1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2016 Intel Corporation.
7 The i40e PMD (**librte_net_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
52 - Identifying your adapter using `Intel Support
53 <http://www.intel.com/support>`_ and get the latest NVM/FW images.
55 - Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup the basic DPDK environment.
57 - To get better performance on Intel platforms, please follow the "How to get best performance with NICs on Intel platforms"
58 section of the :ref:`Getting Started Guide for Linux <linux_gsg>`.
60 - Upgrade the NVM/FW version following the `Intel® Ethernet NVM Update Tool Quick Usage Guide for Linux
61 <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.
63 - For information about supported media, please refer to this document: `Intel® Ethernet Controller X710/XXV710/XL710 Feature Support Matrix
64 <http://www.intel.com/content/dam/www/public/us/en/documents/release-notes/xl710-ethernet-controller-feature-matrix.pdf>`_.
68 * Some adapters based on the Intel(R) Ethernet Controller 700 Series only
69 support Intel Ethernet Optics modules. On these adapters, other modules are not
70 supported and will not function.
72 * For connections based on Intel(R) Ethernet Controller 700 Series,
73 support is dependent on your system board. Please see your vendor for details.
75 * In all cases Intel recommends using Intel Ethernet Optics; other modules
76 may function but are not validated by Intel. Contact Intel for supported media types.
81 - Follow the DPDK `Getting Started Guide for Windows <https://doc.dpdk.org/guides/windows_gsg/index.html>`_ to setup the basic DPDK environment.
83 - Identify the Intel® Ethernet adapter and get the latest NVM/FW version.
85 - To access any Intel® Ethernet hardware, load the NetUIO driver in place of existing built-in (inbox) driver.
87 - To load NetUIO driver, follow the steps mentioned in `dpdk-kmods repository
88 <https://git.dpdk.org/dpdk-kmods/tree/windows/netuio/README.rst>`_.
90 Recommended Matching List
91 -------------------------
93 It is highly recommended to upgrade the i40e kernel driver and firmware to
94 avoid the compatibility issues with i40e PMD. Here is the suggested matching
95 list which has been tested and verified. The detailed information can refer
96 to chapter Tested Platforms/Tested NICs in release notes.
98 For X710/XL710/XXV710,
100 +--------------+-----------------------+------------------+
101 | DPDK version | Kernel driver version | Firmware version |
102 +==============+=======================+==================+
103 | 21.02 | 2.14.13 | 8.00 |
104 +--------------+-----------------------+------------------+
105 | 20.11 | 2.14.13 | 8.00 |
106 +--------------+-----------------------+------------------+
107 | 20.08 | 2.12.6 | 7.30 |
108 +--------------+-----------------------+------------------+
109 | 20.05 | 2.11.27 | 7.30 |
110 +--------------+-----------------------+------------------+
111 | 20.02 | 2.10.19 | 7.20 |
112 +--------------+-----------------------+------------------+
113 | 19.11 | 2.9.21 | 7.00 |
114 +--------------+-----------------------+------------------+
115 | 19.08 | 2.8.43 | 7.00 |
116 +--------------+-----------------------+------------------+
117 | 19.05 | 2.7.29 | 6.80 |
118 +--------------+-----------------------+------------------+
119 | 19.02 | 2.7.26 | 6.80 |
120 +--------------+-----------------------+------------------+
121 | 18.11 | 2.4.6 | 6.01 |
122 +--------------+-----------------------+------------------+
123 | 18.08 | 2.4.6 | 6.01 |
124 +--------------+-----------------------+------------------+
125 | 18.05 | 2.4.6 | 6.01 |
126 +--------------+-----------------------+------------------+
127 | 18.02 | 2.4.3 | 6.01 |
128 +--------------+-----------------------+------------------+
129 | 17.11 | 2.1.26 | 6.01 |
130 +--------------+-----------------------+------------------+
131 | 17.08 | 2.0.19 | 6.01 |
132 +--------------+-----------------------+------------------+
133 | 17.05 | 1.5.23 | 5.05 |
134 +--------------+-----------------------+------------------+
135 | 17.02 | 1.5.23 | 5.05 |
136 +--------------+-----------------------+------------------+
137 | 16.11 | 1.5.23 | 5.05 |
138 +--------------+-----------------------+------------------+
