1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2016 Intel Corporation.
7 The i40e PMD (librte_pmd_i40e) provides poll mode driver support for
8 10/25/40 Gbps Intel® Ethernet 700 Series Network Adapters based on
9 the Intel Ethernet Controller X710/XL710/XXV710 and Intel Ethernet
10 Connection X722 (only support part of features).
16 Features of the i40e PMD are:
18 - Multiple queues for TX and RX
19 - Receiver Side Scaling (RSS)
21 - Packet type information
25 - VLAN/QinQ stripping and inserting
29 - Port hardware statistics
31 - Link state information
33 - Mirror on port, VLAN and VSI
34 - Interrupt mode for RX
35 - Scattered and gather for TX and RX
36 - Vector Poll mode driver
41 - IEEE1588/802.1AS timestamping
42 - VF Daemon (VFD) - EXPERIMENTAL
43 - Dynamic Device Personalization (DDP)
44 - Queue region configuration
45 - Virtual Function Port Representors
46 - Malicious Device Drive event catch and notify
51 - Identifying your adapter using `Intel Support
52 <http://www.intel.com/support>`_ and get the latest NVM/FW images.
54 - Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup the basic DPDK environment.
56 - To get better performance on Intel platforms, please follow the "How to get best performance with NICs on Intel platforms"
57 section of the :ref:`Getting Started Guide for Linux <linux_gsg>`.
59 - Upgrade the NVM/FW version following the `Intel® Ethernet NVM Update Tool Quick Usage Guide for Linux
60 <https://www-ssl.intel.com/content/www/us/en/embedded/products/networking/nvm-update-tool-quick-linux-usage-guide.html>`_ and `Intel® Ethernet NVM Update Tool: Quick Usage Guide for EFI <https://www.intel.com/content/www/us/en/embedded/products/networking/nvm-update-tool-quick-efi-usage-guide.html>`_ if needed.
62 Recommended Matching List
63 -------------------------
65 It is highly recommended to upgrade the i40e kernel driver and firmware to
66 avoid the compatibility issues with i40e PMD. Here is the suggested matching
67 list which has been tested and verified. The detailed information can refer
68 to chapter Tested Platforms/Tested NICs in release notes.
