1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright(c) 2016 Intel Corporation. 3 4I40E Poll Mode Driver 5====================== 6 7The i40e PMD (**librte_net_i40e**) provides poll mode driver support for 810/25/40 Gbps Intel® Ethernet 700 Series Network Adapters based on 9the Intel Ethernet Controller X710/XL710/XXV710 and Intel Ethernet 10Connection X722 (only support part of features). 11 12 13Features 14-------- 15 16Features of the i40e PMD are: 17 18- Multiple queues for TX and RX 19- Receiver Side Scaling (RSS) 20- MAC/VLAN filtering 21- Packet type information 22- Flow director 23- Cloud filter 24- Checksum offload 25- VLAN/QinQ stripping and inserting 26- TSO offload 27- Promiscuous mode 28- Multicast mode 29- Port hardware statistics 30- Jumbo frames 31- Link state information 32- Link flow control 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 37- DCB 38- VMDQ 39- SR-IOV VF 40- Hot plug 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 47- Generic flow API 48 49Linux Prerequisites 50------------------- 51 52- Identifying your adapter using `Intel Support 53 <http://www.intel.com/support>`_ and get the latest NVM/FW images. 54 55- Follow the DPDK :ref:`Getting Started Guide for Linux <linux_gsg>` to setup the basic DPDK environment. 56 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>`. 59 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. 62 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>`_. 65 66 .. Note:: 67 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. 71 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. 74 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. 77 78Windows Prerequisites 79--------------------- 80 81- Follow the :doc:`guide for Windows <../windows_gsg/run_apps>` 82 to setup the basic DPDK environment. 83 84- Identify the Intel® Ethernet adapter and get the latest NVM/FW version. 85 86- To access any Intel® Ethernet hardware, load the NetUIO driver in place of existing built-in (inbox) driver. 87 88- To load NetUIO driver, follow the steps mentioned in `dpdk-kmods repository 89 <https://git.dpdk.org/dpdk-kmods/tree/windows/netuio/README.rst>`_. 90 91Kernel driver and Firmware Matching List 92---------------------------------------- 93 94It is highly recommended to upgrade the i40e kernel driver and firmware 95to avoid the compatibility issues with i40e PMD. 96The table below shows a summary of the DPDK versions 97with corresponding out-of-tree Linux kernel drivers and firmware. 98The full list of in-tree and out-of-tree Linux kernel drivers from kernel.org 99and Linux distributions that were tested and verified 100are listed in the Tested Platforms section of the Release Notes for each release. 101 102For X710/XL710/XXV710, 103 104 +--------------+-----------------------+------------------+ 105 | DPDK version | Kernel driver version | Firmware version | 106 +==============+=======================+==================+ 107 | 24.11 | 2.26.8 | 9.52 | 108 +--------------+-----------------------+------------------+ 109 | 24.07 | 2.25.9 | 9.50 | 110 +--------------+-----------------------+------------------+ 111 | 24.03 | 2.24.6 | 9.40 | 112 +--------------+-----------------------+------------------+ 113 | 23.11 | 2.23.17 | 9.30 | 114 +--------------+-----------------------+------------------+ 115 | 23.07 | 2.22.20 | 9.20 | 116 +--------------+-----------------------+------------------+ 117 | 23.03 | 2.22.18 | 9.20 | 118 +--------------+-----------------------+------------------+ 119 | 22.11 | 2.20.12 | 9.01 | 120 +--------------+-----------------------+------------------+ 121 | 22.07 | 2.19.3 | 8.70 | 122 +--------------+-----------------------+------------------+ 123 | 22.03 | 2.17.15 | 8.30 | 124 +--------------+-----------------------+------------------+ 125 | 21.11 | 2.17.4 | 8.30 | 126 +--------------+-----------------------+------------------+ 127 | 21.08 | 2.15.9 | 8.30 | 128 +--------------+-----------------------+------------------+ 129 | 21.05 | 2.15.9 | 8.30 | 130 +--------------+-----------------------+------------------+ 131 | 21.02 | 2.14.13 | 8.00 | 132 +--------------+-----------------------+------------------+ 133 | 20.11 | 2.14.13 | 8.00 | 134 +--------------+-----------------------+------------------+ 135 | 20.08 | 2.12.6 | 7.30 | 136 +--------------+-----------------------+------------------+ 137 | 20.05 | 2.11.27 | 7.30 | 138 +--------------+-----------------------+------------------+ 139 | 20.02 | 2.10.19 | 7.20 | 140 +--------------+-----------------------+------------------+ 141 | 19.11 | 2.9.21 | 7.00 | 142 +--------------+-----------------------+------------------+ 143 | 19.08 | 2.8.43 | 7.00 | 144 +--------------+-----------------------+------------------+ 145 | 19.05 | 2.7.29 | 6.80 | 146 +--------------+-----------------------+------------------+ 