50. ThunderX NICVF Poll Mode Driver

The ThunderX NICVF PMD (librte_net_thunderx) provides poll mode driver support for the inbuilt NIC found in the Cavium ThunderX SoC family as well as their virtual functions (VF) in SR-IOV context.

More information can be found at Cavium, Inc Official Website.

50.1. Features

Features of the ThunderX PMD are:

  • Multiple queues for TX and RX
  • Receive Side Scaling (RSS)
  • Packet type information
  • Checksum offload
  • Promiscuous mode
  • Multicast mode
  • Port hardware statistics
  • Jumbo frames
  • Link state information
  • Setting up link state.
  • Scattered and gather for TX and RX
  • VLAN stripping
  • NUMA support
  • Multi queue set support (up to 96 queues (12 queue sets)) per port
  • Skip data bytes

50.2. Supported ThunderX SoCs

  • CN88xx
  • CN81xx
  • CN83xx

50.3. Prerequisites

50.4. Driver compilation and testing

Refer to the document compiling and testing a PMD for a NIC for details.

Use config/arm/arm64-thunderx-linux-gcc as a meson cross-file when cross-compiling.

50.5. Linux

50.5.1. SR-IOV: Prerequisites and sample Application Notes

Current ThunderX NIC PF/VF kernel modules maps each physical Ethernet port automatically to virtual function (VF) and presented them as PCIe-like SR-IOV device. This section provides instructions to configure SR-IOV with Linux OS.

  1. Verify PF devices capabilities using lspci:

    lspci -vvv

    Example output:

    0002:01:00.0 Ethernet controller: Cavium Networks Device a01e (rev 01)
    Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)
    Capabilities: [180 v1] Single Root I/O Virtualization (SR-IOV)
    Kernel driver in use: thunder-nic


    Unless thunder-nic driver is in use make sure your kernel config includes CONFIG_THUNDER_NIC_PF setting.

  2. Verify VF devices capabilities and drivers using lspci:

    lspci -vvv

    Example output:

    0002:01:00.1 Ethernet controller: Cavium Networks Device 0011 (rev 01)
    Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)
    Kernel driver in use: thunder-nicvf
    0002:01:00.2 Ethernet controller: Cavium Networks Device 0011 (rev 01)
    Capabilities: [100 v1] Alternative Routing-ID Interpretation (ARI)
    Kernel driver in use: thunder-nicvf


    Unless thunder-nicvf driver is in use make sure your kernel config includes CONFIG_THUNDER_NIC_VF setting.

  3. Pass VF device to VM context (PCIe Passthrough):

    The VF devices may be passed through to the guest VM using qemu or virt-manager or virsh etc.

    Example qemu guest launch command:

    sudo qemu-system-aarch64 -name vm1 \
    -machine virt,gic_version=3,accel=kvm,usb=off \
    -cpu host -m 4096 \
    -smp 4,sockets=1,cores=8,threads=1 \
    -nographic -nodefaults \
    -kernel <kernel image> \
    -append "root=/dev/vda console=ttyAMA0 rw hugepagesz=512M hugepages=3" \
    -device vfio-pci,host=0002:01:00.1 \
    -drive file=<rootfs.ext3>,if=none,id=disk1,format=raw  \
    -device virtio-blk-device,scsi=off,drive=disk1,id=virtio-disk1,bootindex=1 \
    -netdev tap,id=net0,ifname=tap0,script=/etc/qemu-ifup_thunder \
    -device virtio-net-device,netdev=net0 \
    -serial stdio \
    -mem-path /dev/hugepages
  4. Enable VFIO-NOIOMMU mode (optional):

    echo 1 > /sys/module/vfio/parameters/enable_unsafe_noiommu_mode


    VFIO-NOIOMMU is required only when running in VM context and should not be enabled otherwise.

  5. Running testpmd:

    Follow instructions available in the document compiling and testing a PMD for a NIC to run testpmd.

    Example output:

    ./<build_dir>/app/dpdk-testpmd -l 0-3 -n 4 -a 0002:01:00.2 \
      -- -i --no-flush-rx \
    PMD: rte_nicvf_pmd_init(): librte_net_thunderx nicvf version 1.0
    EAL:   probe driver: 177d:11 rte_nicvf_pmd
    EAL:   using IOMMU type 1 (Type 1)
    EAL:   PCI memory mapped at 0x3ffade50000
    EAL: Trying to map BAR 4 that contains the MSI-X table.
         Trying offsets: 0x40000000000:0x0000, 0x10000:0x1f0000
    EAL:   PCI memory mapped at 0x3ffadc60000
    PMD: nicvf_eth_dev_init(): nicvf: device (177d:11) 2:1:0:2
    PMD: nicvf_eth_dev_init(): node=0 vf=1 mode=tns-bypass sqs=false
    PMD: nicvf_eth_dev_init(): Port 0 (177d:11) mac=a6:c6:d9:17:78:01
    Interactive-mode selected
    Configuring Port 0 (socket 0)
    PMD: nicvf_dev_configure(): Configured ethdev port0 hwcap=0x0
    Port 0: A6:C6:D9:17:78:01
    Checking link statuses...
    Port 0 Link Up - speed 10000 Mbps - full-duplex