70 For X710/XL710/XXV710,
72 +--------------+-----------------------+------------------+
73 | DPDK version | Kernel driver version | Firmware version |
74 +==============+=======================+==================+
75 | 19.11 | 2.9.21 | 7.00 |
76 +--------------+-----------------------+------------------+
77 | 19.08 | 2.8.43 | 7.00 |
78 +--------------+-----------------------+------------------+
79 | 19.05 | 2.7.29 | 6.80 |
80 +--------------+-----------------------+------------------+
81 | 19.02 | 2.7.26 | 6.80 |
82 +--------------+-----------------------+------------------+
83 | 18.11 | 2.4.6 | 6.01 |
84 +--------------+-----------------------+------------------+
85 | 18.08 | 2.4.6 | 6.01 |
86 +--------------+-----------------------+------------------+
87 | 18.05 | 2.4.6 | 6.01 |
88 +--------------+-----------------------+------------------+
89 | 18.02 | 2.4.3 | 6.01 |
90 +--------------+-----------------------+------------------+
91 | 17.11 | 2.1.26 | 6.01 |
92 +--------------+-----------------------+------------------+
93 | 17.08 | 2.0.19 | 6.01 |
94 +--------------+-----------------------+------------------+
95 | 17.05 | 1.5.23 | 5.05 |
96 +--------------+-----------------------+------------------+
97 | 17.02 | 1.5.23 | 5.05 |
98 +--------------+-----------------------+------------------+
99 | 16.11 | 1.5.23 | 5.05 |
100 +--------------+-----------------------+------------------+
101 | 16.07 | 1.4.25 | 5.04 |
102 +--------------+-----------------------+------------------+
103 | 16.04 | 1.4.25 | 5.02 |
104 +--------------+-----------------------+------------------+
109 +--------------+-----------------------+------------------+
110 | DPDK version | Kernel driver version | Firmware version |
111 +==============+=======================+==================+
112 | 19.11 | 2.9.21 | 4.10 |
113 +--------------+-----------------------+------------------+
114 | 19.08 | 2.9.21 | 4.10 |
115 +--------------+-----------------------+------------------+
116 | 19.05 | 2.7.29 | 3.33 |
117 +--------------+-----------------------+------------------+
118 | 19.02 | 2.7.26 | 3.33 |
119 +--------------+-----------------------+------------------+
120 | 18.11 | 2.4.6 | 3.33 |
121 +--------------+-----------------------+------------------+
124 Pre-Installation Configuration
125 ------------------------------
130 The following options can be modified in the ``config`` file.
131 Please note that enabling debugging options may affect system performance.
133 - ``CONFIG_RTE_LIBRTE_I40E_PMD`` (default ``y``)
135 Toggle compilation of the ``librte_pmd_i40e`` driver.
137 - ``CONFIG_RTE_LIBRTE_I40E_DEBUG_*`` (default ``n``)
139 Toggle display of generic debugging messages.
141 - ``CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC`` (default ``y``)
143 Toggle bulk allocation for RX.
145 - ``CONFIG_RTE_LIBRTE_I40E_INC_VECTOR`` (default ``n``)
147 Toggle the use of Vector PMD instead of normal RX/TX path.
148 To enable vPMD for RX, bulk allocation for Rx must be allowed.
150 - ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` (default ``n``)
152 Toggle to use a 16-byte RX descriptor, by default the RX descriptor is 32 byte.
154 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``)
156 Number of queues reserved for PF.
158 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``)
160 Number of queues reserved for each VMDQ Pool.
162 Runtime Config Options
163 ~~~~~~~~~~~~~~~~~~~~~~
165 - ``Reserved number of Queues per VF`` (default ``4``)
167 The number of reserved queue per VF is determined by its host PF. If the
168 PCI address of an i40e PF is aaaa:bb.cc, the number of reserved queues per
169 VF can be configured with EAL parameter like -w aaaa:bb.cc,queue-num-per-vf=n.
170 The value n can be 1, 2, 4, 8 or 16. If no such parameter is configured, the
171 number of reserved queues per VF is 4 by default. If VF request more than
172 reserved queues per VF, PF will able to allocate max to 16 queues after a VF
176 - ``Support multiple driver`` (default ``disable``)
178 There was a multiple driver support issue during use of 700 series Ethernet
179 Adapter with both Linux kernel and DPDK PMD. To fix this issue, ``devargs``
180 parameter ``support-multi-driver`` is introduced, for example::
182 -w 84:00.0,support-multi-driver=1
184 With the above configuration, DPDK PMD will not change global registers, and
185 will switch PF interrupt from IntN to Int0 to avoid interrupt conflict between
186 DPDK and Linux Kernel.
188 - ``Support VF Port Representor`` (default ``not enabled``)
190 The i40e PF PMD supports the creation of VF port representors for the control
191 and monitoring of i40e virtual function devices. Each port representor
192 corresponds to a single virtual function of that device. Using the ``devargs``
193 option ``representor`` the user can specify which virtual functions to create
194 port representors for on initialization of the PF PMD by passing the VF IDs of
195 the VFs which are required.::
197 -w DBDF,representor=[0,1,4]
199 Currently hot-plugging of representor ports is not supported so all required
200 representors must be specified on the creation of the PF.
202 - ``Use latest supported vector`` (default ``disable``)
204 Latest supported vector path may not always get the best perf so vector path was
205 recommended to use only on later platform. But users may want the latest vector path
206 since it can get better perf in some real work loading cases. So ``devargs`` param
207 ``use-latest-supported-vec`` is introduced, for example::
209 -w 84:00.0,use-latest-supported-vec=1
211 - ``Enable validation for VF message`` (default ``not enabled``)
213 The PF counts messages from each VF. If in any period of seconds the message
214 statistic from a VF exceeds maximal limitation, the PF will ignore any new message
215 from that VF for some seconds.