139 | 16.07 | 1.4.25 | 5.04 |
140 +--------------+-----------------------+------------------+
141 | 16.04 | 1.4.25 | 5.02 |
142 +--------------+-----------------------+------------------+
147 +--------------+-----------------------+------------------+
148 | DPDK version | Kernel driver version | Firmware version |
149 +==============+=======================+==================+
150 | 21.02 | 2.14.13 | 5.00 |
151 +--------------+-----------------------+------------------+
152 | 20.11 | 2.13.10 | 5.00 |
153 +--------------+-----------------------+------------------+
154 | 20.08 | 2.12.6 | 4.11 |
155 +--------------+-----------------------+------------------+
156 | 20.05 | 2.11.27 | 4.11 |
157 +--------------+-----------------------+------------------+
158 | 20.02 | 2.10.19 | 4.11 |
159 +--------------+-----------------------+------------------+
160 | 19.11 | 2.9.21 | 4.10 |
161 +--------------+-----------------------+------------------+
162 | 19.08 | 2.9.21 | 4.10 |
163 +--------------+-----------------------+------------------+
164 | 19.05 | 2.7.29 | 3.33 |
165 +--------------+-----------------------+------------------+
166 | 19.02 | 2.7.26 | 3.33 |
167 +--------------+-----------------------+------------------+
168 | 18.11 | 2.4.6 | 3.33 |
169 +--------------+-----------------------+------------------+
172 Pre-Installation Configuration
173 ------------------------------
178 The following options can be modified in the ``config/rte_config.h`` file.
180 - ``RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``)
182 Number of queues reserved for PF.
184 - ``RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``)
186 Number of queues reserved for each VMDQ Pool.
188 Runtime Config Options
189 ~~~~~~~~~~~~~~~~~~~~~~
191 - ``Reserved number of Queues per VF`` (default ``4``)
193 The number of reserved queue per VF is determined by its host PF. If the
194 PCI address of an i40e PF is aaaa:bb.cc, the number of reserved queues per
195 VF can be configured with EAL parameter like -a aaaa:bb.cc,queue-num-per-vf=n.
196 The value n can be 1, 2, 4, 8 or 16. If no such parameter is configured, the
197 number of reserved queues per VF is 4 by default. If VF request more than
198 reserved queues per VF, PF will able to allocate max to 16 queues after a VF
202 - ``Support multiple driver`` (default ``disable``)
204 There was a multiple driver support issue during use of 700 series Ethernet
205 Adapter with both Linux kernel and DPDK PMD. To fix this issue, ``devargs``
206 parameter ``support-multi-driver`` is introduced, for example::
208 -a 84:00.0,support-multi-driver=1
210 With the above configuration, DPDK PMD will not change global registers, and
211 will switch PF interrupt from IntN to Int0 to avoid interrupt conflict between
212 DPDK and Linux Kernel.
214 - ``Support VF Port Representor`` (default ``not enabled``)
216 The i40e PF PMD supports the creation of VF port representors for the control
217 and monitoring of i40e virtual function devices. Each port representor
218 corresponds to a single virtual function of that device. Using the ``devargs``
219 option ``representor`` the user can specify which virtual functions to create
220 port representors for on initialization of the PF PMD by passing the VF IDs of
221 the VFs which are required.::
223 -a DBDF,representor=[0,1,4]
225 Currently hot-plugging of representor ports is not supported so all required
226 representors must be specified on the creation of the PF.
228 - ``Enable validation for VF message`` (default ``not enabled``)
230 The PF counts messages from each VF. If in any period of seconds the message
231 statistic from a VF exceeds maximal limitation, the PF will ignore any new message
232 from that VF for some seconds.
233 Format -- "maximal-message@period-seconds:ignore-seconds"
236 -a 84:00.0,vf_msg_cfg=80@120:180
238 Vector RX Pre-conditions
239 ~~~~~~~~~~~~~~~~~~~~~~~~
240 For Vector RX it is assumed that the number of descriptor rings will be a power
241 of 2. With this pre-condition, the ring pointer can easily scroll back to the
242 head after hitting the tail without a conditional check. In addition Vector RX
243 can use this assumption to do a bit mask using ``ring_size - 1``.