147 | 19.02 | 2.7.26 | 6.80 | 148 +--------------+-----------------------+------------------+ 149 | 18.11 | 2.4.6 | 6.01 | 150 +--------------+-----------------------+------------------+ 151 | 18.08 | 2.4.6 | 6.01 | 152 +--------------+-----------------------+------------------+ 153 | 18.05 | 2.4.6 | 6.01 | 154 +--------------+-----------------------+------------------+ 155 | 18.02 | 2.4.3 | 6.01 | 156 +--------------+-----------------------+------------------+ 157 | 17.11 | 2.1.26 | 6.01 | 158 +--------------+-----------------------+------------------+ 159 | 17.08 | 2.0.19 | 6.01 | 160 +--------------+-----------------------+------------------+ 161 | 17.05 | 1.5.23 | 5.05 | 162 +--------------+-----------------------+------------------+ 163 | 17.02 | 1.5.23 | 5.05 | 164 +--------------+-----------------------+------------------+ 165 | 16.11 | 1.5.23 | 5.05 | 166 +--------------+-----------------------+------------------+ 167 | 16.07 | 1.4.25 | 5.04 | 168 +--------------+-----------------------+------------------+ 169 | 16.04 | 1.4.25 | 5.02 | 170 +--------------+-----------------------+------------------+ 171 172 173For X722, 174 175 +--------------+-----------------------+------------------+ 176 | DPDK version | Kernel driver version | Firmware version | 177 +==============+=======================+==================+ 178 | 24.11 | 2.26.8 | 6.50 | 179 +--------------+-----------------------+------------------+ 180 | 24.07 | 2.25.9 | 6.50 | 181 +--------------+-----------------------+------------------+ 182 | 24.03 | 2.24.6 | 6.20 | 183 +--------------+-----------------------+------------------+ 184 | 23.11 | 2.23.17 | 6.20 | 185 +--------------+-----------------------+------------------+ 186 | 23.07 | 2.22.20 | 6.20 | 187 +--------------+-----------------------+------------------+ 188 | 23.03 | 2.22.18 | 6.20 | 189 +--------------+-----------------------+------------------+ 190 | 22.11 | 2.20.12 | 6.00 | 191 +--------------+-----------------------+------------------+ 192 | 22.07 | 2.19.3 | 5.60 | 193 +--------------+-----------------------+------------------+ 194 | 22.03 | 2.17.15 | 5.50 | 195 +--------------+-----------------------+------------------+ 196 | 21.11 | 2.17.4 | 5.30 | 197 +--------------+-----------------------+------------------+ 198 | 21.08 | 2.15.9 | 5.30 | 199 +--------------+-----------------------+------------------+ 200 | 21.05 | 2.15.9 | 5.30 | 201 +--------------+-----------------------+------------------+ 202 | 21.02 | 2.14.13 | 5.00 | 203 +--------------+-----------------------+------------------+ 204 | 20.11 | 2.13.10 | 5.00 | 205 +--------------+-----------------------+------------------+ 206 | 20.08 | 2.12.6 | 4.11 | 207 +--------------+-----------------------+------------------+ 208 | 20.05 | 2.11.27 | 4.11 | 209 +--------------+-----------------------+------------------+ 210 | 20.02 | 2.10.19 | 4.11 | 211 +--------------+-----------------------+------------------+ 212 | 19.11 | 2.9.21 | 4.10 | 213 +--------------+-----------------------+------------------+ 214 | 19.08 | 2.9.21 | 4.10 | 215 +--------------+-----------------------+------------------+ 216 | 19.05 | 2.7.29 | 3.33 | 217 +--------------+-----------------------+------------------+ 218 | 19.02 | 2.7.26 | 3.33 | 219 +--------------+-----------------------+------------------+ 220 | 18.11 | 2.4.6 | 3.33 | 221 +--------------+-----------------------+------------------+ 222 223 224Configuration 225------------- 226 227Compilation Options 228~~~~~~~~~~~~~~~~~~~ 229 230The following options can be modified in the ``config/rte_config.h`` file. 231 232- ``RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF`` (default ``64``) 233 234 Number of queues reserved for PF. 235 236- ``RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM`` (default ``4``) 237 238 Number of queues reserved for each VMDQ Pool. 239 240Runtime Configuration 241~~~~~~~~~~~~~~~~~~~~~ 242 243- ``Reserved number of Queues per VF`` (default ``4``) 244 245 The number of reserved queue per VF is determined by its host PF. If the 246 PCI address of an i40e PF is aaaa:bb.cc, the number of reserved queues per 247 VF can be configured with EAL parameter like -a aaaa:bb.cc,queue-num-per-vf=n. 248 The value n can be 1, 2, 4, 8 or 16. If no such parameter is configured, the 249 number of reserved queues per VF is 4 by default. If VF request more than 250 reserved queues per VF, PF will able to allocate max to 16 queues after a VF 251 reset. 252 253 254- ``Support multiple driver`` (default ``disable``) 255 256 There was a multiple driver support issue during use of 700 series Ethernet 257 Adapter with both Linux kernel and DPDK PMD. To fix this issue, ``devargs`` 258 parameter ``support-multi-driver`` is introduced, for example:: 259 260 -a 84:00.0,support-multi-driver=1 261 262 With the above configuration, DPDK PMD will not change global registers, and 263 will switch PF interrupt from IntN to Int0 to avoid interrupt conflict between 264 DPDK and Linux Kernel. 