216 Format -- "maximal-message@period-seconds:ignore-seconds"
219 -w 84:00.0,vf_msg_cfg=80@120:180
221 Vector RX Pre-conditions
222 ~~~~~~~~~~~~~~~~~~~~~~~~
223 For Vector RX it is assumed that the number of descriptor rings will be a power
224 of 2. With this pre-condition, the ring pointer can easily scroll back to the
225 head after hitting the tail without a conditional check. In addition Vector RX
226 can use this assumption to do a bit mask using ``ring_size - 1``.
228 Driver compilation and testing
229 ------------------------------
231 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
235 SR-IOV: Prerequisites and sample Application Notes
236 --------------------------------------------------
238 #. Load the kernel module:
240 .. code-block:: console
244 Check the output in dmesg:
246 .. code-block:: console
248 i40e 0000:83:00.1 ens802f0: renamed from eth0
250 #. Bring up the PF ports:
252 .. code-block:: console
256 #. Create VF device(s):
258 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry
263 .. code-block:: console
265 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs
268 #. Assign VF MAC address:
270 Assign MAC address to the VF using iproute2 utility. The syntax is:
272 .. code-block:: console
274 ip link set <PF netdev id> vf <VF id> mac <macaddr>
278 .. code-block:: console
280 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0
282 #. Assign VF to VM, and bring up the VM.
283 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*.
287 Follow instructions available in the document
288 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
293 .. code-block:: console
296 EAL: PCI device 0000:83:00.0 on NUMA socket 1
297 EAL: probe driver: 8086:1572 rte_i40e_pmd
298 EAL: PCI memory mapped at 0x7f7f80000000
299 EAL: PCI memory mapped at 0x7f7f80800000
300 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a
301 Interactive-mode selected
302 Configuring Port 0 (socket 0)
305 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are
306 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0.
309 Port 0: 68:05:CA:26:85:84
310 Checking link statuses...
311 Port 0 Link Up - speed 10000 Mbps - full-duplex
317 Sample Application Notes
318 ------------------------
323 Vlan filter only works when Promiscuous mode is off.
325 To start ``testpmd``, and add vlan 10 to port 0:
327 .. code-block:: console
329 ./app/testpmd -l 0-15 -n 4 -- -i --forward-mode=mac
332 testpmd> set promisc 0 off
333 testpmd> rx_vlan add 10 0
339 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
340 The Flow Director filters can match the different fields for different type of packet: flow type, specific input set per flow type and the flexible payload.
342 The default input set of each flow type is::
344 ipv4-other : src_ip_address, dst_ip_address
345 ipv4-frag : src_ip_address, dst_ip_address
346 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port
347 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port
348 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
350 ipv6-other : src_ip_address, dst_ip_address
351 ipv6-frag : src_ip_address, dst_ip_address
352 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port
353 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port
354 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
356 l2_payload : ether_type
358 The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching.
360 Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``:
362 .. code-block:: console
364 ./app/testpmd -l 0-15 -n 4 -- -i --disable-rss --pkt-filter-mode=perfect \
365 --rxq=8 --txq=8 --nb-cores=8 --nb-ports=1
367 Add a rule to direct ``ipv4-udp`` packet whose ``dst_ip=2.2.2.5, src_ip=2.2.2.3, src_port=32, dst_port=32`` to queue 1:
369 .. code-block:: console
371 testpmd> flow_director_filter 0 mode IP add flow ipv4-udp \
372 src 2.2.2.3 32 dst 2.2.2.5 32 vlan 0 flexbytes () \
373 fwd pf queue 1 fd_id 1
375 Check the flow director status:
377 .. code-block:: console
379 testpmd> show port fdir 0
381 ######################## FDIR infos for port 0 ####################
383 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other
384 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other
387 max_len: 16 payload_limit: 480
388 payload_unit: 2 payload_seg: 3
389 bitmask_unit: 2 bitmask_num: 2
392 src_ipv4: 0x00000000,
393 dst_ipv4: 0x00000000,
396 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000,
397 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000
398 FLEX PAYLOAD SRC OFFSET:
399 L2_PAYLOAD: 0 1 2 3 4 5 6 ...
400 L3_PAYLOAD: 0 1 2 3 4 5 6 ...
401 L4_PAYLOAD: 0 1 2 3 4 5 6 ...
403 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
404 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
405 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
406 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
407 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
408 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
409 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
410 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
411 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
412 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
413 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
414 guarant_count: 1 best_count: 0
415 guarant_space: 512 best_space: 7168
422 Delete all flow director rules on a port:
424 .. code-block:: console
426 testpmd> flush_flow_director 0
431 The Intel® Ethernet 700 Series support a feature called
434 A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term
435 for functionality that allows local switching between virtual endpoints within
436 a physical endpoint and also with an external bridge/network.
