dir=subdir`
414 - Adding vhost-user ports to VM
418 Pass the following parameters to QEMU to attach a vhost-user device:
421 -chardev socket,id=char1,path=/usr/local/var/run/openvswitch/vhost-user-1
422 -netdev type=vhost-user,id=mynet1,chardev=char1,vhostforce
423 -device virtio-net-pci,mac=00:00:00:00:00:01,netdev=mynet1
426 where vhost-user-1 is the name of the vhost-user port added
428 Repeat the above parameters for multiple devices, changing the
429 chardev path and id as necessary. Note that a separate and different
430 chardev path needs to be specified for each vhost-user device. For
431 example you have a second vhost-user port named 'vhost-user-2', you
432 append your QEMU command line with an additional set of parameters:
435 -chardev socket,id=char2,path=/usr/local/var/run/openvswitch/vhost-user-2
436 -netdev type=vhost-user,id=mynet2,chardev=char2,vhostforce
437 -device virtio-net-pci,mac=00:00:00:00:00:02,netdev=mynet2
440 2. Configure huge pages.
442 QEMU must allocate the VM's memory on hugetlbfs. vhost-user ports access
443 a virtio-net device's virtual rings and packet buffers mapping the VM's
444 physical memory on hugetlbfs. To enable vhost-user ports to map the VM's
445 memory into their process address space, pass the following parameters
449 -object memory-backend-file,id=mem,size=4096M,mem-path=/dev/hugepages,
450 share=on -numa node,memdev=mem -mem-prealloc
453 3. Enable multiqueue support(OPTIONAL)
455 QEMU needs to be configured to use multiqueue.
456 The $q below is the number of queues.
457 The $v is the number of vectors, which is '$q x 2 + 2'.
460 -chardev socket,id=char2,path=/usr/local/var/run/openvswitch/vhost-user-2
461 -netdev type=vhost-user,id=mynet2,chardev=char2,vhostforce,queues=$q
462 -device virtio-net-pci,mac=00:00:00:00:00:02,netdev=mynet2,mq=on,vectors=$v
465 The vhost-user interface will be automatically reconfigured with required
466 number of rx and tx queues after connection of virtio device.
467 Manual configuration of `n_rxq` is not supported because OVS will work
468 properly only if `n_rxq` will match number of queues configured in QEMU.
470 A least 2 PMDs should be configured for the vswitch when using multiqueue.
471 Using a single PMD will cause traffic to be enqueued to the same vhost
472 queue rather than being distributed among different vhost queues for a
473 vhost-user interface.
475 If traffic destined for a VM configured with multiqueue arrives to the
476 vswitch via a physical DPDK port, then the number of rxqs should also be
477 set to at least 2 for that physical DPDK port. This is required to increase
478 the probability that a different PMD will handle the multiqueue
479 transmission to the guest using a different vhost queue.
481 If one wishes to use multiple queues for an interface in the guest, the
482 driver in the guest operating system must be configured to do so. It is
483 recommended that the number of queues configured be equal to '$q'.
485 For example, this can be done for the Linux kernel virtio-net driver with:
488 ethtool -L <DEV> combined <$q>
490 where `-L`: Changes the numbers of channels of the specified network device
491 and `combined`: Changes the number of multi-purpose channels.
493 - VM Configuration with libvirt
495 * change the user/group, access control policty and restart libvirtd.
497 - In `/etc/libvirt/qemu.conf` add/edit the following lines
504 - Disable SELinux or set to permissive mode
508 - Restart the libvirtd process, For example, on Fedora
510 `systemctl restart libvirtd.service`
514 - Copy the xml configuration from [Guest VM using libvirt] in to workspace.
518 `virsh create demovm.xml`
520 - Connect to the guest console
522 `virsh console demovm`
526 The demovm xml configuration is aimed at achieving out of box performance
529 - The vcpus are pinned to the cores of the CPU socket 0 using vcpupin.
531 - Configure NUMA cell and memory shared using memAccess='shared'.
533 - Disable mrg_rxbuf='off'.
535 Note: For information on libvirt and further tuning refer [libvirt].
543 - Enable vhost-cuse support
545 1. Enable vhost cuse support in DPDK
547 Set `CONFIG_RTE_LIBRTE_VHOST_USER=n` in config/common_linuxapp and follow the
548 steps in 2.2 section of INSTALL.DPDK guide to build DPDK with cuse support.