265 266- ``Support VF Port Representor`` (default ``not enabled``) 267 268 The i40e PF PMD supports the creation of VF port representors for the control 269 and monitoring of i40e virtual function devices. Each port representor 270 corresponds to a single virtual function of that device. Using the ``devargs`` 271 option ``representor`` the user can specify which virtual functions to create 272 port representors for on initialization of the PF PMD by passing the VF IDs of 273 the VFs which are required.:: 274 275 -a DBDF,representor=[0,1,4] 276 277 Currently hot-plugging of representor ports is not supported so all required 278 representors must be specified on the creation of the PF. 279 280- ``Enable validation for VF message`` (default ``not enabled``) 281 282 The PF counts messages from each VF. If in any period of seconds the message 283 statistic from a VF exceeds maximal limitation, the PF will ignore any new message 284 from that VF for some seconds. 285 Format -- "maximal-message@period-seconds:ignore-seconds" 286 For example:: 287 288 -a 84:00.0,vf_msg_cfg=80@120:180 289 290- ``Support Tx diagnostics`` (default ``not enabled``) 291 292 Set the ``devargs`` parameter ``mbuf_check`` to enable Tx diagnostics. 293 For example, ``-a 18:01.0,mbuf_check=<case>`` or ``-a 18:01.0,mbuf_check=[<case1>,<case2>...]``. 294 Also, ``xstats_get`` can be used to get the error counts, 295 which are collected in ``tx_mbuf_error_packets`` xstats. 296 For example, to show the statistics in testpmd, use: ``testpmd> show port xstats all``. 297 Supported values for the ``<case>`` parameter: 298 299 * ``mbuf``: Check for corrupted mbuf. 300 * ``size``: Check min/max packet length according to HW spec. 301 * ``segment``: Check number of mbuf segments not exceed hw limitation. 302 * ``offload``: Check any unsupported offload flag. 303 304Vector RX Pre-conditions 305~~~~~~~~~~~~~~~~~~~~~~~~ 306For Vector RX it is assumed that the number of descriptor rings will be a power 307of 2. With this pre-condition, the ring pointer can easily scroll back to the 308head after hitting the tail without a conditional check. In addition Vector RX 309can use this assumption to do a bit mask using ``ring_size - 1``. 310 311Driver compilation and testing 312------------------------------ 313 314Refer to the document :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>` 315for details. 316 317 318SR-IOV: Prerequisites and sample Application Notes 319-------------------------------------------------- 320 321#. Load the kernel module: 322 323 .. code-block:: console 324 325 modprobe i40e 326 327 Check the output in dmesg: 328 329 .. code-block:: console 330 331 i40e 0000:83:00.1 ens802f0: renamed from eth0 332 333#. Bring up the PF ports: 334 335 .. code-block:: console 336 337 ifconfig ens802f0 up 338 339#. Create VF device(s): 340 341 Echo the number of VFs to be created into the ``sriov_numvfs`` sysfs entry 342 of the parent PF. 343 344 Example: 345 346 .. code-block:: console 347 348 echo 2 > /sys/devices/pci0000:00/0000:00:03.0/0000:81:00.0/sriov_numvfs 349 350 351#. Assign VF MAC address: 352 353 Assign MAC address to the VF using iproute2 utility. The syntax is: 354 355 .. code-block:: console 356 357 ip link set <PF netdev id> vf <VF id> mac <macaddr> 358 359 Example: 360 361 .. code-block:: console 362 363 ip link set ens802f0 vf 0 mac a0:b0:c0:d0:e0:f0 364 365#. Assign VF to VM, and bring up the VM. 366 Please see the documentation for the *I40E/IXGBE/IGB Virtual Function Driver*. 367 368#. Running testpmd: 369 370 Follow instructions available in the document 371 :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>` 372 to run testpmd. 373 374 Example output: 375 376 .. code-block:: console 377 378 ... 379 EAL: PCI device 0000:83:00.0 on NUMA socket 1 380 EAL: probe driver: 8086:1572 rte_i40e_pmd 381 EAL: PCI memory mapped at 0x7f7f80000000 382 EAL: PCI memory mapped at 0x7f7f80800000 383 PMD: eth_i40e_dev_init(): FW 5.0 API 1.5 NVM 05.00.02 eetrack 8000208a 384 Interactive-mode selected 385 Configuring Port 0 (socket 0) 386 ... 387 388 PMD: i40e_dev_rx_queue_setup(): Rx Burst Bulk Alloc Preconditions are 389 satisfied.Rx Burst Bulk Alloc function will be used on port=0, queue=0. 390 391 ... 392 Port 0: 68:05:CA:26:85:84 393 Checking link statuses... 394 Port 0 Link Up - speed 10000 Mbps - full-duplex 395 Done 396 397 testpmd> 398 399 400Sample Application Notes 401------------------------ 402 403Vlan filter 404~~~~~~~~~~~ 405 406Vlan filter only works when Promiscuous mode is off. 407 408To start ``testpmd``, and add vlan 10 to port 0: 409 410.. code-block:: console 411 412 ./<build_dir>/app/dpdk-testpmd -l 0-15 -n 4 -- -i --forward-mode=mac 413 ... 