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.
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 | 20.08 | 2.12.6 | 7.30 |
92 +--------------+-----------------------+------------------+
93 | 20.05 | 2.11.27 | 7.30 |
94 +--------------+-----------------------+------------------+
95 | 20.02 | 2.10.19 | 7.20 |
96 +--------------+-----------------------+------------------+
97 | 19.11 | 2.9.21 | 7.00 |
98 +--------------+-----------------------+------------------+
99 | 19.08 | 2.8.43 | 7.00 |
100 +--------------+-----------------------+------------------+
101 | 19.05 | 2.7.29 | 6.80 |
102 +--------------+-----------------------+------------------+
103 | 19.02 | 2.7.26 | 6.80 |
104 +--------------+-----------------------+------------------+
105 | 18.11 | 2.4.6 | 6.01 |
106 +--------------+-----------------------+------------------+
107 | 18.08 | 2.4.6 | 6.01 |
108 +--------------+-----------------------+------------------+
109 | 18.05 | 2.4.6 | 6.01 |
110 +--------------+-----------------------+------------------+
111 | 18.02 | 2.4.3 | 6.01 |
112 +--------------+-----------------------+------------------+
113 | 17.11 | 2.1.26 | 6.01 |
114 +--------------+-----------------------+------------------+
115 | 17.08 | 2.0.19 | 6.01 |
116 +--------------+-----------------------+------------------+
117 | 17.05 | 1.5.23 | 5.05 |
118 +--------------+-----------------------+------------------+
119 | 17.02 | 1.5.23 | 5.05 |
120 +--------------+-----------------------+------------------+
121 | 16.11 | 1.5.23 | 5.05 |
122 +--------------+-----------------------+------------------+
123 | 16.07 | 1.4.25 | 5.04 |
124 +--------------+-----------------------+------------------+
125 | 16.04 | 1.4.25 | 5.02 |
126 +--------------+-----------------------+------------------+
131 +--------------+-----------------------+------------------+
132 | DPDK version | Kernel driver version | Firmware version |
133 +==============+=======================+==================+
134 | 20.08 | 2.12.6 | 4.11 |
135 +--------------+-----------------------+------------------+
136 | 20.05 | 2.11.27 | 4.11 |
137 +--------------+-----------------------+------------------+
138 | 20.02 | 2.10.19 | 4.11 |
139 +--------------+-----------------------+------------------+
140 | 19.11 | 2.9.21 | 4.10 |
141 +--------------+-----------------------+------------------+
142 | 19.08 | 2.9.21 | 4.10 |
143 +--------------+-----------------------+------------------+
144 | 19.05 | 2.7.29 | 3.33 |
145 +--------------+-----------------------+------------------+
146 | 19.02 | 2.7.26 | 3.33 |
147 +--------------+-----------------------+------------------+
148 | 18.11 | 2.4.6 | 3.33 |
149 +--------------+-----------------------+------------------+
152 Pre-Installation Configuration
153 ------------------------------
158 The following options can be modified in the ``config/rte_config.h`` file.
160 - ``RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``)
162 Number of queues reserved for PF.
164 - ``RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``)
166 Number of queues reserved for each VMDQ Pool.
168 Runtime Config Options
169 ~~~~~~~~~~~~~~~~~~~~~~
171 - ``Reserved number of Queues per VF`` (default ``4``)
173 The number of reserved queue per VF is determined by its host PF. If the
174 PCI address of an i40e PF is aaaa:bb.cc, the number of reserved queues per
175 VF can be configured with EAL parameter like -w aaaa:bb.cc,queue-num-per-vf=n.
176 The value n can be 1, 2, 4, 8 or 16. If no such parameter is configured, the
177 number of reserved queues per VF is 4 by default. If VF request more than
178 reserved queues per VF, PF will able to allocate max to 16 queues after a VF
182 - ``Support multiple driver`` (default ``disable``)
184 There was a multiple driver support issue during use of 700 series Ethernet
185 Adapter with both Linux kernel and DPDK PMD. To fix this issue, ``devargs``
186 parameter ``support-multi-driver`` is introduced, for example::
188 -w 84:00.0,support-multi-driver=1
190 With the above configuration, DPDK PMD will not change global registers, and
191 will switch PF interrupt from IntN to Int0 to avoid interrupt conflict between
192 DPDK and Linux Kernel.
194 - ``Support VF Port Representor`` (default ``not enabled``)
196 The i40e PF PMD supports the creation of VF port representors for the control
197 and monitoring of i40e virtual function devices. Each port representor
198 corresponds to a single virtual function of that device. Using the ``devargs``
199 option ``representor`` the user can specify which virtual functions to create
200 port representors for on initialization of the PF PMD by passing the VF IDs of
201 the VFs which are required.::
203 -w DBDF,representor=[0,1,4]
205 Currently hot-plugging of representor ports is not supported so all required
206 representors must be specified on the creation of the PF.
