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 | 20.02 | 2.10.19 | 7.20 |
76 +--------------+-----------------------+------------------+
77 | 19.11 | 2.9.21 | 7.00 |
78 +--------------+-----------------------+------------------+
79 | 19.08 | 2.8.43 | 7.00 |
80 +--------------+-----------------------+------------------+
81 | 19.05 | 2.7.29 | 6.80 |
82 +--------------+-----------------------+------------------+
83 | 19.02 | 2.7.26 | 6.80 |
84 +--------------+-----------------------+------------------+
85 | 18.11 | 2.4.6 | 6.01 |
86 +--------------+-----------------------+------------------+
87 | 18.08 | 2.4.6 | 6.01 |
88 +--------------+-----------------------+------------------+
89 | 18.05 | 2.4.6 | 6.01 |
90 +--------------+-----------------------+------------------+
91 | 18.02 | 2.4.3 | 6.01 |
92 +--------------+-----------------------+------------------+
93 | 17.11 | 2.1.26 | 6.01 |
94 +--------------+-----------------------+------------------+
95 | 17.08 | 2.0.19 | 6.01 |
96 +--------------+-----------------------+------------------+
97 | 17.05 | 1.5.23 | 5.05 |
98 +--------------+-----------------------+------------------+
99 | 17.02 | 1.5.23 | 5.05 |
100 +--------------+-----------------------+------------------+
101 | 16.11 | 1.5.23 | 5.05 |
102 +--------------+-----------------------+------------------+
103 | 16.07 | 1.4.25 | 5.04 |
104 +--------------+-----------------------+------------------+
105 | 16.04 | 1.4.25 | 5.02 |
106 +--------------+-----------------------+------------------+
111 +--------------+-----------------------+------------------+
112 | DPDK version | Kernel driver version | Firmware version |
113 +==============+=======================+==================+
114 | 20.02 | 2.10.19 | 4.11 |
115 +--------------+-----------------------+------------------+
116 | 19.11 | 2.9.21 | 4.10 |
117 +--------------+-----------------------+------------------+
118 | 19.08 | 2.9.21 | 4.10 |
119 +--------------+-----------------------+------------------+
120 | 19.05 | 2.7.29 | 3.33 |
121 +--------------+-----------------------+------------------+
122 | 19.02 | 2.7.26 | 3.33 |
123 +--------------+-----------------------+------------------+
124 | 18.11 | 2.4.6 | 3.33 |
125 +--------------+-----------------------+------------------+
128 Pre-Installation Configuration
129 ------------------------------
134 The following options can be modified in the ``config`` file.
135 Please note that enabling debugging options may affect system performance.
137 - ``CONFIG_RTE_LIBRTE_I40E_PMD`` (default ``y``)
139 Toggle compilation of the ``librte_pmd_i40e`` driver.
141 - ``CONFIG_RTE_LIBRTE_I40E_DEBUG_*`` (default ``n``)
143 Toggle display of generic debugging messages.
145 - ``CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC`` (default ``y``)
147 Toggle bulk allocation for RX.
149 - ``CONFIG_RTE_LIBRTE_I40E_INC_VECTOR`` (default ``n``)
151 Toggle the use of Vector PMD instead of normal RX/TX path.
152 To enable vPMD for RX, bulk allocation for Rx must be allowed.
154 - ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` (default ``n``)
156 Toggle to use a 16-byte RX descriptor, by default the RX descriptor is 32 byte.
158 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``)
160 Number of queues reserved for PF.
162 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``)
164 Number of queues reserved for each VMDQ Pool.
166 Runtime Config Options
167 ~~~~~~~~~~~~~~~~~~~~~~
169 - ``Reserved number of Queues per VF`` (default ``4``)
