1.. SPDX-License-Identifier: BSD-3-Clause 2 Copyright(c) 2010-2016 Intel Corporation. 3 4.. _testpmd_runtime: 5 6Testpmd Runtime Functions 7========================= 8 9Where the testpmd application is started in interactive mode, (``-i|--interactive``), 10it displays a prompt that can be used to start and stop forwarding, 11configure the application, display statistics (including the extended NIC 12statistics aka xstats) , set the Flow Director and other tasks:: 13 14 testpmd> 15 16The testpmd prompt has some, limited, readline support. 17Common bash command-line functions such as ``Ctrl+a`` and ``Ctrl+e`` to go to the start and end of the prompt line are supported 18as well as access to the command history via the up-arrow. 19 20There is also support for tab completion. 21If you type a partial command and hit ``<TAB>`` you get a list of the available completions: 22 23.. code-block:: console 24 25 testpmd> show port <TAB> 26 27 info [Mul-choice STRING]: show|clear port info|stats|xstats|fdir|stat_qmap|dcb_tc|cap X 28 info [Mul-choice STRING]: show|clear port info|stats|xstats|fdir|stat_qmap|dcb_tc|cap all 29 stats [Mul-choice STRING]: show|clear port info|stats|xstats|fdir|stat_qmap|dcb_tc|cap X 30 stats [Mul-choice STRING]: show|clear port info|stats|xstats|fdir|stat_qmap|dcb_tc|cap all 31 ... 32 33 34.. note:: 35 36 Some examples in this document are too long to fit on one line are are shown wrapped at `"\\"` for display purposes:: 37 38 testpmd> set flow_ctrl rx (on|off) tx (on|off) (high_water) (low_water) \ 39 (pause_time) (send_xon) (port_id) 40 41In the real ``testpmd>`` prompt these commands should be on a single line. 42 43Help Functions 44-------------- 45 46The testpmd has on-line help for the functions that are available at runtime. 47These are divided into sections and can be accessed using help, help section or help all: 48 49.. code-block:: console 50 51 testpmd> help 52 53 help control : Start and stop forwarding. 54 help display : Displaying port, stats and config information. 55 help config : Configuration information. 56 help ports : Configuring ports. 57 help registers : Reading and setting port registers. 58 help filters : Filters configuration help. 59 help all : All of the above sections. 60 61 62Command File Functions 63---------------------- 64 65To facilitate loading large number of commands or to avoid cutting and pasting where not 66practical or possible testpmd supports alternative methods for executing commands. 67 68* If started with the ``--cmdline-file=FILENAME`` command line argument testpmd 69 will execute all CLI commands contained within the file immediately before 70 starting packet forwarding or entering interactive mode. 71 72.. code-block:: console 73 74 ./testpmd -n4 -r2 ... -- -i --cmdline-file=/home/ubuntu/flow-create-commands.txt 75 Interactive-mode selected 76 CLI commands to be read from /home/ubuntu/flow-create-commands.txt 77 Configuring Port 0 (socket 0) 78 Port 0: 7C:FE:90:CB:74:CE 79 Configuring Port 1 (socket 0) 80 Port 1: 7C:FE:90:CB:74:CA 81 Checking link statuses... 82 Port 0 Link Up - speed 10000 Mbps - full-duplex 83 Port 1 Link Up - speed 10000 Mbps - full-duplex 84 Done 85 Flow rule #0 created 86 Flow rule #1 created 87 ... 88 ... 89 Flow rule #498 created 90 Flow rule #499 created 91 Read all CLI commands from /home/ubuntu/flow-create-commands.txt 92 testpmd> 93 94 95* At run-time additional commands can be loaded in bulk by invoking the ``load FILENAME`` 96 command. 97 98.. code-block:: console 99 100 testpmd> load /home/ubuntu/flow-create-commands.txt 101 Flow rule #0 created 102 Flow rule #1 created 103 ... 104 ... 105 Flow rule #498 created 106 Flow rule #499 created 107 Read all CLI commands from /home/ubuntu/flow-create-commands.txt 108 testpmd> 109 110 111In all cases output from any included command will be displayed as standard output. 112Execution will continue until the end of the file is reached regardless of 113whether any errors occur. The end user must examine the output to determine if 114any failures occurred. 115 116 117Control Functions 118----------------- 119 120start 121~~~~~ 122 123Start packet forwarding with current configuration:: 124 125 testpmd> start 126 127start tx_first 128~~~~~~~~~~~~~~ 129 130Start packet forwarding with current configuration after sending specified number of bursts of packets:: 131 132 testpmd> start tx_first (""|burst_num) 133 134The default burst number is 1 when ``burst_num`` not presented. 135 136stop 137~~~~ 138 139Stop packet forwarding, and display accumulated statistics:: 140 141 testpmd> stop 142 143quit 144~~~~ 145 146Quit to prompt:: 147 148 testpmd> quit 149 150 151Display Functions 152----------------- 153 154The functions in the following sections are used to display information about the 155testpmd configuration or the NIC status. 156 157show port 158~~~~~~~~~ 159 160Display information for a given port or all ports:: 161 162 testpmd> show port (info|stats|xstats|fdir|stat_qmap|dcb_tc|cap) (port_id|all) 163 164The available information categories are: 165 166* ``info``: General port information such as MAC address. 167 168* ``stats``: RX/TX statistics. 169 170* ``xstats``: RX/TX extended NIC statistics. 171 172* ``fdir``: Flow Director information and statistics. 173 174* ``stat_qmap``: Queue statistics mapping. 175 176* ``dcb_tc``: DCB information such as TC mapping. 177 178* ``cap``: Supported offload capabilities. 179 180For example: 181 182.. code-block:: console 183 184 testpmd> show port info 0 185 186 ********************* Infos for port 0 ********************* 187 188 MAC address: XX:XX:XX:XX:XX:XX 189 Connect to socket: 0 190 memory allocation on the socket: 0 191 Link status: up 192 Link speed: 40000 Mbps 193 Link duplex: full-duplex 194 Promiscuous mode: enabled 195 Allmulticast mode: disabled 196 Maximum number of MAC addresses: 64 197 Maximum number of MAC addresses of hash filtering: 0 198 VLAN offload: 199 strip on 200 filter on 201 qinq(extend) off 202 Redirection table size: 512 203 Supported flow types: 204 ipv4-frag 205 ipv4-tcp 206 ipv4-udp 207 ipv4-sctp 208 ipv4-other 209 ipv6-frag 210 ipv6-tcp 211 ipv6-udp 212 ipv6-sctp 213 ipv6-other 214 l2_payload 215 port 216 vxlan 217 geneve 218 nvgre 219 220show port rss reta 221~~~~~~~~~~~~~~~~~~ 222 223Display the rss redirection table entry indicated by masks on port X:: 224 225 testpmd> show port (port_id) rss reta (size) (mask0, mask1...) 226 227size is used to indicate the hardware supported reta size 228 229show port rss-hash 230~~~~~~~~~~~~~~~~~~ 231 232Display the RSS hash functions and RSS hash key of a port:: 233 234 testpmd> show port (port_id) rss-hash ipv4|ipv4-frag|ipv4-tcp|ipv4-udp|ipv4-sctp|ipv4-other|ipv6|ipv6-frag|ipv6-tcp|ipv6-udp|ipv6-sctp|ipv6-other|l2-payload|ipv6-ex|ipv6-tcp-ex|ipv6-udp-ex [key] 235 236clear port 237~~~~~~~~~~ 238 239Clear the port statistics for a given port or for all ports:: 240 241 testpmd> clear port (info|stats|xstats|fdir|stat_qmap) (port_id|all) 242 243For example:: 244 245 testpmd> clear port stats all 246 247show (rxq|txq) 248~~~~~~~~~~~~~~ 249 250Display information for a given port's RX/TX queue:: 251 252 testpmd> show (rxq|txq) info (port_id) (queue_id) 253 254show config 255~~~~~~~~~~~ 256 257Displays the configuration of the application. 258The configuration comes from the command-line, the runtime or the application defaults:: 259 260 testpmd> show config (rxtx|cores|fwd|txpkts) 261 262The available information categories are: 263 264* ``rxtx``: RX/TX configuration items. 265 266* ``cores``: List of forwarding cores. 267 268* ``fwd``: Packet forwarding configuration. 269 270* ``txpkts``: Packets to TX configuration. 271 272For example: 273 274.. code-block:: console 275 276 testpmd> show config rxtx 277 278 io packet forwarding - CRC stripping disabled - packets/burst=16 279 nb forwarding cores=2 - nb forwarding ports=1 280 RX queues=1 - RX desc=128 - RX free threshold=0 281 RX threshold registers: pthresh=8 hthresh=8 wthresh=4 282 TX queues=1 - TX desc=512 - TX free threshold=0 283 TX threshold registers: pthresh=36 hthresh=0 wthresh=0 284 TX RS bit threshold=0 - TXQ flags=0x0 285 286set fwd 287~~~~~~~ 288 289Set the packet forwarding mode:: 290 291 testpmd> set fwd (io|mac|macswap|flowgen| \ 292 rxonly|txonly|csum|icmpecho) (""|retry) 293 294``retry`` can be specified for forwarding engines except ``rx_only``. 295 296The available information categories are: 297 298* ``io``: Forwards packets "as-is" in I/O mode. 299 This is the fastest possible forwarding operation as it does not access packets data. 300 This is the default mode. 301 302* ``mac``: Changes the source and the destination Ethernet addresses of packets before forwarding them. 303 Default application behaviour is to set source Ethernet address to that of the transmitting interface, and destination 304 address to a dummy value (set during init). The user may specify a target destination Ethernet address via the 'eth-peer' or 305 'eth-peer-configfile' command-line options. It is not currently possible to specify a specific source Ethernet address. 306 307* ``macswap``: MAC swap forwarding mode. 308 Swaps the source and the destination Ethernet addresses of packets before forwarding them. 309 310* ``flowgen``: Multi-flow generation mode. 311 Originates a number of flows (with varying destination IP addresses), and terminate receive traffic. 312 313* ``rxonly``: Receives packets but doesn't transmit them. 314 315* ``txonly``: Generates and transmits packets without receiving any. 316 317* ``csum``: Changes the checksum field with hardware or software methods depending on the offload flags on the packet. 318 319* ``icmpecho``: Receives a burst of packets, lookup for IMCP echo requests and, if any, send back ICMP echo replies. 320 321* ``ieee1588``: Demonstrate L2 IEEE1588 V2 PTP timestamping for RX and TX. Requires ``CONFIG_RTE_LIBRTE_IEEE1588=y``. 322 323* ``tm``: Traffic Management forwarding mode 324 Demonstrates the use of ethdev traffic management APIs and softnic PMD for 325 QoS traffic management. In this mode, 5-level hierarchical QoS scheduler is 326 available as an default option that can be enabled through CLI. The user can 327 also modify the default hierarchy or specify the new hierarchy through CLI for 328 implementing QoS scheduler. Requires ``CONFIG_RTE_LIBRTE_PMD_SOFTNIC=y`` ``CONFIG_RTE_LIBRTE_SCHED=y``. 329 330Example:: 331 332 testpmd> set fwd rxonly 333 334 Set rxonly packet forwarding mode 335 336 337read rxd 338~~~~~~~~ 339 340Display an RX descriptor for a port RX queue:: 341 342 testpmd> read rxd (port_id) (queue_id) (rxd_id) 343 344For example:: 345 346 testpmd> read rxd 0 0 4 347 0x0000000B - 0x001D0180 / 0x0000000B - 0x001D0180 348 349read txd 350~~~~~~~~ 351 352Display a TX descriptor for a port TX queue:: 353 354 testpmd> read txd (port_id) (queue_id) (txd_id) 355 356For example:: 357 358 testpmd> read txd 0 0 4 359 0x00000001 - 0x24C3C440 / 0x000F0000 - 0x2330003C 360 361ddp get list 362~~~~~~~~~~~~ 363 364Get loaded dynamic device personalization (DDP) package info list:: 365 366 testpmd> ddp get list (port_id) 367 368ddp get info 369~~~~~~~~~~~~ 370 371Display information about dynamic device personalization (DDP) profile:: 372 373 testpmd> ddp get info (profile_path) 374 375show vf stats 376~~~~~~~~~~~~~ 377 378Display VF statistics:: 379 380 testpmd> show vf stats (port_id) (vf_id) 381 382clear vf stats 383~~~~~~~~~~~~~~ 384 385Reset VF statistics:: 386 387 testpmd> clear vf stats (port_id) (vf_id) 388 389show port pctype mapping 390~~~~~~~~~~~~~~~~~~~~~~~~ 391 392List all items from the pctype mapping table:: 393 394 testpmd> show port (port_id) pctype mapping 395 396show rx offloading capabilities 397~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 398 399List all per queue and per port Rx offloading capabilities of a port:: 400 401 testpmd> show port (port_id) rx_offload capabilities 402 403show rx offloading configuration 404~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 405 406List port level and all queue level Rx offloading configuration:: 407 408 testpmd> show port (port_id) rx_offload configuration 409 410show tx offloading capabilities 411~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 412 413List all per queue and per port Tx offloading capabilities of a port:: 414 415 testpmd> show port (port_id) tx_offload capabilities 416 417show tx offloading configuration 418~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 419 420List port level and all queue level Tx offloading configuration:: 421 422 testpmd> show port (port_id) tx_offload configuration 423 424 425Configuration Functions 426----------------------- 427 428The testpmd application can be configured from the runtime as well as from the command-line. 429 430This section details the available configuration functions that are available. 431 432.. note:: 433 434 Configuration changes only become active when forwarding is started/restarted. 435 436set default 437~~~~~~~~~~~ 438 439Reset forwarding to the default configuration:: 440 441 testpmd> set default 442 443set verbose 444~~~~~~~~~~~ 445 446Set the debug verbosity level:: 447 448 testpmd> set verbose (level) 449 450Currently the only available levels are 0 (silent except for error) and 1 (fully verbose). 451 452set log 453~~~~~~~ 454 455Set the log level for a log type:: 456 457 testpmd> set log global|(type) (level) 458 459Where: 460 461* ``type`` is the log name. 462 463* ``level`` is the log level. 464 465For example, to change the global log level:: 466 testpmd> set log global (level) 467 468Regexes can also be used for type. To change log level of user1, user2 and user3:: 469 testpmd> set log user[1-3] (level) 470 471set nbport 472~~~~~~~~~~ 473 474Set the number of ports used by the application: 475 476set nbport (num) 477 478This is equivalent to the ``--nb-ports`` command-line option. 479 480set nbcore 481~~~~~~~~~~ 482 483Set the number of cores used by the application:: 484 485 testpmd> set nbcore (num) 486 487This is equivalent to the ``--nb-cores`` command-line option. 488 489.. note:: 490 491 The number of cores used must not be greater than number of ports used multiplied by the number of queues per port. 492 493set coremask 494~~~~~~~~~~~~ 495 496Set the forwarding cores hexadecimal mask:: 497 498 testpmd> set coremask (mask) 499 500This is equivalent to the ``--coremask`` command-line option. 501 502.. note:: 503 504 The master lcore is reserved for command line parsing only and cannot be masked on for packet forwarding. 505 506set portmask 507~~~~~~~~~~~~ 508 509Set the forwarding ports hexadecimal mask:: 510 511 testpmd> set portmask (mask) 512 513This is equivalent to the ``--portmask`` command-line option. 514 515set burst 516~~~~~~~~~ 517 518Set number of packets per burst:: 519 520 testpmd> set burst (num) 521 522This is equivalent to the ``--burst command-line`` option. 523 524When retry is enabled, the transmit delay time and number of retries can also be set:: 525 526 testpmd> set burst tx delay (microseconds) retry (num) 527 528set txpkts 529~~~~~~~~~~ 530 531Set the length of each segment of the TX-ONLY packets or length of packet for FLOWGEN mode:: 532 533 testpmd> set txpkts (x[,y]*) 534 535Where x[,y]* represents a CSV list of values, without white space. 536 537set txsplit 538~~~~~~~~~~~ 539 540Set the split policy for the TX packets, applicable for TX-ONLY and CSUM forwarding modes:: 541 542 testpmd> set txsplit (off|on|rand) 543 544Where: 545 546* ``off`` disable packet copy & split for CSUM mode. 547 548* ``on`` split outgoing packet into multiple segments. Size of each segment 549 and number of segments per packet is determined by ``set txpkts`` command 550 (see above). 