414 415 testpmd> set promisc 0 off 416 testpmd> rx_vlan add 10 0 417 418 419Flow Director 420~~~~~~~~~~~~~ 421 422The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues. 423The 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. 424 425The default input set of each flow type is:: 426 427 ipv4-other : src_ip_address, dst_ip_address 428 ipv4-frag : src_ip_address, dst_ip_address 429 ipv4-tcp : src_ip_address, dst_ip_address, src_port, dst_port 430 ipv4-udp : src_ip_address, dst_ip_address, src_port, dst_port 431 ipv4-sctp : src_ip_address, dst_ip_address, src_port, dst_port, 432 verification_tag 433 ipv6-other : src_ip_address, dst_ip_address 434 ipv6-frag : src_ip_address, dst_ip_address 435 ipv6-tcp : src_ip_address, dst_ip_address, src_port, dst_port 436 ipv6-udp : src_ip_address, dst_ip_address, src_port, dst_port 437 ipv6-sctp : src_ip_address, dst_ip_address, src_port, dst_port, 438 verification_tag 439 l2_payload : ether_type 440 441The flex payload is selected from offset 0 to 15 of packet's payload by default, while it is masked out from matching. 442 443Start ``testpmd`` with ``--disable-rss`` and ``--pkt-filter-mode=perfect``: 444 445.. code-block:: console 446 447 ./<build_dir>/app/dpdk-testpmd -l 0-15 -n 4 -- -i --disable-rss \ 448 --pkt-filter-mode=perfect --rxq=8 --txq=8 --nb-cores=8 \ 449 --nb-ports=1 450 451Add 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: 452 453.. code-block:: console 454 455 testpmd> flow create 0 ingress pattern eth / ipv4 src is 2.2.2.3 \ 456 dst is 2.2.2.5 / udp src is 32 dst is 32 / end \ 457 actions mark id 1 / queue index 1 / end 458 459Check the flow director status: 460 461.. code-block:: console 462 463 testpmd> show port fdir 0 464 465 ######################## FDIR infos for port 0 #################### 466 MODE: PERFECT 467 SUPPORTED FLOW TYPE: ipv4-frag ipv4-tcp ipv4-udp ipv4-sctp ipv4-other 468 ipv6-frag ipv6-tcp ipv6-udp ipv6-sctp ipv6-other 469 l2_payload 470 FLEX PAYLOAD INFO: 471 max_len: 16 payload_limit: 480 472 payload_unit: 2 payload_seg: 3 473 bitmask_unit: 2 bitmask_num: 2 474 MASK: 475 vlan_tci: 0x0000, 476 src_ipv4: 0x00000000, 477 dst_ipv4: 0x00000000, 478 src_port: 0x0000, 479 dst_port: 0x0000 480 src_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000, 481 dst_ipv6: 0x00000000,0x00000000,0x00000000,0x00000000 482 FLEX PAYLOAD SRC OFFSET: 483 L2_PAYLOAD: 0 1 2 3 4 5 6 ... 484 L3_PAYLOAD: 0 1 2 3 4 5 6 ... 485 L4_PAYLOAD: 0 1 2 3 4 5 6 ... 486 FLEX MASK CFG: 487 ipv4-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 488 ipv4-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 489 ipv4-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 490 ipv4-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 491 ipv4-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 492 ipv6-udp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 493 ipv6-tcp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 494 ipv6-sctp: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 495 ipv6-other: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 496 ipv6-frag: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 497 l2_payload: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 498 guarant_count: 1 best_count: 0 499 guarant_space: 512 best_space: 7168 500 collision: 0 free: 0 501 maxhash: 0 maxlen: 0 502 add: 0 remove: 0 503 f_add: 0 f_remove: 0 504 505 506Floating VEB 507~~~~~~~~~~~~~ 508 509The Intel® Ethernet 700 Series support a feature called 510"Floating VEB". 511 512A Virtual Ethernet Bridge (VEB) is an IEEE Edge Virtual Bridging (EVB) term 513for functionality that allows local switching between virtual endpoints within 514a physical endpoint and also with an external bridge/network. 515 516A "Floating" VEB doesn't have an uplink connection to the outside world so all 517switching is done internally and remains within the host. As such, this 518feature provides security benefits. 519 520In addition, a Floating VEB overcomes a limitation of normal VEBs where they 521cannot forward packets when the physical link is down. Floating VEBs don't need 522to connect to the NIC port so they can still forward traffic from VF to VF 523even when the physical link is down. 524 525Therefore, with this feature enabled VFs can be limited to communicating with 526each other but not an outside network, and they can do so even when there is 527no physical uplink on the associated NIC port. 528 529To enable this feature, the user should pass a ``devargs`` parameter to the 530EAL, for example:: 531 532 -a 84:00.0,enable_floating_veb=1 533 534In this configuration the PMD will use the floating VEB feature for all the 535VFs created by this PF device. 