jdenisco | 1511a4e | 2018-10-30 08:46:02 -0400 | [diff] [blame] | 1 | .. _iperf31: |
| 2 | |
| 3 | Using VPP with Iperf3 |
| 4 | ===================== |
| 5 | |
| 6 | First, disable kernel IP forward in *csp2s22c03* to ensure the host cannot use |
| 7 | kernel forwarding (all the settings in *net2s22c05* and *csp2s22c04* remain unchanged): |
| 8 | |
| 9 | .. code-block:: console |
| 10 | |
| 11 | csp2s22c03$ echo 0 | sudo tee /proc/sys/net/ipv4/ip_forward |
| 12 | 0 |
| 13 | csp2s22c03$ sysctl net.ipv4.ip_forward |
| 14 | net.ipv4.ip_forward = 0 |
| 15 | |
| 16 | You can use DPDK’s device binding utility (./install-vpp-native/dpdk/sbin/dpdk-devbind) |
| 17 | to list network devices and bind/unbind them from specific drivers. The flag “-s/--status” |
| 18 | shows the status of devices; the flag “-b/--bind” selects the driver to bind. The |
| 19 | status of devices in our system indicates that the two 40-GbE XL710 devices are located |
| 20 | at 82:00.0 and 82:00.1. Use the device’s slots to bind them to the driver uio_pci_generic: |
| 21 | |
| 22 | .. code-block:: console |
| 23 | |
| 24 | csp2s22c03$ ./install-vpp-native/dpdk/sbin/dpdk-devbind -s |
| 25 | |
| 26 | Network devices using DPDK-compatible driver |
| 27 | ============================================ |
| 28 | <none> |
| 29 | |
| 30 | Network devices using kernel driver |
| 31 | =================================== |
| 32 | 0000:03:00.0 'Ethernet Controller 10-Gigabit X540-AT2' if=enp3s0f0 drv=ixgbe unused=vfio-pci,uio_pci_generic *Active* |
| 33 | 0000:03:00.1 'Ethernet Controller 10-Gigabit X540-AT2' if=enp3s0f1 drv=ixgbe unused=vfio-pci,uio_pci_generic *Active* |
| 34 | 0000:82:00.0 'Ethernet Controller XL710 for 40GbE QSFP+' if=ens802f0d1,ens802f0 drv=i40e unused=uio_pci_generic |
| 35 | 0000:82:00.1 'Ethernet Controller XL710 for 40GbE QSFP+' if=ens802f1d1,ens802f1 drv=i40e unused=uio_pci_generic |
| 36 | |
| 37 | Other network devices |
| 38 | ===================== |
| 39 | <none> |
| 40 | |
| 41 | csp2s22c03$ sudo modprobe uio_pci_generic |
| 42 | csp2s22c03$ sudo ./install-vpp-native/dpdk/sbin/dpdk-devbind --bind uio_pci_generic 82:00.0 |
| 43 | csp2s22c03$ sudo ./install-vpp-native/dpdk/sbin/dpdk-devbind --bind uio_pci_generic 82:00.1 |
| 44 | |
| 45 | csp2s22c03$ sudo ./install-vpp-native/dpdk/sbin/dpdk-devbind -s |
| 46 | |
| 47 | Network devices using DPDK-compatible driver |
| 48 | ============================================ |
| 49 | 0000:82:00.0 'Ethernet Controller XL710 for 40GbE QSFP+' drv=uio_pci_generic unused=i40e,vfio-pci |
| 50 | 0000:82:00.1 'Ethernet Controller XL710 for 40GbE QSFP+' drv=uio_pci_generic unused=i40e,vfio-pci |
| 51 | |
| 52 | Network devices using kernel driver |
| 53 | =================================== |
| 54 | 0000:03:00.0 'Ethernet Controller 10-Gigabit X540-AT2' if=enp3s0f0 drv=ixgbe unused=vfio-pci,uio_pci_generic *Active* |
| 55 | 0000:03:00.1 'Ethernet Controller 10-Gigabit X540-AT2' if=enp3s0f1 drv=ixgbe unused=vfio-pci,uio_pci_generic *Active* |
| 56 | |
| 57 | Start the VPP service, and verify that VPP is running: |
| 58 | |
| 59 | .. code-block:: console |
| 60 | |
| 61 | csp2s22c03$ sudo service vpp start |
