Open vSwitch hardware offloading

The purpose of this page is to describe how to enable Open vSwitch hardware offloading functionality available in OpenStack (using OpenStack Networking). This functionality was first introduced in the OpenStack Pike release. This page intends to serve as a guide for how to configure OpenStack Networking and OpenStack Compute to enable Open vSwitch hardware offloading.

The basics

Open vSwitch is a production quality, multilayer virtual switch licensed under the open source Apache 2.0 license. It is designed to enable massive network automation through programmatic extension, while still supporting standard management interfaces and protocols. Open vSwitch (OVS) allows Virtual Machines (VM) to communicate with each other and with the outside world. The OVS software based solution is CPU intensive, affecting system performance and preventing fully utilizing available bandwidth.

Term

Definition

PF

Physical Function. The physical Ethernet controller that supports SR-IOV.

VF

Virtual Function. The virtual PCIe device created from a physical Ethernet controller.

Representor Port

Virtual network interface similar to SR-IOV port that represents Nova instance.

First Compute Node

OpenStack Compute Node that can host Compute instances (Virtual Machines).

Second Compute Node

OpenStack Compute Node that can host Compute instances (Virtual Machines).

Supported Ethernet controllers

The following manufacturers are known to work:

  • Mellanox ConnectX-4 NIC (VLAN Offload)

  • Mellanox ConnectX-4 Lx/ConnectX-5 NICs (VLAN/VXLAN Offload)

  • Broadcom NetXtreme-S series NICs

  • Broadcom NetXtreme-E series NICs

For information on Mellanox Ethernet Cards, see Mellanox: Ethernet Cards - Overview.

Prerequisites

  • Linux Kernel >= 4.13

  • Open vSwitch >= 2.8

  • iproute >= 4.12

  • Mellanox or Broadcom NIC

    Note

    Mellanox NIC FW that supports Open vSwitch hardware offloading:

    ConnectX-5 >= 16.21.0338

    ConnectX-4 >= 12.18.2000

    ConnectX-4 Lx >= 14.21.0338

Using Open vSwitch hardware offloading

In order to enable Open vSwitch hardware offloading, the following steps are required:

  1. Enable SR-IOV

  2. Configure NIC to switchdev mode (relevant Nodes)

  3. Enable Open vSwitch hardware offloading

Note

Throughout this guide, enp3s0f0 is used as the PF and eth3 is used as the representor port. These ports may vary in different environments.

Note

Throughout this guide, we use systemctl to restart OpenStack services. This is correct for systemd OS. Other methods to restart services should be used in other environments.

Create Compute virtual functions

Create the VFs for the network interface that will be used for SR-IOV. We use enp3s0f0 as PF, which is also used as the interface for the VLAN provider network and has access to the private networks of all nodes.

Note

The following steps detail how to create VFs using Mellanox ConnectX-4 and SR-IOV Ethernet cards on an Intel system. Steps may be different for the hardware of your choice.

  1. Ensure SR-IOV and VT-d are enabled on the system. Enable IOMMU in Linux by adding intel_iommu=on to kernel parameters, for example, using GRUB.

  2. On each Compute node, create the VFs:

    # echo '4' > /sys/class/net/enp3s0f0/device/sriov_numvfs

    Note

    A network interface can be used both for PCI passthrough, using the PF, and SR-IOV, using the VFs. If the PF is used, the VF number stored in the sriov_numvfs file is lost. If the PF is attached again to the operating system, the number of VFs assigned to this interface will be zero. To keep the number of VFs always assigned to this interface, update a relevant file according to your OS. See some examples below:

    In Ubuntu, modifying the /etc/network/interfaces file:

    auto enp3s0f0
iface enp3s0f0 inet dhcp
pre-up echo '4' > /sys/class/net/enp3s0f0/device/sriov_numvfs

    In Red Hat, modifying the /sbin/ifup-local file:

    #!/bin/sh
if [[ "$1" == "enp3s0f0" ]]
then
    echo '4' > /sys/class/net/enp3s0f0/device/sriov_numvfs
fi

    Warning

    Alternatively, you can create VFs by passing the max_vfs to the kernel module of your network interface. However, the max_vfs parameter has been deprecated, so the PCI /sys interface is the preferred method.

