1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright(c) 2010-2016 Intel Corporation.
6 Testpmd Runtime Functions
7 =========================
9 Where the testpmd application is started in interactive mode, (``-i|--interactive``),
10 it displays a prompt that can be used to start and stop forwarding,
11 configure the application, display statistics (including the extended NIC
12 statistics aka xstats) , set the Flow Director and other tasks::
16 The testpmd prompt has some, limited, readline support.
17 Common bash command-line functions such as ``Ctrl+a`` and ``Ctrl+e`` to go to the start and end of the prompt line are supported
18 as well as access to the command history via the up-arrow.
20 There is also support for tab completion.
21 If you type a partial command and hit ``<TAB>`` you get a list of the available completions:
23 .. code-block:: console
25 testpmd> show port <TAB>
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
36 Some examples in this document are too long to fit on one line are are shown wrapped at `"\\"` for display purposes::
38 testpmd> set flow_ctrl rx (on|off) tx (on|off) (high_water) (low_water) \
39 (pause_time) (send_xon) (port_id)
41 In the real ``testpmd>`` prompt these commands should be on a single line.
46 The testpmd has on-line help for the functions that are available at runtime.
47 These are divided into sections and can be accessed using help, help section or help all:
49 .. code-block:: console
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.
62 Command File Functions
63 ----------------------
65 To facilitate loading large number of commands or to avoid cutting and pasting where not
66 practical or possible testpmd supports alternative methods for executing commands.
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.
72 .. code-block:: console
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
89 Flow rule #498 created
90 Flow rule #499 created
91 Read all CLI commands from /home/ubuntu/flow-create-commands.txt
95 * At run-time additional commands can be loaded in bulk by invoking the ``load FILENAME``
98 .. code-block:: console
100 testpmd> load /home/ubuntu/flow-create-commands.txt
105 Flow rule #498 created
106 Flow rule #499 created
107 Read all CLI commands from /home/ubuntu/flow-create-commands.txt
111 In all cases output from any included command will be displayed as standard output.
112 Execution will continue until the end of the file is reached regardless of
113 whether any errors occur. The end user must examine the output to determine if
114 any failures occurred.
123 Start packet forwarding with current configuration::
130 Start packet forwarding with current configuration after sending specified number of bursts of packets::
132 testpmd> start tx_first (""|burst_num)
134 The default burst number is 1 when ``burst_num`` not presented.
139 Stop packet forwarding, and display accumulated statistics::
154 The functions in the following sections are used to display information about the
155 testpmd configuration or the NIC status.
160 Display information for a given port or all ports::
162 testpmd> show port (info|stats|xstats|fdir|stat_qmap|dcb_tc|cap) (port_id|all)
164 The available information categories are:
166 * ``info``: General port information such as MAC address.
168 * ``stats``: RX/TX statistics.
170 * ``xstats``: RX/TX extended NIC statistics.
172 * ``fdir``: Flow Director information and statistics.
174 * ``stat_qmap``: Queue statistics mapping.
176 * ``dcb_tc``: DCB information such as TC mapping.
178 * ``cap``: Supported offload capabilities.
182 .. code-block:: console
184 testpmd> show port info 0
186 ********************* Infos for port 0 *********************
188 MAC address: XX:XX:XX:XX:XX:XX
190 memory allocation on the socket: 0
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
202 Redirection table size: 512
203 Supported flow types:
223 Display the rss redirection table entry indicated by masks on port X::
225 testpmd> show port (port_id) rss reta (size) (mask0, mask1...)
227 size is used to indicate the hardware supported reta size
232 Display the RSS hash functions and RSS hash key of a port::
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]
239 Clear the port statistics for a given port or for all ports::
241 testpmd> clear port (info|stats|xstats|fdir|stat_qmap) (port_id|all)
245 testpmd> clear port stats all
250 Display information for a given port's RX/TX queue::
252 testpmd> show (rxq|txq) info (port_id) (queue_id)
257 Displays the configuration of the application.
258 The configuration comes from the command-line, the runtime or the application defaults::
260 testpmd> show config (rxtx|cores|fwd|txpkts)
262 The available information categories are:
264 * ``rxtx``: RX/TX configuration items.
266 * ``cores``: List of forwarding cores.
268 * ``fwd``: Packet forwarding configuration.
270 * ``txpkts``: Packets to TX configuration.
