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
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.
78 Recommended Matching List
79 -------------------------
81 It is highly recommended to upgrade the i40e kernel driver and firmware to
82 avoid the compatibility issues with i40e PMD. Here is the suggested matching
83 list which has been tested and verified. The detailed information can refer
84 to chapter Tested Platforms/Tested NICs in release notes.
86 For X710/XL710/XXV710,
88 +--------------+-----------------------+------------------+
89 | DPDK version | Kernel driver version | Firmware version |
90 +==============+=======================+==================+
91 | 20.05 | 2.11.27 | 7.30 |
92 +--------------+-----------------------+------------------+
93 | 20.02 | 2.10.19 | 7.20 |
94 +--------------+-----------------------+------------------+
95 | 19.11 | 2.9.21 | 7.00 |
96 +--------------+-----------------------+------------------+
97 | 19.08 | 2.8.43 | 7.00 |
98 +--------------+-----------------------+------------------+
99 | 19.05 | 2.7.29 | 6.80 |
100 +--------------+-----------------------+------------------+
101 | 19.02 | 2.7.26 | 6.80 |
102 +--------------+-----------------------+------------------+
103 | 18.11 | 2.4.6 | 6.01 |
104 +--------------+-----------------------+------------------+
105 | 18.08 | 2.4.6 | 6.01 |
106 +--------------+-----------------------+------------------+
107 | 18.05 | 2.4.6 | 6.01 |
108 +--------------+-----------------------+------------------+
109 | 18.02 | 2.4.3 | 6.01 |
110 +--------------+-----------------------+------------------+
111 | 17.11 | 2.1.26 | 6.01 |
112 +--------------+-----------------------+------------------+
113 | 17.08 | 2.0.19 | 6.01 |
114 +--------------+-----------------------+------------------+
115 | 17.05 | 1.5.23 | 5.05 |
116 +--------------+-----------------------+------------------+
117 | 17.02 | 1.5.23 | 5.05 |
118 +--------------+-----------------------+------------------+
119 | 16.11 | 1.5.23 | 5.05 |
120 +--------------+-----------------------+------------------+
121 | 16.07 | 1.4.25 | 5.04 |
122 +--------------+-----------------------+------------------+
123 | 16.04 | 1.4.25 | 5.02 |
124 +--------------+-----------------------+------------------+
129 +--------------+-----------------------+------------------+
130 | DPDK version | Kernel driver version | Firmware version |
131 +==============+=======================+==================+
132 | 20.05 | 2.11.27 | 4.11 |
133 +--------------+-----------------------+------------------+
134 | 20.02 | 2.10.19 | 4.11 |
135 +--------------+-----------------------+------------------+
136 | 19.11 | 2.9.21 | 4.10 |
137 +--------------+-----------------------+------------------+
138 | 19.08 | 2.9.21 | 4.10 |
139 +--------------+-----------------------+------------------+
140 | 19.05 | 2.7.29 | 3.33 |
141 +--------------+-----------------------+------------------+
142 | 19.02 | 2.7.26 | 3.33 |
143 +--------------+-----------------------+------------------+
144 | 18.11 | 2.4.6 | 3.33 |
145 +--------------+-----------------------+------------------+
148 Pre-Installation Configuration
149 ------------------------------
154 The following options can be modified in the ``config`` file.
155 Please note that enabling debugging options may affect system performance.
157 - ``CONFIG_RTE_LIBRTE_I40E_PMD`` (default ``y``)
159 Toggle compilation of the ``librte_pmd_i40e`` driver.
161 - ``CONFIG_RTE_LIBRTE_I40E_DEBUG_*`` (default ``n``)
163 Toggle display of generic debugging messages.
165 - ``CONFIG_RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC`` (default ``y``)
167 Toggle bulk allocation for RX.
169 - ``CONFIG_RTE_LIBRTE_I40E_INC_VECTOR`` (default ``n``)