551 552* ``rand`` same as 'on', but number of segments per each packet is a random value between 1 and total number of segments. 553 554set corelist 555~~~~~~~~~~~~ 556 557Set the list of forwarding cores:: 558 559 testpmd> set corelist (x[,y]*) 560 561For example, to change the forwarding cores: 562 563.. code-block:: console 564 565 testpmd> set corelist 3,1 566 testpmd> show config fwd 567 568 io packet forwarding - ports=2 - cores=2 - streams=2 - NUMA support disabled 569 Logical Core 3 (socket 0) forwards packets on 1 streams: 570 RX P=0/Q=0 (socket 0) -> TX P=1/Q=0 (socket 0) peer=02:00:00:00:00:01 571 Logical Core 1 (socket 0) forwards packets on 1 streams: 572 RX P=1/Q=0 (socket 0) -> TX P=0/Q=0 (socket 0) peer=02:00:00:00:00:00 573 574.. note:: 575 576 The cores are used in the same order as specified on the command line. 577 578set portlist 579~~~~~~~~~~~~ 580 581Set the list of forwarding ports:: 582 583 testpmd> set portlist (x[,y]*) 584 585For example, to change the port forwarding: 586 587.. code-block:: console 588 589 testpmd> set portlist 0,2,1,3 590 testpmd> show config fwd 591 592 io packet forwarding - ports=4 - cores=1 - streams=4 593 Logical Core 3 (socket 0) forwards packets on 4 streams: 594 RX P=0/Q=0 (socket 0) -> TX P=2/Q=0 (socket 0) peer=02:00:00:00:00:01 595 RX P=2/Q=0 (socket 0) -> TX P=0/Q=0 (socket 0) peer=02:00:00:00:00:00 596 RX P=1/Q=0 (socket 0) -> TX P=3/Q=0 (socket 0) peer=02:00:00:00:00:03 597 RX P=3/Q=0 (socket 0) -> TX P=1/Q=0 (socket 0) peer=02:00:00:00:00:02 598 599set tx loopback 600~~~~~~~~~~~~~~~ 601 602Enable/disable tx loopback:: 603 604 testpmd> set tx loopback (port_id) (on|off) 605 606set drop enable 607~~~~~~~~~~~~~~~ 608 609set drop enable bit for all queues:: 610 611 testpmd> set all queues drop (port_id) (on|off) 612 613set split drop enable (for VF) 614~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 615 616set split drop enable bit for VF from PF:: 617 618 testpmd> set vf split drop (port_id) (vf_id) (on|off) 619 620set mac antispoof (for VF) 621~~~~~~~~~~~~~~~~~~~~~~~~~~ 622 623Set mac antispoof for a VF from the PF:: 624 625 testpmd> set vf mac antispoof (port_id) (vf_id) (on|off) 626 627set macsec offload 628~~~~~~~~~~~~~~~~~~ 629 630Enable/disable MACsec offload:: 631 632 testpmd> set macsec offload (port_id) on encrypt (on|off) replay-protect (on|off) 633 testpmd> set macsec offload (port_id) off 634 635set macsec sc 636~~~~~~~~~~~~~ 637 638Configure MACsec secure connection (SC):: 639 640 testpmd> set macsec sc (tx|rx) (port_id) (mac) (pi) 641 642.. note:: 643 644 The pi argument is ignored for tx. 645 Check the NIC Datasheet for hardware limits. 646 647set macsec sa 648~~~~~~~~~~~~~ 649 650Configure MACsec secure association (SA):: 651 652 testpmd> set macsec sa (tx|rx) (port_id) (idx) (an) (pn) (key) 653 654.. note:: 655 656 The IDX value must be 0 or 1. 657 Check the NIC Datasheet for hardware limits. 658 659set broadcast mode (for VF) 660~~~~~~~~~~~~~~~~~~~~~~~~~~~ 661 662Set broadcast mode for a VF from the PF:: 663 664 testpmd> set vf broadcast (port_id) (vf_id) (on|off) 665 666vlan set strip 667~~~~~~~~~~~~~~ 668 669Set the VLAN strip on a port:: 670 671 testpmd> vlan set strip (on|off) (port_id) 672 673vlan set stripq 674~~~~~~~~~~~~~~~ 675 676Set the VLAN strip for a queue on a port:: 677 678 testpmd> vlan set stripq (on|off) (port_id,queue_id) 679 680vlan set stripq (for VF) 681~~~~~~~~~~~~~~~~~~~~~~~~ 682 683Set VLAN strip for all queues in a pool for a VF from the PF:: 684 685 testpmd> set vf vlan stripq (port_id) (vf_id) (on|off) 686 687vlan set insert (for VF) 688~~~~~~~~~~~~~~~~~~~~~~~~ 689 690Set VLAN insert for a VF from the PF:: 691 692 testpmd> set vf vlan insert (port_id) (vf_id) (vlan_id) 693 694vlan set tag (for VF) 695~~~~~~~~~~~~~~~~~~~~~ 696 697Set VLAN tag for a VF from the PF:: 698 699 testpmd> set vf vlan tag (port_id) (vf_id) (on|off) 700 701vlan set antispoof (for VF) 702~~~~~~~~~~~~~~~~~~~~~~~~~~~ 703 704Set VLAN antispoof for a VF from the PF:: 705 706 testpmd> set vf vlan antispoof (port_id) (vf_id) (on|off) 707 708vlan set filter 709~~~~~~~~~~~~~~~ 710 711Set the VLAN filter on a port:: 712 713 testpmd> vlan set filter (on|off) (port_id) 714 715vlan set qinq 716~~~~~~~~~~~~~ 717 718Set the VLAN QinQ (extended queue in queue) on for a port:: 719 720 testpmd> vlan set qinq (on|off) (port_id) 721 722vlan set tpid 723~~~~~~~~~~~~~ 724 725Set the inner or outer VLAN TPID for packet filtering on a port:: 726 727 testpmd> vlan set (inner|outer) tpid (value) (port_id) 728 729.. note:: 730 731 TPID value must be a 16-bit number (value <= 65536). 732 733rx_vlan add 734~~~~~~~~~~~ 735 736Add a VLAN ID, or all identifiers, to the set of VLAN identifiers filtered by port ID:: 737 738 testpmd> rx_vlan add (vlan_id|all) (port_id) 739 740.. note:: 741 742 VLAN filter must be set on that port. VLAN ID < 4096. 743 Depending on the NIC used, number of vlan_ids may be limited to the maximum entries 744 in VFTA table. This is important if enabling all vlan_ids. 745 746rx_vlan rm 747~~~~~~~~~~ 748 749Remove a VLAN ID, or all identifiers, from the set of VLAN identifiers filtered by port ID:: 750 751 testpmd> rx_vlan rm (vlan_id|all) (port_id) 752 753rx_vlan add (for VF) 754~~~~~~~~~~~~~~~~~~~~ 755 756Add a VLAN ID, to the set of VLAN identifiers filtered for VF(s) for port ID:: 757 758 testpmd> rx_vlan add (vlan_id) port (port_id) vf (vf_mask) 759 760rx_vlan rm (for VF) 761~~~~~~~~~~~~~~~~~~~ 762 763Remove a VLAN ID, from the set of VLAN identifiers filtered for VF(s) for port ID:: 764 765 testpmd> rx_vlan rm (vlan_id) port (port_id) vf (vf_mask) 766 767tunnel_filter add 768~~~~~~~~~~~~~~~~~ 769 770Add a tunnel filter on a port:: 771 772 testpmd> tunnel_filter add (port_id) (outer_mac) (inner_mac) (ip_addr) \ 773 (inner_vlan) (vxlan|nvgre|ipingre) (imac-ivlan|imac-ivlan-tenid|\ 774 imac-tenid|imac|omac-imac-tenid|oip|iip) (tenant_id) (queue_id) 775 776The available information categories are: 777 778* ``vxlan``: Set tunnel type as VXLAN. 779 780* ``nvgre``: Set tunnel type as NVGRE. 781 782* ``ipingre``: Set tunnel type as IP-in-GRE. 783 784* ``imac-ivlan``: Set filter type as Inner MAC and VLAN. 785 786* ``imac-ivlan-tenid``: Set filter type as Inner MAC, VLAN and tenant ID. 787 788* ``imac-tenid``: Set filter type as Inner MAC and tenant ID. 789 790* ``imac``: Set filter type as Inner MAC. 791 792* ``omac-imac-tenid``: Set filter type as Outer MAC, Inner MAC and tenant ID. 793 794* ``oip``: Set filter type as Outer IP. 795 796* ``iip``: Set filter type as Inner IP. 797 798Example:: 799 800 testpmd> tunnel_filter add 0 68:05:CA:28:09:82 00:00:00:00:00:00 \ 801 192.168.2.2 0 ipingre oip 1 1 802 803 Set an IP-in-GRE tunnel on port 0, and the filter type is Outer IP. 804 805tunnel_filter remove 806~~~~~~~~~~~~~~~~~~~~ 807 808Remove a tunnel filter on a port:: 809 810 testpmd> tunnel_filter rm (port_id) (outer_mac) (inner_mac) (ip_addr) \ 811 (inner_vlan) (vxlan|nvgre|ipingre) (imac-ivlan|imac-ivlan-tenid|\ 812 imac-tenid|imac|omac-imac-tenid|oip|iip) (tenant_id) (queue_id) 813 814rx_vxlan_port add 815~~~~~~~~~~~~~~~~~ 816 817Add an UDP port for VXLAN packet filter on a port:: 818 819 testpmd> rx_vxlan_port add (udp_port) (port_id) 820 821rx_vxlan_port remove 822~~~~~~~~~~~~~~~~~~~~ 823 824Remove an UDP port for VXLAN packet filter on a port:: 825 826 testpmd> rx_vxlan_port rm (udp_port) (port_id) 827 828tx_vlan set 829~~~~~~~~~~~ 830 831Set hardware insertion of VLAN IDs in packets sent on a port:: 832 833 testpmd> tx_vlan set (port_id) vlan_id[, vlan_id_outer] 834 835For example, set a single VLAN ID (5) insertion on port 0:: 836 837 tx_vlan set 0 5 838 839Or, set double VLAN ID (inner: 2, outer: 3) insertion on port 1:: 840 841 tx_vlan set 1 2 3 842 843 844tx_vlan set pvid 845~~~~~~~~~~~~~~~~ 846 847Set port based hardware insertion of VLAN ID in packets sent on a port:: 848 849 testpmd> tx_vlan set pvid (port_id) (vlan_id) (on|off) 850 851tx_vlan reset 852~~~~~~~~~~~~~ 853 854Disable hardware insertion of a VLAN header in packets sent on a port:: 855 856 testpmd> tx_vlan reset (port_id) 857 858csum set 859~~~~~~~~ 860 861Select hardware or software calculation of the checksum when 862transmitting a packet using the ``csum`` forwarding engine:: 863 864 testpmd> csum set (ip|udp|tcp|sctp|outer-ip) (hw|sw) (port_id) 865 866Where: 867 868* ``ip|udp|tcp|sctp`` always relate to the inner layer. 869 870* ``outer-ip`` relates to the outer IP layer (only for IPv4) in the case where the packet is recognized 871 as a tunnel packet by the forwarding engine (vxlan, gre and ipip are 872 supported). See also the ``csum parse-tunnel`` command. 873 874.. note:: 875 876 Check the NIC Datasheet for hardware limits. 877 878RSS queue region 879~~~~~~~~~~~~~~~~ 880 881Set RSS queue region span on a port:: 882 883 testpmd> set port (port_id) queue-region region_id (value) \ 884 queue_start_index (value) queue_num (value) 885 886Set flowtype mapping on a RSS queue region on a port:: 887 888 testpmd> set port (port_id) queue-region region_id (value) flowtype (value) 889 890where: 891 892* For the flowtype(pctype) of packet,the specific index for each type has 893 been defined in file i40e_type.h as enum i40e_filter_pctype. 894 895Set user priority mapping on a RSS queue region on a port:: 896 897 testpmd> set port (port_id) queue-region UP (value) region_id (value) 898 899Flush all queue region related configuration on a port:: 900 901 testpmd> set port (port_id) queue-region flush (on|off) 902 903where: 904 905* "on"is just an enable function which server for other configuration, 906 it is for all configuration about queue region from up layer, 907 at first will only keep in DPDK softwarestored in driver, 908 only after "flush on", it commit all configuration to HW. 909 "off" is just clean all configuration about queue region just now, 910 and restore all to DPDK i40e driver default config when start up. 911 912Show all queue region related configuration info on a port:: 913 914 testpmd> show port (port_id) queue-region 915 916.. note:: 917 918 Queue region only support on PF by now, so these command is 919 only for configuration of queue region on PF port. 920 921csum parse-tunnel 922~~~~~~~~~~~~~~~~~ 923 924Define how tunneled packets should be handled by the csum forward 925engine:: 926 927 testpmd> csum parse-tunnel (on|off) (tx_port_id) 928 929If enabled, the csum forward engine will try to recognize supported 930tunnel headers (vxlan, gre, ipip). 931 932If disabled, treat tunnel packets as non-tunneled packets (a inner 933header is handled as a packet payload). 934 935.. note:: 936 937 The port argument is the TX port like in the ``csum set`` command. 938 939Example: 940 941Consider a packet in packet like the following:: 942 943 eth_out/ipv4_out/udp_out/vxlan/eth_in/ipv4_in/tcp_in 944 945* If parse-tunnel is enabled, the ``ip|udp|tcp|sctp`` parameters of ``csum set`` 946 command relate to the inner headers (here ``ipv4_in`` and ``tcp_in``), and the 947 ``outer-ip parameter`` relates to the outer headers (here ``ipv4_out``). 948 949* If parse-tunnel is disabled, the ``ip|udp|tcp|sctp`` parameters of ``csum set`` 950 command relate to the outer headers, here ``ipv4_out`` and ``udp_out``. 951 952csum show 953~~~~~~~~~ 954 955Display tx checksum offload configuration:: 956 957 testpmd> csum show (port_id) 958 959tso set 960~~~~~~~ 961 962Enable TCP Segmentation Offload (TSO) in the ``csum`` forwarding engine:: 963 964 testpmd> tso set (segsize) (port_id) 965 966.. note:: 967 968 Check the NIC datasheet for hardware limits. 969 970tso show 971~~~~~~~~ 972 973Display the status of TCP Segmentation Offload:: 974 975 testpmd> tso show (port_id) 976 977set port - gro 978~~~~~~~~~~~~~~ 979 980Enable or disable GRO in ``csum`` forwarding engine:: 981 982 testpmd> set port <port_id> gro on|off 983 984If enabled, the csum forwarding engine will perform GRO on the TCP/IPv4 985packets received from the given port. 986 987If disabled, packets received from the given port won't be performed 988GRO. By default, GRO is disabled for all ports. 989 990.. note:: 991 992 When enable GRO for a port, TCP/IPv4 packets received from the port 993 will be performed GRO. After GRO, all merged packets have bad 994 checksums, since the GRO library doesn't re-calculate checksums for 995 the merged packets. Therefore, if users want the merged packets to 996 have correct checksums, please select HW IP checksum calculation and 997 HW TCP checksum calculation for the port which the merged packets are 998 transmitted to. 999 1000show port - gro 1001~~~~~~~~~~~~~~~ 1002 1003Display GRO configuration for a given port:: 1004 1005 testpmd> show port <port_id> gro 1006 1007set gro flush 1008~~~~~~~~~~~~~ 1009 1010Set the cycle to flush the GROed packets from reassembly tables:: 1011 1012 testpmd> set gro flush <cycles> 1013 1014When enable GRO, the csum forwarding engine performs GRO on received 1015packets, and the GROed packets are stored in reassembly tables. Users 1016can use this command to determine when the GROed packets are flushed 1017from the reassembly tables. 1018 1019The ``cycles`` is measured in GRO operation times. The csum forwarding 1020engine flushes the GROed packets from the tables every ``cycles`` GRO 1021operations. 1022 1023By default, the value of ``cycles`` is 1, which means flush GROed packets 1024from the reassembly tables as soon as one GRO operation finishes. The value 1025of ``cycles`` should be in the range of 1 to ``GRO_MAX_FLUSH_CYCLES``. 1026 1027Please note that the large value of ``cycles`` may cause the poor TCP/IP 1028stack performance. Because the GROed packets are delayed to arrive the 1029stack, thus causing more duplicated ACKs and TCP retransmissions. 1030 1031set port - gso 1032~~~~~~~~~~~~~~ 1033 1034Toggle per-port GSO support in ``csum`` forwarding engine:: 1035 1036 testpmd> set port <port_id> gso on|off 1037 1038If enabled, the csum forwarding engine will perform GSO on supported IPv4 1039packets, transmitted on the given port. 1040 1041If disabled, packets transmitted on the given port will not undergo GSO. 1042By default, GSO is disabled for all ports. 1043 1044.. note:: 1045 1046 When GSO is enabled on a port, supported IPv4 packets transmitted on that 1047 port undergo GSO. Afterwards, the segmented packets are represented by 1048 multi-segment mbufs; however, the csum forwarding engine doesn't calculation 1049 of checksums for GSO'd segments in SW. As a result, if users want correct 1050 checksums in GSO segments, they should enable HW checksum calculation for 1051 GSO-enabled ports. 1052 1053 For example, HW checksum calculation for VxLAN GSO'd packets may be enabled 1054 by setting the following options in the csum forwarding engine: 1055 1056 testpmd> csum set outer_ip hw <port_id> 1057 1058 testpmd> csum set ip hw <port_id> 1059 1060 testpmd> csum set tcp hw <port_id> 1061 1062set gso segsz 1063~~~~~~~~~~~~~ 1064 1065Set the maximum GSO segment size (measured in bytes), which includes the 1066packet header and the packet payload for GSO-enabled ports (global):: 1067 1068 testpmd> set gso segsz <length> 1069 1070show port - gso 1071~~~~~~~~~~~~~~~ 1072 1073Display the status of Generic Segmentation Offload for a given port:: 1074 1075 testpmd> show port <port_id> gso 1076 1077mac_addr add 1078~~~~~~~~~~~~ 1079 1080Add an alternative MAC address to a port:: 1081 1082 testpmd> mac_addr add (port_id) (XX:XX:XX:XX:XX:XX) 1083 1084mac_addr remove 1085~~~~~~~~~~~~~~~ 1086 1087Remove a MAC address from a port:: 1088 1089 testpmd> mac_addr remove (port_id) (XX:XX:XX:XX:XX:XX) 1090 1091mac_addr add (for VF) 1092~~~~~~~~~~~~~~~~~~~~~ 1093 1094Add an alternative MAC address for a VF to a port:: 1095 1096 testpmd> mac_add add port (port_id) vf (vf_id) (XX:XX:XX:XX:XX:XX) 1097 1098mac_addr set 1099~~~~~~~~~~~~ 1100 1101Set the default MAC address for a port:: 1102 1103 testpmd> mac_addr set (port_id) (XX:XX:XX:XX:XX:XX) 1104 1105mac_addr set (for VF) 1106~~~~~~~~~~~~~~~~~~~~~ 1107 1108Set the MAC address for a VF from the PF:: 1109 1110 testpmd> set vf mac addr (port_id) (vf_id) (XX:XX:XX:XX:XX:XX) 1111 1112set eth-peer 1113~~~~~~~~~~~~ 1114 1115Set the forwarding peer address for certain port:: 1116 1117 testpmd> set eth-peer (port_id) (perr_addr) 1118 1119This is equivalent to the ``--eth-peer`` command-line option. 