171 The number of reserved queue per VF is determined by its host PF. If the
172 PCI address of an i40e PF is aaaa:bb.cc, the number of reserved queues per
173 VF can be configured with EAL parameter like -w aaaa:bb.cc,queue-num-per-vf=n.
174 The value n can be 1, 2, 4, 8 or 16. If no such parameter is configured, the
175 number of reserved queues per VF is 4 by default. If VF request more than
176 reserved queues per VF, PF will able to allocate max to 16 queues after a VF
180 - ``Support multiple driver`` (default ``disable``)
182 There was a multiple driver support issue during use of 700 series Ethernet
183 Adapter with both Linux kernel and DPDK PMD. To fix this issue, ``devargs``
184 parameter ``support-multi-driver`` is introduced, for example::
186 -w 84:00.0,support-multi-driver=1
188 With the above configuration, DPDK PMD will not change global registers, and
189 will switch PF interrupt from IntN to Int0 to avoid interrupt conflict between
190 DPDK and Linux Kernel.
192 - ``Support VF Port Representor`` (default ``not enabled``)
194 The i40e PF PMD supports the creation of VF port representors for the control
195 and monitoring of i40e virtual function devices. Each port representor
196 corresponds to a single virtual function of that device. Using the ``devargs``
197 option ``representor`` the user can specify which virtual functions to create
198 port representors for on initialization of the PF PMD by passing the VF IDs of
199 the VFs which are required.::
201 -w DBDF,representor=[0,1,4]
203 Currently hot-plugging of representor ports is not supported so all required
204 representors must be specified on the creation of the PF.
206 - ``Use latest supported vector`` (default ``disable``)
208 Latest supported vector path may not always get the best perf so vector path was
209 recommended to use only on later platform. But users may want the latest vector path
210 since it can get better perf in some real work loading cases. So ``devargs`` param
211 ``use-latest-supported-vec`` is introduced, for example::
213 -w 84:00.0,use-latest-supported-vec=1
215 - ``Enable validation for VF message`` (default ``not enabled``)
217 The PF counts messages from each VF. If in any period of seconds the message
218 statistic from a VF exceeds maximal limitation, the PF will ignore any new message
219 from that VF for some seconds.
220 Format -- "maximal-message@period-seconds:ignore-seconds"
223 -w 84:00.0,vf_msg_cfg=80@120:180
225 Vector RX Pre-conditions
226 ~~~~~~~~~~~~~~~~~~~~~~~~
227 For Vector RX it is assumed that the number of descriptor rings will be a power
228 of 2. With this pre-condition, the ring pointer can easily scroll back to the
229 head after hitting the tail without a conditional check. In addition Vector RX
230 can use this assumption to do a bit mask using ``ring_size - 1``.
232 Driver compilation and testing
233 ------------------------------
235 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
239 SR-IOV: Prerequisites and sample Application Notes
240 --------------------------------------------------
242 #. Load the kernel module:
244 .. code-block:: console
248 Check the output in dmesg:
250 .. code-block:: console
252 i40e 0000:83:00.1 ens802f0: renamed from eth0
254 #. Bring up the PF ports:
256 .. code-block:: console
260 #. Create VF device(s):
262 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry
267 .. code-block:: console
269 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs
272 #. Assign VF MAC address:
274 Assign MAC address to the VF using iproute2 utility. The syntax is:
276 .. code-block:: console
278 ip link set <PF netdev id> vf <VF id> mac <macaddr>
282 .. code-block:: console
284 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0
286 #. Assign VF to VM, and bring up the VM.
287 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*.
291 Follow instructions available in the document
292 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
297 .. code-block:: console
300 EAL: PCI device 0000:83:00.0 on NUMA socket 1
301 EAL: probe driver: 8086:1572 rte_i40e_pmd
302 EAL: PCI memory mapped at 0x7f7f80000000
303 EAL: PCI memory mapped at 0x7f7f80800000
304 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a
305 Interactive-mode selected
306 Configuring Port 0 (socket 0)
309 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are
310 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0.
313 Port 0: 68:05:CA:26:85:84
314 Checking link statuses...
315 Port 0 Link Up - speed 10000 Mbps - full-duplex
321 Sample Application Notes
322 ------------------------
327 Vlan filter only works when Promiscuous mode is off.
329 To start ``testpmd``, and add vlan 10 to port 0:
331 .. code-block:: console
333 ./app/testpmd -l 0-15 -n 4 -- -i --forward-mode=mac
336 testpmd> set promisc 0 off
337 testpmd> rx_vlan add 10 0
343 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
344 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.