171 Toggle the use of Vector PMD instead of normal RX/TX path.
172 To enable vPMD for RX, bulk allocation for Rx must be allowed.
174 - ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` (default ``n``)
176 Toggle to use a 16-byte RX descriptor, by default the RX descriptor is 32 byte.
178 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``)
180 Number of queues reserved for PF.
182 - ``CONFIG_RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``)
184 Number of queues reserved for each VMDQ Pool.
186 Runtime Config Options
187 ~~~~~~~~~~~~~~~~~~~~~~
189 - ``Reserved number of Queues per VF`` (default ``4``)
191 The number of reserved queue per VF is determined by its host PF. If the
192 PCI address of an i40e PF is aaaa:bb.cc, the number of reserved queues per
193 VF can be configured with EAL parameter like -w aaaa:bb.cc,queue-num-per-vf=n.
194 The value n can be 1, 2, 4, 8 or 16. If no such parameter is configured, the
195 number of reserved queues per VF is 4 by default. If VF request more than
196 reserved queues per VF, PF will able to allocate max to 16 queues after a VF
200 - ``Support multiple driver`` (default ``disable``)
202 There was a multiple driver support issue during use of 700 series Ethernet
203 Adapter with both Linux kernel and DPDK PMD. To fix this issue, ``devargs``
204 parameter ``support-multi-driver`` is introduced, for example::
206 -w 84:00.0,support-multi-driver=1
208 With the above configuration, DPDK PMD will not change global registers, and
209 will switch PF interrupt from IntN to Int0 to avoid interrupt conflict between
210 DPDK and Linux Kernel.
212 - ``Support VF Port Representor`` (default ``not enabled``)
214 The i40e PF PMD supports the creation of VF port representors for the control
215 and monitoring of i40e virtual function devices. Each port representor
216 corresponds to a single virtual function of that device. Using the ``devargs``
217 option ``representor`` the user can specify which virtual functions to create
218 port representors for on initialization of the PF PMD by passing the VF IDs of
219 the VFs which are required.::
221 -w DBDF,representor=[0,1,4]
223 Currently hot-plugging of representor ports is not supported so all required
224 representors must be specified on the creation of the PF.
226 - ``Use latest supported vector`` (default ``disable``)
228 Latest supported vector path may not always get the best perf so vector path was
229 recommended to use only on later platform. But users may want the latest vector path
230 since it can get better perf in some real work loading cases. So ``devargs`` param
231 ``use-latest-supported-vec`` is introduced, for example::
233 -w 84:00.0,use-latest-supported-vec=1
235 - ``Enable validation for VF message`` (default ``not enabled``)
237 The PF counts messages from each VF. If in any period of seconds the message
238 statistic from a VF exceeds maximal limitation, the PF will ignore any new message
239 from that VF for some seconds.
240 Format -- "maximal-message@period-seconds:ignore-seconds"
243 -w 84:00.0,vf_msg_cfg=80@120:180
245 Vector RX Pre-conditions
246 ~~~~~~~~~~~~~~~~~~~~~~~~
247 For Vector RX it is assumed that the number of descriptor rings will be a power
248 of 2. With this pre-condition, the ring pointer can easily scroll back to the
249 head after hitting the tail without a conditional check. In addition Vector RX
250 can use this assumption to do a bit mask using ``ring_size - 1``.
252 Driver compilation and testing
253 ------------------------------
255 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
259 SR-IOV: Prerequisites and sample Application Notes
260 --------------------------------------------------
262 #. Load the kernel module:
264 .. code-block:: console
268 Check the output in dmesg:
270 .. code-block:: console
272 i40e 0000:83:00.1 ens802f0: renamed from eth0
274 #. Bring up the PF ports:
276 .. code-block:: console
280 #. Create VF device(s):
282 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry
287 .. code-block:: console
289 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs
292 #. Assign VF MAC address:
294 Assign MAC address to the VF using iproute2 utility. The syntax is:
296 .. code-block:: console
298 ip link set <PF netdev id> vf <VF id> mac <macaddr>
302 .. code-block:: console
304 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0
306 #. Assign VF to VM, and bring up the VM.