536 537Alternatively, the user can specify which VFs need to connect to this floating 538VEB using the ``floating_veb_list`` argument:: 539 540 -a 84:00.0,enable_floating_veb=1,floating_veb_list=1;3-4 541 542In this example ``VF1``, ``VF3`` and ``VF4`` connect to the floating VEB, 543while other VFs connect to the normal VEB. 544 545The current implementation only supports one floating VEB and one regular 546VEB. VFs can connect to a floating VEB or a regular VEB according to the 547configuration passed on the EAL command line. 548 549The floating VEB functionality requires a NIC firmware version of 5.0 550or greater. 551 552Dynamic Device Personalization (DDP) 553~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 554 555The Intel® Ethernet 700 Series except for the Intel Ethernet Connection 556X722 support a feature called "Dynamic Device Personalization (DDP)", 557which is used to configure hardware by downloading a profile to support 558protocols/filters which are not supported by default. The DDP 559functionality requires a NIC firmware version of 6.0 or greater. 560 561Current implementation supports GTP-C/GTP-U/PPPoE/PPPoL2TP/ESP, 562steering can be used with rte_flow API. 563 564GTPv1 package is released, and it can be downloaded from 565https://downloadcenter.intel.com/download/27587. 566 567PPPoE package is released, and it can be downloaded from 568https://downloadcenter.intel.com/download/28040. 569 570ESP-AH package is released, and it can be downloaded from 571https://downloadcenter.intel.com/download/29446. 572 573Load a profile which supports GTP and store backup profile: 574 575.. code-block:: console 576 577 testpmd> ddp add 0 ./gtp.pkgo,./backup.pkgo 578 579Delete a GTP profile and restore backup profile: 580 581.. code-block:: console 582 583 testpmd> ddp del 0 ./backup.pkgo 584 585Get loaded DDP package info list: 586 587.. code-block:: console 588 589 testpmd> ddp get list 0 590 591Display information about a GTP profile: 592 593.. code-block:: console 594 595 testpmd> ddp get info ./gtp.pkgo 596 597Input set configuration 598~~~~~~~~~~~~~~~~~~~~~~~ 599Input set for any PCTYPE can be configured with user defined configuration, 600For example, to use only 48bit prefix for IPv6 src address for IPv6 TCP RSS: 601 602.. code-block:: console 603 604 testpmd> port config 0 pctype 43 hash_inset clear all 605 testpmd> port config 0 pctype 43 hash_inset set field 13 606 testpmd> port config 0 pctype 43 hash_inset set field 14 607 testpmd> port config 0 pctype 43 hash_inset set field 15 608 609Queue region configuration 610~~~~~~~~~~~~~~~~~~~~~~~~~~~ 611The Intel® Ethernet 700 Series supports a feature of queue regions 612configuration for RSS in the PF, so that different traffic classes or 613different packet classification types can be separated to different 614queues in different queue regions. There is an API for configuration 615of queue regions in RSS with a command line. It can parse the parameters 616of the region index, queue number, queue start index, user priority, traffic 617classes and so on. Depending on commands from the command line, it will call 618i40e private APIs and start the process of setting or flushing the queue 619region configuration. As this feature is specific for i40e only private 620APIs are used. 621 622.. code-block:: console 623 624 testpmd> set port (port_id) queue-region region_id (value) \ 625 queue_start_index (value) queue_num (value) 626 testpmd> set port (port_id) queue-region region_id (value) flowtype (value) 627 testpmd> set port (port_id) queue-region UP (value) region_id (value) 628 testpmd> set port (port_id) queue-region flush (on|off) 629 testpmd> show port (port_id) queue-region 630 631Generic flow API 632~~~~~~~~~~~~~~~~~~~ 633 634- ``RSS Flow`` 635 636 RSS Flow supports to set hash input set, hash function, enable hash 637 and configure queues. 638 For example: 639 Configure queues as queue 0, 1, 2, 3. 640 641 .. code-block:: console 642 643 testpmd> flow create 0 ingress pattern end actions rss types end \ 644 queues 0 1 2 3 end / end 645 646 Enable hash and set input set for ipv4-tcp. 647 648 .. code-block:: console 649 650 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \ 651 actions rss types ipv4-tcp l3-src-only end queues end / end 652 653 Set symmetric hash enable for flow type ipv4-tcp. 654 655 .. code-block:: console 656 657 testpmd> flow create 0 ingress pattern eth / ipv4 / tcp / end \ 658 actions rss types ipv4-tcp end queues end func symmetric_toeplitz / end 659 660 Set hash function as simple xor. 661 662 .. code-block:: console 663 664 testpmd> flow create 0 ingress pattern end actions rss types end \ 665 queues end func simple_xor / end 666 667Limitations or Known issues 668--------------------------- 669 670MPLS packet classification 