1120 1121set port-uta 1122~~~~~~~~~~~~ 1123 1124Set the unicast hash filter(s) on/off for a port:: 1125 1126 testpmd> set port (port_id) uta (XX:XX:XX:XX:XX:XX|all) (on|off) 1127 1128set promisc 1129~~~~~~~~~~~ 1130 1131Set the promiscuous mode on for a port or for all ports. 1132In promiscuous mode packets are not dropped if they aren't for the specified MAC address:: 1133 1134 testpmd> set promisc (port_id|all) (on|off) 1135 1136set allmulti 1137~~~~~~~~~~~~ 1138 1139Set the allmulti mode for a port or for all ports:: 1140 1141 testpmd> set allmulti (port_id|all) (on|off) 1142 1143Same as the ifconfig (8) option. Controls how multicast packets are handled. 1144 1145set promisc (for VF) 1146~~~~~~~~~~~~~~~~~~~~ 1147 1148Set the unicast promiscuous mode for a VF from PF. 1149It's supported by Intel i40e NICs now. 1150In promiscuous mode packets are not dropped if they aren't for the specified MAC address:: 1151 1152 testpmd> set vf promisc (port_id) (vf_id) (on|off) 1153 1154set allmulticast (for VF) 1155~~~~~~~~~~~~~~~~~~~~~~~~~ 1156 1157Set the multicast promiscuous mode for a VF from PF. 1158It's supported by Intel i40e NICs now. 1159In promiscuous mode packets are not dropped if they aren't for the specified MAC address:: 1160 1161 testpmd> set vf allmulti (port_id) (vf_id) (on|off) 1162 1163set tx max bandwidth (for VF) 1164~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1165 1166Set TX max absolute bandwidth (Mbps) for a VF from PF:: 1167 1168 testpmd> set vf tx max-bandwidth (port_id) (vf_id) (max_bandwidth) 1169 1170set tc tx min bandwidth (for VF) 1171~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1172 1173Set all TCs' TX min relative bandwidth (%) for a VF from PF:: 1174 1175 testpmd> set vf tc tx min-bandwidth (port_id) (vf_id) (bw1, bw2, ...) 1176 1177set tc tx max bandwidth (for VF) 1178~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1179 1180Set a TC's TX max absolute bandwidth (Mbps) for a VF from PF:: 1181 1182 testpmd> set vf tc tx max-bandwidth (port_id) (vf_id) (tc_no) (max_bandwidth) 1183 1184set tc strict link priority mode 1185~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1186 1187Set some TCs' strict link priority mode on a physical port:: 1188 1189 testpmd> set tx strict-link-priority (port_id) (tc_bitmap) 1190 1191set tc tx min bandwidth 1192~~~~~~~~~~~~~~~~~~~~~~~ 1193 1194Set all TCs' TX min relative bandwidth (%) globally for all PF and VFs:: 1195 1196 testpmd> set tc tx min-bandwidth (port_id) (bw1, bw2, ...) 1197 1198set flow_ctrl rx 1199~~~~~~~~~~~~~~~~ 1200 1201Set the link flow control parameter on a port:: 1202 1203 testpmd> set flow_ctrl rx (on|off) tx (on|off) (high_water) (low_water) \ 1204 (pause_time) (send_xon) mac_ctrl_frame_fwd (on|off) \ 1205 autoneg (on|off) (port_id) 1206 1207Where: 1208 1209* ``high_water`` (integer): High threshold value to trigger XOFF. 1210 1211* ``low_water`` (integer): Low threshold value to trigger XON. 1212 1213* ``pause_time`` (integer): Pause quota in the Pause frame. 1214 1215* ``send_xon`` (0/1): Send XON frame. 1216 1217* ``mac_ctrl_frame_fwd``: Enable receiving MAC control frames. 1218 1219* ``autoneg``: Change the auto-negotiation parameter. 1220 1221set pfc_ctrl rx 1222~~~~~~~~~~~~~~~ 1223 1224Set the priority flow control parameter on a port:: 1225 1226 testpmd> set pfc_ctrl rx (on|off) tx (on|off) (high_water) (low_water) \ 1227 (pause_time) (priority) (port_id) 1228 1229Where: 1230 1231* ``high_water`` (integer): High threshold value. 1232 1233* ``low_water`` (integer): Low threshold value. 1234 1235* ``pause_time`` (integer): Pause quota in the Pause frame. 1236 1237* ``priority`` (0-7): VLAN User Priority. 1238 1239set stat_qmap 1240~~~~~~~~~~~~~ 1241 1242Set statistics mapping (qmapping 0..15) for RX/TX queue on port:: 1243 1244 testpmd> set stat_qmap (tx|rx) (port_id) (queue_id) (qmapping) 1245 1246For example, to set rx queue 2 on port 0 to mapping 5:: 1247 1248 testpmd>set stat_qmap rx 0 2 5 1249 1250set xstats-hide-zero 1251~~~~~~~~~~~~~~~~~~~~ 1252 1253Set the option to hide zero values for xstats display:: 1254 1255 testpmd> set xstats-hide-zero on|off 1256 1257.. note:: 1258 1259 By default, the zero values are displayed for xstats. 1260 1261set port - rx/tx (for VF) 1262~~~~~~~~~~~~~~~~~~~~~~~~~ 1263 1264Set VF receive/transmit from a port:: 1265 1266 testpmd> set port (port_id) vf (vf_id) (rx|tx) (on|off) 1267 1268set port - mac address filter (for VF) 1269~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1270 1271Add/Remove unicast or multicast MAC addr filter for a VF:: 1272 1273 testpmd> set port (port_id) vf (vf_id) (mac_addr) \ 1274 (exact-mac|exact-mac-vlan|hashmac|hashmac-vlan) (on|off) 1275 1276set port - rx mode(for VF) 1277~~~~~~~~~~~~~~~~~~~~~~~~~~ 1278 1279Set the VF receive mode of a port:: 1280 1281 testpmd> set port (port_id) vf (vf_id) \ 1282 rxmode (AUPE|ROPE|BAM|MPE) (on|off) 1283 1284The available receive modes are: 1285 1286* ``AUPE``: Accepts untagged VLAN. 1287 1288* ``ROPE``: Accepts unicast hash. 1289 1290* ``BAM``: Accepts broadcast packets. 1291 1292* ``MPE``: Accepts all multicast packets. 1293 1294set port - tx_rate (for Queue) 1295~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1296 1297Set TX rate limitation for a queue on a port:: 1298 1299 testpmd> set port (port_id) queue (queue_id) rate (rate_value) 1300 1301set port - tx_rate (for VF) 1302~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1303 1304Set TX rate limitation for queues in VF on a port:: 1305 1306 testpmd> set port (port_id) vf (vf_id) rate (rate_value) queue_mask (queue_mask) 1307 1308set port - mirror rule 1309~~~~~~~~~~~~~~~~~~~~~~ 1310 1311Set pool or vlan type mirror rule for a port:: 1312 1313 testpmd> set port (port_id) mirror-rule (rule_id) \ 1314 (pool-mirror-up|pool-mirror-down|vlan-mirror) \ 1315 (poolmask|vlanid[,vlanid]*) dst-pool (pool_id) (on|off) 1316 1317Set link mirror rule for a port:: 1318 1319 testpmd> set port (port_id) mirror-rule (rule_id) \ 1320 (uplink-mirror|downlink-mirror) dst-pool (pool_id) (on|off) 1321 1322For example to enable mirror traffic with vlan 0,1 to pool 0:: 1323 1324 set port 0 mirror-rule 0 vlan-mirror 0,1 dst-pool 0 on 1325 1326reset port - mirror rule 1327~~~~~~~~~~~~~~~~~~~~~~~~ 1328 1329Reset a mirror rule for a port:: 1330 1331 testpmd> reset port (port_id) mirror-rule (rule_id) 1332 1333set flush_rx 1334~~~~~~~~~~~~ 1335 1336Set the flush on RX streams before forwarding. 1337The default is flush ``on``. 1338Mainly used with PCAP drivers to turn off the default behavior of flushing the first 512 packets on RX streams:: 1339 1340 testpmd> set flush_rx off 1341 1342set bypass mode 1343~~~~~~~~~~~~~~~ 1344 1345Set the bypass mode for the lowest port on bypass enabled NIC:: 1346 1347 testpmd> set bypass mode (normal|bypass|isolate) (port_id) 1348 1349set bypass event 1350~~~~~~~~~~~~~~~~ 1351 1352Set the event required to initiate specified bypass mode for the lowest port on a bypass enabled:: 1353 1354 testpmd> set bypass event (timeout|os_on|os_off|power_on|power_off) \ 1355 mode (normal|bypass|isolate) (port_id) 1356 1357Where: 1358 1359* ``timeout``: Enable bypass after watchdog timeout. 1360 1361* ``os_on``: Enable bypass when OS/board is powered on. 1362 1363* ``os_off``: Enable bypass when OS/board is powered off. 1364 1365* ``power_on``: Enable bypass when power supply is turned on. 1366 1367* ``power_off``: Enable bypass when power supply is turned off. 1368 1369 1370set bypass timeout 1371~~~~~~~~~~~~~~~~~~ 1372 1373Set the bypass watchdog timeout to ``n`` seconds where 0 = instant:: 1374 1375 testpmd> set bypass timeout (0|1.5|2|3|4|8|16|32) 1376 1377show bypass config 1378~~~~~~~~~~~~~~~~~~ 1379 1380Show the bypass configuration for a bypass enabled NIC using the lowest port on the NIC:: 1381 1382 testpmd> show bypass config (port_id) 1383 1384set link up 1385~~~~~~~~~~~ 1386 1387Set link up for a port:: 1388 1389 testpmd> set link-up port (port id) 1390 1391set link down 1392~~~~~~~~~~~~~ 1393 1394Set link down for a port:: 1395 1396 testpmd> set link-down port (port id) 1397 1398E-tag set 1399~~~~~~~~~ 1400 1401Enable E-tag insertion for a VF on a port:: 1402 1403 testpmd> E-tag set insertion on port-tag-id (value) port (port_id) vf (vf_id) 1404 1405Disable E-tag insertion for a VF on a port:: 1406 1407 testpmd> E-tag set insertion off port (port_id) vf (vf_id) 1408 1409Enable/disable E-tag stripping on a port:: 1410 1411 testpmd> E-tag set stripping (on|off) port (port_id) 1412 1413Enable/disable E-tag based forwarding on a port:: 1414 1415 testpmd> E-tag set forwarding (on|off) port (port_id) 1416 1417Add an E-tag forwarding filter on a port:: 1418 1419 testpmd> E-tag set filter add e-tag-id (value) dst-pool (pool_id) port (port_id) 1420 1421Delete an E-tag forwarding filter on a port:: 1422 testpmd> E-tag set filter del e-tag-id (value) port (port_id) 1423 1424ddp add 1425~~~~~~~ 1426 1427Load a dynamic device personalization (DDP) profile and store backup profile:: 1428 1429 testpmd> ddp add (port_id) (profile_path[,backup_profile_path]) 1430 1431ddp del 1432~~~~~~~ 1433 1434Delete a dynamic device personalization profile and restore backup profile:: 1435 1436 testpmd> ddp del (port_id) (backup_profile_path) 1437 1438ptype mapping 1439~~~~~~~~~~~~~ 1440 1441List all items from the ptype mapping table:: 1442 1443 testpmd> ptype mapping get (port_id) (valid_only) 1444 1445Where: 1446 1447* ``valid_only``: A flag indicates if only list valid items(=1) or all itemss(=0). 1448 1449Replace a specific or a group of software defined ptype with a new one:: 1450 1451 testpmd> ptype mapping replace (port_id) (target) (mask) (pkt_type) 1452 1453where: 1454 1455* ``target``: A specific software ptype or a mask to represent a group of software ptypes. 1456 1457* ``mask``: A flag indicate if "target" is a specific software ptype(=0) or a ptype mask(=1). 1458 1459* ``pkt_type``: The new software ptype to replace the old ones. 1460 1461Update hardware defined ptype to software defined packet type mapping table:: 1462 1463 testpmd> ptype mapping update (port_id) (hw_ptype) (sw_ptype) 1464 1465where: 1466 1467* ``hw_ptype``: hardware ptype as the index of the ptype mapping table. 1468 1469* ``sw_ptype``: software ptype as the value of the ptype mapping table. 1470 1471Reset ptype mapping table:: 1472 1473 testpmd> ptype mapping reset (port_id) 1474 1475config per port Rx offloading 1476~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1477 1478Enable or disable a per port Rx offloading on all Rx queues of a port:: 1479 1480 testpmd> port config (port_id) rx_offload (offloading) on|off 1481 1482* ``offloading``: can be any of these offloading capability: 1483 vlan_strip, ipv4_cksum, udp_cksum, tcp_cksum, tcp_lro, 1484 qinq_strip, outer_ipv4_cksum, macsec_strip, 1485 header_split, vlan_filter, vlan_extend, jumbo_frame, 1486 crc_strip, scatter, timestamp, security 1487 1488This command should be run when the port is stopped, or else it will fail. 1489 1490config per queue Rx offloading 1491~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1492 1493Enable or disable a per queue Rx offloading only on a specific Rx queue:: 1494 1495 testpmd> port (port_id) rxq (queue_id) rx_offload (offloading) on|off 1496 1497* ``offloading``: can be any of these offloading capability: 1498 vlan_strip, ipv4_cksum, udp_cksum, tcp_cksum, tcp_lro, 1499 qinq_strip, outer_ipv4_cksum, macsec_strip, 1500 header_split, vlan_filter, vlan_extend, jumbo_frame, 1501 crc_strip, scatter, timestamp, security 1502 1503This command should be run when the port is stopped, or else it will fail. 1504 1505config per port Tx offloading 1506~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1507 1508Enable or disable a per port Tx offloading on all Tx queues of a port:: 1509 1510 testpmd> port config (port_id) tx_offload (offloading) on|off 1511 1512* ``offloading``: can be any of these offloading capability: 1513 vlan_insert, ipv4_cksum, udp_cksum, udp_cksum, 1514 sctp_cksum, tcp_tso, udp_tso, outer_ipv4_cksum, 1515 qinq_insert, vxlan_tnl_tso, gre_tnl_tso, 1516 ipip_tnl_tso, geneve_tnl_tso, macsec_insert, 1517 mt_lockfree, multi_segs, fast_free, security 1518 1519This command should be run when the port is stopped, or else it will fail. 1520 1521config per queue Tx offloading 1522~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1523 1524Enable or disable a per queue Tx offloading only on a specific Tx queue:: 1525 1526 testpmd> port (port_id) txq (queue_id) tx_offload (offloading) on|off 1527 1528* ``offloading``: can be any of these offloading capability: 1529 vlan_insert, ipv4_cksum, udp_cksum, udp_cksum, 1530 sctp_cksum, tcp_tso, udp_tso, outer_ipv4_cksum, 1531 qinq_insert, vxlan_tnl_tso, gre_tnl_tso, 1532 ipip_tnl_tso, geneve_tnl_tso, macsec_insert, 1533 mt_lockfree, multi_segs, fast_free, security 1534 1535This command should be run when the port is stopped, or else it will fail. 1536 1537 1538Port Functions 1539-------------- 1540 1541The following sections show functions for configuring ports. 1542 1543.. note:: 1544 1545 Port configuration changes only become active when forwarding is started/restarted. 1546 1547port attach 1548~~~~~~~~~~~ 1549 1550Attach a port specified by pci address or virtual device args:: 1551 1552 testpmd> port attach (identifier) 1553 1554To attach a new pci device, the device should be recognized by kernel first. 1555Then it should be moved under DPDK management. 1556Finally the port can be attached to testpmd. 1557 1558For example, to move a pci device using ixgbe under DPDK management: 1559 1560.. code-block:: console 1561 1562 # Check the status of the available devices. 1563 ./usertools/dpdk-devbind.py --status 1564 1565 Network devices using DPDK-compatible driver 1566 ============================================ 1567 <none> 1568 1569 Network devices using kernel driver 1570 =================================== 1571 0000:0a:00.0 '82599ES 10-Gigabit' if=eth2 drv=ixgbe unused= 1572 1573 1574 # Bind the device to igb_uio. 1575 sudo ./usertools/dpdk-devbind.py -b igb_uio 0000:0a:00.0 1576 1577 1578 # Recheck the status of the devices. 1579 ./usertools/dpdk-devbind.py --status 1580 Network devices using DPDK-compatible driver 1581 ============================================ 1582 0000:0a:00.0 '82599ES 10-Gigabit' drv=igb_uio unused= 1583 1584To attach a port created by virtual device, above steps are not needed. 1585 1586For example, to attach a port whose pci address is 0000:0a:00.0. 1587 1588.. code-block:: console 1589 1590 testpmd> port attach 0000:0a:00.0 1591 Attaching a new port... 1592 EAL: PCI device 0000:0a:00.0 on NUMA socket -1 1593 EAL: probe driver: 8086:10fb rte_ixgbe_pmd 1594 EAL: PCI memory mapped at 0x7f83bfa00000 1595 EAL: PCI memory mapped at 0x7f83bfa80000 1596 PMD: eth_ixgbe_dev_init(): MAC: 2, PHY: 18, SFP+: 5 1597 PMD: eth_ixgbe_dev_init(): port 0 vendorID=0x8086 deviceID=0x10fb 1598 Port 0 is attached. Now total ports is 1 1599 Done 1600 1601For example, to attach a port created by pcap PMD. 1602 1603.. code-block:: console 1604 1605 testpmd> port attach net_pcap0 1606 Attaching a new port... 1607 PMD: Initializing pmd_pcap for net_pcap0 1608 PMD: Creating pcap-backed ethdev on numa socket 0 1609 Port 0 is attached. Now total ports is 1 1610 Done 1611 1612In this case, identifier is ``net_pcap0``. 