274 .. code-block:: console
276 testpmd> show config rxtx
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
289 Set the packet forwarding mode::
291 testpmd> set fwd (io|mac|macswap|flowgen| \
292 rxonly|txonly|csum|icmpecho) (""|retry)
294 ``retry`` can be specified for forwarding engines except ``rx_only``.
296 The available information categories are:
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.
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.
307 * ``macswap``: MAC swap forwarding mode.
308 Swaps the source and the destination Ethernet addresses of packets before forwarding them.
310 * ``flowgen``: Multi-flow generation mode.
311 Originates a number of flows (with varying destination IP addresses), and terminate receive traffic.
313 * ``rxonly``: Receives packets but doesn't transmit them.
315 * ``txonly``: Generates and transmits packets without receiving any.
317 * ``csum``: Changes the checksum field with hardware or software methods depending on the offload flags on the packet.
319 * ``icmpecho``: Receives a burst of packets, lookup for IMCP echo requests and, if any, send back ICMP echo replies.
321 * ``ieee1588``: Demonstrate L2 IEEE1588 V2 PTP timestamping for RX and TX. Requires ``CONFIG_RTE_LIBRTE_IEEE1588=y``.
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``.
332 testpmd> set fwd rxonly
334 Set rxonly packet forwarding mode
340 Display an RX descriptor for a port RX queue::
342 testpmd> read rxd (port_id) (queue_id) (rxd_id)
346 testpmd> read rxd 0 0 4
347 0x0000000B - 0x001D0180 / 0x0000000B - 0x001D0180
352 Display a TX descriptor for a port TX queue::
354 testpmd> read txd (port_id) (queue_id) (txd_id)
358 testpmd> read txd 0 0 4
359 0x00000001 - 0x24C3C440 / 0x000F0000 - 0x2330003C
364 Get loaded dynamic device personalization (DDP) package info list::
366 testpmd> ddp get list (port_id)
371 Display information about dynamic device personalization (DDP) profile::
373 testpmd> ddp get info (profile_path)
378 Display VF statistics::
380 testpmd> show vf stats (port_id) (vf_id)
385 Reset VF statistics::
387 testpmd> clear vf stats (port_id) (vf_id)
389 show port pctype mapping
390 ~~~~~~~~~~~~~~~~~~~~~~~~
392 List all items from the pctype mapping table::
394 testpmd> show port (port_id) pctype mapping
396 show rx offloading capabilities
397 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
399 List all per queue and per port Rx offloading capabilities of a port::
401 testpmd> show port (port_id) rx_offload capabilities
403 show rx offloading configuration
404 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
406 List port level and all queue level Rx offloading configuration::
408 testpmd> show port (port_id) rx_offload configuration
410 show tx offloading capabilities
411 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
413 List all per queue and per port Tx offloading capabilities of a port::
415 testpmd> show port (port_id) tx_offload capabilities
417 show tx offloading configuration
418 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
420 List port level and all queue level Tx offloading configuration::
422 testpmd> show port (port_id) tx_offload configuration
425 Configuration Functions
426 -----------------------
428 The testpmd application can be configured from the runtime as well as from the command-line.
430 This section details the available configuration functions that are available.
434 Configuration changes only become active when forwarding is started/restarted.
439 Reset forwarding to the default configuration::
446 Set the debug verbosity level::
448 testpmd> set verbose (level)
450 Currently the only available levels are 0 (silent except for error) and 1 (fully verbose).
455 Set the log level for a log type::
457 testpmd> set log global|(type) (level)
461 * ``type`` is the log name.
463 * ``level`` is the log level.
465 For example, to change the global log level::
466 testpmd> set log global (level)
468 Regexes can also be used for type. To change log level of user1, user2 and user3::
469 testpmd> set log user[1-3] (level)
474 Set the number of ports used by the application:
478 This is equivalent to the ``--nb-ports`` command-line option.
483 Set the number of cores used by the application::
485 testpmd> set nbcore (num)
487 This is equivalent to the ``--nb-cores`` command-line option.
491 The number of cores used must not be greater than number of ports used multiplied by the number of queues per port.
496 Set the forwarding cores hexadecimal mask::
498 testpmd> set coremask (mask)