245 Driver compilation and testing
246 ------------------------------
248 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
252 SR-IOV: Prerequisites and sample Application Notes
253 --------------------------------------------------
255 #. Load the kernel module:
257 .. code-block:: console
261 Check the output in dmesg:
263 .. code-block:: console
265 i40e 0000:83:00.1 ens802f0: renamed from eth0
267 #. Bring up the PF ports:
269 .. code-block:: console
273 #. Create VF device(s):
275 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry
280 .. code-block:: console
282 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs
285 #. Assign VF MAC address:
287 Assign MAC address to the VF using iproute2 utility. The syntax is:
289 .. code-block:: console
291 ip link set <PF netdev id> vf <VF id> mac <macaddr>
295 .. code-block:: console
297 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0
299 #. Assign VF to VM, and bring up the VM.
300 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*.
304 Follow instructions available in the document
305 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
310 .. code-block:: console
313 EAL: PCI device 0000:83:00.0 on NUMA socket 1
314 EAL: probe driver: 8086:1572 rte_i40e_pmd
315 EAL: PCI memory mapped at 0x7f7f80000000
316 EAL: PCI memory mapped at 0x7f7f80800000
317 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a
318 Interactive-mode selected
319 Configuring Port 0 (socket 0)
322 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are
323 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0.
326 Port 0: 68:05:CA:26:85:84
327 Checking link statuses...
328 Port 0 Link Up - speed 10000 Mbps - full-duplex
334 Sample Application Notes
335 ------------------------
340 Vlan filter only works when Promiscuous mode is off.
342 To start ``testpmd``, and add vlan 10 to port 0:
344 .. code-block:: console
346 ./<build_dir>/app/dpdk-testpmd -l 0-15 -n 4 -- -i --forward-mode=mac
349 testpmd> set promisc 0 off
350 testpmd> rx_vlan add 10 0
356 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
357 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.
359 The default input set of each flow type is::
361 ipv4-other : src_ip_address, dst_ip_address
362 ipv4-frag : src_ip_address, dst_ip_address
363 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port
364 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port
365 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
367 ipv6-other : src_ip_address, dst_ip_address
368 ipv6-frag : src_ip_address, dst_ip_address
369 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port
370 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port
371 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
373 l2_payload : ether_type
375 The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching.
377 Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``:
379 .. code-block:: console
381 ./<build_dir>/app/dpdk-testpmd -l 0-15 -n 4 -- -i --disable-rss \
382 --pkt-filter-mode=perfect --rxq=8 --txq=8 --nb-cores=8 \
385 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:
387 .. code-block:: console
389 testpmd> flow create 0 ingress pattern eth / ipv4 src is 2.2.2.3 \
390 dst is 2.2.2.5 / udp src is 32 dst is 32 / end \
391 actions mark id 1 / queue index 1 / end
393 Check the flow director status:
395 .. code-block:: console
397 testpmd> show port fdir 0
399 ######################## FDIR infos for port 0 ####################
401 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other
402 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other
405 max_len: 16 payload_limit: 480
406 payload_unit: 2 payload_seg: 3
407 bitmask_unit: 2 bitmask_num: 2
410 src_ipv4: 0x00000000,
411 dst_ipv4: 0x00000000,
414 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000,
415 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000
416 FLEX PAYLOAD SRC OFFSET:
417 L2_PAYLOAD: 0 1 2 3 4 5 6 ...
418 L3_PAYLOAD: 0 1 2 3 4 5 6 ...
419 L4_PAYLOAD: 0 1 2 3 4 5 6 ...
421 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
422 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
423 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
424 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
425 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
426 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
427 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
428 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
429 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
430 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
431 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
432 guarant_count: 1 best_count: 0
433 guarant_space: 512 best_space: 7168
443 The Intel® Ethernet 700 Series support a feature called
446 A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term
447 for functionality that allows local switching between virtual endpoints within
448 a physical endpoint and also with an external bridge/network.
450 A "Floating" VEB doesn't have an uplink connection to the outside world so all
451 switching is done internally and remains within the host. As such, this
452 feature provides security benefits.
454 In addition, a Floating VEB overcomes a limitation of normal VEBs where they
455 cannot forward packets when the physical link is down. Floating VEBs don't need
456 to connect to the NIC port so they can still forward traffic from VF to VF
457 even when the physical link is down.