1613This identifier format is the same as ``--vdev`` format of DPDK applications. 1614 1615For example, to re-attach a bonded port which has been previously detached, 1616the mode and slave parameters must be given. 1617 1618.. code-block:: console 1619 1620 testpmd> port attach net_bond_0,mode=0,slave=1 1621 Attaching a new port... 1622 EAL: Initializing pmd_bond for net_bond_0 1623 EAL: Create bonded device net_bond_0 on port 0 in mode 0 on socket 0. 1624 Port 0 is attached. Now total ports is 1 1625 Done 1626 1627 1628port detach 1629~~~~~~~~~~~ 1630 1631Detach a specific port:: 1632 1633 testpmd> port detach (port_id) 1634 1635Before detaching a port, the port should be stopped and closed. 1636 1637For example, to detach a pci device port 0. 1638 1639.. code-block:: console 1640 1641 testpmd> port stop 0 1642 Stopping ports... 1643 Done 1644 testpmd> port close 0 1645 Closing ports... 1646 Done 1647 1648 testpmd> port detach 0 1649 Detaching a port... 1650 EAL: PCI device 0000:0a:00.0 on NUMA socket -1 1651 EAL: remove driver: 8086:10fb rte_ixgbe_pmd 1652 EAL: PCI memory unmapped at 0x7f83bfa00000 1653 EAL: PCI memory unmapped at 0x7f83bfa80000 1654 Done 1655 1656 1657For example, to detach a virtual device port 0. 1658 1659.. code-block:: console 1660 1661 testpmd> port stop 0 1662 Stopping ports... 1663 Done 1664 testpmd> port close 0 1665 Closing ports... 1666 Done 1667 1668 testpmd> port detach 0 1669 Detaching a port... 1670 PMD: Closing pcap ethdev on numa socket 0 1671 Port 'net_pcap0' is detached. Now total ports is 0 1672 Done 1673 1674To remove a pci device completely from the system, first detach the port from testpmd. 1675Then the device should be moved under kernel management. 1676Finally the device can be removed using kernel pci hotplug functionality. 1677 1678For example, to move a pci device under kernel management: 1679 1680.. code-block:: console 1681 1682 sudo ./usertools/dpdk-devbind.py -b ixgbe 0000:0a:00.0 1683 1684 ./usertools/dpdk-devbind.py --status 1685 1686 Network devices using DPDK-compatible driver 1687 ============================================ 1688 <none> 1689 1690 Network devices using kernel driver 1691 =================================== 1692 0000:0a:00.0 '82599ES 10-Gigabit' if=eth2 drv=ixgbe unused=igb_uio 1693 1694To remove a port created by a virtual device, above steps are not needed. 1695 1696port start 1697~~~~~~~~~~ 1698 1699Start all ports or a specific port:: 1700 1701 testpmd> port start (port_id|all) 1702 1703port stop 1704~~~~~~~~~ 1705 1706Stop all ports or a specific port:: 1707 1708 testpmd> port stop (port_id|all) 1709 1710port close 1711~~~~~~~~~~ 1712 1713Close all ports or a specific port:: 1714 1715 testpmd> port close (port_id|all) 1716 1717port config - queue ring size 1718~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1719 1720Configure a rx/tx queue ring size:: 1721 1722 testpmd> port (port_id) (rxq|txq) (queue_id) ring_size (value) 1723 1724Only take effect after command that (re-)start the port or command that setup specific queue. 1725 1726port start/stop queue 1727~~~~~~~~~~~~~~~~~~~~~ 1728 1729Start/stop a rx/tx queue on a specific port:: 1730 1731 testpmd> port (port_id) (rxq|txq) (queue_id) (start|stop) 1732 1733port setup queue 1734~~~~~~~~~~~~~~~~~~~~~ 1735 1736Setup a rx/tx queue on a specific port:: 1737 1738 testpmd> port (port_id) (rxq|txq) (queue_id) setup 1739 1740Only take effect when port is started. 1741 1742port config - speed 1743~~~~~~~~~~~~~~~~~~~ 1744 1745Set the speed and duplex mode for all ports or a specific port:: 1746 1747 testpmd> port config (port_id|all) speed (10|100|1000|10000|25000|40000|50000|100000|auto) \ 1748 duplex (half|full|auto) 1749 1750port config - queues/descriptors 1751~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1752 1753Set number of queues/descriptors for rxq, txq, rxd and txd:: 1754 1755 testpmd> port config all (rxq|txq|rxd|txd) (value) 1756 1757This is equivalent to the ``--rxq``, ``--txq``, ``--rxd`` and ``--txd`` command-line options. 1758 1759port config - max-pkt-len 1760~~~~~~~~~~~~~~~~~~~~~~~~~ 1761 1762Set the maximum packet length:: 1763 1764 testpmd> port config all max-pkt-len (value) 1765 1766This is equivalent to the ``--max-pkt-len`` command-line option. 1767 1768port config - CRC Strip 1769~~~~~~~~~~~~~~~~~~~~~~~ 1770 1771Set hardware CRC stripping on or off for all ports:: 1772 1773 testpmd> port config all crc-strip (on|off) 1774 1775CRC stripping is on by default. 1776 1777The ``off`` option is equivalent to the ``--disable-crc-strip`` command-line option. 1778 1779port config - scatter 1780~~~~~~~~~~~~~~~~~~~~~~~ 1781 1782Set RX scatter mode on or off for all ports:: 1783 1784 testpmd> port config all scatter (on|off) 1785 1786RX scatter mode is off by default. 1787 1788The ``on`` option is equivalent to the ``--enable-scatter`` command-line option. 1789 1790port config - RX Checksum 1791~~~~~~~~~~~~~~~~~~~~~~~~~ 1792 1793Set hardware RX checksum offload to on or off for all ports:: 1794 1795 testpmd> port config all rx-cksum (on|off) 1796 1797Checksum offload is off by default. 1798 1799The ``on`` option is equivalent to the ``--enable-rx-cksum`` command-line option. 1800 1801port config - VLAN 1802~~~~~~~~~~~~~~~~~~ 1803 1804Set hardware VLAN on or off for all ports:: 1805 1806 testpmd> port config all hw-vlan (on|off) 1807 1808Hardware VLAN is off by default. 1809 1810The ``on`` option is equivalent to the ``--enable-hw-vlan`` command-line option. 1811 1812port config - VLAN filter 1813~~~~~~~~~~~~~~~~~~~~~~~~~ 1814 1815Set hardware VLAN filter on or off for all ports:: 1816 1817 testpmd> port config all hw-vlan-filter (on|off) 1818 1819Hardware VLAN filter is off by default. 1820 1821The ``on`` option is equivalent to the ``--enable-hw-vlan-filter`` command-line option. 1822 1823port config - VLAN strip 1824~~~~~~~~~~~~~~~~~~~~~~~~ 1825 1826Set hardware VLAN strip on or off for all ports:: 1827 1828 testpmd> port config all hw-vlan-strip (on|off) 1829 1830Hardware VLAN strip is off by default. 1831 1832The ``on`` option is equivalent to the ``--enable-hw-vlan-strip`` command-line option. 1833 1834port config - VLAN extend 1835~~~~~~~~~~~~~~~~~~~~~~~~~ 1836 1837Set hardware VLAN extend on or off for all ports:: 1838 1839 testpmd> port config all hw-vlan-extend (on|off) 1840 1841Hardware VLAN extend is off by default. 1842 1843The ``on`` option is equivalent to the ``--enable-hw-vlan-extend`` command-line option. 1844 1845port config - Drop Packets 1846~~~~~~~~~~~~~~~~~~~~~~~~~~ 1847 1848Set packet drop for packets with no descriptors on or off for all ports:: 1849 1850 testpmd> port config all drop-en (on|off) 1851 1852Packet dropping for packets with no descriptors is off by default. 1853 1854The ``on`` option is equivalent to the ``--enable-drop-en`` command-line option. 1855 1856port config - RSS 1857~~~~~~~~~~~~~~~~~ 1858 1859Set the RSS (Receive Side Scaling) mode on or off:: 1860 1861 testpmd> port config all rss (all|default|ip|tcp|udp|sctp|ether|port|vxlan|geneve|nvgre|none) 1862 1863RSS is on by default. 1864 1865The ``all`` option is equivalent to ip|tcp|udp|sctp|ether. 1866The ``default`` option enables all supported RSS types reported by device info. 1867The ``none`` option is equivalent to the ``--disable-rss`` command-line option. 1868 1869port config - RSS Reta 1870~~~~~~~~~~~~~~~~~~~~~~ 1871 1872Set the RSS (Receive Side Scaling) redirection table:: 1873 1874 testpmd> port config all rss reta (hash,queue)[,(hash,queue)] 1875 1876port config - DCB 1877~~~~~~~~~~~~~~~~~ 1878 1879Set the DCB mode for an individual port:: 1880 1881 testpmd> port config (port_id) dcb vt (on|off) (traffic_class) pfc (on|off) 1882 1883The traffic class should be 4 or 8. 1884 1885port config - Burst 1886~~~~~~~~~~~~~~~~~~~ 1887 1888Set the number of packets per burst:: 1889 1890 testpmd> port config all burst (value) 1891 1892This is equivalent to the ``--burst`` command-line option. 1893 1894port config - Threshold 1895~~~~~~~~~~~~~~~~~~~~~~~ 1896 1897Set thresholds for TX/RX queues:: 1898 1899 testpmd> port config all (threshold) (value) 1900 1901Where the threshold type can be: 1902 1903* ``txpt:`` Set the prefetch threshold register of the TX rings, 0 <= value <= 255. 1904 1905* ``txht:`` Set the host threshold register of the TX rings, 0 <= value <= 255. 1906 1907* ``txwt:`` Set the write-back threshold register of the TX rings, 0 <= value <= 255. 1908 1909* ``rxpt:`` Set the prefetch threshold register of the RX rings, 0 <= value <= 255. 1910 1911* ``rxht:`` Set the host threshold register of the RX rings, 0 <= value <= 255. 1912 1913* ``rxwt:`` Set the write-back threshold register of the RX rings, 0 <= value <= 255. 1914 1915* ``txfreet:`` Set the transmit free threshold of the TX rings, 0 <= value <= txd. 1916 1917* ``rxfreet:`` Set the transmit free threshold of the RX rings, 0 <= value <= rxd. 1918 1919* ``txrst:`` Set the transmit RS bit threshold of TX rings, 0 <= value <= txd. 1920 1921These threshold options are also available from the command-line. 1922 1923port config - E-tag 1924~~~~~~~~~~~~~~~~~~~ 1925 1926Set the value of ether-type for E-tag:: 1927 1928 testpmd> port config (port_id|all) l2-tunnel E-tag ether-type (value) 1929 1930Enable/disable the E-tag support:: 1931 1932 testpmd> port config (port_id|all) l2-tunnel E-tag (enable|disable) 1933 1934port config pctype mapping 1935~~~~~~~~~~~~~~~~~~~~~~~~~~ 1936 1937Reset pctype mapping table:: 1938 1939 testpmd> port config (port_id) pctype mapping reset 1940 1941Update hardware defined pctype to software defined flow type mapping table:: 1942 1943 testpmd> port config (port_id) pctype mapping update (pctype_id_0[,pctype_id_1]*) (flow_type_id) 1944 1945where: 1946 1947* ``pctype_id_x``: hardware pctype id as index of bit in bitmask value of the pctype mapping table. 1948 1949* ``flow_type_id``: software flow type id as the index of the pctype mapping table. 1950 1951port config input set 1952~~~~~~~~~~~~~~~~~~~~~ 1953 1954Config RSS/FDIR/FDIR flexible payload input set for some pctype:: 1955 testpmd> port config (port_id) pctype (pctype_id) \ 1956 (hash_inset|fdir_inset|fdir_flx_inset) \ 1957 (get|set|clear) field (field_idx) 1958 1959Clear RSS/FDIR/FDIR flexible payload input set for some pctype:: 1960 testpmd> port config (port_id) pctype (pctype_id) \ 1961 (hash_inset|fdir_inset|fdir_flx_inset) clear all 1962 1963where: 1964 1965* ``pctype_id``: hardware packet classification types. 1966* ``field_idx``: hardware field index. 1967 1968port config udp_tunnel_port 1969~~~~~~~~~~~~~~~~~~~~~~~~~~~ 1970 1971Add/remove UDP tunnel port for VXLAN/GENEVE tunneling protocols:: 1972 testpmd> port config (port_id) udp_tunnel_port add|rm vxlan|geneve (udp_port) 1973 1974Link Bonding Functions 1975---------------------- 1976 1977The Link Bonding functions make it possible to dynamically create and 1978manage link bonding devices from within testpmd interactive prompt. 1979 1980create bonded device 1981~~~~~~~~~~~~~~~~~~~~ 1982 1983Create a new bonding device:: 1984 1985 testpmd> create bonded device (mode) (socket) 1986 1987For example, to create a bonded device in mode 1 on socket 0:: 1988 1989 testpmd> create bonded 1 0 1990 created new bonded device (port X) 1991 1992add bonding slave 1993~~~~~~~~~~~~~~~~~ 1994 1995Adds Ethernet device to a Link Bonding device:: 1996 1997 testpmd> add bonding slave (slave id) (port id) 1998 1999For example, to add Ethernet device (port 6) to a Link Bonding device (port 10):: 2000 2001 testpmd> add bonding slave 6 10 2002 2003 2004remove bonding slave 2005~~~~~~~~~~~~~~~~~~~~ 2006 2007Removes an Ethernet slave device from a Link Bonding device:: 2008 2009 testpmd> remove bonding slave (slave id) (port id) 2010 2011For example, to remove Ethernet slave device (port 6) to a Link Bonding device (port 10):: 2012 2013 testpmd> remove bonding slave 6 10 2014 2015set bonding mode 2016~~~~~~~~~~~~~~~~ 2017 2018Set the Link Bonding mode of a Link Bonding device:: 2019 2020 testpmd> set bonding mode (value) (port id) 2021 2022For example, to set the bonding mode of a Link Bonding device (port 10) to broadcast (mode 3):: 2023 2024 testpmd> set bonding mode 3 10 2025 2026set bonding primary 2027~~~~~~~~~~~~~~~~~~~ 2028 2029Set an Ethernet slave device as the primary device on a Link Bonding device:: 2030 2031 testpmd> set bonding primary (slave id) (port id) 2032 2033For example, to set the Ethernet slave device (port 6) as the primary port of a Link Bonding device (port 10):: 2034 2035 testpmd> set bonding primary 6 10 2036 2037set bonding mac 2038~~~~~~~~~~~~~~~ 2039 2040Set the MAC address of a Link Bonding device:: 2041 2042 testpmd> set bonding mac (port id) (mac) 2043 2044For example, to set the MAC address of a Link Bonding device (port 10) to 00:00:00:00:00:01:: 2045 2046 testpmd> set bonding mac 10 00:00:00:00:00:01 2047 2048set bonding xmit_balance_policy 2049~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2050 2051Set the transmission policy for a Link Bonding device when it is in Balance XOR mode:: 2052 2053 testpmd> set bonding xmit_balance_policy (port_id) (l2|l23|l34) 2054 2055For example, set a Link Bonding device (port 10) to use a balance policy of layer 3+4 (IP addresses & UDP ports):: 2056 2057 testpmd> set bonding xmit_balance_policy 10 l34 2058 2059 2060set bonding mon_period 2061~~~~~~~~~~~~~~~~~~~~~~ 2062 2063Set the link status monitoring polling period in milliseconds for a bonding device. 2064 2065This adds support for PMD slave devices which do not support link status interrupts. 2066When the mon_period is set to a value greater than 0 then all PMD's which do not support 2067link status ISR will be queried every polling interval to check if their link status has changed:: 2068 2069 testpmd> set bonding mon_period (port_id) (value) 2070 2071For example, to set the link status monitoring polling period of bonded device (port 5) to 150ms:: 2072 2073 testpmd> set bonding mon_period 5 150 2074 2075 2076set bonding lacp dedicated_queue 2077~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2078 2079Enable dedicated tx/rx queues on bonding devices slaves to handle LACP control plane traffic 2080when in mode 4 (link-aggregration-802.3ad):: 2081 2082 testpmd> set bonding lacp dedicated_queues (port_id) (enable|disable) 2083 2084 2085set bonding agg_mode 2086~~~~~~~~~~~~~~~~~~~~ 2087 2088Enable one of the specific aggregators mode when in mode 4 (link-aggregration-802.3ad):: 2089 2090 testpmd> set bonding agg_mode (port_id) (bandwidth|count|stable) 2091 2092 2093show bonding config 2094~~~~~~~~~~~~~~~~~~~ 2095 2096Show the current configuration of a Link Bonding device:: 2097 2098 testpmd> show bonding config (port id) 2099 2100For example, 2101to show the configuration a Link Bonding device (port 9) with 3 slave devices (1, 3, 4) 2102in balance mode with a transmission policy of layer 2+3:: 2103 2104 testpmd> show bonding config 9 2105 Bonding mode: 2 2106 Balance Xmit Policy: BALANCE_XMIT_POLICY_LAYER23 2107 Slaves (3): [1 3 4] 2108 Active Slaves (3): [1 3 4] 2109 Primary: [3] 2110 2111 2112Register Functions 2113------------------ 2114 2115The Register Functions can be used to read from and write to registers on the network card referenced by a port number. 2116This is mainly useful for debugging purposes. 2117Reference should be made to the appropriate datasheet for the network card for details on the register addresses 2118and fields that can be accessed. 