500 This is equivalent to the ``--coremask`` command-line option.
504 The master lcore is reserved for command line parsing only and cannot be masked on for packet forwarding.
509 Set the forwarding ports hexadecimal mask::
511 testpmd> set portmask (mask)
513 This is equivalent to the ``--portmask`` command-line option.
518 Set number of packets per burst::
520 testpmd> set burst (num)
522 This is equivalent to the ``--burst command-line`` option.
524 When retry is enabled, the transmit delay time and number of retries can also be set::
526 testpmd> set burst tx delay (microseconds) retry (num)
531 Set the length of each segment of the TX-ONLY packets or length of packet for FLOWGEN mode::
533 testpmd> set txpkts (x[,y]*)
535 Where x[,y]* represents a CSV list of values, without white space.
540 Set the split policy for the TX packets, applicable for TX-ONLY and CSUM forwarding modes::
542 testpmd> set txsplit (off|on|rand)
546 * ``off`` disable packet copy & split for CSUM mode.
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
552 * ``rand`` same as 'on', but number of segments per each packet is a random value between 1 and total number of segments.
557 Set the list of forwarding cores::
559 testpmd> set corelist (x[,y]*)
561 For example, to change the forwarding cores:
563 .. code-block:: console
565 testpmd> set corelist 3,1
566 testpmd> show config fwd
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
576 The cores are used in the same order as specified on the command line.
581 Set the list of forwarding ports::
583 testpmd> set portlist (x[,y]*)
585 For example, to change the port forwarding:
587 .. code-block:: console
589 testpmd> set portlist 0,2,1,3
590 testpmd> show config fwd
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
602 Enable/disable tx loopback::
604 testpmd> set tx loopback (port_id) (on|off)
609 set drop enable bit for all queues::
611 testpmd> set all queues drop (port_id) (on|off)
613 set split drop enable (for VF)
614 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
616 set split drop enable bit for VF from PF::
618 testpmd> set vf split drop (port_id) (vf_id) (on|off)
620 set mac antispoof (for VF)
621 ~~~~~~~~~~~~~~~~~~~~~~~~~~
623 Set mac antispoof for a VF from the PF::
625 testpmd> set vf mac antispoof (port_id) (vf_id) (on|off)
630 Enable/disable MACsec offload::
632 testpmd> set macsec offload (port_id) on encrypt (on|off) replay-protect (on|off)
633 testpmd> set macsec offload (port_id) off
638 Configure MACsec secure connection (SC)::
640 testpmd> set macsec sc (tx|rx) (port_id) (mac) (pi)
644 The pi argument is ignored for tx.
645 Check the NIC Datasheet for hardware limits.
650 Configure MACsec secure association (SA)::
652 testpmd> set macsec sa (tx|rx) (port_id) (idx) (an) (pn) (key)
656 The IDX value must be 0 or 1.
657 Check the NIC Datasheet for hardware limits.
659 set broadcast mode (for VF)
660 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
662 Set broadcast mode for a VF from the PF::
664 testpmd> set vf broadcast (port_id) (vf_id) (on|off)
669 Set the VLAN strip on a port::
671 testpmd> vlan set strip (on|off) (port_id)
676 Set the VLAN strip for a queue on a port::
678 testpmd> vlan set stripq (on|off) (port_id,queue_id)
680 vlan set stripq (for VF)
681 ~~~~~~~~~~~~~~~~~~~~~~~~
683 Set VLAN strip for all queues in a pool for a VF from the PF::
685 testpmd> set vf vlan stripq (port_id) (vf_id) (on|off)
687 vlan set insert (for VF)
688 ~~~~~~~~~~~~~~~~~~~~~~~~
690 Set VLAN insert for a VF from the PF::
692 testpmd> set vf vlan insert (port_id) (vf_id) (vlan_id)
694 vlan set tag (for VF)
695 ~~~~~~~~~~~~~~~~~~~~~
697 Set VLAN tag for a VF from the PF::
699 testpmd> set vf vlan tag (port_id) (vf_id) (on|off)
701 vlan set antispoof (for VF)
702 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
704 Set VLAN antispoof for a VF from the PF::
706 testpmd> set vf vlan antispoof (port_id) (vf_id) (on|off)
711 Set the VLAN filter on a port::
713 testpmd> vlan set filter (on|off) (port_id)
718 Set the VLAN QinQ (extended queue in queue) on for a port::
720 testpmd> vlan set qinq (on|off) (port_id)
725 Set the inner or outer VLAN TPID for packet filtering on a port::
727 testpmd> vlan set (inner|outer) tpid (value) (port_id)
731 TPID value must be a 16-bit number (value <= 65536).
736 Add a VLAN ID, or all identifiers, to the set of VLAN identifiers filtered by port ID::
738 testpmd> rx_vlan add (vlan_id|all) (port_id)
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.