549 OVS will detect that DPDK has vhost-cuse libraries compiled and in turn will enable
550 support for it in the switch and disable vhost-user support.
552 2. Insert the Cuse module
556 3. Build and insert the `eventfd_link` module
559 cd $DPDK_DIR/lib/librte_vhost/eventfd_link/
561 insmod $DPDK_DIR/lib/librte_vhost/eventfd_link.ko
564 - Adding vhost-cuse ports to Switch
566 Unlike DPDK ring ports, DPDK vhost-cuse ports can have arbitrary names.
567 For vhost-cuse, the name of the port type is `dpdkvhostcuse`
570 ovs-vsctl add-port br0 vhost-cuse-1 -- set Interface vhost-cuse-1
574 When attaching vhost-cuse ports to QEMU, the name provided during the
575 add-port operation must match the ifname parameter on the QEMU cmd line.
577 - Adding vhost-cuse ports to VM
579 vhost-cuse ports use a Linux* character device to communicate with QEMU.
580 By default it is set to `/dev/vhost-net`. It is possible to reuse this
581 standard device for DPDK vhost, which makes setup a little simpler but it
582 is better practice to specify an alternative character device in order to
583 avoid any conflicts if kernel vhost is to be used in parallel.
585 1. This step is only needed if using an alternative character device.
588 ./utilities/ovs-vsctl --no-wait set Open_vSwitch . \
589 other_config:cuse-dev-name=my-vhost-net
592 In the example above, the character device to be used will be
595 2. In case of reusing kernel vhost character device, there would be conflict
596 user should remove it.
598 `rm -rf /dev/vhost-net`
600 3. Configure virtio-net adapters
602 The following parameters must be passed to the QEMU binary, repeat
603 the below parameters for multiple devices.
606 -netdev tap,id=<id>,script=no,downscript=no,ifname=<name>,vhost=on
607 -device virtio-net-pci,netdev=net1,mac=<mac>
610 The DPDK vhost library will negotiate its own features, so they
611 need not be passed in as command line params. Note that as offloads
612 are disabled this is the equivalent of setting
614 `csum=off,gso=off,guest_tso4=off,guest_tso6=off,guest_ecn=off`
616 When using an alternative character device, it must be explicitly
617 passed to QEMU using the `vhostfd` argument
620 -netdev tap,id=<id>,script=no,downscript=no,ifname=<name>,vhost=on,
621 vhostfd=<open_fd> -device virtio-net-pci,netdev=net1,mac=<mac>
624 The open file descriptor must be passed to QEMU running as a child
625 process. This could be done with a simple python script.
629 fd = os.open("/dev/usvhost", os.O_RDWR)
630 subprocess.call("qemu-system-x86_64 .... -netdev tap,id=vhostnet0,\
631 vhost=on,vhostfd=" + fd +"...", shell=True)
634 4. Configure huge pages
636 QEMU must allocate the VM's memory on hugetlbfs. Vhost ports access a
637 virtio-net device's virtual rings and packet buffers mapping the VM's
638 physical memory on hugetlbfs. To enable vhost-ports to map the VM's
639 memory into their process address space, pass the following parameters
642 `-object memory-backend-file,id=mem,size=4096M,mem-path=/dev/hugepages,
643 share=on -numa node,memdev=mem -mem-prealloc`
645 - VM Configuration with QEMU wrapper
647 The QEMU wrapper script automatically detects and calls QEMU with the
648 necessary parameters. It performs the following actions:
650 * Automatically detects the location of the hugetlbfs and inserts this
651 into the command line parameters.
652 * Automatically open file descriptors for each virtio-net device and
653 inserts this into the command line parameters.
654 * Calls QEMU passing both the command line parameters passed to the
655 script itself and those it has auto-detected.
657 Before use, you **must** edit the configuration parameters section of the
658 script to point to the correct emulator location and set additional
659 settings. Of these settings, `emul_path` and `us_vhost_path` **must** be
660 set. All other settings are optional.