671~~~~~~~~~~~~~~~~~~~~~~~~~~ 672 673For firmware versions prior to 5.0, MPLS packets are not recognized by the NIC. 674The L2 Payload flow type in flow director can be used to classify MPLS packet 675by using a command in testpmd like: 676 677 testpmd> flow_director_filter 0 mode IP add flow l2_payload ether \ 678 0x8847 flexbytes () fwd pf queue <N> fd_id <M> 679 680With the NIC firmware version 5.0 or greater, some limited MPLS support 681is added: Native MPLS (MPLS in Ethernet) skip is implemented, while no 682new packet type, no classification or offload are possible. With this change, 683L2 Payload flow type in flow director cannot be used to classify MPLS packet 684as with previous firmware versions. Meanwhile, the Ethertype filter can be 685used to classify MPLS packet by using a command in testpmd like: 686 687 testpmd> flow create 0 ingress pattern eth type is 0x8847 / end \ 688 actions queue index <M> / end 689 690Receive packets with Ethertype 0x88A8 691~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 692 693Due to the FW limitation, PF can receive packets with Ethertype 0x88A8 694only when floating VEB is disabled. 695 696Incorrect Rx statistics when packet is oversize 697~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 698 699When a packet is over maximum frame size, the packet is dropped. 700However, the Rx statistics, when calling `rte_eth_stats_get` incorrectly 701shows it as received. 702 703RX/TX statistics may be incorrect when register overflowed 704~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 705 706The rx_bytes/tx_bytes statistics register is 48 bit length. 707Although this limitation is enlarged to 64 bit length on the software side, 708but there is no way to detect if the overflow occurred more than once. 709So rx_bytes/tx_bytes statistics data is correct when statistics are 710updated at least once between two overflows. 711 712VF & TC max bandwidth setting 713~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 714 715The per VF max bandwidth and per TC max bandwidth cannot be enabled in parallel. 716The behavior is different when handling per VF and per TC max bandwidth setting. 717When enabling per VF max bandwidth, SW will check if per TC max bandwidth is 718enabled. If so, return failure. 719When enabling per TC max bandwidth, SW will check if per VF max bandwidth 720is enabled. If so, disable per VF max bandwidth and continue with per TC max 721bandwidth setting. 722 723TC TX scheduling mode setting 724~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 725 726There are 2 TX scheduling modes for TCs, round robin and strict priority mode. 727If a TC is set to strict priority mode, it can consume unlimited bandwidth. 728It means if APP has set the max bandwidth for that TC, it comes to no 729effect. 730It's suggested to set the strict priority mode for a TC that is latency 731sensitive but no consuming much bandwidth. 732 733DCB function 734~~~~~~~~~~~~ 735 736DCB works only when RSS is enabled. 737 738Global configuration warning 739~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 740 741I40E PMD will set some global registers to enable some function or set some 742configure. Then when using different ports of the same NIC with Linux kernel 743and DPDK, the port with Linux kernel will be impacted by the port with DPDK. 744For example, register I40E_GL_SWT_L2TAGCTRL is used to control L2 tag, i40e 745PMD uses I40E_GL_SWT_L2TAGCTRL to set vlan TPID. If setting TPID in port A 746with DPDK, then the configuration will also impact port B in the NIC with 747kernel driver, which don't want to use the TPID. 748So PMD reports warning to clarify what is changed by writing global register. 749 750Cloud Filter 751~~~~~~~~~~~~ 752 753When programming cloud filters for IPv4/6_UDP/TCP/SCTP with SRC port only or DST port only, 754it will make any cloud filter using inner_vlan or tunnel key invalid. Default configuration will be 755recovered only by NIC core reset. 756 757Mirror rule limitation for X722 758~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 759 760Due to firmware restriction of X722, the same VSI cannot have more than one mirror rule. 761 762.. _net_i40e_testpmd_commands: 763 764Testpmd driver specific commands 765-------------------------------- 766 767Some i40e driver specific features are integrated in testpmd. 768 769RSS queue region 770~~~~~~~~~~~~~~~~ 771 772Set RSS queue region span on a port:: 773 774 testpmd> set port (port_id) queue-region region_id (value) \ 775 queue_start_index (value) queue_num (value) 776 777Set flowtype mapping on a RSS queue region on a port:: 778 779 testpmd> set port (port_id) queue-region region_id (value) flowtype (value) 780 781where: 782 783* For the flowtype(pctype) of packet,the specific index for each type has 784 been defined in file i40e_type.h as enum i40e_filter_pctype. 