438 A "Floating" VEB doesn't have an uplink connection to the outside world so all
439 switching is done internally and remains within the host. As such, this
440 feature provides security benefits.
442 In addition, a Floating VEB overcomes a limitation of normal VEBs where they
443 cannot forward packets when the physical link is down. Floating VEBs don't need
444 to connect to the NIC port so they can still forward traffic from VF to VF
445 even when the physical link is down.
447 Therefore, with this feature enabled VFs can be limited to communicating with
448 each other but not an outside network, and they can do so even when there is
449 no physical uplink on the associated NIC port.
451 To enable this feature, the user should pass a ``devargs`` parameter to the
454 -w 84:00.0,enable_floating_veb=1
456 In this configuration the PMD will use the floating VEB feature for all the
457 VFs created by this PF device.
459 Alternatively, the user can specify which VFs need to connect to this floating
460 VEB using the ``floating_veb_list`` argument::
462 -w 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4
464 In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB,
465 while other VFs connect to the normal VEB.
467 The current implementation only supports one floating VEB and one regular
468 VEB. VFs can connect to a floating VEB or a regular VEB according to the
469 configuration passed on the EAL command line.
471 The floating VEB functionality requires a NIC firmware version of 5.0
474 Dynamic Device Personalization (DDP)
475 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
477 The Intel® Ethernet 700 Series except for the Intel Ethernet Connection
478 X722 support a feature called "Dynamic Device Personalization (DDP)",
479 which is used to configure hardware by downloading a profile to support
480 protocols/filters which are not supported by default. The DDP
481 functionality requires a NIC firmware version of 6.0 or greater.
483 Current implementation supports GTP-C/GTP-U/PPPoE/PPPoL2TP/ESP,
484 steering can be used with rte_flow API.
486 GTPv1 package is released, and it can be downloaded from
487 https://downloadcenter.intel.com/download/27587.
489 PPPoE package is released, and it can be downloaded from
490 https://downloadcenter.intel.com/download/28040.
492 ESP-AH package is not released yet.
494 Load a profile which supports GTP and store backup profile:
496 .. code-block:: console
498 testpmd> ddp add 0 ./gtp.pkgo,./backup.pkgo
500 Delete a GTP profile and restore backup profile:
502 .. code-block:: console
504 testpmd> ddp del 0 ./backup.pkgo
506 Get loaded DDP package info list:
508 .. code-block:: console
510 testpmd> ddp get list 0
512 Display information about a GTP profile:
514 .. code-block:: console
516 testpmd> ddp get info ./gtp.pkgo
518 Input set configuration
519 ~~~~~~~~~~~~~~~~~~~~~~~
520 Input set for any PCTYPE can be configured with user defined configuration,
521 For example, to use only 48bit prefix for IPv6 src address for IPv6 TCP RSS:
523 .. code-block:: console
525 testpmd> port config 0 pctype 43 hash_inset clear all
526 testpmd> port config 0 pctype 43 hash_inset set field 13
527 testpmd> port config 0 pctype 43 hash_inset set field 14
528 testpmd> port config 0 pctype 43 hash_inset set field 15
530 Queue region configuration
531 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
532 The Intel® Ethernet 700 Series supports a feature of queue regions
533 configuration for RSS in the PF, so that different traffic classes or
534 different packet classification types can be separated to different
535 queues in different queue regions. There is an API for configuration
536 of queue regions in RSS with a command line. It can parse the parameters
537 of the region index, queue number, queue start index, user priority, traffic
538 classes and so on. Depending on commands from the command line, it will call
539 i40e private APIs and start the process of setting or flushing the queue
540 region configuration. As this feature is specific for i40e only private
541 APIs are used. These new ``test_pmd`` commands are as shown below. For
542 details please refer to :doc:`../testpmd_app_ug/index`.