459 Therefore, with this feature enabled VFs can be limited to communicating with
460 each other but not an outside network, and they can do so even when there is
461 no physical uplink on the associated NIC port.
463 To enable this feature, the user should pass a ``devargs`` parameter to the
466 -a 84:00.0,enable_floating_veb=1
468 In this configuration the PMD will use the floating VEB feature for all the
469 VFs created by this PF device.
471 Alternatively, the user can specify which VFs need to connect to this floating
472 VEB using the ``floating_veb_list`` argument::
474 -a 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4
476 In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB,
477 while other VFs connect to the normal VEB.
479 The current implementation only supports one floating VEB and one regular
480 VEB. VFs can connect to a floating VEB or a regular VEB according to the
481 configuration passed on the EAL command line.
483 The floating VEB functionality requires a NIC firmware version of 5.0
486 Dynamic Device Personalization (DDP)
487 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
489 The Intel® Ethernet 700 Series except for the Intel Ethernet Connection
490 X722 support a feature called "Dynamic Device Personalization (DDP)",
491 which is used to configure hardware by downloading a profile to support
492 protocols/filters which are not supported by default. The DDP
493 functionality requires a NIC firmware version of 6.0 or greater.
495 Current implementation supports GTP-C/GTP-U/PPPoE/PPPoL2TP/ESP,
496 steering can be used with rte_flow API.
498 GTPv1 package is released, and it can be downloaded from
499 https://downloadcenter.intel.com/download/27587.
501 PPPoE package is released, and it can be downloaded from
502 https://downloadcenter.intel.com/download/28040.
504 ESP-AH package is released, and it can be downloaded from
505 https://downloadcenter.intel.com/download/29446.
507 Load a profile which supports GTP and store backup profile:
509 .. code-block:: console
511 testpmd> ddp add 0 ./gtp.pkgo,./backup.pkgo
513 Delete a GTP profile and restore backup profile:
515 .. code-block:: console
517 testpmd> ddp del 0 ./backup.pkgo
519 Get loaded DDP package info list:
521 .. code-block:: console
523 testpmd> ddp get list 0
525 Display information about a GTP profile:
527 .. code-block:: console
529 testpmd> ddp get info ./gtp.pkgo
531 Input set configuration
532 ~~~~~~~~~~~~~~~~~~~~~~~
533 Input set for any PCTYPE can be configured with user defined configuration,
534 For example, to use only 48bit prefix for IPv6 src address for IPv6 TCP RSS:
536 .. code-block:: console
538 testpmd> port config 0 pctype 43 hash_inset clear all
539 testpmd> port config 0 pctype 43 hash_inset set field 13
540 testpmd> port config 0 pctype 43 hash_inset set field 14
541 testpmd> port config 0 pctype 43 hash_inset set field 15
543 Queue region configuration
544 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
545 The Intel® Ethernet 700 Series supports a feature of queue regions
546 configuration for RSS in the PF, so that different traffic classes or
547 different packet classification types can be separated to different
548 queues in different queue regions. There is an API for configuration
549 of queue regions in RSS with a command line. It can parse the parameters
550 of the region index, queue number, queue start index, user priority, traffic
551 classes and so on. Depending on commands from the command line, it will call
552 i40e private APIs and start the process of setting or flushing the queue
553 region configuration. As this feature is specific for i40e only private
554 APIs are used. These new ``test_pmd`` commands are as shown below. For
555 details please refer to :doc:`../testpmd_app_ug/index`.
557 .. code-block:: console
559 testpmd> set port (port_id) queue-region region_id (value) \
560 queue_start_index (value) queue_num (value)
561 testpmd> set port (port_id) queue-region region_id (value) flowtype (value)
562 testpmd> set port (port_id) queue-region UP (value) region_id (value)
563 testpmd> set port (port_id) queue-region flush (on|off)
564 testpmd> show port (port_id) queue-region
571 RSS Flow supports to set hash input set, hash function, enable hash
572 and configure queues.