    You can determine the maximum number of VFs a PF can support:

    # cat /sys/class/net/enp3s0f0/device/sriov_totalvfs
8

  3. Verify that the VFs have been created and are in up state:

    Note

    The PCI bus number of the PF (03:00.0) and VFs (03:00.2 .. 03:00.5) will be used later.

    # ip link show enp3s0f0
8: enp3s0f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP mode DEFAULT qlen 1000
   link/ether a0:36:9f:8f:3f:b8 brd ff:ff:ff:ff:ff:ff
   vf 0 MAC 00:00:00:00:00:00, spoof checking on, link-state auto
   vf 1 MAC 00:00:00:00:00:00, spoof checking on, link-state auto
   vf 2 MAC 00:00:00:00:00:00, spoof checking on, link-state auto
   vf 3 MAC 00:00:00:00:00:00, spoof checking on, link-state auto

    If the interfaces are down, set them to up before launching a guest, otherwise the instance will fail to spawn:

    # ip link set enp3s0f0 up

Configure Open vSwitch hardware offloading

  1. Change the e-switch mode from legacy to switchdev on the PF device. This will also create the VF representor network devices in the host OS.

    # echo 0000:03:00.2 > /sys/bus/pci/drivers/mlx5_core/unbind

    This tells the driver to unbind VF 03:00.2

    Note

    This should be done for all relevant VFs (in this example 0000:03:00.2 .. 0000:03:00.5)

  2. Enable Open vSwitch hardware offloading, set PF to switchdev mode and bind VFs back.

    # sudo devlink dev eswitch set pci/0000:03:00.0 mode switchdev
# sudo ethtool -K enp3s0f0 hw-tc-offload on
# echo 0000:03:00.2 > /sys/bus/pci/drivers/mlx5_core/bind

    Note

    This should be done for all relevant VFs (in this example 0000:03:00.2 .. 0000:03:00.5)

  3. Restart Open vSwitch

    # sudo systemctl enable openvswitch.service
# sudo ovs-vsctl set Open_vSwitch . other_config:hw-offload=true
# sudo systemctl restart openvswitch.service

    Note

    The given aging of OVS is given in milliseconds and can be controlled with:

    # ovs-vsctl set Open_vSwitch . other_config:max-idle=30000

Configure Nodes (VLAN Configuration)

  1. Update /etc/neutron/plugins/ml2/ml2_conf.ini on Controller nodes

    [ml2]
tenant_network_types = vlan
type_drivers = vlan
mechanism_drivers = openvswitch

  2. Update /etc/neutron/neutron.conf on Controller nodes

    [DEFAULT]
core_plugin = ml2

  3. Update /etc/nova/nova.conf on Controller nodes

    [filter_scheduler]
enabled_filters = PciPassthroughFilter

  4. Update /etc/nova/nova.conf on Compute nodes

    [pci]
#VLAN Configuration passthrough_whitelist example
passthrough_whitelist ={"'"address"'":"'"*:'"03:00"'.*"'","'"physical_network"'":"'"physnet2"'"}

Configure Nodes (VXLAN Configuration)

  1. Update /etc/neutron/plugins/ml2/ml2_conf.ini on Controller nodes

    [ml2]
tenant_network_types = vxlan
type_drivers = vxlan
mechanism_drivers = openvswitch

  2. Update /etc/neutron/neutron.conf on Controller nodes

    [DEFAULT]
core_plugin = ml2

  3. Update /etc/nova/nova.conf on Controller nodes

    [filter_scheduler]
enabled_filters = PciPassthroughFilter

  4. Update /etc/nova/nova.conf on Compute nodes

    Note

    VXLAN configuration requires physical_network to be null.

    [pci]
#VLAN Configuration passthrough_whitelist example
passthrough_whitelist ={"'"address"'":"'"*:'"03:00"'.*"'","'"physical_network"'":null}

  5. Restart nova and neutron services

    # sudo systemctl restart openstack-nova-compute.service
# sudo systemctl restart openstack-nova-scheduler.service
# sudo systemctl restart neutron-server.service

Validate Open vSwitch hardware offloading

Note

In this example we will bring up two instances on different Compute nodes and send ICMP echo packets between them. Then we will check TCP packets on a representor port and we will see that only the first packet will be shown there. All the rest will be offloaded.