346 The default input set of each flow type is::
348 ipv4-other : src_ip_address, dst_ip_address
349 ipv4-frag : src_ip_address, dst_ip_address
350 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port
351 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port
352 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
354 ipv6-other : src_ip_address, dst_ip_address
355 ipv6-frag : src_ip_address, dst_ip_address
356 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port
357 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port
358 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
360 l2_payload : ether_type
362 The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching.
364 Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``:
366 .. code-block:: console
368 ./app/testpmd -l 0-15 -n 4 -- -i --disable-rss --pkt-filter-mode=perfect \
369 --rxq=8 --txq=8 --nb-cores=8 --nb-ports=1
371 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:
373 .. code-block:: console
375 testpmd> flow_director_filter 0 mode IP add flow ipv4-udp \
376 src 2.2.2.3 32 dst 2.2.2.5 32 vlan 0 flexbytes () \
377 fwd pf queue 1 fd_id 1
379 Check the flow director status:
381 .. code-block:: console
383 testpmd> show port fdir 0
385 ######################## FDIR infos for port 0 ####################
387 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other
388 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other
391 max_len: 16 payload_limit: 480
392 payload_unit: 2 payload_seg: 3
393 bitmask_unit: 2 bitmask_num: 2
396 src_ipv4: 0x00000000,
397 dst_ipv4: 0x00000000,
400 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000,
401 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000
402 FLEX PAYLOAD SRC OFFSET:
403 L2_PAYLOAD: 0 1 2 3 4 5 6 ...
404 L3_PAYLOAD: 0 1 2 3 4 5 6 ...
405 L4_PAYLOAD: 0 1 2 3 4 5 6 ...
407 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
408 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
409 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
410 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
411 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
412 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
413 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
414 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
415 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
416 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
417 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
418 guarant_count: 1 best_count: 0
419 guarant_space: 512 best_space: 7168
426 Delete all flow director rules on a port:
428 .. code-block:: console
430 testpmd> flush_flow_director 0
435 The Intel® Ethernet 700 Series support a feature called
438 A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term
439 for functionality that allows local switching between virtual endpoints within
440 a physical endpoint and also with an external bridge/network.
442 A "Floating" VEB doesn't have an uplink connection to the outside world so all
443 switching is done internally and remains within the host. As such, this
444 feature provides security benefits.
446 In addition, a Floating VEB overcomes a limitation of normal VEBs where they
447 cannot forward packets when the physical link is down. Floating VEBs don't need
448 to connect to the NIC port so they can still forward traffic from VF to VF
449 even when the physical link is down.
451 Therefore, with this feature enabled VFs can be limited to communicating with
452 each other but not an outside network, and they can do so even when there is
453 no physical uplink on the associated NIC port.
455 To enable this feature, the user should pass a ``devargs`` parameter to the
458 -w 84:00.0,enable_floating_veb=1
460 In this configuration the PMD will use the floating VEB feature for all the
461 VFs created by this PF device.
463 Alternatively, the user can specify which VFs need to connect to this floating
464 VEB using the ``floating_veb_list`` argument::
466 -w 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4
468 In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB,
469 while other VFs connect to the normal VEB.
471 The current implementation only supports one floating VEB and one regular
472 VEB. VFs can connect to a floating VEB or a regular VEB according to the
473 configuration passed on the EAL command line.
475 The floating VEB functionality requires a NIC firmware version of 5.0
478 Dynamic Device Personalization (DDP)
479 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
481 The Intel® Ethernet 700 Series except for the Intel Ethernet Connection
482 X722 support a feature called "Dynamic Device Personalization (DDP)",
483 which is used to configure hardware by downloading a profile to support
484 protocols/filters which are not supported by default. The DDP
485 functionality requires a NIC firmware version of 6.0 or greater.
487 Current implementation supports GTP-C/GTP-U/PPPoE/PPPoL2TP/ESP,
488 steering can be used with rte_flow API.
490 GTPv1 package is released, and it can be downloaded from
491 https://downloadcenter.intel.com/download/27587.
493 PPPoE package is released, and it can be downloaded from
494 https://downloadcenter.intel.com/download/28040.