574 Configure queues as queue 0, 1, 2, 3.
576 .. code-block:: console
578 testpmd> flow create 0 ingress pattern end actions rss types end \
579 queues 0 1 2 3 end / end
581 Enable hash and set input set for ipv4-tcp.
583 .. code-block:: console
585 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \
586 actions rss types ipv4-tcp l3-src-only end queues end / end
588 Set symmetric hash enable for flow type ipv4-tcp.
590 .. code-block:: console
592 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \
593 actions rss types ipv4-tcp end queues end func symmetric_toeplitz / end
595 Set hash function as simple xor.
597 .. code-block:: console
599 testpmd> flow create 0 ingress pattern end actions rss types end \
600 queues end func simple_xor / end
602 Limitations or Known issues
603 ---------------------------
605 MPLS packet classification
606 ~~~~~~~~~~~~~~~~~~~~~~~~~~
608 For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC.
609 The L2 Payload flow type in flow director can be used to classify MPLS packet
610 by using a command in testpmd like:
612 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \
613 0x8847 flexbytes () fwd pf queue <N> fd_id <M>
615 With the NIC firmware version 5.0 or greater, some limited MPLS support
616 is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no
617 new packet type, no classification or offload are possible. With this change,
618 L2 Payload flow type in flow director cannot be used to classify MPLS packet
619 as with previous firmware versions. Meanwhile, the Ethertype filter can be
620 used to classify MPLS packet by using a command in testpmd like:
622 testpmd> flow create 0 ingress pattern eth type is 0x8847 / end \
623 actions queue index <M> / end
625 16 Byte RX Descriptor setting on DPDK VF
626 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
628 Currently the VF's RX descriptor mode is decided by PF. There's no PF-VF
629 interface for VF to request the RX descriptor mode, also no interface to notify
630 VF its own RX descriptor mode.
631 For all available versions of the i40e driver, these drivers don't support 16
632 byte RX descriptor. If the Linux i40e kernel driver is used as host driver,
633 while DPDK i40e PMD is used as the VF driver, DPDK cannot choose 16 byte receive
634 descriptor. The reason is that the RX descriptor is already set to 32 byte by
635 the i40e kernel driver.
636 In the future, if the Linux i40e driver supports 16 byte RX descriptor, user
637 should make sure the DPDK VF uses the same RX descriptor mode, 16 byte or 32
638 byte, as the PF driver.
640 The same rule for DPDK PF + DPDK VF. The PF and VF should use the same RX
641 descriptor mode. Or the VF RX will not work.
643 Receive packets with Ethertype 0x88A8
644 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
646 Due to the FW limitation, PF can receive packets with Ethertype 0x88A8
647 only when floating VEB is disabled.
649 Incorrect Rx statistics when packet is oversize
650 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
652 When a packet is over maximum frame size, the packet is dropped.
653 However, the Rx statistics, when calling `rte_eth_stats_get` incorrectly
654 shows it as received.
656 RX/TX statistics may be incorrect when register overflowed
657 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
659 The rx_bytes/tx_bytes statistics register is 48 bit length.
660 Although this limitation is enlarged to 64 bit length on the software side,
661 but there is no way to detect if the overflow occurred more than once.
662 So rx_bytes/tx_bytes statistics data is correct when statistics are
663 updated at least once between two overflows.
665 VF & TC max bandwidth setting
666 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
668 The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel.
669 The behavior is different when handling per VF and per TC max bandwidth setting.
670 When enabling per VF max bandwidth, SW will check if per TC max bandwidth is
671 enabled. If so, return failure.
672 When enabling per TC max bandwidth, SW will check if per VF max bandwidth
673 is enabled. If so, disable per VF max bandwidth and continue with per TC max
676 TC TX scheduling mode setting
677 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
679 There are 2 TX scheduling modes for TCs, round robin and strict priority mode.
680 If a TC is set to strict priority mode, it can consume unlimited bandwidth.
681 It means if APP has set the max bandwidth for that TC, it comes to no
683 It's suggested to set the strict priority mode for a TC that is latency
684 sensitive but no consuming much bandwidth.
686 VF performance is impacted by PCI extended tag setting
687 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
689 To reach maximum NIC performance in the VF the PCI extended tag must be
690 enabled. The DPDK i40e PF driver will set this feature during initialization,
691 but the kernel PF driver does not. So when running traffic on a VF which is
692 managed by the kernel PF driver, a significant NIC performance downgrade has
693 been observed (for 64 byte packets, there is about 25% line-rate downgrade for
694 a 25GbE device and about 35% for a 40GbE device).