307 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*.
311 Follow instructions available in the document
312 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
317 .. code-block:: console
320 EAL: PCI device 0000:83:00.0 on NUMA socket 1
321 EAL: probe driver: 8086:1572 rte_i40e_pmd
322 EAL: PCI memory mapped at 0x7f7f80000000
323 EAL: PCI memory mapped at 0x7f7f80800000
324 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a
325 Interactive-mode selected
326 Configuring Port 0 (socket 0)
329 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are
330 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0.
333 Port 0: 68:05:CA:26:85:84
334 Checking link statuses...
335 Port 0 Link Up - speed 10000 Mbps - full-duplex
341 Sample Application Notes
342 ------------------------
347 Vlan filter only works when Promiscuous mode is off.
349 To start ``testpmd``, and add vlan 10 to port 0:
351 .. code-block:: console
353 ./app/testpmd -l 0-15 -n 4 -- -i --forward-mode=mac
356 testpmd> set promisc 0 off
357 testpmd> rx_vlan add 10 0
363 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
364 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.
366 The default input set of each flow type is::
368 ipv4-other : src_ip_address, dst_ip_address
369 ipv4-frag : src_ip_address, dst_ip_address
370 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port
371 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port
372 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
374 ipv6-other : src_ip_address, dst_ip_address
375 ipv6-frag : src_ip_address, dst_ip_address
376 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port
377 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port
378 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
380 l2_payload : ether_type
382 The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching.
384 Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``:
386 .. code-block:: console
388 ./app/testpmd -l 0-15 -n 4 -- -i --disable-rss --pkt-filter-mode=perfect \
389 --rxq=8 --txq=8 --nb-cores=8 --nb-ports=1
391 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:
393 .. code-block:: console
395 testpmd> flow_director_filter 0 mode IP add flow ipv4-udp \
396 src 2.2.2.3 32 dst 2.2.2.5 32 vlan 0 flexbytes () \
397 fwd pf queue 1 fd_id 1
399 Check the flow director status:
401 .. code-block:: console
403 testpmd> show port fdir 0
405 ######################## FDIR infos for port 0 ####################
407 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other
408 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other
411 max_len: 16 payload_limit: 480
412 payload_unit: 2 payload_seg: 3
413 bitmask_unit: 2 bitmask_num: 2
416 src_ipv4: 0x00000000,
417 dst_ipv4: 0x00000000,
420 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000,
421 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000
422 FLEX PAYLOAD SRC OFFSET:
423 L2_PAYLOAD: 0 1 2 3 4 5 6 ...
424 L3_PAYLOAD: 0 1 2 3 4 5 6 ...
425 L4_PAYLOAD: 0 1 2 3 4 5 6 ...
427 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
428 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
429 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
430 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
431 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
432 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
433 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
434 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
435 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
436 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
437 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
438 guarant_count: 1 best_count: 0
439 guarant_space: 512 best_space: 7168
446 Delete all flow director rules on a port:
448 .. code-block:: console
450 testpmd> flush_flow_director 0
455 The Intel® Ethernet 700 Series support a feature called
458 A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term
459 for functionality that allows local switching between virtual endpoints within
460 a physical endpoint and also with an external bridge/network.
462 A "Floating" VEB doesn't have an uplink connection to the outside world so all
463 switching is done internally and remains within the host. As such, this
464 feature provides security benefits.
466 In addition, a Floating VEB overcomes a limitation of normal VEBs where they
467 cannot forward packets when the physical link is down. Floating VEBs don't need
468 to connect to the NIC port so they can still forward traffic from VF to VF
469 even when the physical link is down.
471 Therefore, with this feature enabled VFs can be limited to communicating with
472 each other but not an outside network, and they can do so even when there is
473 no physical uplink on the associated NIC port.