544 .. code-block:: console
546 testpmd> set port (port_id) queue-region region_id (value) \
547 queue_start_index (value) queue_num (value)
548 testpmd> set port (port_id) queue-region region_id (value) flowtype (value)
549 testpmd> set port (port_id) queue-region UP (value) region_id (value)
550 testpmd> set port (port_id) queue-region flush (on|off)
551 testpmd> show port (port_id) queue-region
553 Limitations or Known issues
554 ---------------------------
556 MPLS packet classification
557 ~~~~~~~~~~~~~~~~~~~~~~~~~~
559 For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC.
560 The L2 Payload flow type in flow director can be used to classify MPLS packet
561 by using a command in testpmd like:
563 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \
564 0x8847 flexbytes () fwd pf queue <N> fd_id <M>
566 With the NIC firmware version 5.0 or greater, some limited MPLS support
567 is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no
568 new packet type, no classification or offload are possible. With this change,
569 L2 Payload flow type in flow director cannot be used to classify MPLS packet
570 as with previous firmware versions. Meanwhile, the Ethertype filter can be
571 used to classify MPLS packet by using a command in testpmd like:
573 testpmd> ethertype_filter 0 add mac_ignr 00:00:00:00:00:00 ethertype \
576 16 Byte RX Descriptor setting on DPDK VF
577 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
579 Currently the VF's RX descriptor mode is decided by PF. There's no PF-VF
580 interface for VF to request the RX descriptor mode, also no interface to notify
581 VF its own RX descriptor mode.
582 For all available versions of the i40e driver, these drivers don't support 16
583 byte RX descriptor. If the Linux i40e kernel driver is used as host driver,
584 while DPDK i40e PMD is used as the VF driver, DPDK cannot choose 16 byte receive
585 descriptor. The reason is that the RX descriptor is already set to 32 byte by
586 the i40e kernel driver. That is to say, user should keep
587 ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n`` in config file.
588 In the future, if the Linux i40e driver supports 16 byte RX descriptor, user
589 should make sure the DPDK VF uses the same RX descriptor mode, 16 byte or 32
590 byte, as the PF driver.
592 The same rule for DPDK PF + DPDK VF. The PF and VF should use the same RX
593 descriptor mode. Or the VF RX will not work.
595 Receive packets with Ethertype 0x88A8
596 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
598 Due to the FW limitation, PF can receive packets with Ethertype 0x88A8
599 only when floating VEB is disabled.
601 Incorrect Rx statistics when packet is oversize
602 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
604 When a packet is over maximum frame size, the packet is dropped.
605 However, the Rx statistics, when calling `rte_eth_stats_get` incorrectly
606 shows it as received.
608 VF & TC max bandwidth setting
609 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
611 The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel.
612 The behavior is different when handling per VF and per TC max bandwidth setting.
613 When enabling per VF max bandwidth, SW will check if per TC max bandwidth is
614 enabled. If so, return failure.
615 When enabling per TC max bandwidth, SW will check if per VF max bandwidth
616 is enabled. If so, disable per VF max bandwidth and continue with per TC max
619 TC TX scheduling mode setting
620 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
622 There are 2 TX scheduling modes for TCs, round robin and strict priority mode.
623 If a TC is set to strict priority mode, it can consume unlimited bandwidth.
624 It means if APP has set the max bandwidth for that TC, it comes to no
626 It's suggested to set the strict priority mode for a TC that is latency
627 sensitive but no consuming much bandwidth.
629 VF performance is impacted by PCI extended tag setting
630 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
632 To reach maximum NIC performance in the VF the PCI extended tag must be
633 enabled. The DPDK i40e PF driver will set this feature during initialization,
634 but the kernel PF driver does not. So when running traffic on a VF which is
635 managed by the kernel PF driver, a significant NIC performance downgrade has
636 been observed (for 64 byte packets, there is about 25% line-rate downgrade for
637 a 25GbE device and about 35% for a 40GbE device).