208 - ``Use latest supported vector`` (default ``disable``)
210 Latest supported vector path may not always get the best perf so vector path was
211 recommended to use only on later platform. But users may want the latest vector path
212 since it can get better perf in some real work loading cases. So ``devargs`` param
213 ``use-latest-supported-vec`` is introduced, for example::
215 -w 84:00.0,use-latest-supported-vec=1
217 - ``Enable validation for VF message`` (default ``not enabled``)
219 The PF counts messages from each VF. If in any period of seconds the message
220 statistic from a VF exceeds maximal limitation, the PF will ignore any new message
221 from that VF for some seconds.
222 Format -- "maximal-message@period-seconds:ignore-seconds"
225 -w 84:00.0,vf_msg_cfg=80@120:180
227 Vector RX Pre-conditions
228 ~~~~~~~~~~~~~~~~~~~~~~~~
229 For Vector RX it is assumed that the number of descriptor rings will be a power
230 of 2. With this pre-condition, the ring pointer can easily scroll back to the
231 head after hitting the tail without a conditional check. In addition Vector RX
232 can use this assumption to do a bit mask using ``ring_size - 1``.
234 Driver compilation and testing
235 ------------------------------
237 Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
241 SR-IOV: Prerequisites and sample Application Notes
242 --------------------------------------------------
244 #. Load the kernel module:
246 .. code-block:: console
250 Check the output in dmesg:
252 .. code-block:: console
254 i40e 0000:83:00.1 ens802f0: renamed from eth0
256 #. Bring up the PF ports:
258 .. code-block:: console
262 #. Create VF device(s):
264 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry
269 .. code-block:: console
271 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs
274 #. Assign VF MAC address:
276 Assign MAC address to the VF using iproute2 utility. The syntax is:
278 .. code-block:: console
280 ip link set <PF netdev id> vf <VF id> mac <macaddr>
284 .. code-block:: console
286 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0
288 #. Assign VF to VM, and bring up the VM.
289 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*.
293 Follow instructions available in the document
294 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`
299 .. code-block:: console
302 EAL: PCI device 0000:83:00.0 on NUMA socket 1
303 EAL: probe driver: 8086:1572 rte_i40e_pmd
304 EAL: PCI memory mapped at 0x7f7f80000000
305 EAL: PCI memory mapped at 0x7f7f80800000
306 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a
307 Interactive-mode selected
308 Configuring Port 0 (socket 0)
311 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are
312 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0.
315 Port 0: 68:05:CA:26:85:84
316 Checking link statuses...
317 Port 0 Link Up - speed 10000 Mbps - full-duplex
323 Sample Application Notes
324 ------------------------
329 Vlan filter only works when Promiscuous mode is off.
331 To start ``testpmd``, and add vlan 10 to port 0:
333 .. code-block:: console
335 ./<build_dir>/app/dpdk-testpmd -l 0-15 -n 4 -- -i --forward-mode=mac
338 testpmd> set promisc 0 off
339 testpmd> rx_vlan add 10 0
345 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
346 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.
348 The default input set of each flow type is::
350 ipv4-other : src_ip_address, dst_ip_address
351 ipv4-frag : src_ip_address, dst_ip_address
352 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port
353 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port
354 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
356 ipv6-other : src_ip_address, dst_ip_address
357 ipv6-frag : src_ip_address, dst_ip_address
358 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port
359 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port
360 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port,
362 l2_payload : ether_type
364 The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching.
366 Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``:
368 .. code-block:: console
370 ./<build_dir>/app/dpdk-testpmd -l 0-15 -n 4 -- -i --disable-rss \
371 --pkt-filter-mode=perfect --rxq=8 --txq=8 --nb-cores=8 \
374 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:
