9. IXGBE Driver

9.1. Vector PMD for IXGBE

Vector PMD uses Intel® SIMD instructions to optimize packet I/O. It improves load/store bandwidth efficiency of L1 data cache by using a wider SSE/AVX register 1 (1). The wider register gives space to hold multiple packet buffers so as to save instruction number when processing bulk of packets.

There is no change to PMD API. The RX/TX handler are the only two entries for vPMD packet I/O. They are transparently registered at runtime RX/TX execution if all condition checks pass.

  1. To date, only an SSE version of IX GBE vPMD is available. To ensure that vPMD is in the binary code, ensure that the option CONFIG_RTE_IXGBE_INC_VECTOR=y is in the configure file.

Some constraints apply as pre-conditions for specific optimizations on bulk packet transfers. The following sections explain RX and TX constraints in the vPMD.

9.1.1. RX Constraints

9.1.1.1. Prerequisites and Pre-conditions

The following prerequisites apply:

  • To enable vPMD to work for RX, bulk allocation for Rx must be allowed.

Ensure that the following pre-conditions are satisfied:

  • rxq->rx_free_thresh >= RTE_PMD_IXGBE_RX_MAX_BURST
  • rxq->rx_free_thresh < rxq->nb_rx_desc
  • (rxq->nb_rx_desc % rxq->rx_free_thresh) == 0
  • rxq->nb_rx_desc < (IXGBE_MAX_RING_DESC - RTE_PMD_IXGBE_RX_MAX_BURST)

These conditions are checked in the code.

Scattered packets are not supported in this mode. If an incoming packet is greater than the maximum acceptable length of one “mbuf” data size (by default, the size is 2 KB), vPMD for RX would be disabled.

By default, IXGBE_MAX_RING_DESC is set to 4096 and RTE_PMD_IXGBE_RX_MAX_BURST is set to 32.

9.1.1.2. Feature not Supported by RX Vector PMD

Some features are not supported when trying to increase the throughput in vPMD. They are:

  • IEEE1588
  • FDIR
  • Header split
  • RX checksum off load

Other features are supported using optional MACRO configuration. They include:

  • HW VLAN strip
  • HW extend dual VLAN
  • Enabled by RX_OLFLAGS (RTE_IXGBE_RX_OLFLAGS_ENABLE=y)

To guarantee the constraint, configuration flags in dev_conf.rxmode will be checked:

  • hw_vlan_strip
  • hw_vlan_extend
  • hw_ip_checksum
  • header_split
  • dev_conf

fdir_conf->mode will also be checked.

9.1.1.3. RX Burst Size

As vPMD is focused on high throughput, it assumes that the RX burst size is equal to or greater than 32 per burst. It returns zero if using nb_pkt < 32 as the expected packet number in the receive handler.

9.1.2. TX Constraint

9.1.2.1. Prerequisite

The only prerequisite is related to tx_rs_thresh. The tx_rs_thresh value must be greater than or equal to RTE_PMD_IXGBE_TX_MAX_BURST, but less or equal to RTE_IXGBE_TX_MAX_FREE_BUF_SZ. Consequently, by default the tx_rs_thresh value is in the range 32 to 64.

9.1.2.2. Feature not Supported by RX Vector PMD

TX vPMD only works when txq_flags is set to IXGBE_SIMPLE_FLAGS.

This means that it does not support TX multi-segment, VLAN offload and TX csum offload. The following MACROs are used for these three features:

  • ETH_TXQ_FLAGS_NOMULTSEGS
  • ETH_TXQ_FLAGS_NOVLANOFFL
  • ETH_TXQ_FLAGS_NOXSUMSCTP
  • ETH_TXQ_FLAGS_NOXSUMUDP
  • ETH_TXQ_FLAGS_NOXSUMTCP

9.1.3. Sample Application Notes

9.1.3.1. testpmd

By default, using CONFIG_RTE_IXGBE_RX_OLFLAGS_ENABLE=y:

./x86_64-native-linuxapp-gcc/app/testpmd -c 300 -n 4 -- -i --burst=32 --rxfreet=32 --mbcache=250 --txpt=32 --rxht=8 --rxwt=0 --txfreet=32 --txrst=32 --txqflags=0xf01

When CONFIG_RTE_IXGBE_RX_OLFLAGS_ENABLE=n, better performance can be achieved:

./x86_64-native-linuxapp-gcc/app/testpmd -c 300 -n 4 -- -i --burst=32 --rxfreet=32 --mbcache=250 --txpt=32 --rxht=8 --rxwt=0 --txfreet=32 --txrst=32 --txqflags=0xf01 --disable-hw-vlan

9.1.3.2. l3fwd

When running l3fwd with vPMD, there is one thing to note. In the configuration, ensure that port_conf.rxmode.hw_ip_checksum=0. Otherwise, by default, RX vPMD is disabled.

9.1.3.3. load_balancer

As in the case of l3fwd, set configure port_conf.rxmode.hw_ip_checksum=0 to enable vPMD. In addition, for improved performance, use -bsz “(32,32),(64,64),(32,32)” in load_balancer to avoid using the default burst size of 144.

9.2. Malicious Driver Detection not Supported

The Intel x550 series NICs support a feature called MDD (Malicious Driver Detection) which checks the behavior of the VF driver. If this feature is enabled, the VF must use the advanced context descriptor correctly and set the CC (Check Context) bit. DPDK PF doesn’t support MDD, but kernel PF does. We may hit problem in this scenario kernel PF + DPDK VF. If user enables MDD in kernel PF, DPDK VF will not work. Because kernel PF thinks the VF is malicious. But actually it’s not. The only reason is the VF doesn’t act as MDD required. There’s significant performance impact to support MDD. DPDK should check if the advanced context descriptor should be set and set it. And DPDK has to ask the info about the header length from the upper layer, because parsing the packet itself is not acceptable. So, it’s too expensive to support MDD. When using kernel PF + DPDK VF on x550, please make sure using the kernel driver that disables MDD or can disable MDD. (Some kernel driver can use this CLI ‘insmod ixgbe.ko MDD=0,0’ to disable MDD. Some kernel driver disables it by default.)

9.3. Statistics

The statistics of ixgbe hardware must be polled regularly in order for it to remain consistent. Running a DPDK application without polling the statistics will cause registers on hardware to count to the maximum value, and “stick” at that value.

In order to avoid statistic registers every reaching the maximum value, read the statistics from the hardware using rte_eth_stats_get() or rte_eth_xstats_get().

The maximum time between statistics polls that ensures consistent results can be calculated as follows:

max_read_interval = UINT_MAX / max_packets_per_second
max_read_interval = 4294967295 / 14880952
max_read_interval = 288.6218096127183 (seconds)
max_read_interval = ~4 mins 48 sec.

In order to ensure valid results, it is recommended to poll every 4 minutes.