| 62 | csp2s22c03$ ps -ef | grep vpp |
| 63 | root 105655 1 98 17:34 ? 00:00:02 /usr/bin/vpp -c /etc/vpp/startup.conf |
| 64 | :w |
| 65 | 105675 105512 0 17:34 pts/4 00:00:00 grep --color=auto vpp |
| 66 | |
| 67 | To access the VPP CLI, issue the command sudo vppctl . From the VPP interface, list |
| 68 | all interfaces that are bound to DPDK using the command show interface: |
| 69 | |
| 70 | VPP shows that the two 40-Gbps ports located at 82:0:0 and 82:0:1 are bound. Next, |
| 71 | you need to assign IP addresses to those interfaces, bring them up, and verify: |
| 72 | |
| 73 | .. code-block:: console |
| 74 | |
| 75 | vpp# set interface ip address FortyGigabitEthernet82/0/0 10.10.1.1/24 |
| 76 | vpp# set interface ip address FortyGigabitEthernet82/0/1 10.10.2.1/24 |
| 77 | vpp# set interface state FortyGigabitEthernet82/0/0 up |
| 78 | vpp# set interface state FortyGigabitEthernet82/0/1 up |
| 79 | vpp# show interface address |
| 80 | FortyGigabitEthernet82/0/0 (up): |
| 81 | 10.10.1.1/24 |
| 82 | FortyGigabitEthernet82/0/1 (up): |
| 83 | 10.10.2.1/24 |
| 84 | local0 (dn): |
| 85 | |
| 86 | At this point VPP is operational. You can ping these interfaces either from *net2s22c05* |
| 87 | or *csp2s22c04*. Moreover, VPP can forward packets whose IP address are 10.10.1.0/24 and |
| 88 | 10.10.2.0/24, so you can ping between *net2s22c05* and *csp2s22c04*. Also, you can |
| 89 | run iperf3 as illustrated in the previous example, and the result from running iperf3 |
| 90 | between *net2s22c05* and *csp2s22c04* increases to 20.3 Gbits per second. |
| 91 | |
| 92 | .. code-block:: console |
| 93 | |
| 94 | ET2S22C05$ iperf3 -c 10.10.1.2 |
| 95 | Connecting to host 10.10.1.2, port 5201 |
| 96 | [ 4] local 10.10.2.2 port 54078 connected to 10.10.1.2 port 5201 |
| 97 | [ ID] Interval Transfer Bandwidth Retr Cwnd |
| 98 | [ 4] 0.00-1.00 sec 2.02 GBytes 17.4 Gbits/sec 460 1.01 MBytes |
| 99 | [ 4] 1.00-2.00 sec 3.28 GBytes 28.2 Gbits/sec 0 1.53 MBytes |
| 100 | [ 4] 2.00-3.00 sec 2.38 GBytes 20.4 Gbits/sec 486 693 KBytes |
| 101 | [ 4] 3.00-4.00 sec 2.06 GBytes 17.7 Gbits/sec 1099 816 KBytes |
| 102 | [ 4] 4.00-5.00 sec 2.07 GBytes 17.8 Gbits/sec 614 1.04 MBytes |
| 103 | [ 4] 5.00-6.00 sec 2.25 GBytes 19.3 Gbits/sec 2869 716 KBytes |
| 104 | [ 4] 6.00-7.00 sec 2.26 GBytes 19.4 Gbits/sec 3321 683 KBytes |
| 105 | [ 4] 7.00-8.00 sec 2.33 GBytes 20.0 Gbits/sec 2322 594 KBytes |
| 106 | [ 4] 8.00-9.00 sec 2.28 GBytes 19.6 Gbits/sec 1690 1.23 MBytes |
| 107 | [ 4] 9.00-10.00 sec 2.73 GBytes 23.5 Gbits/sec 573 680 KBytes |
| 108 | - - - - - - - - - - - - - - - - - - - - - - - - - |
| 109 | [ ID] Interval Transfer Bandwidth Retr |
| 110 | [ 4] 0.00-10.00 sec 23.7 GBytes 20.3 Gbits/sec 13434 sender |
| 111 | [ 4] 0.00-10.00 sec 23.7 GBytes 20.3 Gbits/sec receiver |
| 112 | |
| 113 | iperf Done. |
| 114 | |
| 115 | The **show run** command displays the graph runtime statistics. Observe that the |
| 116 | average vector per node is 6.76, which means on average, a vector of 6.76 packets |
| 117 | is handled in a graph node. |
| 118 | |
| 119 | .. figure:: /_images/build-a-fast-network-stack-terminal.png |