696 For kernel version >= 4.11, the kernel's PCI driver will enable the extended
697 tag if it detects that the device supports it. So by default, this is not an
698 issue. For kernels <= 4.11 or when the PCI extended tag is disabled it can be
699 enabled using the steps below.
701 #. Get the current value of the PCI configure register::
703 setpci -s <XX:XX.X> a8.w
707 value = value | 0x100
709 #. Set the PCI configure register with new value::
711 setpci -s <XX:XX.X> a8.w=<value>
716 The VF vlan strip function is only supported in the i40e kernel driver >= 2.1.26.
721 DCB works only when RSS is enabled.
723 Global configuration warning
724 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
726 I40E PMD will set some global registers to enable some function or set some
727 configure. Then when using different ports of the same NIC with Linux kernel
728 and DPDK, the port with Linux kernel will be impacted by the port with DPDK.
729 For example, register I40E_GL_SWT_L2TAGCTRL is used to control L2 tag, i40e
730 PMD uses I40E_GL_SWT_L2TAGCTRL to set vlan TPID. If setting TPID in port A
731 with DPDK, then the configuration will also impact port B in the NIC with
732 kernel driver, which don't want to use the TPID.
733 So PMD reports warning to clarify what is changed by writing global register.
738 When programming cloud filters for IPv4/6_UDP/TCP/SCTP with SRC port only or DST port only,
739 it will make any cloud filter using inner_vlan or tunnel key invalid. Default configuration will be
740 recovered only by NIC core reset.
742 Mirror rule limitation for X722
743 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
745 Due to firmware restriction of X722, the same VSI cannot have more than one mirror rule.
747 High Performance of Small Packets on 40GbE NIC
748 ----------------------------------------------
750 As there might be firmware fixes for performance enhancement in latest version
751 of firmware image, the firmware update might be needed for getting high performance.
752 Check the Intel support website for the latest firmware updates.
753 Users should consult the release notes specific to a DPDK release to identify
754 the validated firmware version for a NIC using the i40e driver.
756 Use 16 Bytes RX Descriptor Size
757 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
759 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.
760 In ``config/rte_config.h`` set the following to use 16 bytes size RX descriptors::
762 #define RTE_LIBRTE_I40E_16BYTE_RX_DESC 1
764 Input set requirement of each pctype for FDIR
765 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
767 Each PCTYPE can only have one specific FDIR input set at one time.
768 For example, if creating 2 rte_flow rules with different input set for one PCTYPE,
769 it will fail and return the info "Conflict with the first rule's input set",
770 which means the current rule's input set conflicts with the first rule's.
771 Remove the first rule if want to change the input set of the PCTYPE.
773 Example of getting best performance with l3fwd example
774 ------------------------------------------------------
776 The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with a
777 server with Intel Xeon processors and Intel Ethernet CNA XL710.
779 The example scenario is to get best performance with two Intel Ethernet CNA XL710 40GbE ports.
780 See :numref:`figure_intel_perf_test_setup` for the performance test setup.
782 .. _figure_intel_perf_test_setup:
784 .. figure:: img/intel_perf_test_setup.*
786 Performance Test Setup
789 1. Add two Intel Ethernet CNA XL710 to the platform, and use one port per card to get best performance.
790 The reason for using two NICs is to overcome a PCIe v3.0 limitation since it cannot provide 80GbE bandwidth
791 for two 40GbE ports, but two different PCIe v3.0 x8 slot can.
792 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports::
794 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
795 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
797 2. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator.
799 3. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id.
800 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
802 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
803 cores from different cores (e.g core18 and core19).
805 4. Bind these two ports to igb_uio.
807 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
808 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets.
810 6. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding.
811 Compile the ``l3fwd sample`` with the default lpm mode.
813 7. The command line of running l3fwd would be something like the following::
815 ./dpdk-l3fwd -l 18-21 -n 4 -a 82:00.0 -a 85:00.0 \
816 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)'
818 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding,
819 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding.
821 8. Configure the traffic at a traffic generator.
823 * Start creating a stream on packet generator.
825 * Set the Ethernet II type to 0x0800.
827 Tx bytes affected by the link status change
828 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
830 For firmware versions prior to 6.01 for X710 series and 3.33 for X722 series, the tx_bytes statistics data is affected by
831 the link down event. Each time the link status changes to down, the tx_bytes decreases 110 bytes.