2119 2120read reg 2121~~~~~~~~ 2122 2123Display the value of a port register:: 2124 2125 testpmd> read reg (port_id) (address) 2126 2127For example, to examine the Flow Director control register (FDIRCTL, 0x0000EE000) on an Intel 82599 10 GbE Controller:: 2128 2129 testpmd> read reg 0 0xEE00 2130 port 0 PCI register at offset 0xEE00: 0x4A060029 (1241907241) 2131 2132read regfield 2133~~~~~~~~~~~~~ 2134 2135Display a port register bit field:: 2136 2137 testpmd> read regfield (port_id) (address) (bit_x) (bit_y) 2138 2139For example, reading the lowest two bits from the register in the example above:: 2140 2141 testpmd> read regfield 0 0xEE00 0 1 2142 port 0 PCI register at offset 0xEE00: bits[0, 1]=0x1 (1) 2143 2144read regbit 2145~~~~~~~~~~~ 2146 2147Display a single port register bit:: 2148 2149 testpmd> read regbit (port_id) (address) (bit_x) 2150 2151For example, reading the lowest bit from the register in the example above:: 2152 2153 testpmd> read regbit 0 0xEE00 0 2154 port 0 PCI register at offset 0xEE00: bit 0=1 2155 2156write reg 2157~~~~~~~~~ 2158 2159Set the value of a port register:: 2160 2161 testpmd> write reg (port_id) (address) (value) 2162 2163For example, to clear a register:: 2164 2165 testpmd> write reg 0 0xEE00 0x0 2166 port 0 PCI register at offset 0xEE00: 0x00000000 (0) 2167 2168write regfield 2169~~~~~~~~~~~~~~ 2170 2171Set bit field of a port register:: 2172 2173 testpmd> write regfield (port_id) (address) (bit_x) (bit_y) (value) 2174 2175For example, writing to the register cleared in the example above:: 2176 2177 testpmd> write regfield 0 0xEE00 0 1 2 2178 port 0 PCI register at offset 0xEE00: 0x00000002 (2) 2179 2180write regbit 2181~~~~~~~~~~~~ 2182 2183Set single bit value of a port register:: 2184 2185 testpmd> write regbit (port_id) (address) (bit_x) (value) 2186 2187For example, to set the high bit in the register from the example above:: 2188 2189 testpmd> write regbit 0 0xEE00 31 1 2190 port 0 PCI register at offset 0xEE00: 0x8000000A (2147483658) 2191 2192Traffic Metering and Policing 2193----------------------------- 2194 2195The following section shows functions for configuring traffic metering and 2196policing on the ethernet device through the use of generic ethdev API. 2197 2198show port traffic management capability 2199~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2200 2201Show traffic metering and policing capability of the port:: 2202 2203 testpmd> show port meter cap (port_id) 2204 2205add port meter profile (srTCM rfc2967) 2206~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2207 2208Add meter profile (srTCM rfc2697) to the ethernet device:: 2209 2210 testpmd> add port meter profile srtcm_rfc2697 (port_id) (profile_id) \ 2211 (cir) (cbs) (ebs) 2212 2213where: 2214 2215* ``profile_id``: ID for the meter profile. 2216* ``cir``: Committed Information Rate (CIR) (bytes/second). 2217* ``cbs``: Committed Burst Size (CBS) (bytes). 2218* ``ebs``: Excess Burst Size (EBS) (bytes). 2219 2220add port meter profile (trTCM rfc2968) 2221~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2222 2223Add meter profile (srTCM rfc2698) to the ethernet device:: 2224 2225 testpmd> add port meter profile trtcm_rfc2698 (port_id) (profile_id) \ 2226 (cir) (pir) (cbs) (pbs) 2227 2228where: 2229 2230* ``profile_id``: ID for the meter profile. 2231* ``cir``: Committed information rate (bytes/second). 2232* ``pir``: Peak information rate (bytes/second). 2233* ``cbs``: Committed burst size (bytes). 2234* ``pbs``: Peak burst size (bytes). 2235 2236add port meter profile (trTCM rfc4115) 2237~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2238 2239Add meter profile (trTCM rfc4115) to the ethernet device:: 2240 2241 testpmd> add port meter profile trtcm_rfc4115 (port_id) (profile_id) \ 2242 (cir) (eir) (cbs) (ebs) 2243 2244where: 2245 2246* ``profile_id``: ID for the meter profile. 2247* ``cir``: Committed information rate (bytes/second). 2248* ``eir``: Excess information rate (bytes/second). 2249* ``cbs``: Committed burst size (bytes). 2250* ``ebs``: Excess burst size (bytes). 2251 2252delete port meter profile 2253~~~~~~~~~~~~~~~~~~~~~~~~~ 2254 2255Delete meter profile from the ethernet device:: 2256 2257 testpmd> del port meter profile (port_id) (profile_id) 2258 2259create port meter 2260~~~~~~~~~~~~~~~~~ 2261 2262Create new meter object for the ethernet device:: 2263 2264 testpmd> create port meter (port_id) (mtr_id) (profile_id) \ 2265 (meter_enable) (g_action) (y_action) (r_action) (stats_mask) (shared) \ 2266 (use_pre_meter_color) [(dscp_tbl_entry0) (dscp_tbl_entry1)...\ 2267 (dscp_tbl_entry63)] 2268 2269where: 2270 2271* ``mtr_id``: meter object ID. 2272* ``profile_id``: ID for the meter profile. 2273* ``meter_enable``: When this parameter has a non-zero value, the meter object 2274 gets enabled at the time of creation, otherwise remains disabled. 2275* ``g_action``: Policer action for the packet with green color. 2276* ``y_action``: Policer action for the packet with yellow color. 2277* ``r_action``: Policer action for the packet with red color. 2278* ``stats_mask``: Mask of statistics counter types to be enabled for the 2279 meter object. 2280* ``shared``: When this parameter has a non-zero value, the meter object is 2281 shared by multiple flows. Otherwise, meter object is used by single flow. 2282* ``use_pre_meter_color``: When this parameter has a non-zero value, the 2283 input color for the current meter object is determined by the latest meter 2284 object in the same flow. Otherwise, the current meter object uses the 2285 *dscp_table* to determine the input color. 2286* ``dscp_tbl_entryx``: DSCP table entry x providing meter providing input 2287 color, 0 <= x <= 63. 2288 2289enable port meter 2290~~~~~~~~~~~~~~~~~ 2291 2292Enable meter for the ethernet device:: 2293 2294 testpmd> enable port meter (port_id) (mtr_id) 2295 2296disable port meter 2297~~~~~~~~~~~~~~~~~~ 2298 2299Disable meter for the ethernet device:: 2300 2301 testpmd> disable port meter (port_id) (mtr_id) 2302 2303delete port meter 2304~~~~~~~~~~~~~~~~~ 2305 2306Delete meter for the ethernet device:: 2307 2308 testpmd> del port meter (port_id) (mtr_id) 2309 2310Set port meter profile 2311~~~~~~~~~~~~~~~~~~~~~~ 2312 2313Set meter profile for the ethernet device:: 2314 2315 testpmd> set port meter profile (port_id) (mtr_id) (profile_id) 2316 2317set port meter dscp table 2318~~~~~~~~~~~~~~~~~~~~~~~~~ 2319 2320Set meter dscp table for the ethernet device:: 2321 2322 testpmd> set port meter dscp table (port_id) (mtr_id) [(dscp_tbl_entry0) \ 2323 (dscp_tbl_entry1)...(dscp_tbl_entry63)] 2324 2325set port meter policer action 2326~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2327 2328Set meter policer action for the ethernet device:: 2329 2330 testpmd> set port meter policer action (port_id) (mtr_id) (action_mask) \ 2331 (action0) [(action1) (action1)] 2332 2333where: 2334 2335* ``action_mask``: Bit mask indicating which policer actions need to be 2336 updated. One or more policer actions can be updated in a single function 2337 invocation. To update the policer action associated with color C, bit 2338 (1 << C) needs to be set in *action_mask* and element at position C 2339 in the *actions* array needs to be valid. 2340* ``actionx``: Policer action for the color x, 2341 RTE_MTR_GREEN <= x < RTE_MTR_COLORS 2342 2343set port meter stats mask 2344~~~~~~~~~~~~~~~~~~~~~~~~~ 2345 2346Set meter stats mask for the ethernet device:: 2347 2348 testpmd> set port meter stats mask (port_id) (mtr_id) (stats_mask) 2349 2350where: 2351 2352* ``stats_mask``: Bit mask indicating statistics counter types to be enabled. 2353 2354show port meter stats 2355~~~~~~~~~~~~~~~~~~~~~ 2356 2357Show meter stats of the ethernet device:: 2358 2359 testpmd> show port meter stats (port_id) (mtr_id) (clear) 2360 2361where: 2362 2363* ``clear``: Flag that indicates whether the statistics counters should 2364 be cleared (i.e. set to zero) immediately after they have been read or not. 2365 2366Traffic Management 2367------------------ 2368 2369The following section shows functions for configuring traffic management on 2370on the ethernet device through the use of generic TM API. 2371 2372show port traffic management capability 2373~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2374 2375Show traffic management capability of the port:: 2376 2377 testpmd> show port tm cap (port_id) 2378 2379show port traffic management capability (hierarchy level) 2380~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2381 2382Show traffic management hierarchy level capability of the port:: 2383 2384 testpmd> show port tm level cap (port_id) (level_id) 2385 2386show port traffic management capability (hierarchy node level) 2387~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2388 2389Show the traffic management hierarchy node capability of the port:: 2390 2391 testpmd> show port tm node cap (port_id) (node_id) 2392 2393show port traffic management hierarchy node type 2394~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2395 2396Show the port traffic management hierarchy node type:: 2397 2398 testpmd> show port tm node type (port_id) (node_id) 2399 2400show port traffic management hierarchy node stats 2401~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2402 2403Show the port traffic management hierarchy node statistics:: 2404 2405 testpmd> show port tm node stats (port_id) (node_id) (clear) 2406 2407where: 2408 2409* ``clear``: When this parameter has a non-zero value, the statistics counters 2410 are cleared (i.e. set to zero) immediately after they have been read, 2411 otherwise the statistics counters are left untouched. 2412 2413Add port traffic management private shaper profile 2414~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2415 2416Add the port traffic management private shaper profile:: 2417 2418 testpmd> add port tm node shaper profile (port_id) (shaper_profile_id) \ 2419 (tb_rate) (tb_size) (packet_length_adjust) 2420 2421where: 2422 2423* ``shaper_profile id``: Shaper profile ID for the new profile. 2424* ``tb_rate``: Token bucket rate (bytes per second). 2425* ``tb_size``: Token bucket size (bytes). 2426* ``packet_length_adjust``: The value (bytes) to be added to the length of 2427 each packet for the purpose of shaping. This parameter value can be used to 2428 correct the packet length with the framing overhead bytes that are consumed 2429 on the wire. 2430 2431Delete port traffic management private shaper profile 2432~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2433 2434Delete the port traffic management private shaper:: 2435 2436 testpmd> del port tm node shaper profile (port_id) (shaper_profile_id) 2437 2438where: 2439 2440* ``shaper_profile id``: Shaper profile ID that needs to be deleted. 2441 2442Add port traffic management shared shaper 2443~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2444 2445Create the port traffic management shared shaper:: 2446 2447 testpmd> add port tm node shared shaper (port_id) (shared_shaper_id) \ 2448 (shaper_profile_id) 2449 2450where: 2451 2452* ``shared_shaper_id``: Shared shaper ID to be created. 2453* ``shaper_profile id``: Shaper profile ID for shared shaper. 2454 2455Set port traffic management shared shaper 2456~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2457 2458Update the port traffic management shared shaper:: 2459 2460 testpmd> set port tm node shared shaper (port_id) (shared_shaper_id) \ 2461 (shaper_profile_id) 2462 2463where: 2464 2465* ``shared_shaper_id``: Shared shaper ID to be update. 2466* ``shaper_profile id``: Shaper profile ID for shared shaper. 2467 2468Delete port traffic management shared shaper 2469~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2470 2471Delete the port traffic management shared shaper:: 2472 2473 testpmd> del port tm node shared shaper (port_id) (shared_shaper_id) 2474 2475where: 2476 2477* ``shared_shaper_id``: Shared shaper ID to be deleted. 2478 2479Set port traffic management hiearchy node private shaper 2480~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2481 2482set the port traffic management hierarchy node private shaper:: 2483 2484 testpmd> set port tm node shaper profile (port_id) (node_id) \ 2485 (shaper_profile_id) 2486 2487where: 2488 2489* ``shaper_profile id``: Private shaper profile ID to be enabled on the 2490 hierarchy node. 2491 2492Add port traffic management WRED profile 2493~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2494 2495Create a new WRED profile:: 2496 2497 testpmd> add port tm node wred profile (port_id) (wred_profile_id) \ 2498 (color_g) (min_th_g) (max_th_g) (maxp_inv_g) (wq_log2_g) \ 2499 (color_y) (min_th_y) (max_th_y) (maxp_inv_y) (wq_log2_y) \ 2500 (color_r) (min_th_r) (max_th_r) (maxp_inv_r) (wq_log2_r) 2501 2502where: 2503 2504* ``wred_profile id``: Identifier for the newly create WRED profile 2505* ``color_g``: Packet color (green) 2506* ``min_th_g``: Minimum queue threshold for packet with green color 2507* ``max_th_g``: Minimum queue threshold for packet with green color 2508* ``maxp_inv_g``: Inverse of packet marking probability maximum value (maxp) 2509* ``wq_log2_g``: Negated log2 of queue weight (wq) 2510* ``color_y``: Packet color (yellow) 2511* ``min_th_y``: Minimum queue threshold for packet with yellow color 2512* ``max_th_y``: Minimum queue threshold for packet with yellow color 2513* ``maxp_inv_y``: Inverse of packet marking probability maximum value (maxp) 2514* ``wq_log2_y``: Negated log2 of queue weight (wq) 2515* ``color_r``: Packet color (red) 2516* ``min_th_r``: Minimum queue threshold for packet with yellow color 2517* ``max_th_r``: Minimum queue threshold for packet with yellow color 2518* ``maxp_inv_r``: Inverse of packet marking probability maximum value (maxp) 2519* ``wq_log2_r``: Negated log2 of queue weight (wq) 2520 2521Delete port traffic management WRED profile 2522~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2523 2524Delete the WRED profile:: 2525 2526 testpmd> del port tm node wred profile (port_id) (wred_profile_id) 2527 2528Add port traffic management hierarchy nonleaf node 2529~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2530 2531Add nonleaf node to port traffic management hiearchy:: 2532 2533 testpmd> add port tm nonleaf node (port_id) (node_id) (parent_node_id) \ 2534 (priority) (weight) (level_id) (shaper_profile_id) \ 2535 (n_sp_priorities) (stats_mask) (n_shared_shapers) \ 2536 [(shared_shaper_0) (shared_shaper_1) ...] \ 2537 2538where: 2539 2540* ``parent_node_id``: Node ID of the parent. 2541* ``priority``: Node priority (highest node priority is zero). This is used by 2542 the SP algorithm running on the parent node for scheduling this node. 2543* ``weight``: Node weight (lowest weight is one). The node weight is relative 2544 to the weight sum of all siblings that have the same priority. It is used by 2545 the WFQ algorithm running on the parent node for scheduling this node. 2546* ``level_id``: Hiearchy level of the node. 2547* ``shaper_profile_id``: Shaper profile ID of the private shaper to be used by 2548 the node. 2549* ``n_sp_priorities``: Number of strict priorities. 2550* ``stats_mask``: Mask of statistics counter types to be enabled for this node. 2551* ``n_shared_shapers``: Number of shared shapers. 2552* ``shared_shaper_id``: Shared shaper id. 2553 2554Add port traffic management hierarchy leaf node 2555~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2556 2557Add leaf node to port traffic management hiearchy:: 2558 2559 testpmd> add port tm leaf node (port_id) (node_id) (parent_node_id) \ 2560 (priority) (weight) (level_id) (shaper_profile_id) \ 2561 (cman_mode) (wred_profile_id) (stats_mask) (n_shared_shapers) \ 2562 [(shared_shaper_id) (shared_shaper_id) ...] \ 2563 2564where: 2565 2566* ``parent_node_id``: Node ID of the parent. 