376 .. code-block:: console
378 testpmd> flow create 0 ingress pattern eth / ipv4 src is 2.2.2.3 \
379 dst is 2.2.2.5 / udp src is 32 dst is 32 / end \
380 actions mark id 1 / queue index 1 / end
382 Check the flow director status:
384 .. code-block:: console
386 testpmd> show port fdir 0
388 ######################## FDIR infos for port 0 ####################
390 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other
391 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other
394 max_len: 16 payload_limit: 480
395 payload_unit: 2 payload_seg: 3
396 bitmask_unit: 2 bitmask_num: 2
399 src_ipv4: 0x00000000,
400 dst_ipv4: 0x00000000,
403 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000,
404 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000
405 FLEX PAYLOAD SRC OFFSET:
406 L2_PAYLOAD: 0 1 2 3 4 5 6 ...
407 L3_PAYLOAD: 0 1 2 3 4 5 6 ...
408 L4_PAYLOAD: 0 1 2 3 4 5 6 ...
410 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
411 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
412 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
413 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
414 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
415 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
416 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
417 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
418 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
419 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
420 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
421 guarant_count: 1 best_count: 0
422 guarant_space: 512 best_space: 7168
432 The Intel® Ethernet 700 Series support a feature called
435 A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term
436 for functionality that allows local switching between virtual endpoints within
437 a physical endpoint and also with an external bridge/network.
439 A "Floating" VEB doesn't have an uplink connection to the outside world so all
440 switching is done internally and remains within the host. As such, this
441 feature provides security benefits.
443 In addition, a Floating VEB overcomes a limitation of normal VEBs where they
444 cannot forward packets when the physical link is down. Floating VEBs don't need
445 to connect to the NIC port so they can still forward traffic from VF to VF
446 even when the physical link is down.
448 Therefore, with this feature enabled VFs can be limited to communicating with
449 each other but not an outside network, and they can do so even when there is
450 no physical uplink on the associated NIC port.
452 To enable this feature, the user should pass a ``devargs`` parameter to the
455 -w 84:00.0,enable_floating_veb=1
457 In this configuration the PMD will use the floating VEB feature for all the
458 VFs created by this PF device.
460 Alternatively, the user can specify which VFs need to connect to this floating
461 VEB using the ``floating_veb_list`` argument::
463 -w 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4
465 In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB,
466 while other VFs connect to the normal VEB.
468 The current implementation only supports one floating VEB and one regular
469 VEB. VFs can connect to a floating VEB or a regular VEB according to the
470 configuration passed on the EAL command line.
472 The floating VEB functionality requires a NIC firmware version of 5.0
475 Dynamic Device Personalization (DDP)
476 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
478 The Intel® Ethernet 700 Series except for the Intel Ethernet Connection
479 X722 support a feature called "Dynamic Device Personalization (DDP)",
480 which is used to configure hardware by downloading a profile to support
481 protocols/filters which are not supported by default. The DDP
482 functionality requires a NIC firmware version of 6.0 or greater.
484 Current implementation supports GTP-C/GTP-U/PPPoE/PPPoL2TP/ESP,
485 steering can be used with rte_flow API.
487 GTPv1 package is released, and it can be downloaded from
488 https://downloadcenter.intel.com/download/27587.
490 PPPoE package is released, and it can be downloaded from
491 https://downloadcenter.intel.com/download/28040.
493 ESP-AH package is released, and it can be downloaded from
494 https://downloadcenter.intel.com/download/29446.
496 Load a profile which supports GTP and store backup profile:
498 .. code-block:: console
500 testpmd> ddp add 0 ./gtp.pkgo,./backup.pkgo
502 Delete a GTP profile and restore backup profile:
504 .. code-block:: console
506 testpmd> ddp del 0 ./backup.pkgo
508 Get loaded DDP package info list:
510 .. code-block:: console
512 testpmd> ddp get list 0
514 Display information about a GTP profile:
516 .. code-block:: console
518 testpmd> ddp get info ./gtp.pkgo
520 Input set configuration
521 ~~~~~~~~~~~~~~~~~~~~~~~
522 Input set for any PCTYPE can be configured with user defined configuration,
523 For example, to use only 48bit prefix for IPv6 src address for IPv6 TCP RSS:
525 .. code-block:: console
527 testpmd> port config 0 pctype 43 hash_inset clear all
528 testpmd> port config 0 pctype 43 hash_inset set field 13
529 testpmd> port config 0 pctype 43 hash_inset set field 14
530 testpmd> port config 0 pctype 43 hash_inset set field 15
532 Queue region configuration
533 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
534 The Intel® Ethernet 700 Series supports a feature of queue regions
535 configuration for RSS in the PF, so that different traffic classes or
536 different packet classification types can be separated to different
537 queues in different queue regions. There is an API for configuration
538 of queue regions in RSS with a command line. It can parse the parameters
539 of the region index, queue number, queue start index, user priority, traffic
540 classes and so on. Depending on commands from the command line, it will call
541 i40e private APIs and start the process of setting or flushing the queue
542 region configuration. As this feature is specific for i40e only private
543 APIs are used. These new ``test_pmd`` commands are as shown below. For
544 details please refer to :doc:`../testpmd_app_ug/index`.