475 To enable this feature, the user should pass a ``devargs`` parameter to the
478 -w 84:00.0,enable_floating_veb=1
480 In this configuration the PMD will use the floating VEB feature for all the
481 VFs created by this PF device.
483 Alternatively, the user can specify which VFs need to connect to this floating
484 VEB using the ``floating_veb_list`` argument::
486 -w 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4
488 In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB,
489 while other VFs connect to the normal VEB.
491 The current implementation only supports one floating VEB and one regular
492 VEB. VFs can connect to a floating VEB or a regular VEB according to the
493 configuration passed on the EAL command line.
495 The floating VEB functionality requires a NIC firmware version of 5.0
498 Dynamic Device Personalization (DDP)
499 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
501 The Intel® Ethernet 700 Series except for the Intel Ethernet Connection
502 X722 support a feature called "Dynamic Device Personalization (DDP)",
503 which is used to configure hardware by downloading a profile to support
504 protocols/filters which are not supported by default. The DDP
505 functionality requires a NIC firmware version of 6.0 or greater.
507 Current implementation supports GTP-C/GTP-U/PPPoE/PPPoL2TP/ESP,
508 steering can be used with rte_flow API.
510 GTPv1 package is released, and it can be downloaded from
511 https://downloadcenter.intel.com/download/27587.
513 PPPoE package is released, and it can be downloaded from
514 https://downloadcenter.intel.com/download/28040.
516 ESP-AH package is released, and it can be downloaded from
517 https://downloadcenter.intel.com/download/29446.
519 Load a profile which supports GTP and store backup profile:
521 .. code-block:: console
523 testpmd> ddp add 0 ./gtp.pkgo,./backup.pkgo
525 Delete a GTP profile and restore backup profile:
527 .. code-block:: console
529 testpmd> ddp del 0 ./backup.pkgo
531 Get loaded DDP package info list:
533 .. code-block:: console
535 testpmd> ddp get list 0
537 Display information about a GTP profile:
539 .. code-block:: console
541 testpmd> ddp get info ./gtp.pkgo
543 Input set configuration
544 ~~~~~~~~~~~~~~~~~~~~~~~
545 Input set for any PCTYPE can be configured with user defined configuration,
546 For example, to use only 48bit prefix for IPv6 src address for IPv6 TCP RSS:
548 .. code-block:: console
550 testpmd> port config 0 pctype 43 hash_inset clear all
551 testpmd> port config 0 pctype 43 hash_inset set field 13
552 testpmd> port config 0 pctype 43 hash_inset set field 14
553 testpmd> port config 0 pctype 43 hash_inset set field 15
555 Queue region configuration
556 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
557 The Intel® Ethernet 700 Series supports a feature of queue regions
558 configuration for RSS in the PF, so that different traffic classes or
559 different packet classification types can be separated to different
560 queues in different queue regions. There is an API for configuration
561 of queue regions in RSS with a command line. It can parse the parameters
562 of the region index, queue number, queue start index, user priority, traffic
563 classes and so on. Depending on commands from the command line, it will call
564 i40e private APIs and start the process of setting or flushing the queue
565 region configuration. As this feature is specific for i40e only private
566 APIs are used. These new ``test_pmd`` commands are as shown below. For
567 details please refer to :doc:`../testpmd_app_ug/index`.
569 .. code-block:: console
571 testpmd> set port (port_id) queue-region region_id (value) \
572 queue_start_index (value) queue_num (value)
573 testpmd> set port (port_id) queue-region region_id (value) flowtype (value)
574 testpmd> set port (port_id) queue-region UP (value) region_id (value)
575 testpmd> set port (port_id) queue-region flush (on|off)
576 testpmd> show port (port_id) queue-region
583 RSS Flow supports to set hash input set, hash function, enable hash
584 and configure queue region.