2567* ``priority``: Node priority (highest node priority is zero). This is used by 2568 the SP algorithm running on the parent node for scheduling this node. 2569* ``weight``: Node weight (lowest weight is one). The node weight is relative 2570 to the weight sum of all siblings that have the same priority. It is used by 2571 the WFQ algorithm running on the parent node for scheduling this node. 2572* ``level_id``: Hiearchy level of the node. 2573* ``shaper_profile_id``: Shaper profile ID of the private shaper to be used by 2574 the node. 2575* ``cman_mode``: Congestion management mode to be enabled for this node. 2576* ``wred_profile_id``: WRED profile id to be enabled for this node. 2577* ``stats_mask``: Mask of statistics counter types to be enabled for this node. 2578* ``n_shared_shapers``: Number of shared shapers. 2579* ``shared_shaper_id``: Shared shaper id. 2580 2581Delete port traffic management hierarchy node 2582~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2583 2584Delete node from port traffic management hiearchy:: 2585 2586 testpmd> del port tm node (port_id) (node_id) 2587 2588Update port traffic management hierarchy parent node 2589~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2590 2591Update port traffic management hierarchy parent node:: 2592 2593 testpmd> set port tm node parent (port_id) (node_id) (parent_node_id) \ 2594 (priority) (weight) 2595 2596This function can only be called after the hierarchy commit invocation. Its 2597success depends on the port support for this operation, as advertised through 2598the port capability set. This function is valid for all nodes of the traffic 2599management hierarchy except root node. 2600 2601Suspend port traffic management hierarchy node 2602~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2603 2604 testpmd> suspend port tm node (port_id) (node_id) 2605 2606Resume port traffic management hierarchy node 2607~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2608 2609 testpmd> resume port tm node (port_id) (node_id) 2610 2611Commit port traffic management hierarchy 2612~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2613 2614Commit the traffic management hierarchy on the port:: 2615 2616 testpmd> port tm hierarchy commit (port_id) (clean_on_fail) 2617 2618where: 2619 2620* ``clean_on_fail``: When set to non-zero, hierarchy is cleared on function 2621 call failure. On the other hand, hierarchy is preserved when this parameter 2622 is equal to zero. 2623 2624Set port traffic management default hierarchy (tm forwarding mode) 2625~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 2626 2627set the traffic management default hierarchy on the port:: 2628 2629 testpmd> set port tm hierarchy default (port_id) 2630 2631Filter Functions 2632---------------- 2633 2634This section details the available filter functions that are available. 2635 2636Note these functions interface the deprecated legacy filtering framework, 2637superseded by *rte_flow*. See `Flow rules management`_. 2638 2639ethertype_filter 2640~~~~~~~~~~~~~~~~~~~~ 2641 2642Add or delete a L2 Ethertype filter, which identify packets by their L2 Ethertype mainly assign them to a receive queue:: 2643 2644 ethertype_filter (port_id) (add|del) (mac_addr|mac_ignr) (mac_address) \ 2645 ethertype (ether_type) (drop|fwd) queue (queue_id) 2646 2647The available information parameters are: 2648 2649* ``port_id``: The port which the Ethertype filter assigned on. 2650 2651* ``mac_addr``: Compare destination mac address. 2652 2653* ``mac_ignr``: Ignore destination mac address match. 2654 2655* ``mac_address``: Destination mac address to match. 2656 2657* ``ether_type``: The EtherType value want to match, 2658 for example 0x0806 for ARP packet. 0x0800 (IPv4) and 0x86DD (IPv6) are invalid. 2659 2660* ``queue_id``: The receive queue associated with this EtherType filter. 2661 It is meaningless when deleting or dropping. 2662 2663Example, to add/remove an ethertype filter rule:: 2664 2665 testpmd> ethertype_filter 0 add mac_ignr 00:11:22:33:44:55 \ 2666 ethertype 0x0806 fwd queue 3 2667 2668 testpmd> ethertype_filter 0 del mac_ignr 00:11:22:33:44:55 \ 2669 ethertype 0x0806 fwd queue 3 2670 26712tuple_filter 2672~~~~~~~~~~~~~~~~~ 2673 2674Add or delete a 2-tuple filter, 2675which identifies packets by specific protocol and destination TCP/UDP port 2676and forwards packets into one of the receive queues:: 2677 2678 2tuple_filter (port_id) (add|del) dst_port (dst_port_value) \ 2679 protocol (protocol_value) mask (mask_value) \ 2680 tcp_flags (tcp_flags_value) priority (prio_value) \ 2681 queue (queue_id) 2682 2683The available information parameters are: 2684 2685* ``port_id``: The port which the 2-tuple filter assigned on. 2686 2687* ``dst_port_value``: Destination port in L4. 2688 2689* ``protocol_value``: IP L4 protocol. 2690 2691* ``mask_value``: Participates in the match or not by bit for field above, 1b means participate. 2692 2693* ``tcp_flags_value``: TCP control bits. The non-zero value is invalid, when the pro_value is not set to 0x06 (TCP). 2694 2695* ``prio_value``: Priority of this filter. 2696 2697* ``queue_id``: The receive queue associated with this 2-tuple filter. 2698 2699Example, to add/remove an 2tuple filter rule:: 2700 2701 testpmd> 2tuple_filter 0 add dst_port 32 protocol 0x06 mask 0x03 \ 2702 tcp_flags 0x02 priority 3 queue 3 2703 2704 testpmd> 2tuple_filter 0 del dst_port 32 protocol 0x06 mask 0x03 \ 2705 tcp_flags 0x02 priority 3 queue 3 2706 27075tuple_filter 2708~~~~~~~~~~~~~~~~~ 2709 2710Add or delete a 5-tuple filter, 2711which consists of a 5-tuple (protocol, source and destination IP addresses, source and destination TCP/UDP/SCTP port) 2712and routes packets into one of the receive queues:: 2713 2714 5tuple_filter (port_id) (add|del) dst_ip (dst_address) src_ip \ 2715 (src_address) dst_port (dst_port_value) \ 2716 src_port (src_port_value) protocol (protocol_value) \ 2717 mask (mask_value) tcp_flags (tcp_flags_value) \ 2718 priority (prio_value) queue (queue_id) 2719 2720The available information parameters are: 2721 2722* ``port_id``: The port which the 5-tuple filter assigned on. 2723 2724* ``dst_address``: Destination IP address. 2725 2726* ``src_address``: Source IP address. 2727 2728* ``dst_port_value``: TCP/UDP destination port. 2729 2730* ``src_port_value``: TCP/UDP source port. 2731 2732* ``protocol_value``: L4 protocol. 2733 2734* ``mask_value``: Participates in the match or not by bit for field above, 1b means participate 2735 2736* ``tcp_flags_value``: TCP control bits. The non-zero value is invalid, when the protocol_value is not set to 0x06 (TCP). 2737 2738* ``prio_value``: The priority of this filter. 2739 2740* ``queue_id``: The receive queue associated with this 5-tuple filter. 2741 2742Example, to add/remove an 5tuple filter rule:: 2743 2744 testpmd> 5tuple_filter 0 add dst_ip 2.2.2.5 src_ip 2.2.2.4 \ 2745 dst_port 64 src_port 32 protocol 0x06 mask 0x1F \ 2746 flags 0x0 priority 3 queue 3 2747 2748 testpmd> 5tuple_filter 0 del dst_ip 2.2.2.5 src_ip 2.2.2.4 \ 2749 dst_port 64 src_port 32 protocol 0x06 mask 0x1F \ 2750 flags 0x0 priority 3 queue 3 2751 2752syn_filter 2753~~~~~~~~~~ 2754 2755Using the SYN filter, TCP packets whose *SYN* flag is set can be forwarded to a separate queue:: 2756 2757 syn_filter (port_id) (add|del) priority (high|low) queue (queue_id) 2758 2759The available information parameters are: 2760 2761* ``port_id``: The port which the SYN filter assigned on. 2762 2763* ``high``: This SYN filter has higher priority than other filters. 2764 2765* ``low``: This SYN filter has lower priority than other filters. 2766 2767* ``queue_id``: The receive queue associated with this SYN filter 2768 2769Example:: 2770 2771 testpmd> syn_filter 0 add priority high queue 3 2772 2773flex_filter 2774~~~~~~~~~~~ 2775 2776With flex filter, packets can be recognized by any arbitrary pattern within the first 128 bytes of the packet 2777and routed into one of the receive queues:: 2778 2779 flex_filter (port_id) (add|del) len (len_value) bytes (bytes_value) \ 2780 mask (mask_value) priority (prio_value) queue (queue_id) 2781 2782The available information parameters are: 2783 2784* ``port_id``: The port which the Flex filter is assigned on. 2785 2786* ``len_value``: Filter length in bytes, no greater than 128. 2787 2788* ``bytes_value``: A string in hexadecimal, means the value the flex filter needs to match. 2789 2790* ``mask_value``: A string in hexadecimal, bit 1 means corresponding byte participates in the match. 2791 2792* ``prio_value``: The priority of this filter. 2793 2794* ``queue_id``: The receive queue associated with this Flex filter. 2795 2796Example:: 2797 2798 testpmd> flex_filter 0 add len 16 bytes 0x00000000000000000000000008060000 \ 2799 mask 000C priority 3 queue 3 2800 2801 testpmd> flex_filter 0 del len 16 bytes 0x00000000000000000000000008060000 \ 2802 mask 000C priority 3 queue 3 2803 2804 2805.. _testpmd_flow_director: 2806 2807flow_director_filter 2808~~~~~~~~~~~~~~~~~~~~ 2809 2810The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues. 2811 2812Four types of filtering are supported which are referred to as Perfect Match, Signature, Perfect-mac-vlan and 2813Perfect-tunnel filters, the match mode is set by the ``--pkt-filter-mode`` command-line parameter: 2814 2815* Perfect match filters. 2816 The hardware checks a match between the masked fields of the received packets and the programmed filters. 2817 The masked fields are for IP flow. 2818 2819* Signature filters. 2820 The hardware checks a match between a hash-based signature of the masked fields of the received packet. 2821 2822* Perfect-mac-vlan match filters. 2823 The hardware checks a match between the masked fields of the received packets and the programmed filters. 2824 The masked fields are for MAC VLAN flow. 2825 2826* Perfect-tunnel match filters. 2827 The hardware checks a match between the masked fields of the received packets and the programmed filters. 2828 The masked fields are for tunnel flow. 2829 2830* Perfect-raw-flow-type match filters. 2831 The hardware checks a match between the masked fields of the received packets and pre-loaded raw (template) packet. 2832 The masked fields are specified by input sets. 2833 2834The Flow Director filters can match the different fields for different type of packet: flow type, specific input set 2835per flow type and the flexible payload. 2836 2837The Flow Director can also mask out parts of all of these fields so that filters 2838are only applied to certain fields or parts of the fields. 2839 2840Note that for raw flow type mode the source and destination fields in the 2841raw packet buffer need to be presented in a reversed order with respect 2842to the expected received packets. 2843For example: IP source and destination addresses or TCP/UDP/SCTP 2844source and destination ports 2845 2846Different NICs may have different capabilities, command show port fdir (port_id) can be used to acquire the information. 2847 2848# Commands to add flow director filters of different flow types:: 2849 2850 flow_director_filter (port_id) mode IP (add|del|update) \ 2851 flow (ipv4-other|ipv4-frag|ipv6-other|ipv6-frag) \ 2852 src (src_ip_address) dst (dst_ip_address) \ 2853 tos (tos_value) proto (proto_value) ttl (ttl_value) \ 2854 vlan (vlan_value) flexbytes (flexbytes_value) \ 2855 (drop|fwd) pf|vf(vf_id) queue (queue_id) \ 2856 fd_id (fd_id_value) 2857 2858 flow_director_filter (port_id) mode IP (add|del|update) \ 2859 flow (ipv4-tcp|ipv4-udp|ipv6-tcp|ipv6-udp) \ 2860 src (src_ip_address) (src_port) \ 2861 dst (dst_ip_address) (dst_port) \ 2862 tos (tos_value) ttl (ttl_value) \ 2863 vlan (vlan_value) flexbytes (flexbytes_value) \ 2864 (drop|fwd) queue pf|vf(vf_id) (queue_id) \ 2865 fd_id (fd_id_value) 2866 2867 flow_director_filter (port_id) mode IP (add|del|update) \ 2868 flow (ipv4-sctp|ipv6-sctp) \ 2869 src (src_ip_address) (src_port) \ 2870 dst (dst_ip_address) (dst_port) \ 2871 tos (tos_value) ttl (ttl_value) \ 2872 tag (verification_tag) vlan (vlan_value) \ 2873 flexbytes (flexbytes_value) (drop|fwd) \ 2874 pf|vf(vf_id) queue (queue_id) fd_id (fd_id_value) 2875 2876 flow_director_filter (port_id) mode IP (add|del|update) flow l2_payload \ 2877 ether (ethertype) flexbytes (flexbytes_value) \ 2878 (drop|fwd) pf|vf(vf_id) queue (queue_id) 2879 fd_id (fd_id_value) 2880 2881 flow_director_filter (port_id) mode MAC-VLAN (add|del|update) \ 2882 mac (mac_address) vlan (vlan_value) \ 2883 flexbytes (flexbytes_value) (drop|fwd) \ 2884 queue (queue_id) fd_id (fd_id_value) 2885 2886 flow_director_filter (port_id) mode Tunnel (add|del|update) \ 2887 mac (mac_address) vlan (vlan_value) \ 2888 tunnel (NVGRE|VxLAN) tunnel-id (tunnel_id_value) \ 2889 flexbytes (flexbytes_value) (drop|fwd) \ 2890 queue (queue_id) fd_id (fd_id_value) 2891 2892 flow_director_filter (port_id) mode raw (add|del|update) flow (flow_id) \ 2893 (drop|fwd) queue (queue_id) fd_id (fd_id_value) \ 2894 packet (packet file name) 2895 2896For example, to add an ipv4-udp flow type filter:: 2897 2898 testpmd> flow_director_filter 0 mode IP add flow ipv4-udp src 2.2.2.3 32 \ 2899 dst 2.2.2.5 33 tos 2 ttl 40 vlan 0x1 flexbytes (0x88,0x48) \ 2900 fwd pf queue 1 fd_id 1 2901 2902For example, add an ipv4-other flow type filter:: 2903 2904 testpmd> flow_director_filter 0 mode IP add flow ipv4-other src 2.2.2.3 \ 2905 dst 2.2.2.5 tos 2 proto 20 ttl 40 vlan 0x1 \ 2906 flexbytes (0x88,0x48) fwd pf queue 1 fd_id 1 2907 2908flush_flow_director 2909~~~~~~~~~~~~~~~~~~~ 2910 2911Flush all flow director filters on a device:: 2912 2913 testpmd> flush_flow_director (port_id) 2914 2915Example, to flush all flow director filter on port 0:: 2916 2917 testpmd> flush_flow_director 0 2918 2919flow_director_mask 2920~~~~~~~~~~~~~~~~~~ 2921 2922Set flow director's input masks:: 2923 2924 flow_director_mask (port_id) mode IP vlan (vlan_value) \ 2925 src_mask (ipv4_src) (ipv6_src) (src_port) \ 2926 dst_mask (ipv4_dst) (ipv6_dst) (dst_port) 2927 2928 flow_director_mask (port_id) mode MAC-VLAN vlan (vlan_value) 2929 2930 flow_director_mask (port_id) mode Tunnel vlan (vlan_value) \ 2931 mac (mac_value) tunnel-type (tunnel_type_value) \ 2932 tunnel-id (tunnel_id_value) 2933 2934Example, to set flow director mask on port 0:: 2935 2936 testpmd> flow_director_mask 0 mode IP vlan 0xefff \ 2937 src_mask 255.255.255.255 \ 2938 FFFF:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF 0xFFFF \ 2939 dst_mask 255.255.255.255 \ 2940 FFFF:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF:FFFF 0xFFFF 2941 2942flow_director_flex_mask 2943~~~~~~~~~~~~~~~~~~~~~~~ 2944 2945set masks of flow director's flexible payload based on certain flow type:: 2946 2947 testpmd> flow_director_flex_mask (port_id) \ 2948 flow (none|ipv4-other|ipv4-frag|ipv4-tcp|ipv4-udp|ipv4-sctp| \ 2949 ipv6-other|ipv6-frag|ipv6-tcp|ipv6-udp|ipv6-sctp| \ 2950 l2_payload|all) (mask) 2951 2952Example, to set flow director's flex mask for all flow type on port 0:: 2953 2954 testpmd> flow_director_flex_mask 0 flow all \ 2955 (0xff,0xff,0,0,0,0,0,0,0,0,0,0,0,0,0,0) 2956 2957 2958flow_director_flex_payload 2959~~~~~~~~~~~~~~~~~~~~~~~~~~ 2960 2961Configure flexible payload selection:: 2962 2963 flow_director_flex_payload (port_id) (raw|l2|l3|l4) (config) 2964 2965For example, to select the first 16 bytes from the offset 4 (bytes) of packet's payload as flexible payload:: 2966 2967 testpmd> flow_director_flex_payload 0 l4 \ 2968 (4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19) 2969 2970get_sym_hash_ena_per_port 2971~~~~~~~~~~~~~~~~~~~~~~~~~ 2972 2973Get symmetric hash enable configuration per port:: 2974 2975 get_sym_hash_ena_per_port (port_id) 2976 2977For example, to get symmetric hash enable configuration of port 1:: 2978 2979 testpmd> get_sym_hash_ena_per_port 1 2980 2981set_sym_hash_ena_per_port 2982~~~~~~~~~~~~~~~~~~~~~~~~~ 2983 2984Set symmetric hash enable configuration per port to enable or disable:: 2985 2986 