785 786Set user priority mapping on a RSS queue region on a port:: 787 788 testpmd> set port (port_id) queue-region UP (value) region_id (value) 789 790Flush all queue region related configuration on a port:: 791 792 testpmd> set port (port_id) queue-region flush (on|off) 793 794where: 795 796* ``on``: is just an enable function which server for other configuration, 797 it is for all configuration about queue region from up layer, 798 at first will only keep in DPDK software stored in driver, 799 only after "flush on", it commit all configuration to HW. 800 801* ``"off``: is just clean all configuration about queue region just now, 802 and restore all to DPDK i40e driver default config when start up. 803 804Show all queue region related configuration info on a port:: 805 806 testpmd> show port (port_id) queue-region 807 808.. note:: 809 810 Queue region only support on PF by now, so these command is 811 only for configuration of queue region on PF port. 812 813set promisc (for VF) 814~~~~~~~~~~~~~~~~~~~~ 815 816Set the unicast promiscuous mode for a VF from PF. 817It's supported by Intel i40e NICs now. 818In promiscuous mode packets are not dropped if they aren't for the specified MAC address:: 819 820 testpmd> set vf promisc (port_id) (vf_id) (on|off) 821 822set allmulticast (for VF) 823~~~~~~~~~~~~~~~~~~~~~~~~~ 824 825Set the multicast promiscuous mode for a VF from PF. 826It's supported by Intel i40e NICs now. 827In promiscuous mode packets are not dropped if they aren't for the specified MAC address:: 828 829 testpmd> set vf allmulti (port_id) (vf_id) (on|off) 830 831set broadcast mode (for VF) 832~~~~~~~~~~~~~~~~~~~~~~~~~~~ 833 834Set broadcast mode for a VF from the PF:: 835 836 testpmd> set vf broadcast (port_id) (vf_id) (on|off) 837 838vlan set tag (for VF) 839~~~~~~~~~~~~~~~~~~~~~ 840 841Set VLAN tag for a VF from the PF:: 842 843 testpmd> set vf vlan tag (port_id) (vf_id) (on|off) 844 845set tx max bandwidth (for VF) 846~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 847 848Set TX max absolute bandwidth (Mbps) for a VF from PF:: 849 850 testpmd> set vf tx max-bandwidth (port_id) (vf_id) (max_bandwidth) 851 852set tc tx min bandwidth (for VF) 853~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 854 855Set all TCs' TX min relative bandwidth (%) for a VF from PF:: 856 857 testpmd> set vf tc tx min-bandwidth (port_id) (vf_id) (bw1, bw2, ...) 858 859set tc tx max bandwidth (for VF) 860~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 861 862Set a TC's TX max absolute bandwidth (Mbps) for a VF from PF:: 863 864 testpmd> set vf tc tx max-bandwidth (port_id) (vf_id) (tc_no) (max_bandwidth) 865 866set tc strict link priority mode 867~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 868 869Set some TCs' strict link priority mode on a physical port:: 870 871 testpmd> set tx strict-link-priority (port_id) (tc_bitmap) 872 873ddp add 874~~~~~~~ 875 876Load a dynamic device personalization (DDP) profile and store backup profile:: 877 878 testpmd> ddp add (port_id) (profile_path[,backup_profile_path]) 879 880ddp del 881~~~~~~~ 882 883Delete a dynamic device personalization profile and restore backup profile:: 884 885 testpmd> ddp del (port_id) (backup_profile_path) 886 887ddp get list 888~~~~~~~~~~~~ 889 890Get loaded dynamic device personalization (DDP) package info list:: 891 892 testpmd> ddp get list (port_id) 893 894ddp get info 895~~~~~~~~~~~~ 896 897Display information about dynamic device personalization (DDP) profile:: 898 899 testpmd> ddp get info (profile_path) 900 901ptype mapping 902~~~~~~~~~~~~~ 903 904List all items from the ptype mapping table:: 905 906 testpmd> ptype mapping get (port_id) (valid_only) 907 908Where: 909 910* ``valid_only``: A flag indicates if only list valid items(=1) or all items(=0). 911 912Replace a specific or a group of software defined ptype with a new one:: 913 914 testpmd> ptype mapping replace (port_id) (target) (mask) (pkt_type) 915 916where: 917 918* ``target``: A specific software ptype or a mask to represent a group of software ptypes. 919 920* ``mask``: A flag indicate if "target" is a specific software ptype(=0) or a ptype mask(=1). 921 922* ``pkt_type``: The new software ptype to replace the old ones. 923 924Update hardware defined ptype to software defined packet type mapping table:: 925 926 testpmd> ptype mapping update (port_id) (hw_ptype) (sw_ptype) 927 928where: 929 930* ``hw_ptype``: hardware ptype as the index of the ptype mapping table. 931 932* ``sw_ptype``: software ptype as the value of the ptype mapping table. 