639 For kernel version >= 4.11, the kernel's PCI driver will enable the extended
640 tag if it detects that the device supports it. So by default, this is not an
641 issue. For kernels <= 4.11 or when the PCI extended tag is disabled it can be
642 enabled using the steps below.
644 #. Get the current value of the PCI configure register::
646 setpci -s <XX:XX.X> a8.w
650 value = value | 0x100
652 #. Set the PCI configure register with new value::
654 setpci -s <XX:XX.X> a8.w=<value>
659 The VF vlan strip function is only supported in the i40e kernel driver >= 2.1.26.
664 DCB works only when RSS is enabled.
666 Global configuration warning
667 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
669 I40E PMD will set some global registers to enable some function or set some
670 configure. Then when using different ports of the same NIC with Linux kernel
671 and DPDK, the port with Linux kernel will be impacted by the port with DPDK.
672 For example, register I40E_GL_SWT_L2TAGCTRL is used to control L2 tag, i40e
673 PMD uses I40E_GL_SWT_L2TAGCTRL to set vlan TPID. If setting TPID in port A
674 with DPDK, then the configuration will also impact port B in the NIC with
675 kernel driver, which don't want to use the TPID.
676 So PMD reports warning to clarify what is changed by writing global register.
678 High Performance of Small Packets on 40GbE NIC
679 ----------------------------------------------
681 As there might be firmware fixes for performance enhancement in latest version
682 of firmware image, the firmware update might be needed for getting high performance.
683 Check the Intel support website for the latest firmware updates.
684 Users should consult the release notes specific to a DPDK release to identify
685 the validated firmware version for a NIC using the i40e driver.
687 Use 16 Bytes RX Descriptor Size
688 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
690 As i40e PMD supports both 16 and 32 bytes RX descriptor sizes, and 16 bytes size can provide helps to high performance of small packets.
691 Configuration of ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` in config files can be changed to use 16 bytes size RX descriptors.
693 Example of getting best performance with l3fwd example
694 ------------------------------------------------------
696 The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with a
697 server with Intel Xeon processors and Intel Ethernet CNA XL710.
699 The example scenario is to get best performance with two Intel Ethernet CNA XL710 40GbE ports.
700 See :numref:`figure_intel_perf_test_setup` for the performance test setup.
702 .. _figure_intel_perf_test_setup:
704 .. figure:: img/intel_perf_test_setup.*
706 Performance Test Setup
709 1. Add two Intel Ethernet CNA XL710 to the platform, and use one port per card to get best performance.
710 The reason for using two NICs is to overcome a PCIe v3.0 limitation since it cannot provide 80GbE bandwidth
711 for two 40GbE ports, but two different PCIe v3.0 x8 slot can.
712 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports::
714 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
715 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
717 2. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator.
719 3. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id.
720 In this case, ``82:00.0`` and ``85:00.0`` are both in socket 1, and the cores on socket 1 in the referenced platform
722 Note: Don't use 2 logical cores on the same core (e.g core18 has 2 logical cores, core18 and core54), instead, use 2 logical
723 cores from different cores (e.g core18 and core19).
725 4. Bind these two ports to igb_uio.
727 5. As to Intel Ethernet CNA XL710 40GbE port, we need at least two queue pairs to achieve best performance, then two queues per port
728 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets.
730 6. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding.
731 Compile the ``l3fwd sample`` with the default lpm mode.
733 7. The command line of running l3fwd would be something like the following::
735 ./l3fwd -l 18-21 -n 4 -w 82:00.0 -w 85:00.0 \
736 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)'
738 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding,
739 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding.
741 8. Configure the traffic at a traffic generator.
743 * Start creating a stream on packet generator.
745 * Set the Ethernet II type to 0x0800.
747 Tx bytes affected by the link status change
748 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
750 For firmware versions prior to 6.01 for X710 series and 3.33 for X722 series, the tx_bytes statistics data is affected by
751 the link down event. Each time the link status changes to down, the tx_bytes decreases 110 bytes.