  1. Create a port direct on private network

    # openstack port create --network private --vnic-type=direct --binding-profile '{"capabilities": ["switchdev"]}' direct_port1

  2. Create an instance using the direct port on ‘First Compute Node’

    # openstack server create --flavor m1.small --image cloud_image --nic port-id=direct_port1 vm1

  3. Repeat steps above and create a second instance on ‘Second Compute Node’

    # openstack port create --network private --vnic-type=direct --binding-profile '{"capabilities": ["switchdev"]}' direct_port2
# openstack server create --flavor m1.small --image mellanox_fedora --nic port-id=direct_port2 vm2

    Note

    You can use –availability-zone nova:compute_node_1 option to set the desired Compute Node

  4. Connect to instance1 and send ICMP Echo Request packets to instance2

    # vncviewer localhost:5900
vm_1# ping vm2

  5. Connect to ‘Second Compute Node’ and find representor port of the instance

    Note

    Find a representor port first, in our case it’s eth3

    compute_node2# ip link show enp3s0f0
6: enp3s0f0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq master ovs-system state UP mode DEFAULT group default qlen 1000
   link/ether ec:0d:9a:46:9e:84 brd ff:ff:ff:ff:ff:ff
   vf 0 MAC 00:00:00:00:00:00, spoof checking off, link-state enable, trust off, query_rss off
   vf 1 MAC 00:00:00:00:00:00, spoof checking off, link-state enable, trust off, query_rss off
   vf 2 MAC 00:00:00:00:00:00, spoof checking off, link-state enable, trust off, query_rss off
   vf 3 MAC fa:16:3e:b9:b8:ce, vlan 57, spoof checking on, link-state enable, trust off, query_rss off

compute_node2# ls -l /sys/class/net/
lrwxrwxrwx 1 root root 0 Sep 11 10:54 eth0 -> ../../devices/virtual/net/eth0
lrwxrwxrwx 1 root root 0 Sep 11 10:54 eth1 -> ../../devices/virtual/net/eth1
lrwxrwxrwx 1 root root 0 Sep 11 10:54 eth2 -> ../../devices/virtual/net/eth2
lrwxrwxrwx 1 root root 0 Sep 11 10:54 eth3 -> ../../devices/virtual/net/eth3

compute_node2# sudo ovs-dpctl show
system@ovs-system:
  lookups: hit:1684 missed:1465 lost:0
  flows: 0
  masks: hit:8420 total:1 hit/pkt:2.67
  port 0: ovs-system (internal)
  port 1: br-enp3s0f0 (internal)
  port 2: br-int (internal)
  port 3: br-ex (internal)
  port 4: enp3s0f0
  port 5: tapfdc744bb-61 (internal)
  port 6: qr-a7b1e843-4f (internal)
  port 7: qg-79a77e6d-8f (internal)
  port 8: qr-f55e4c5f-f3 (internal)
  port 9: eth3

  6. Check traffic on the representor port. Verify that only the first ICMP packet appears.

    compute_node2# tcpdump -nnn -i eth3

tcpdump: verbose output suppressed, use -v or -vv for full protocol decode
listening on eth3, link-type EN10MB (Ethernet), capture size 262144 bytes
17:12:41.220447 ARP, Request who-has 172.0.0.10 tell 172.0.0.13, length 46
17:12:41.220684 ARP, Reply 172.0.0.10 is-at fa:16:3e:f2:8b:23, length 42
17:12:41.260487 IP 172.0.0.13 > 172.0.0.10: ICMP echo request, id 1263, seq 1, length 64
17:12:41.260778 IP 172.0.0.10 > 172.0.0.13: ICMP echo reply, id 1263, seq 1, length 64
17:12:46.268951 ARP, Request who-has 172.0.0.13 tell 172.0.0.10, length 42
17:12:46.271771 ARP, Reply 172.0.0.13 is-at fa:16:3e:1a:10:05, length 46
17:12:55.354737 IP6 fe80::f816:3eff:fe29:8118 > ff02::1: ICMP6, router advertisement, length 64
17:12:56.106705 IP 0.0.0.0.68 > 255.255.255.255.67: BOOTP/DHCP, Request from 62:21:f0:89:40:73, length 300