933 934Reset ptype mapping table:: 935 936 testpmd> ptype mapping reset (port_id) 937 938show port pctype mapping 939~~~~~~~~~~~~~~~~~~~~~~~~ 940 941List all items from the pctype mapping table:: 942 943 testpmd> show port (port_id) pctype mapping 944 945High Performance of Small Packets on 40GbE NIC 946---------------------------------------------- 947 948As there might be firmware fixes for performance enhancement in latest version 949of firmware image, the firmware update might be needed for getting high performance. 950Check the Intel support website for the latest firmware updates. 951Users should consult the release notes specific to a DPDK release to identify 952the validated firmware version for a NIC using the i40e driver. 953 954Use 16 Bytes RX Descriptor Size 955~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 956 957As i40e PMD supports both 16 and 32 bytes RX descriptor sizes, and 16 bytes size can provide helps to high performance of small packets. 958In ``config/rte_config.h`` set the following to use 16 bytes size RX descriptors:: 959 960 #define RTE_LIBRTE_I40E_16BYTE_RX_DESC 1 961 962Input set requirement of each pctype for FDIR 963~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 964 965Each PCTYPE can only have one specific FDIR input set at one time. 966For example, if creating 2 rte_flow rules with different input set for one PCTYPE, 967it will fail and return the info "Conflict with the first rule's input set", 968which means the current rule's input set conflicts with the first rule's. 969Remove the first rule if want to change the input set of the PCTYPE. 970 971Vlan related Features miss when FW >= 8.4 972~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 973 974If FW version >= 8.4, there'll be some Vlan related issues: 975 976#. TCI input set for QinQ is invalid. 977#. Fail to configure TPID for QinQ. 978#. Need to enable QinQ before enabling Vlan filter. 979#. Fail to strip outer Vlan. 980 981Example of getting best performance with l3fwd example 982------------------------------------------------------ 983 984The following is an example of running the DPDK ``l3fwd`` sample application to get high performance with a 985server with Intel Xeon processors and Intel Ethernet CNA XL710. 986 987The example scenario is to get best performance with two Intel Ethernet CNA XL710 40GbE ports. 988See :numref:`figure_intel_perf_test_setup` for the performance test setup. 989 990.. _figure_intel_perf_test_setup: 991 992.. figure:: img/intel_perf_test_setup.* 993 994 Performance Test Setup 995 996 997#. Add two Intel Ethernet CNA XL710 to the platform, and use one port per card to get best performance. 998 The reason for using two NICs is to overcome a PCIe v3.0 limitation since it cannot provide 80GbE bandwidth 999 for two 40GbE ports, but two different PCIe v3.0 x8 slot can. 1000 Refer to the sample NICs output above, then we can select ``82:00.0`` and ``85:00.0`` as test ports:: 1001 1002 82:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583] 1003 85:00.0 Ethernet [0200]: Intel XL710 for 40GbE QSFP+ [8086:1583] 1004 1005#. Connect the ports to the traffic generator. For high speed testing, it's best to use a hardware traffic generator. 1006 1007#. Check the PCI devices numa node (socket id) and get the cores number on the exact socket id. 1008 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 1009 are 18-35 and 54-71. 1010 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 1011 cores from different cores (e.g core18 and core19). 1012 1013#. Bind these two ports to igb_uio. 1014 1015#. As to Intel Ethernet CNA XL710 40GbE port, we need at least two queue pairs to achieve best performance, then two queues per port 1016 will be required, and each queue pair will need a dedicated CPU core for receiving/transmitting packets. 1017 1018#. The DPDK sample application ``l3fwd`` will be used for performance testing, with using two ports for bi-directional forwarding. 1019 Compile the ``l3fwd sample`` with the default lpm mode. 1020 1021#. The command line of running l3fwd would be something like the following:: 1022 1023 ./dpdk-l3fwd -l 18-21 -n 4 -a 82:00.0 -a 85:00.0 \ 1024 -- -p 0x3 --config '(0,0,18),(0,1,19),(1,0,20),(1,1,21)' 1025 1026 This means that the application uses core 18 for port 0, queue pair 0 forwarding, core 19 for port 0, queue pair 1 forwarding, 1027 core 20 for port 1, queue pair 0 forwarding, and core 21 for port 1, queue pair 1 forwarding. 1028 1029#. Configure the traffic at a traffic generator. 1030 1031 * Start creating a stream on packet generator. 1032 1033 * Set the Ethernet II type to 0x0800. 1034 1035Tx bytes affected by the link status change 1036~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1037 1038For firmware versions prior to 6.01 for X710 series and 3.33 for X722 series, the tx_bytes statistics data is affected by 1039the link down event. Each time the link status changes to down, the tx_bytes decreases 110 bytes. 1040