586 Configure queue region as queue 0, 1, 2, 3.
588 .. code-block:: console
590 testpmd> flow create 0 ingress pattern end actions rss types end \
591 queues 0 1 2 3 end / end
593 Enable hash and set input set for ipv4-tcp.
595 .. code-block:: console
597 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \
598 actions rss types ipv4-tcp l3-src-only end queues end / end
600 Set symmetric hash enable for flow type ipv4-tcp.
602 .. code-block:: console
604 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \
605 actions rss types ipv4-tcp end queues end func symmetric_toeplitz / end
607 Set hash function as simple xor.
609 .. code-block:: console
611 testpmd> flow create 0 ingress pattern end actions rss types end \
612 queues end func simple_xor / end
614 Limitations or Known issues
615 ---------------------------
617 MPLS packet classification
618 ~~~~~~~~~~~~~~~~~~~~~~~~~~
620 For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC.
621 The L2 Payload flow type in flow director can be used to classify MPLS packet
622 by using a command in testpmd like:
624 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \
625 0x8847 flexbytes () fwd pf queue <N> fd_id <M>
627 With the NIC firmware version 5.0 or greater, some limited MPLS support
628 is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no
629 new packet type, no classification or offload are possible. With this change,
630 L2 Payload flow type in flow director cannot be used to classify MPLS packet
631 as with previous firmware versions. Meanwhile, the Ethertype filter can be
632 used to classify MPLS packet by using a command in testpmd like:
634 testpmd> ethertype_filter 0 add mac_ignr 00:00:00:00:00:00 ethertype \
637 16 Byte RX Descriptor setting on DPDK VF
638 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
640 Currently the VF's RX descriptor mode is decided by PF. There's no PF-VF
641 interface for VF to request the RX descriptor mode, also no interface to notify
642 VF its own RX descriptor mode.
643 For all available versions of the i40e driver, these drivers don't support 16
644 byte RX descriptor. If the Linux i40e kernel driver is used as host driver,
645 while DPDK i40e PMD is used as the VF driver, DPDK cannot choose 16 byte receive
646 descriptor. The reason is that the RX descriptor is already set to 32 byte by
647 the i40e kernel driver. That is to say, user should keep
648 ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC=n`` in config file.
649 In the future, if the Linux i40e driver supports 16 byte RX descriptor, user
650 should make sure the DPDK VF uses the same RX descriptor mode, 16 byte or 32
651 byte, as the PF driver.
653 The same rule for DPDK PF + DPDK VF. The PF and VF should use the same RX
654 descriptor mode. Or the VF RX will not work.
656 Receive packets with Ethertype 0x88A8
657 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
659 Due to the FW limitation, PF can receive packets with Ethertype 0x88A8
660 only when floating VEB is disabled.
662 Incorrect Rx statistics when packet is oversize
663 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
665 When a packet is over maximum frame size, the packet is dropped.
666 However, the Rx statistics, when calling `rte_eth_stats_get` incorrectly
667 shows it as received.
669 VF & TC max bandwidth setting
670 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
672 The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel.
673 The behavior is different when handling per VF and per TC max bandwidth setting.
674 When enabling per VF max bandwidth, SW will check if per TC max bandwidth is
675 enabled. If so, return failure.
676 When enabling per TC max bandwidth, SW will check if per VF max bandwidth
677 is enabled. If so, disable per VF max bandwidth and continue with per TC max
680 TC TX scheduling mode setting
681 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
683 There are 2 TX scheduling modes for TCs, round robin and strict priority mode.
684 If a TC is set to strict priority mode, it can consume unlimited bandwidth.
685 It means if APP has set the max bandwidth for that TC, it comes to no
687 It's suggested to set the strict priority mode for a TC that is latency
688 sensitive but no consuming much bandwidth.
690 VF performance is impacted by PCI extended tag setting
691 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
693 To reach maximum NIC performance in the VF the PCI extended tag must be
694 enabled. The DPDK i40e PF driver will set this feature during initialization,
695 but the kernel PF driver does not. So when running traffic on a VF which is
696 managed by the kernel PF driver, a significant NIC performance downgrade has
697 been observed (for 64 byte packets, there is about 25% line-rate downgrade for
698 a 25GbE device and about 35% for a 40GbE device).