set_sym_hash_ena_per_port (port_id) (enable|disable) 2987 2988For example, to set symmetric hash enable configuration of port 1 to enable:: 2989 2990 testpmd> set_sym_hash_ena_per_port 1 enable 2991 2992get_hash_global_config 2993~~~~~~~~~~~~~~~~~~~~~~ 2994 2995Get the global configurations of hash filters:: 2996 2997 get_hash_global_config (port_id) 2998 2999For example, to get the global configurations of hash filters of port 1:: 3000 3001 testpmd> get_hash_global_config 1 3002 3003set_hash_global_config 3004~~~~~~~~~~~~~~~~~~~~~~ 3005 3006Set the global configurations of hash filters:: 3007 3008 set_hash_global_config (port_id) (toeplitz|simple_xor|default) \ 3009 (ipv4|ipv4-frag|ipv4-tcp|ipv4-udp|ipv4-sctp|ipv4-other|ipv6|ipv6-frag| \ 3010 ipv6-tcp|ipv6-udp|ipv6-sctp|ipv6-other|l2_payload|<flow_id>) \ 3011 (enable|disable) 3012 3013For example, to enable simple_xor for flow type of ipv6 on port 2:: 3014 3015 testpmd> set_hash_global_config 2 simple_xor ipv6 enable 3016 3017set_hash_input_set 3018~~~~~~~~~~~~~~~~~~ 3019 3020Set the input set for hash:: 3021 3022 set_hash_input_set (port_id) (ipv4-frag|ipv4-tcp|ipv4-udp|ipv4-sctp| \ 3023 ipv4-other|ipv6-frag|ipv6-tcp|ipv6-udp|ipv6-sctp|ipv6-other| \ 3024 l2_payload|<flow_id>) (ovlan|ivlan|src-ipv4|dst-ipv4|src-ipv6|dst-ipv6| \ 3025 ipv4-tos|ipv4-proto|ipv6-tc|ipv6-next-header|udp-src-port|udp-dst-port| \ 3026 tcp-src-port|tcp-dst-port|sctp-src-port|sctp-dst-port|sctp-veri-tag| \ 3027 udp-key|gre-key|fld-1st|fld-2nd|fld-3rd|fld-4th|fld-5th|fld-6th|fld-7th| \ 3028 fld-8th|none) (select|add) 3029 3030For example, to add source IP to hash input set for flow type of ipv4-udp on port 0:: 3031 3032 testpmd> set_hash_input_set 0 ipv4-udp src-ipv4 add 3033 3034set_fdir_input_set 3035~~~~~~~~~~~~~~~~~~ 3036 3037The Flow Director filters can match the different fields for different type of packet, i.e. specific input set 3038on per flow type and the flexible payload. This command can be used to change input set for each flow type. 3039 3040Set the input set for flow director:: 3041 3042 set_fdir_input_set (port_id) (ipv4-frag|ipv4-tcp|ipv4-udp|ipv4-sctp| \ 3043 ipv4-other|ipv6|ipv6-frag|ipv6-tcp|ipv6-udp|ipv6-sctp|ipv6-other| \ 3044 l2_payload|<flow_id>) (ivlan|ethertype|src-ipv4|dst-ipv4|src-ipv6|dst-ipv6| \ 3045 ipv4-tos|ipv4-proto|ipv4-ttl|ipv6-tc|ipv6-next-header|ipv6-hop-limits| \ 3046 tudp-src-port|udp-dst-port|cp-src-port|tcp-dst-port|sctp-src-port| \ 3047 sctp-dst-port|sctp-veri-tag|none) (select|add) 3048 3049For example to add source IP to FD input set for flow type of ipv4-udp on port 0:: 3050 3051 testpmd> set_fdir_input_set 0 ipv4-udp src-ipv4 add 3052 3053global_config 3054~~~~~~~~~~~~~ 3055 3056Set different GRE key length for input set:: 3057 3058 global_config (port_id) gre-key-len (number in bytes) 3059 3060For example to set GRE key length for input set to 4 bytes on port 0:: 3061 3062 testpmd> global_config 0 gre-key-len 4 3063 3064 3065.. _testpmd_rte_flow: 3066 3067Flow rules management 3068--------------------- 3069 3070Control of the generic flow API (*rte_flow*) is fully exposed through the 3071``flow`` command (validation, creation, destruction, queries and operation 3072modes). 3073 3074Considering *rte_flow* overlaps with all `Filter Functions`_, using both 3075features simultaneously may cause undefined side-effects and is therefore 3076not recommended. 3077 3078``flow`` syntax 3079~~~~~~~~~~~~~~~ 3080 3081Because the ``flow`` command uses dynamic tokens to handle the large number 3082of possible flow rules combinations, its behavior differs slightly from 3083other commands, in particular: 3084 3085- Pressing *?* or the *<tab>* key displays contextual help for the current 3086 token, not that of the entire command. 3087 3088- Optional and repeated parameters are supported (provided they are listed 3089 in the contextual help). 3090 3091The first parameter stands for the operation mode. Possible operations and 3092their general syntax are described below. They are covered in detail in the 3093following sections. 3094 3095- Check whether a flow rule can be created:: 3096 3097 flow validate {port_id} 3098 [group {group_id}] [priority {level}] [ingress] [egress] [transfer] 3099 pattern {item} [/ {item} [...]] / end 3100 actions {action} [/ {action} [...]] / end 3101 3102- Create a flow rule:: 3103 3104 flow create {port_id} 3105 [group {group_id}] [priority {level}] [ingress] [egress] [transfer] 3106 pattern {item} [/ {item} [...]] / end 3107 actions {action} [/ {action} [...]] / end 3108 3109- Destroy specific flow rules:: 3110 3111 flow destroy {port_id} rule {rule_id} [...] 3112 3113- Destroy all flow rules:: 3114 3115 flow flush {port_id} 3116 3117- Query an existing flow rule:: 3118 3119 flow query {port_id} {rule_id} {action} 3120 3121- List existing flow rules sorted by priority, filtered by group 3122 identifiers:: 3123 3124 flow list {port_id} [group {group_id}] [...] 3125 3126- Restrict ingress traffic to the defined flow rules:: 3127 3128 flow isolate {port_id} {boolean} 3129 3130Validating flow rules 3131~~~~~~~~~~~~~~~~~~~~~ 3132 3133``flow validate`` reports whether a flow rule would be accepted by the 3134underlying device in its current state but stops short of creating it. It is 3135bound to ``rte_flow_validate()``:: 3136 3137 flow validate {port_id} 3138 [group {group_id}] [priority {level}] [ingress] [egress] [transfer] 3139 pattern {item} [/ {item} [...]] / end 3140 actions {action} [/ {action} [...]] / end 3141 3142If successful, it will show:: 3143 3144 Flow rule validated 3145 3146Otherwise it will show an error message of the form:: 3147 3148 Caught error type [...] ([...]): [...] 3149 3150This command uses the same parameters as ``flow create``, their format is 3151described in `Creating flow rules`_. 3152 3153Check whether redirecting any Ethernet packet received on port 0 to RX queue 3154index 6 is supported:: 3155 3156 testpmd> flow validate 0 ingress pattern eth / end 3157 actions queue index 6 / end 3158 Flow rule validated 3159 testpmd> 3160 3161Port 0 does not support TCPv6 rules:: 3162 3163 testpmd> flow validate 0 ingress pattern eth / ipv6 / tcp / end 3164 actions drop / end 3165 Caught error type 9 (specific pattern item): Invalid argument 3166 testpmd> 3167 3168Creating flow rules 3169~~~~~~~~~~~~~~~~~~~ 3170 3171``flow create`` validates and creates the specified flow rule. It is bound 3172to ``rte_flow_create()``:: 3173 3174 flow create {port_id} 3175 [group {group_id}] [priority {level}] [ingress] [egress] [transfer] 3176 pattern {item} [/ {item} [...]] / end 3177 actions {action} [/ {action} [...]] / end 3178 3179If successful, it will return a flow rule ID usable with other commands:: 3180 3181 Flow rule #[...] created 3182 3183Otherwise it will show an error message of the form:: 3184 3185 Caught error type [...] ([...]): [...] 3186 3187Parameters describe in the following order: 3188 3189- Attributes (*group*, *priority*, *ingress*, *egress*, *transfer* tokens). 3190- A matching pattern, starting with the *pattern* token and terminated by an 3191 *end* pattern item. 3192- Actions, starting with the *actions* token and terminated by an *end* 3193 action. 3194 3195These translate directly to *rte_flow* objects provided as-is to the 3196underlying functions. 3197 3198The shortest valid definition only comprises mandatory tokens:: 3199 3200 testpmd> flow create 0 pattern end actions end 3201 3202Note that PMDs may refuse rules that essentially do nothing such as this 3203one. 3204 3205**All unspecified object values are automatically initialized to 0.** 3206 3207Attributes 3208^^^^^^^^^^ 3209 3210These tokens affect flow rule attributes (``struct rte_flow_attr``) and are 3211specified before the ``pattern`` token. 3212 3213- ``group {group id}``: priority group. 3214- ``priority {level}``: priority level within group. 3215- ``ingress``: rule applies to ingress traffic. 3216- ``egress``: rule applies to egress traffic. 3217- ``transfer``: apply rule directly to endpoints found in pattern. 3218 3219Each instance of an attribute specified several times overrides the previous 3220value as shown below (group 4 is used):: 3221 3222 testpmd> flow create 0 group 42 group 24 group 4 [...] 3223 3224Note that once enabled, ``ingress`` and ``egress`` cannot be disabled. 3225 3226While not specifying a direction is an error, some rules may allow both 3227simultaneously. 3228 3229Most rules affect RX therefore contain the ``ingress`` token:: 3230 3231 testpmd> flow create 0 ingress pattern [...] 3232 3233Matching pattern 3234^^^^^^^^^^^^^^^^ 3235 3236A matching pattern starts after the ``pattern`` token. It is made of pattern 3237items and is terminated by a mandatory ``end`` item. 3238 3239Items are named after their type (*RTE_FLOW_ITEM_TYPE_* from ``enum 3240rte_flow_item_type``). 3241 3242The ``/`` token is used as a separator between pattern items as shown 3243below:: 3244 3245 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end [...] 3246 3247Note that protocol items like these must be stacked from lowest to highest 3248layer to make sense. For instance, the following rule is either invalid or 3249unlikely to match any packet:: 3250 3251 testpmd> flow create 0 ingress pattern eth / udp / ipv4 / end [...] 3252 3253More information on these restrictions can be found in the *rte_flow* 3254documentation. 3255 3256Several items support additional specification structures, for example 3257``ipv4`` allows specifying source and destination addresses as follows:: 3258 3259 testpmd> flow create 0 ingress pattern eth / ipv4 src is 10.1.1.1 3260 dst is 10.2.0.0 / end [...] 3261 3262This rule matches all IPv4 traffic with the specified properties. 3263 3264In this example, ``src`` and ``dst`` are field names of the underlying 3265``struct rte_flow_item_ipv4`` object. All item properties can be specified 3266in a similar fashion. 3267 3268The ``is`` token means that the subsequent value must be matched exactly, 3269and assigns ``spec`` and ``mask`` fields in ``struct rte_flow_item`` 3270accordingly. Possible assignment tokens are: 3271 3272- ``is``: match value perfectly (with full bit-mask). 3273- ``spec``: match value according to configured bit-mask. 3274- ``last``: specify upper bound to establish a range. 3275- ``mask``: specify bit-mask with relevant bits set to one. 3276- ``prefix``: generate bit-mask from a prefix length. 3277 3278These yield identical results:: 3279 3280 ipv4 src is 10.1.1.1 3281 3282:: 3283 3284 ipv4 src spec 10.1.1.1 src mask 255.255.255.255 3285 3286:: 3287 3288 ipv4 src spec 10.1.1.1 src prefix 32 3289 3290:: 3291 3292 ipv4 src is 10.1.1.1 src last 10.1.1.1 # range with a single value 3293 3294:: 3295 3296 ipv4 src is 10.1.1.1 src last 0 # 0 disables range 3297 3298Inclusive ranges can be defined with ``last``:: 3299 3300 ipv4 src is 10.1.1.1 src last 10.2.3.4 # 10.1.1.1 to 10.2.3.4 3301 3302Note that ``mask`` affects both ``spec`` and ``last``:: 3303 3304 ipv4 src is 10.1.1.1 src last 10.2.3.4 src mask 255.255.0.0 3305 # matches 10.1.0.0 to 10.2.255.255 3306 3307Properties can be modified multiple times:: 3308 3309 ipv4 src is 10.1.1.1 src is 10.1.2.3 src is 10.2.3.4 # matches 10.2.3.4 3310 3311:: 3312 3313 ipv4 src is 10.1.1.1 src prefix 24 src prefix 16 # matches 10.1.0.0/16 3314 3315Pattern items 3316^^^^^^^^^^^^^ 3317 3318This section lists supported pattern items and their attributes, if any. 3319 3320- ``end``: end list of pattern items. 3321 3322- ``void``: no-op pattern item. 3323 3324- ``invert``: perform actions when pattern does not match. 3325 3326- ``any``: match any protocol for the current layer. 3327 3328 - ``num {unsigned}``: number of layers covered. 3329 3330- ``pf``: match traffic from/to the physical function. 3331 3332- ``vf``: match traffic from/to a virtual function ID. 3333 3334 - ``id {unsigned}``: VF ID. 3335 3336- ``phy_port``: match traffic from/to a specific physical port. 3337 3338 - ``index {unsigned}``: physical port index. 3339 3340- ``port_id``: match traffic from/to a given DPDK port ID. 3341 3342 - ``id {unsigned}``: DPDK port ID. 3343 3344- ``mark``: match value set in previously matched flow rule using the mark action. 3345 3346 - ``id {unsigned}``: arbitrary integer value. 3347 3348- ``raw``: match an arbitrary byte string. 3349 3350 - ``relative {boolean}``: look for pattern after the previous item. 3351 - ``search {boolean}``: search pattern from offset (see also limit). 3352 - ``offset {integer}``: absolute or relative offset for pattern. 3353 - ``limit {unsigned}``: search area limit for start of pattern. 3354 - ``pattern {string}``: byte string to look for. 3355 3356- ``eth``: match Ethernet header. 3357 3358 - ``dst {MAC-48}``: destination MAC. 3359 - ``src {MAC-48}``: source MAC. 3360 - ``type {unsigned}``: EtherType or TPID. 3361 3362- ``vlan``: match 802.1Q/ad VLAN tag. 3363 3364 - ``tci {unsigned}``: tag control information. 3365 - ``pcp {unsigned}``: priority code point. 3366 - ``dei {unsigned}``: drop eligible indicator. 3367 - ``vid {unsigned}``: VLAN identifier. 3368 - ``inner_type {unsigned}``: inner EtherType or TPID. 3369 3370- ``ipv4``: match IPv4 header. 3371 3372 - ``tos {unsigned}``: type of service. 3373 - ``ttl {unsigned}``: time to live. 3374 - ``proto {unsigned}``: next protocol ID. 3375 - ``src {ipv4 address}``: source address. 3376 - ``dst {ipv4 address}``: destination address. 3377 3378- ``ipv6``: match IPv6 header. 3379 3380 - ``tc {unsigned}``: traffic class. 3381 - ``flow {unsigned}``: flow label. 3382 - ``proto {unsigned}``: protocol (next header). 3383 - ``hop {unsigned}``: hop limit. 3384 - ``src {ipv6 address}``: source address. 3385 - ``dst {ipv6 address}``: destination address. 3386 3387- ``icmp``: match ICMP header. 3388 3389 - ``type {unsigned}``: ICMP packet type. 3390 - ``code {unsigned}``: ICMP packet code. 3391 3392- ``udp``: match UDP header. 3393 3394 - ``src {unsigned}``: UDP source port. 3395 - ``dst {unsigned}``: UDP destination port. 3396 3397- ``tcp``: match TCP header. 3398 3399 - ``src {unsigned}``: TCP source port. 3400 - ``dst {unsigned}``: TCP destination port. 3401 3402- ``sctp``: match SCTP header. 3403 3404 - ``src {unsigned}``: SCTP source port. 3405 - ``dst {unsigned}``: SCTP destination port. 3406 - ``tag {unsigned}``: validation tag. 3407 - ``cksum {unsigned}``: checksum. 3408 3409- ``vxlan``: match VXLAN header. 3410 3411 - ``vni {unsigned}``: VXLAN identifier. 3412 3413- ``e_tag``: match IEEE 802.1BR E-Tag header. 3414 3415 - ``grp_ecid_b {unsigned}``: GRP and E-CID base. 3416 3417- ``nvgre``: match NVGRE header. 3418 3419 - ``tni {unsigned}``: virtual subnet ID. 3420 3421- ``mpls``: match MPLS header. 3422 3423 - ``label {unsigned}``: MPLS label. 3424 3425- ``gre``: match GRE header. 3426 3427 - ``protocol {unsigned}``: protocol type. 3428 3429- ``fuzzy``: fuzzy pattern match, expect faster than default. 3430 3431 - ``thresh {unsigned}``: accuracy threshold. 3432 3433- ``gtp``, ``gtpc``, ``gtpu``: match GTPv1 header. 3434 3435 - ``teid {unsigned}``: tunnel endpoint identifier. 3436 3437- ``geneve``: match GENEVE header. 3438 3439 - ``vni {unsigned}``: virtual network identifier. 3440 - ``protocol {unsigned}``: protocol type. 3441 3442- ``vxlan-gpe``: match VXLAN-GPE header. 3443 3444 - ``vni {unsigned}``: VXLAN-GPE identifier. 3445 3446- ``arp_eth_ipv4``: match ARP header for Ethernet/IPv4. 3447 3448 - ``sha {MAC-48}``: sender hardware address. 3449 - ``spa {ipv4 address}``: sender IPv4 address. 3450 - ``tha {MAC-48}``: target hardware address. 3451 - ``tpa {ipv4 address}``: target IPv4 address. 3452 3453- ``ipv6_ext``: match presence of any IPv6 extension header. 3454 3455 - ``next_hdr {unsigned}``: next header. 3456 3457- ``icmp6``: match any ICMPv6 header. 