662 To use directly from the command line simply pass the wrapper some of the
663 QEMU parameters: it will configure the rest. For example:
666 qemu-wrap.py -cpu host -boot c -hda <disk image> -m 4096 -smp 4
667 --enable-kvm -nographic -vnc none -net none -netdev tap,id=net1,
668 script=no,downscript=no,ifname=if1,vhost=on -device virtio-net-pci,
669 netdev=net1,mac=00:00:00:00:00:01
672 - VM Configuration with libvirt
674 If you are using libvirt, you must enable libvirt to access the character
675 device by adding it to controllers cgroup for libvirtd using the following
678 1. In `/etc/libvirt/qemu.conf` add/edit the following lines:
681 clear_emulator_capabilities = 0
684 cgroup_device_acl = [
685 "/dev/null", "/dev/full", "/dev/zero",
686 "/dev/random", "/dev/urandom",
687 "/dev/ptmx", "/dev/kvm", "/dev/kqemu",
688 "/dev/rtc", "/dev/hpet", "/dev/net/tun",
689 "/dev/<my-vhost-device>",
693 <my-vhost-device> refers to "vhost-net" if using the `/dev/vhost-net`
694 device. If you have specificed a different name in the database
695 using the "other_config:cuse-dev-name" parameter, please specify that
698 2. Disable SELinux or set to permissive mode
700 3. Restart the libvirtd process
701 For example, on Fedora:
703 `systemctl restart libvirtd.service`
705 After successfully editing the configuration, you may launch your
706 vhost-enabled VM. The XML describing the VM can be configured like so
707 within the <qemu:commandline> section:
709 1. Set up shared hugepages:
712 <qemu:arg value='-object'/>
713 <qemu:arg value='memory-backend-file,id=mem,size=4096M,mem-path=/dev/hugepages,share=on'/>
714 <qemu:arg value='-numa'/>
715 <qemu:arg value='node,memdev=mem'/>
716 <qemu:arg value='-mem-prealloc'/>
719 2. Set up your tap devices:
722 <qemu:arg value='-netdev'/>
723 <qemu:arg value='type=tap,id=net1,script=no,downscript=no,ifname=vhost0,vhost=on'/>
724 <qemu:arg value='-device'/>
725 <qemu:arg value='virtio-net-pci,netdev=net1,mac=00:00:00:00:00:01'/>
728 Repeat for as many devices as are desired, modifying the id, ifname
729 and mac as necessary.
731 Again, if you are using an alternative character device (other than
732 `/dev/vhost-net`), please specify the file descriptor like so:
734 `<qemu:arg value='type=tap,id=net3,script=no,downscript=no,ifname=vhost0,vhost=on,vhostfd=<open_fd>'/>`
736 Where <open_fd> refers to the open file descriptor of the character device.
737 Instructions of how to retrieve the file descriptor can be found in the
738 "DPDK vhost VM configuration" section.
739 Alternatively, the process is automated with the qemu-wrap.py script,
740 detailed in the next section.
742 Now you may launch your VM using virt-manager, or like so:
744 `virsh create my_vhost_vm.xml`
746 - VM Configuration with libvirt & QEMU wrapper
748 To use the qemu-wrapper script in conjuntion with libvirt, follow the
749 steps in the previous section before proceeding with the following steps:
751 1. Place `qemu-wrap.py` in libvirtd binary search PATH ($PATH)
752 Ideally in the same directory that the QEMU binary is located.
754 2. Ensure that the script has the same owner/group and file permissions
757 3. Update the VM xml file using "virsh edit VM.xml"
759 Set the VM to use the launch script.
760 Set the emulator path contained in the `<emulator><emulator/>` tags.
761 For example, replace `<emulator>/usr/bin/qemu-kvm<emulator/>` with
762 `<emulator>/usr/bin/qemu-wrap.py<emulator/>`
764 4. Edit the Configuration Parameters section of the script to point to
765 the correct emulator location and set any additional options. If you are
766 using a alternative character device name, please set "us_vhost_path" to the
767 location of that device. The script will automatically detect and insert
768 the correct "vhostfd" value in the QEMU command line arguments.
770 5. Use virt-manager to launch the VM
772 ### 6.3 DPDK backend inside VM
774 Please note that additional configuration is required if you want to run
775 ovs-vswitchd with DPDK backend inside a QEMU virtual machine. Ovs-vswitchd
776 creates separate DPDK TX queues for each CPU core available. This operation
777 fails inside QEMU virtual machine because, by default, VirtIO NIC provided
778 to the guest is configured to support only single TX queue and single RX
779 queue. To change this behavior, you need to turn on 'mq' (multiqueue)
780 property of all virtio-net-pci devices emulated by QEMU and used by DPDK.