50.5.2. Multiple Queue Set per DPDK port configuration

There are two types of VFs:

  • Primary VF
  • Secondary VF

Each port consists of a primary VF and n secondary VF(s). Each VF provides 8 Tx/Rx queues to a port. When a given port is configured to use more than 8 queues, it requires one (or more) secondary VF. Each secondary VF adds 8 additional queues to the queue set.

During PMD driver initialization, the primary VF’s are enumerated by checking the specific flag (see sqs message in DPDK boot log - sqs indicates secondary queue set). They are at the beginning of VF list (the remain ones are secondary VF’s).

The primary VFs are used as master queue sets. Secondary VFs provide additional queue sets for primary ones. If a port is configured for more then 8 queues than it will request for additional queues from secondary VFs.

Secondary VFs cannot be shared between primary VFs.

Primary VFs are present on the beginning of the ‘Network devices using kernel driver’ list, secondary VFs are on the remaining on the remaining part of the list.


The VNIC driver in the multiqueue setup works differently than other drivers like ixgbe. We need to bind separately each specific queue set device with the usertools/dpdk-devbind.py utility.


Depending on the hardware used, the kernel driver sets a threshold vf_id. VFs that try to attached with an id below or equal to this boundary are considered primary VFs. VFs that try to attach with an id above this boundary are considered secondary VFs.

50.5.3. LBK HW Access

Loopback HW Unit (LBK) receives packets from NIC-RX and sends packets back to NIC-TX. The loopback block has N channels and contains data buffering that is shared across all channels. Four primary VFs are reserved as loopback ports.

50.5.4. Example device binding

If a system has three interfaces, a total of 18 VF devices will be created on a non-NUMA machine.


NUMA systems have 12 VFs per port and non-NUMA 6 VFs per port.

# usertools/dpdk-devbind.py --status

Network devices using DPDK-compatible driver

Network devices using kernel driver
0000:01:10.0 'THUNDERX BGX (Common Ethernet Interface) a026' if= drv=thunder-BGX unused=vfio-pci
0000:01:10.1 'THUNDERX BGX (Common Ethernet Interface) a026' if= drv=thunder-BGX unused=vfio-pci
0001:01:00.0 'THUNDERX Network Interface Controller a01e' if= drv=thunder-nic unused=vfio-pci
0001:01:00.1 'Device a034' if=eth0 drv=thunder-nicvf unused=vfio-pci
0001:01:00.2 'Device a034' if=eth1 drv=thunder-nicvf unused=vfio-pci
0001:01:00.3 'Device a034' if=eth2 drv=thunder-nicvf unused=vfio-pci
0001:01:00.4 'Device a034' if=eth3 drv=thunder-nicvf unused=vfio-pci
0001:01:00.5 'Device a034' if=eth4 drv=thunder-nicvf unused=vfio-pci
0001:01:00.6 'Device a034' if=lbk0 drv=thunder-nicvf unused=vfio-pci
0001:01:00.7 'Device a034' if=lbk1 drv=thunder-nicvf unused=vfio-pci
0001:01:01.0 'Device a034' if=lbk2 drv=thunder-nicvf unused=vfio-pci
0001:01:01.1 'Device a034' if=lbk3 drv=thunder-nicvf unused=vfio-pci
0001:01:01.2 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:01.3 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:01.4 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:01.5 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:01.6 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:01.7 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:02.0 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:02.1 'Device a034' if= drv=thunder-nicvf unused=vfio-pci
0001:01:02.2 'Device a034' if= drv=thunder-nicvf unused=vfio-pci

Other network devices
0002:00:03.0 'Device a01f' unused=vfio-pci,uio_pci_generic


Here total no of primary VFs = 5 (variable, depends on no of ethernet ports present) + 4 (fixed, loopback ports). Ethernet ports are indicated as if=eth0 while loopback ports as if=lbk0.

We want to bind two physical interfaces with 24 queues each device, we attach two primary VFs and four secondary VFs. In our example we choose two 10G interfaces eth1 (0002:01:00.2) and eth2 (0002:01:00.3). We will choose four secondary queue sets from the ending of the list (0001:01:01.2-0002:01:02.2).