3458 3459 - ``type {unsigned}``: ICMPv6 type. 3460 - ``code {unsigned}``: ICMPv6 code. 3461 3462- ``icmp6_nd_ns``: match ICMPv6 neighbor discovery solicitation. 3463 3464 - ``target_addr {ipv6 address}``: target address. 3465 3466- ``icmp6_nd_na``: match ICMPv6 neighbor discovery advertisement. 3467 3468 - ``target_addr {ipv6 address}``: target address. 3469 3470- ``icmp6_nd_opt``: match presence of any ICMPv6 neighbor discovery option. 3471 3472 - ``type {unsigned}``: ND option type. 3473 3474- ``icmp6_nd_opt_sla_eth``: match ICMPv6 neighbor discovery source Ethernet 3475 link-layer address option. 3476 3477 - ``sla {MAC-48}``: source Ethernet LLA. 3478 3479- ``icmp6_nd_opt_sla_eth``: match ICMPv6 neighbor discovery target Ethernet 3480 link-layer address option. 3481 3482 - ``tla {MAC-48}``: target Ethernet LLA. 3483 3484Actions list 3485^^^^^^^^^^^^ 3486 3487A list of actions starts after the ``actions`` token in the same fashion as 3488`Matching pattern`_; actions are separated by ``/`` tokens and the list is 3489terminated by a mandatory ``end`` action. 3490 3491Actions are named after their type (*RTE_FLOW_ACTION_TYPE_* from ``enum 3492rte_flow_action_type``). 3493 3494Dropping all incoming UDPv4 packets can be expressed as follows:: 3495 3496 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end 3497 actions drop / end 3498 3499Several actions have configurable properties which must be specified when 3500there is no valid default value. For example, ``queue`` requires a target 3501queue index. 3502 3503This rule redirects incoming UDPv4 traffic to queue index 6:: 3504 3505 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end 3506 actions queue index 6 / end 3507 3508While this one could be rejected by PMDs (unspecified queue index):: 3509 3510 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end 3511 actions queue / end 3512 3513As defined by *rte_flow*, the list is not ordered, all actions of a given 3514rule are performed simultaneously. These are equivalent:: 3515 3516 queue index 6 / void / mark id 42 / end 3517 3518:: 3519 3520 void / mark id 42 / queue index 6 / end 3521 3522All actions in a list should have different types, otherwise only the last 3523action of a given type is taken into account:: 3524 3525 queue index 4 / queue index 5 / queue index 6 / end # will use queue 6 3526 3527:: 3528 3529 drop / drop / drop / end # drop is performed only once 3530 3531:: 3532 3533 mark id 42 / queue index 3 / mark id 24 / end # mark will be 24 3534 3535Considering they are performed simultaneously, opposite and overlapping 3536actions can sometimes be combined when the end result is unambiguous:: 3537 3538 drop / queue index 6 / end # drop has no effect 3539 3540:: 3541 3542 queue index 6 / rss queues 6 7 8 / end # queue has no effect 3543 3544:: 3545 3546 drop / passthru / end # drop has no effect 3547 3548Note that PMDs may still refuse such combinations. 3549 3550Actions 3551^^^^^^^ 3552 3553This section lists supported actions and their attributes, if any. 3554 3555- ``end``: end list of actions. 3556 3557- ``void``: no-op action. 3558 3559- ``passthru``: let subsequent rule process matched packets. 3560 3561- ``jump``: redirect traffic to group on device. 3562 3563 - ``group {unsigned}``: group to redirect to. 3564 3565- ``mark``: attach 32 bit value to packets. 3566 3567 - ``id {unsigned}``: 32 bit value to return with packets. 3568 3569- ``flag``: flag packets. 3570 3571- ``queue``: assign packets to a given queue index. 3572 3573 - ``index {unsigned}``: queue index to use. 3574 3575- ``drop``: drop packets (note: passthru has priority). 3576 3577- ``count``: enable counters for this rule. 3578 3579- ``rss``: spread packets among several queues. 3580 3581 - ``func {hash function}``: RSS hash function to apply, allowed tokens are 3582 the same as `set_hash_global_config`_. 3583 3584 - ``level {unsigned}``: encapsulation level for ``types``. 3585 3586 - ``types [{RSS hash type} [...]] end``: specific RSS hash types, allowed 3587 tokens are the same as `set_hash_input_set`_, except that an empty list 3588 does not disable RSS but instead requests unspecified "best-effort" 3589 settings. 3590 3591 - ``key {string}``: RSS hash key, overrides ``key_len``. 3592 3593 - ``key_len {unsigned}``: RSS hash key length in bytes, can be used in 3594 conjunction with ``key`` to pad or truncate it. 3595 3596 - ``queues [{unsigned} [...]] end``: queue indices to use. 3597 3598- ``pf``: direct traffic to physical function. 3599 3600- ``vf``: direct traffic to a virtual function ID. 3601 3602 - ``original {boolean}``: use original VF ID if possible. 3603 - ``id {unsigned}``: VF ID. 3604 3605- ``phy_port``: direct packets to physical port index. 3606 3607 - ``original {boolean}``: use original port index if possible. 3608 - ``index {unsigned}``: physical port index. 3609 3610- ``port_id``: direct matching traffic to a given DPDK port ID. 3611 3612 - ``original {boolean}``: use original DPDK port ID if possible. 3613 - ``id {unsigned}``: DPDK port ID. 3614 3615- ``of_set_mpls_ttl``: OpenFlow's ``OFPAT_SET_MPLS_TTL``. 3616 3617 - ``mpls_ttl``: MPLS TTL. 3618 3619- ``of_dec_mpls_ttl``: OpenFlow's ``OFPAT_DEC_MPLS_TTL``. 3620 3621- ``of_set_nw_ttl``: OpenFlow's ``OFPAT_SET_NW_TTL``. 3622 3623 - ``nw_ttl``: IP TTL. 3624 3625- ``of_dec_nw_ttl``: OpenFlow's ``OFPAT_DEC_NW_TTL``. 3626 3627- ``of_copy_ttl_out``: OpenFlow's ``OFPAT_COPY_TTL_OUT``. 3628 3629- ``of_copy_ttl_in``: OpenFlow's ``OFPAT_COPY_TTL_IN``. 3630 3631- ``of_pop_vlan``: OpenFlow's ``OFPAT_POP_VLAN``. 3632 3633- ``of_push_vlan``: OpenFlow's ``OFPAT_PUSH_VLAN``. 3634 3635 - ``ethertype``: Ethertype. 3636 3637- ``of_set_vlan_vid``: OpenFlow's ``OFPAT_SET_VLAN_VID``. 3638 3639 - ``vlan_vid``: VLAN id. 3640 3641- ``of_set_vlan_pcp``: OpenFlow's ``OFPAT_SET_VLAN_PCP``. 3642 3643 - ``vlan_pcp``: VLAN priority. 3644 3645- ``of_pop_mpls``: OpenFlow's ``OFPAT_POP_MPLS``. 3646 3647 - ``ethertype``: Ethertype. 3648 3649- ``of_push_mpls``: OpenFlow's ``OFPAT_PUSH_MPLS``. 3650 3651 - ``ethertype``: Ethertype. 3652 3653Destroying flow rules 3654~~~~~~~~~~~~~~~~~~~~~ 3655 3656``flow destroy`` destroys one or more rules from their rule ID (as returned 3657by ``flow create``), this command calls ``rte_flow_destroy()`` as many 3658times as necessary:: 3659 3660 flow destroy {port_id} rule {rule_id} [...] 3661 3662If successful, it will show:: 3663 3664 Flow rule #[...] destroyed 3665 3666It does not report anything for rule IDs that do not exist. The usual error 3667message is shown when a rule cannot be destroyed:: 3668 3669 Caught error type [...] ([...]): [...] 3670 3671``flow flush`` destroys all rules on a device and does not take extra 3672arguments. It is bound to ``rte_flow_flush()``:: 3673 3674 flow flush {port_id} 3675 3676Any errors are reported as above. 3677 3678Creating several rules and destroying them:: 3679 3680 testpmd> flow create 0 ingress pattern eth / ipv6 / end 3681 actions queue index 2 / end 3682 Flow rule #0 created 3683 testpmd> flow create 0 ingress pattern eth / ipv4 / end 3684 actions queue index 3 / end 3685 Flow rule #1 created 3686 testpmd> flow destroy 0 rule 0 rule 1 3687 Flow rule #1 destroyed 3688 Flow rule #0 destroyed 3689 testpmd> 3690 3691The same result can be achieved using ``flow flush``:: 3692 3693 testpmd> flow create 0 ingress pattern eth / ipv6 / end 3694 actions queue index 2 / end 3695 Flow rule #0 created 3696 testpmd> flow create 0 ingress pattern eth / ipv4 / end 3697 actions queue index 3 / end 3698 Flow rule #1 created 3699 testpmd> flow flush 0 3700 testpmd> 3701 3702Non-existent rule IDs are ignored:: 3703 3704 testpmd> flow create 0 ingress pattern eth / ipv6 / end 3705 actions queue index 2 / end 3706 Flow rule #0 created 3707 testpmd> flow create 0 ingress pattern eth / ipv4 / end 3708 actions queue index 3 / end 3709 Flow rule #1 created 3710 testpmd> flow destroy 0 rule 42 rule 10 rule 2 3711 testpmd> 3712 testpmd> flow destroy 0 rule 0 3713 Flow rule #0 destroyed 3714 testpmd> 3715 3716Querying flow rules 3717~~~~~~~~~~~~~~~~~~~ 3718 3719``flow query`` queries a specific action of a flow rule having that 3720ability. Such actions collect information that can be reported using this 3721command. It is bound to ``rte_flow_query()``:: 3722 3723 flow query {port_id} {rule_id} {action} 3724 3725If successful, it will display either the retrieved data for known actions 3726or the following message:: 3727 3728 Cannot display result for action type [...] ([...]) 3729 3730Otherwise, it will complain either that the rule does not exist or that some 3731error occurred:: 3732 3733 Flow rule #[...] not found 3734 3735:: 3736 3737 Caught error type [...] ([...]): [...] 3738 3739Currently only the ``count`` action is supported. This action reports the 3740number of packets that hit the flow rule and the total number of bytes. Its 3741output has the following format:: 3742 3743 count: 3744 hits_set: [...] # whether "hits" contains a valid value 3745 bytes_set: [...] # whether "bytes" contains a valid value 3746 hits: [...] # number of packets 3747 bytes: [...] # number of bytes 3748 3749Querying counters for TCPv6 packets redirected to queue 6:: 3750 3751 testpmd> flow create 0 ingress pattern eth / ipv6 / tcp / end 3752 actions queue index 6 / count / end 3753 Flow rule #4 created 3754 testpmd> flow query 0 4 count 3755 count: 3756 hits_set: 1 3757 bytes_set: 0 3758 hits: 386446 3759 bytes: 0 3760 testpmd> 3761 3762Listing flow rules 3763~~~~~~~~~~~~~~~~~~ 3764 3765``flow list`` lists existing flow rules sorted by priority and optionally 3766filtered by group identifiers:: 3767 3768 flow list {port_id} [group {group_id}] [...] 3769 3770This command only fails with the following message if the device does not 3771exist:: 3772 3773 Invalid port [...] 3774 3775Output consists of a header line followed by a short description of each 3776flow rule, one per line. There is no output at all when no flow rules are 3777configured on the device:: 3778 3779 ID Group Prio Attr Rule 3780 [...] [...] [...] [...] [...] 3781 3782``Attr`` column flags: 3783 3784- ``i`` for ``ingress``. 3785- ``e`` for ``egress``. 3786 3787Creating several flow rules and listing them:: 3788 3789 testpmd> flow create 0 ingress pattern eth / ipv4 / end 3790 actions queue index 6 / end 3791 Flow rule #0 created 3792 testpmd> flow create 0 ingress pattern eth / ipv6 / end 3793 actions queue index 2 / end 3794 Flow rule #1 created 3795 testpmd> flow create 0 priority 5 ingress pattern eth / ipv4 / udp / end 3796 actions rss queues 6 7 8 end / end 3797 Flow rule #2 created 3798 testpmd> flow list 0 3799 ID Group Prio Attr Rule 3800 0 0 0 i- ETH IPV4 => QUEUE 3801 1 0 0 i- ETH IPV6 => QUEUE 3802 2 0 5 i- ETH IPV4 UDP => RSS 3803 testpmd> 3804 3805Rules are sorted by priority (i.e. group ID first, then priority level):: 3806 3807 testpmd> flow list 1 3808 ID Group Prio Attr Rule 3809 0 0 0 i- ETH => COUNT 3810 6 0 500 i- ETH IPV6 TCP => DROP COUNT 3811 5 0 1000 i- ETH IPV6 ICMP => QUEUE 3812 1 24 0 i- ETH IPV4 UDP => QUEUE 3813 4 24 10 i- ETH IPV4 TCP => DROP 3814 3 24 20 i- ETH IPV4 => DROP 3815 2 24 42 i- ETH IPV4 UDP => QUEUE 3816 7 63 0 i- ETH IPV6 UDP VXLAN => MARK QUEUE 3817 testpmd> 3818 3819Output can be limited to specific groups:: 3820 3821 testpmd> flow list 1 group 0 group 63 3822 ID Group Prio Attr Rule 3823 0 0 0 i- ETH => COUNT 3824 6 0 500 i- ETH IPV6 TCP => DROP COUNT 3825 5 0 1000 i- ETH IPV6 ICMP => QUEUE 3826 7 63 0 i- ETH IPV6 UDP VXLAN => MARK QUEUE 3827 testpmd> 3828 3829Toggling isolated mode 3830~~~~~~~~~~~~~~~~~~~~~~ 3831 3832``flow isolate`` can be used to tell the underlying PMD that ingress traffic 3833must only be injected from the defined flow rules; that no default traffic 3834is expected outside those rules and the driver is free to assign more 3835resources to handle them. It is bound to ``rte_flow_isolate()``:: 3836 3837 flow isolate {port_id} {boolean} 3838 3839If successful, enabling or disabling isolated mode shows either:: 3840 3841 Ingress traffic on port [...] 3842 is now restricted to the defined flow rules 3843 3844Or:: 3845 3846 Ingress traffic on port [...] 3847 is not restricted anymore to the defined flow rules 3848 3849Otherwise, in case of error:: 3850 3851 Caught error type [...] ([...]): [...] 3852 3853Mainly due to its side effects, PMDs supporting this mode may not have the 3854ability to toggle it more than once without reinitializing affected ports 3855first (e.g. by exiting testpmd). 3856 3857Enabling isolated mode:: 3858 3859 testpmd> flow isolate 0 true 3860 Ingress traffic on port 0 is now restricted to the defined flow rules 3861 testpmd> 3862 3863Disabling isolated mode:: 3864 3865 testpmd> flow isolate 0 false 3866 Ingress traffic on port 0 is not restricted anymore to the defined flow rules 3867 testpmd> 3868 3869Sample QinQ flow rules 3870~~~~~~~~~~~~~~~~~~~~~~ 3871 3872Before creating QinQ rule(s) the following commands should be issued to enable QinQ:: 3873 3874 testpmd> port stop 0 3875 testpmd> vlan set qinq on 0 3876 3877The above command sets the inner and outer TPID's to 0x8100. 3878 3879To change the TPID's the following commands should be used:: 3880 3881 testpmd> vlan set outer tpid 0xa100 0 3882 testpmd> vlan set inner tpid 0x9100 0 3883 testpmd> port start 0 3884 3885Validate and create a QinQ rule on port 0 to steer traffic to a VF queue in a VM. 3886 3887:: 3888 3889 testpmd> flow validate 0 ingress pattern eth / vlan tci is 123 / 3890 vlan tci is 456 / end actions vf id 1 / queue index 0 / end 3891 Flow rule #0 validated 3892 3893 testpmd> flow create 0 ingress pattern eth / vlan tci is 4 / 3894 vlan tci is 456 / end actions vf id 123 / queue index 0 / end 3895 Flow rule #0 created 3896 3897 testpmd> flow list 0 3898 ID Group Prio Attr Rule 3899 0 0 0 i- ETH VLAN VLAN=>VF QUEUE 3900 3901Validate and create a QinQ rule on port 0 to steer traffic to a queue on the host. 3902 3903:: 3904 3905 testpmd> flow validate 0 ingress pattern eth / vlan tci is 321 / 3906 vlan tci is 654 / end actions pf / queue index 0 / end 3907 Flow rule #1 validated 3908 3909 testpmd> flow create 0 ingress pattern eth / vlan tci is 321 / 3910 vlan tci is 654 / end actions pf / queue index 1 / end 3911 Flow rule #1 created 3912 3913 testpmd> flow list 0 3914 ID Group Prio Attr Rule 3915 0 0 0 i- ETH VLAN VLAN=>VF QUEUE 3916 1 0 0 i- ETH VLAN VLAN=>PF QUEUE 3917 3918BPF Functions 3919-------------- 3920 3921The following sections show functions to load/unload eBPF based filters. 3922 3923bpf-load 3924~~~~~~~~ 3925 3926Load an eBPF program as a callback for partciular RX/TX queue:: 3927 3928 testpmd> bpf-load rx|tx (portid) (queueid) (load-flags) (bpf-prog-filename) 3929 3930The available load-flags are: 3931 3932* ``J``: use JIT generated native code, otherwise BPF interpreter will be used. 3933 3934* ``M``: assume input parameter is a pointer to rte_mbuf, otherwise assume it is a pointer to first segment's data. 3935 3936* ``-``: none. 3937 3938.. note:: 3939 3940 You'll need clang v3.7 or above to build bpf program you'd like to load 3941 3942For example: 3943 3944.. code-block:: console 3945 3946 cd test/bpf 3947 clang -O2 -target bpf -c t1.c 3948 3949Then to load (and JIT compile) t1.o at RX queue 0, port 1:: 3950 3951.. code-block:: console 3952 3953 testpmd> bpf-load rx 1 0 J ./dpdk.org/test/bpf/t1.o 3954 3955To load (not JITed) t1.o at TX queue 0, port 0:: 3956 3957.. code-block:: console 3958 3959 testpmd> bpf-load tx 0 0 - ./dpdk.org/test/bpf/t1.o 3960 3961bpf-unload 3962~~~~~~~~~~ 3963 3964Unload previously loaded eBPF program for partciular RX/TX queue:: 3965 3966 testpmd> bpf-unload rx|tx (portid) (queueid) 3967 3968For example to unload BPF filter from TX queue 0, port 0: 3969 3970.. code-block:: console 3971 3972 testpmd> bpf-load tx 0 0 - ./dpdk.org/test/bpf/t1.o 3973