700 For kernel version >= 4.11, the kernel's PCI driver will enable the extended
701 tag if it detects that the device supports it. So by default, this is not an
702 issue. For kernels <= 4.11 or when the PCI extended tag is disabled it can be
703 enabled using the steps below.
705 #. Get the current value of the PCI configure register::
707 setpci -s <XX:XX.X> a8.w
711 value = value | 0x100
713 #. Set the PCI configure register with new value::
715 setpci -s <XX:XX.X> a8.w=<value>
720 The VF vlan strip function is only supported in the i40e kernel driver >= 2.1.26.
725 DCB works only when RSS is enabled.
727 Global configuration warning
728 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
730 I40E PMD will set some global registers to enable some function or set some
731 configure. Then when using different ports of the same NIC with Linux kernel
732 and DPDK, the port with Linux kernel will be impacted by the port with DPDK.
733 For example, register I40E_GL_SWT_L2TAGCTRL is used to control L2 tag, i40e
734 PMD uses I40E_GL_SWT_L2TAGCTRL to set vlan TPID. If setting TPID in port A
735 with DPDK, then the configuration will also impact port B in the NIC with
736 kernel driver, which don't want to use the TPID.
737 So PMD reports warning to clarify what is changed by writing global register.
739 High Performance of Small Packets on 40GbE NIC
740 ----------------------------------------------
742 As there might be firmware fixes for performance enhancement in latest version
743 of firmware image, the firmware update might be needed for getting high performance.
744 Check the Intel support website for the latest firmware updates.
745 Users should consult the release notes specific to a DPDK release to identify
746 the validated firmware version for a NIC using the i40e driver.
748 Use 16 Bytes RX Descriptor Size
749 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
751 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.
752 Configuration of ``CONFIG_RTE_LIBRTE_I40E_16BYTE_RX_DESC`` in config files can be changed to use 16 bytes size RX descriptors.
754 Input set requirement of each pctype for FDIR
755 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
757 Each PCTYPE can only have one specific FDIR input set at one time.
758 For example, if creating 2 rte_flow rules with different input set for one PCTYPE,
759 it will fail and return the info "Conflict with the first rule's input set",
760 which means the current rule's input set conflicts with the first rule's.
761 Remove the first rule if want to change the input set of the PCTYPE.
763 Example of getting best performance with l3fwd example
764 ------------------------------------------------------
766 The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with a
767 server with Intel Xeon processors and Intel Ethernet CNA XL710.
769 The example scenario is to get best performance with two Intel Ethernet CNA XL710 40GbE ports.
770 See :numref:`figure_intel_perf_test_setup` for the performance test setup.
772 .. _figure_intel_perf_test_setup:
774 .. figure:: img/intel_perf_test_setup.*
776 Performance Test Setup
779 1. Add two Intel Ethernet CNA XL710 to the platform, and use one port per card to get best performance.
780 The reason for using two NICs is to overcome a PCIe v3.0 limitation since it cannot provide 80GbE bandwidth
781 for two 40GbE ports, but two different PCIe v3.0 x8 slot can.
782 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports::
784 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
785 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
787 2. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator.
789 3. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id.
790 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
792 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
793 cores from different cores (e.g core18 and core19).
795 4. Bind these two ports to igb_uio.
797 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
798 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets.
800 6. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding.
801 Compile the ``l3fwd sample`` with the default lpm mode.
803 7. The command line of running l3fwd would be something like the following::
805 ./l3fwd -l 18-21 -n 4 -w 82:00.0 -w 85:00.0 \
806 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)'
808 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding,
809 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding.
811 8. Configure the traffic at a traffic generator.
813 * Start creating a stream on packet generator.
815 * Set the Ethernet II type to 0x0800.
817 Tx bytes affected by the link status change
818 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
820 For firmware versions prior to 6.01 for X710 series and 3.33 for X722 series, the tx_bytes statistics data is affected by
821 the link down event. Each time the link status changes to down, the tx_bytes decreases 110 bytes.