546 .. code-block:: console
548 testpmd> set port (port_id) queue-region region_id (value) \
549 queue_start_index (value) queue_num (value)
550 testpmd> set port (port_id) queue-region region_id (value) flowtype (value)
551 testpmd> set port (port_id) queue-region UP (value) region_id (value)
552 testpmd> set port (port_id) queue-region flush (on|off)
553 testpmd> show port (port_id) queue-region
560 RSS Flow supports to set hash input set, hash function, enable hash
561 and configure queue region.
563 Configure queue region as queue 0, 1, 2, 3.
565 .. code-block:: console
567 testpmd> flow create 0 ingress pattern end actions rss types end \
568 queues 0 1 2 3 end / end
570 Enable hash and set input set for ipv4-tcp.
572 .. code-block:: console
574 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \
575 actions rss types ipv4-tcp l3-src-only end queues end / end
577 Set symmetric hash enable for flow type ipv4-tcp.
579 .. code-block:: console
581 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \
582 actions rss types ipv4-tcp end queues end func symmetric_toeplitz / end
584 Set hash function as simple xor.
586 .. code-block:: console
588 testpmd> flow create 0 ingress pattern end actions rss types end \
589 queues end func simple_xor / end
591 Limitations or Known issues
592 ---------------------------
594 MPLS packet classification
595 ~~~~~~~~~~~~~~~~~~~~~~~~~~
597 For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC.
598 The L2 Payload flow type in flow director can be used to classify MPLS packet
599 by using a command in testpmd like:
601 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \
602 0x8847 flexbytes () fwd pf queue <N> fd_id <M>
604 With the NIC firmware version 5.0 or greater, some limited MPLS support
605 is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no
606 new packet type, no classification or offload are possible. With this change,
607 L2 Payload flow type in flow director cannot be used to classify MPLS packet
608 as with previous firmware versions. Meanwhile, the Ethertype filter can be
609 used to classify MPLS packet by using a command in testpmd like:
611 testpmd> flow create 0 ingress pattern eth type is 0x8847 / end \
612 actions queue index <M> / end
614 16 Byte RX Descriptor setting on DPDK VF
615 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
617 Currently the VF's RX descriptor mode is decided by PF. There's no PF-VF
618 interface for VF to request the RX descriptor mode, also no interface to notify
619 VF its own RX descriptor mode.
620 For all available versions of the i40e driver, these drivers don't support 16
621 byte RX descriptor. If the Linux i40e kernel driver is used as host driver,
622 while DPDK i40e PMD is used as the VF driver, DPDK cannot choose 16 byte receive
623 descriptor. The reason is that the RX descriptor is already set to 32 byte by
624 the i40e kernel driver.
625 In the future, if the Linux i40e driver supports 16 byte RX descriptor, user
626 should make sure the DPDK VF uses the same RX descriptor mode, 16 byte or 32
627 byte, as the PF driver.
629 The same rule for DPDK PF + DPDK VF. The PF and VF should use the same RX
630 descriptor mode. Or the VF RX will not work.
632 Receive packets with Ethertype 0x88A8
633 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
635 Due to the FW limitation, PF can receive packets with Ethertype 0x88A8
636 only when floating VEB is disabled.
638 Incorrect Rx statistics when packet is oversize
639 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
641 When a packet is over maximum frame size, the packet is dropped.
642 However, the Rx statistics, when calling `rte_eth_stats_get` incorrectly
643 shows it as received.
645 RX/TX statistics may be incorrect when register overflowed
646 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
648 The rx_bytes/tx_bytes statistics register is 48 bit length.
649 Although this limitation is enlarged to 64 bit length on the software side,
650 but there is no way to detect if the overflow occurred more than once.
651 So rx_bytes/tx_bytes statistics data is correct when statistics are
652 updated at least once between two overflows.
654 VF & TC max bandwidth setting
655 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
657 The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel.
658 The behavior is different when handling per VF and per TC max bandwidth setting.
659 When enabling per VF max bandwidth, SW will check if per TC max bandwidth is
660 enabled. If so, return failure.
661 When enabling per TC max bandwidth, SW will check if per VF max bandwidth
662 is enabled. If so, disable per VF max bandwidth and continue with per TC max
665 TC TX scheduling mode setting
666 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
668 There are 2 TX scheduling modes for TCs, round robin and strict priority mode.
669 If a TC is set to strict priority mode, it can consume unlimited bandwidth.
670 It means if APP has set the max bandwidth for that TC, it comes to no
672 It's suggested to set the strict priority mode for a TC that is latency
673 sensitive but no consuming much bandwidth.