496 ESP-AH package is not released yet.
498 Load a profile which supports GTP and store backup profile:
500 .. code-block:: console
502 testpmd> ddp add 0 ./gtp.pkgo,./backup.pkgo
504 Delete a GTP profile and restore backup profile:
506 .. code-block:: console
508 testpmd> ddp del 0 ./backup.pkgo
510 Get loaded DDP package info list:
512 .. code-block:: console
514 testpmd> ddp get list 0
516 Display information about a GTP profile:
518 .. code-block:: console
520 testpmd> ddp get info ./gtp.pkgo
522 Input set configuration
523 ~~~~~~~~~~~~~~~~~~~~~~~
524 Input set for any PCTYPE can be configured with user defined configuration,
525 For example, to use only 48bit prefix for IPv6 src address for IPv6 TCP RSS:
527 .. code-block:: console
529 testpmd> port config 0 pctype 43 hash_inset clear all
530 testpmd> port config 0 pctype 43 hash_inset set field 13
531 testpmd> port config 0 pctype 43 hash_inset set field 14
532 testpmd> port config 0 pctype 43 hash_inset set field 15
534 Queue region configuration
535 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
536 The Intel® Ethernet 700 Series supports a feature of queue regions
537 configuration for RSS in the PF, so that different traffic classes or
538 different packet classification types can be separated to different
539 queues in different queue regions. There is an API for configuration
540 of queue regions in RSS with a command line. It can parse the parameters
541 of the region index, queue number, queue start index, user priority, traffic
542 classes and so on. Depending on commands from the command line, it will call
543 i40e private APIs and start the process of setting or flushing the queue
544 region configuration. As this feature is specific for i40e only private
545 APIs are used. These new ``test_pmd`` commands are as shown below. For
546 details please refer to :doc:`../testpmd_app_ug/index`.
548 .. code-block:: console
550 testpmd> set port (port_id) queue-region region_id (value) \
551 queue_start_index (value) queue_num (value)
552 testpmd> set port (port_id) queue-region region_id (value) flowtype (value)
553 testpmd> set port (port_id) queue-region UP (value) region_id (value)
554 testpmd> set port (port_id) queue-region flush (on|off)
555 testpmd> show port (port_id) queue-region
557 Limitations or Known issues
558 ---------------------------
560 MPLS packet classification
561 ~~~~~~~~~~~~~~~~~~~~~~~~~~
563 For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC.
564 The L2 Payload flow type in flow director can be used to classify MPLS packet
565 by using a command in testpmd like:
567 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \
568 0x8847 flexbytes () fwd pf queue <N> fd_id <M>
570 With the NIC firmware version 5.0 or greater, some limited MPLS support
571 is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no
572 new packet type, no classification or offload are possible. With this change,
573 L2 Payload flow type in flow director cannot be used to classify MPLS packet
574 as with previous firmware versions. Meanwhile, the Ethertype filter can be
575 used to classify MPLS packet by using a command in testpmd like:
577 testpmd> ethertype_filter 0 add mac_ignr 00:00:00:00:00:00 ethertype \
580 16 Byte RX Descriptor setting on DPDK VF
581 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
583 Currently the VF's RX descriptor mode is decided by PF. There's no PF-VF
584 interface for VF to request the RX descriptor mode, also no interface to notify
585 VF its own RX descriptor mode.
586 For all available versions of the i40e driver, these drivers don't support 16
587 byte RX descriptor. If the Linux i40e kernel driver is used as host driver,
588 while DPDK i40e PMD is used as the VF driver, DPDK cannot choose 16 byte receive
589 descriptor. The reason is that the RX descriptor is already set to 32 byte by
590 the i40e kernel driver. That is to say, user should keep
591 ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n`` in config file.
592 In the future, if the Linux i40e driver supports 16 byte RX descriptor, user
593 should make sure the DPDK VF uses the same RX descriptor mode, 16 byte or 32
594 byte, as the PF driver.
596 The same rule for DPDK PF + DPDK VF. The PF and VF should use the same RX
597 descriptor mode. Or the VF RX will not work.
599 Receive packets with Ethertype 0x88A8
600 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
602 Due to the FW limitation, PF can receive packets with Ethertype 0x88A8
603 only when floating VEB is disabled.
605 Incorrect Rx statistics when packet is oversize
606 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
608 When a packet is over maximum frame size, the packet is dropped.
609 However, the Rx statistics, when calling `rte_eth_stats_get` incorrectly
610 shows it as received.
612 VF & TC max bandwidth setting
613 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
615 The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel.
616 The behavior is different when handling per VF and per TC max bandwidth setting.
617 When enabling per VF max bandwidth, SW will check if per TC max bandwidth is
618 enabled. If so, return failure.