781 You may do it manually (by changing QEMU command line) or, if you use
782 Libvirt, by adding the following string:
784 `<driver name='vhost' queues='N'/>`
786 to <interface> sections of all network devices used by DPDK. Parameter 'N'
787 determines how many queues can be used by the guest.This may not work with
788 old versions of QEMU found in some distros and need Qemu version >= 2.2.
790 ## <a name="qos"></a> 7. QOS
792 Here is an example on QOS usage.
793 Assuming you have a vhost-user port transmitting traffic consisting of
794 packets of size 64 bytes, the following command would limit the egress
795 transmission rate of the port to ~1,000,000 packets per second
797 `ovs-vsctl set port vhost-user0 qos=@newqos -- --id=@newqos create qos
798 type=egress-policer other-config:cir=46000000 other-config:cbs=2048`
800 To examine the QoS configuration of the port:
802 `ovs-appctl -t ovs-vswitchd qos/show vhost-user0`
804 To clear the QoS configuration from the port and ovsdb use the following:
806 `ovs-vsctl destroy QoS vhost-user0 -- clear Port vhost-user0 qos`
808 For more details regarding egress-policer parameters please refer to the
811 ## <a name="rl"></a> 8. Rate Limiting
813 Here is an example on Ingress Policing usage.
814 Assuming you have a vhost-user port receiving traffic consisting of
815 packets of size 64 bytes, the following command would limit the reception
816 rate of the port to ~1,000,000 packets per second:
818 `ovs-vsctl set interface vhost-user0 ingress_policing_rate=368000
819 ingress_policing_burst=1000`
821 To examine the ingress policer configuration of the port:
823 `ovs-vsctl list interface vhost-user0`
825 To clear the ingress policer configuration from the port use the following:
827 `ovs-vsctl set interface vhost-user0 ingress_policing_rate=0`
829 For more details regarding ingress-policer see the vswitch.xml.
831 ## <a name="fc"></a> 9. Flow control.
832 Flow control can be enabled only on DPDK physical ports.
833 To enable flow control support at tx side while adding a port, add the
834 'tx-flow-ctrl' option to the 'ovs-vsctl add-port' as in the eg: below.
837 ovs-vsctl add-port br0 dpdk0 -- \
838 set Interface dpdk0 type=dpdk options:tx-flow-ctrl=true
841 Similarly to enable rx flow control,
844 ovs-vsctl add-port br0 dpdk0 -- \
845 set Interface dpdk0 type=dpdk options:rx-flow-ctrl=true
848 And to enable the flow control auto-negotiation,
851 ovs-vsctl add-port br0 dpdk0 -- \
852 set Interface dpdk0 type=dpdk options:flow-ctrl-autoneg=true
855 To turn ON the tx flow control at run time(After the port is being added
856 to OVS), the command-line input will be,
858 `ovs-vsctl set Interface dpdk0 options:tx-flow-ctrl=true`
860 The flow control parameters can be turned off by setting 'false' to the
861 respective parameter. To disable the flow control at tx side,
863 `ovs-vsctl set Interface dpdk0 options:tx-flow-ctrl=false`
865 ## <a name="vsperf"></a> 10. Vsperf
867 Vsperf project goal is to develop vSwitch test framework that can be used to
868 validate the suitability of different vSwitch implementations in a Telco deployment
869 environment. More information can be found in below link.
871 https://wiki.opnfv.org/display/vsperf/VSperf+Home
877 Please report problems to bugs@openvswitch.org.
880 [INSTALL.userspace.md]:INSTALL.userspace.md
881 [INSTALL.md]:INSTALL.md
882 [DPDK Linux GSG]: http://www.dpdk.org/doc/guides/linux_gsg/build_dpdk.html#binding-and-unbinding-network-ports-to-from-the-igb-uioor-vfio-modules
883 [DPDK Docs]: http://dpdk.org/doc
884 [libvirt]: http://libvirt.org/formatdomain.html
885 [Guest VM using libvirt]: INSTALL.DPDK.md#ovstc
886 [INSTALL DPDK]: INSTALL.DPDK.md#build
887 [INSTALL OVS]: INSTALL.DPDK.md#build