  1. Bind two primary VFs to the vfio-pci driver:

    usertools/dpdk-devbind.py -b vfio-pci 0002:01:00.2
    usertools/dpdk-devbind.py -b vfio-pci 0002:01:00.3
  2. Bind four primary VFs to the vfio-pci driver:

    usertools/dpdk-devbind.py -b vfio-pci 0002:01:01.7
    usertools/dpdk-devbind.py -b vfio-pci 0002:01:02.0
    usertools/dpdk-devbind.py -b vfio-pci 0002:01:02.1
    usertools/dpdk-devbind.py -b vfio-pci 0002:01:02.2

The nicvf thunderx driver will make use of attached secondary VFs automatically during the interface configuration stage.

50.5.5. Thunder-nic VF’s

Use sysfs to distinguish thunder-nic primary VFs and secondary VFs.
ls -l /sys/bus/pci/drivers/thunder-nic/
total 0
drwxr-xr-x  2 root root     0 Jan 22 11:19 ./
drwxr-xr-x 86 root root     0 Jan 22 11:07 ../
lrwxrwxrwx  1 root root     0 Jan 22 11:19 0001:01:00.0 -> '../../../../devices/platform/soc@0/849000000000.pci/pci0001:00/0001:00:10.0/0001:01:00.0'/
cat /sys/bus/pci/drivers/thunder-nic/0001\:01\:00.0/sriov_sqs_assignment
0 0001:01:00.1 vfio-pci +: 12 13
1 0001:01:00.2 thunder-nicvf -:
2 0001:01:00.3 thunder-nicvf -:
3 0001:01:00.4 thunder-nicvf -:
4 0001:01:00.5 thunder-nicvf -:
5 0001:01:00.6 thunder-nicvf -:
6 0001:01:00.7 thunder-nicvf -:
7 0001:01:01.0 thunder-nicvf -:
8 0001:01:01.1 thunder-nicvf -:
9 0001:01:01.2 thunder-nicvf -:
10 0001:01:01.3 thunder-nicvf -:
11 0001:01:01.4 thunder-nicvf -:
12 0001:01:01.5 vfio-pci: 0
13 0001:01:01.6 vfio-pci: 0
14 0001:01:01.7 thunder-nicvf: 255
15 0001:01:02.0 thunder-nicvf: 255
16 0001:01:02.1 thunder-nicvf: 255
17 0001:01:02.2 thunder-nicvf: 255
18 0001:01:02.3 thunder-nicvf: 255
19 0001:01:02.4 thunder-nicvf: 255
20 0001:01:02.5 thunder-nicvf: 255
21 0001:01:02.6 thunder-nicvf: 255
22 0001:01:02.7 thunder-nicvf: 255
23 0001:01:03.0 thunder-nicvf: 255
24 0001:01:03.1 thunder-nicvf: 255
25 0001:01:03.2 thunder-nicvf: 255
26 0001:01:03.3 thunder-nicvf: 255
27 0001:01:03.4 thunder-nicvf: 255
28 0001:01:03.5 thunder-nicvf: 255
29 0001:01:03.6 thunder-nicvf: 255
30 0001:01:03.7 thunder-nicvf: 255
31 0001:01:04.0 thunder-nicvf: 255

Every column that ends with ‘thunder-nicvf: number’ can be used as secondary VF. In printout above all entres after ‘14 0001:01:01.7 thunder-nicvf: 255’ can be used as secondary VF.

50.6. Debugging Options

EAL command option to change log level

50.7. Module params

50.7.1. skip_data_bytes

This feature is used to create a hole between HEADROOM and actual data. Size of hole is specified in bytes as module param(“skip_data_bytes”) to pmd. This scheme is useful when application would like to insert vlan header without disturbing HEADROOM.

-a 0002:01:00.2,skip_data_bytes=8

50.8. Limitations

50.8.1. CRC stripping

The ThunderX SoC family NICs strip the CRC for every packets coming into the host interface irrespective of the offload configuration.

50.8.2. Maximum packet length

The ThunderX SoC family NICs support a maximum of a 9K jumbo frame. The value is fixed and cannot be changed. So, even when the rxmode.max_rx_pkt_len member of struct rte_eth_conf is set to a value lower than 9200, frames up to 9200 bytes can still reach the host interface.

50.8.3. Maximum packet segments

The ThunderX SoC family NICs support up to 12 segments per packet when working in scatter/gather mode. So, setting MTU will result with EINVAL when the frame size does not fit in the maximum number of segments.

50.8.4. skip_data_bytes

Maximum limit of skip_data_bytes is 128 bytes and number of bytes should be multiple of 8.