675 VF performance is impacted by PCI extended tag setting
676 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
678 To reach maximum NIC performance in the VF the PCI extended tag must be
679 enabled. The DPDK i40e PF driver will set this feature during initialization,
680 but the kernel PF driver does not. So when running traffic on a VF which is
681 managed by the kernel PF driver, a significant NIC performance downgrade has
682 been observed (for 64 byte packets, there is about 25% line-rate downgrade for
683 a 25GbE device and about 35% for a 40GbE device).
685 For kernel version >= 4.11, the kernel's PCI driver will enable the extended
686 tag if it detects that the device supports it. So by default, this is not an
687 issue. For kernels <= 4.11 or when the PCI extended tag is disabled it can be
688 enabled using the steps below.
690 #. Get the current value of the PCI configure register::
692 setpci -s <XX:XX.X> a8.w
696 value = value | 0x100
698 #. Set the PCI configure register with new value::
700 setpci -s <XX:XX.X> a8.w=<value>
705 The VF vlan strip function is only supported in the i40e kernel driver >= 2.1.26.
710 DCB works only when RSS is enabled.
712 Global configuration warning
713 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
715 I40E PMD will set some global registers to enable some function or set some
716 configure. Then when using different ports of the same NIC with Linux kernel
717 and DPDK, the port with Linux kernel will be impacted by the port with DPDK.
718 For example, register I40E_GL_SWT_L2TAGCTRL is used to control L2 tag, i40e
719 PMD uses I40E_GL_SWT_L2TAGCTRL to set vlan TPID. If setting TPID in port A
720 with DPDK, then the configuration will also impact port B in the NIC with
721 kernel driver, which don't want to use the TPID.
722 So PMD reports warning to clarify what is changed by writing global register.
727 When programming cloud filters for IPv4/6_UDP/TCP/SCTP with SRC port only or DST port only,
728 it will make any cloud filter using inner_vlan or tunnel key invalid. Default configuration will be
729 recovered only by NIC core reset.
731 High Performance of Small Packets on 40GbE NIC
732 ----------------------------------------------
734 As there might be firmware fixes for performance enhancement in latest version
735 of firmware image, the firmware update might be needed for getting high performance.
736 Check the Intel support website for the latest firmware updates.
737 Users should consult the release notes specific to a DPDK release to identify
738 the validated firmware version for a NIC using the i40e driver.
740 Use 16 Bytes RX Descriptor Size
741 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
743 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.
744 In ``config/rte_config.h`` set the following to use 16 bytes size RX descriptors::
746 #define RTE_LIBRTE_I40E_16BYTE_RX_DESC 1
748 Input set requirement of each pctype for FDIR
749 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
751 Each PCTYPE can only have one specific FDIR input set at one time.
752 For example, if creating 2 rte_flow rules with different input set for one PCTYPE,
753 it will fail and return the info "Conflict with the first rule's input set",
754 which means the current rule's input set conflicts with the first rule's.
755 Remove the first rule if want to change the input set of the PCTYPE.
757 Example of getting best performance with l3fwd example
758 ------------------------------------------------------
760 The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with a
761 server with Intel Xeon processors and Intel Ethernet CNA XL710.
763 The example scenario is to get best performance with two Intel Ethernet CNA XL710 40GbE ports.
764 See :numref:`figure_intel_perf_test_setup` for the performance test setup.
766 .. _figure_intel_perf_test_setup:
768 .. figure:: img/intel_perf_test_setup.*
770 Performance Test Setup
773 1. Add two Intel Ethernet CNA XL710 to the platform, and use one port per card to get best performance.
774 The reason for using two NICs is to overcome a PCIe v3.0 limitation since it cannot provide 80GbE bandwidth
775 for two 40GbE ports, but two different PCIe v3.0 x8 slot can.
776 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports::
778 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
779 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583]
781 2. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator.
783 3. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id.
784 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
786 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
787 cores from different cores (e.g core18 and core19).
789 4. Bind these two ports to igb_uio.
791 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
792 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets.
794 6. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding.
795 Compile the ``l3fwd sample`` with the default lpm mode.
797 7. The command line of running l3fwd would be something like the following::
799 ./dpdk-l3fwd -l 18-21 -n 4 -w 82:00.0 -w 85:00.0 \
800 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)'
802 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding,
803 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding.
805 8. Configure the traffic at a traffic generator.
807 * Start creating a stream on packet generator.
809 * Set the Ethernet II type to 0x0800.
811 Tx bytes affected by the link status change
812 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
814 For firmware versions prior to 6.01 for X710 series and 3.33 for X722 series, the tx_bytes statistics data is affected by
815 the link down event. Each time the link status changes to down, the tx_bytes decreases 110 bytes.