619 When enabling per TC max bandwidth, SW will check if per VF max bandwidth
620 is enabled. If so, disable per VF max bandwidth and continue with per TC max
623 TC TX scheduling mode setting
624 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
626 There are 2 TX scheduling modes for TCs, round robin and strict priority mode.
627 If a TC is set to strict priority mode, it can consume unlimited bandwidth.
628 It means if APP has set the max bandwidth for that TC, it comes to no
630 It's suggested to set the strict priority mode for a TC that is latency
631 sensitive but no consuming much bandwidth.
633 VF performance is impacted by PCI extended tag setting
634 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
636 To reach maximum NIC performance in the VF the PCI extended tag must be
637 enabled. The DPDK i40e PF driver will set this feature during initialization,
638 but the kernel PF driver does not. So when running traffic on a VF which is
639 managed by the kernel PF driver, a significant NIC performance downgrade has
640 been observed (for 64 byte packets, there is about 25% line-rate downgrade for
641 a 25GbE device and about 35% for a 40GbE device).
643 For kernel version >= 4.11, the kernel's PCI driver will enable the extended
644 tag if it detects that the device supports it. So by default, this is not an
645 issue. For kernels <= 4.11 or when the PCI extended tag is disabled it can be
646 enabled using the steps below.
648 #. Get the current value of the PCI configure register::
650 setpci -s <XX:XX.X> a8.w
654 value = value | 0x100
656 #. Set the PCI configure register with new value::
658 setpci -s <XX:XX.X> a8.w=<value>
663 The VF vlan strip function is only supported in the i40e kernel driver >= 2.1.26.
668 DCB works only when RSS is enabled.
670 Global configuration warning
671 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
673 I40E PMD will set some global registers to enable some function or set some
674 configure. Then when using different ports of the same NIC with Linux kernel
675 and DPDK, the port with Linux kernel will be impacted by the port with DPDK.
676 For example, register I40E_GL_SWT_L2TAGCTRL is used to control L2 tag, i40e
677 PMD uses I40E_GL_SWT_L2TAGCTRL to set vlan TPID. If setting TPID in port A
678 with DPDK, then the configuration will also impact port B in the NIC with
679 kernel driver, which don't want to use the TPID.
680 So PMD reports warning to clarify what is changed by writing global register.
682 High Performance of Small Packets on 40GbE NIC
683 ----------------------------------------------
685 As there might be firmware fixes for performance enhancement in latest version
686 of firmware image, the firmware update might be needed for getting high performance.
687 Check the Intel support website for the latest firmware updates.
688 Users should consult the release notes specific to a DPDK release to identify
689 the validated firmware version for a NIC using the i40e driver.
691 Use 16 Bytes RX Descriptor Size
692 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
694 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.
695 Configuration of ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` in config files can be changed to use 16 bytes size RX descriptors.
697 Example of getting best performance with l3fwd example
698 ------------------------------------------------------
700 The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with a
701 server with Intel Xeon processors and Intel Ethernet CNA XL710.
703 The example scenario is to get best performance with two Intel Ethernet CNA XL710 40GbE ports.
704 See :numref:`figure_intel_perf_test_setup` for the performance test setup.
706 .. _figure_intel_perf_test_setup:
708 .. figure:: img/intel_perf_test_setup.*
710 Performance Test Setup
713 1. Add two Intel Ethernet CNA XL710 to the platform, and use one port per card to get best performance.
714 The reason for using two NICs is to overcome a PCIe v3.0 limitation since it cannot provide 80GbE bandwidth
715 for two 40GbE ports, but two different PCIe v3.0 x8 slot can.
716 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports::
718 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
719 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
721 2. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator.
723 3. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id.
724 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
726 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
727 cores from different cores (e.g core18 and core19).
729 4. Bind these two ports to igb_uio.
731 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
732 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets.
734 6. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding.
735 Compile the ``l3fwd sample`` with the default lpm mode.
737 7. The command line of running l3fwd would be something like the following::
739 ./l3fwd -l 18-21 -n 4 -w 82:00.0 -w 85:00.0 \
740 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)'
742 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding,
743 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding.
745 8. Configure the traffic at a traffic generator.
747 * Start creating a stream on packet generator.
749 * Set the Ethernet II type to 0x0800.
751 Tx bytes affected by the link status change
752 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
754 For firmware versions prior to 6.01 for X710 series and 3.33 for X722 series, the tx_bytes statistics data is affected by
755 the link down event. Each time the link status changes to down, the tx_bytes decreases 110 bytes.