749 Remove a VLAN ID, or all identifiers, from the set of VLAN identifiers filtered by port ID::
751 testpmd> rx_vlan rm (vlan_id|all) (port_id)
756 Add a VLAN ID, to the set of VLAN identifiers filtered for VF(s) for port ID::
758 testpmd> rx_vlan add (vlan_id) port (port_id) vf (vf_mask)
763 Remove a VLAN ID, from the set of VLAN identifiers filtered for VF(s) for port ID::
765 testpmd> rx_vlan rm (vlan_id) port (port_id) vf (vf_mask)
770 Add a tunnel filter on a port::
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)
776 The available information categories are:
778 * ``vxlan``: Set tunnel type as VXLAN.
780 * ``nvgre``: Set tunnel type as NVGRE.
782 * ``ipingre``: Set tunnel type as IP-in-GRE.
784 * ``imac-ivlan``: Set filter type as Inner MAC and VLAN.
786 * ``imac-ivlan-tenid``: Set filter type as Inner MAC, VLAN and tenant ID.
788 * ``imac-tenid``: Set filter type as Inner MAC and tenant ID.
790 * ``imac``: Set filter type as Inner MAC.
792 * ``omac-imac-tenid``: Set filter type as Outer MAC, Inner MAC and tenant ID.
794 * ``oip``: Set filter type as Outer IP.
796 * ``iip``: Set filter type as Inner IP.
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
803 Set an IP-in-GRE tunnel on port 0, and the filter type is Outer IP.
808 Remove a tunnel filter on a port::
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)
817 Add an UDP port for VXLAN packet filter on a port::
819 testpmd> rx_vxlan_port add (udp_port) (port_id)
824 Remove an UDP port for VXLAN packet filter on a port::
826 testpmd> rx_vxlan_port rm (udp_port) (port_id)
831 Set hardware insertion of VLAN IDs in packets sent on a port::
833 testpmd> tx_vlan set (port_id) vlan_id[, vlan_id_outer]
835 For example, set a single VLAN ID (5) insertion on port 0::
839 Or, set double VLAN ID (inner: 2, outer: 3) insertion on port 1::
847 Set port based hardware insertion of VLAN ID in packets sent on a port::
849 testpmd> tx_vlan set pvid (port_id) (vlan_id) (on|off)
854 Disable hardware insertion of a VLAN header in packets sent on a port::
856 testpmd> tx_vlan reset (port_id)
861 Select hardware or software calculation of the checksum when
862 transmitting a packet using the ``csum`` forwarding engine::
864 testpmd> csum set (ip|udp|tcp|sctp|outer-ip) (hw|sw) (port_id)
868 * ``ip|udp|tcp|sctp`` always relate to the inner layer.
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.
876 Check the NIC Datasheet for hardware limits.
881 Set RSS queue region span on a port::
883 testpmd> set port (port_id) queue-region region_id (value) \
884 queue_start_index (value) queue_num (value)
886 Set flowtype mapping on a RSS queue region on a port::
888 testpmd> set port (port_id) queue-region region_id (value) flowtype (value)
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.
895 Set user priority mapping on a RSS queue region on a port::
897 testpmd> set port (port_id) queue-region UP (value) region_id (value)
899 Flush all queue region related configuration on a port::
901 testpmd> set port (port_id) queue-region flush (on|off)
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.
912 Show all queue region related configuration info on a port::
914 testpmd> show port (port_id) queue-region
918 Queue region only support on PF by now, so these command is
919 only for configuration of queue region on PF port.
924 Define how tunneled packets should be handled by the csum forward
927 testpmd> csum parse-tunnel (on|off) (tx_port_id)
929 If enabled, the csum forward engine will try to recognize supported
930 tunnel headers (vxlan, gre, ipip).
932 If disabled, treat tunnel packets as non-tunneled packets (a inner
933 header is handled as a packet payload).
937 The port argument is the TX port like in the ``csum set`` command.
941 Consider a packet in packet like the following::
943 eth_out/ipv4_out/udp_out/vxlan/eth_in/ipv4_in/tcp_in
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``).
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``.
955 Display tx checksum offload configuration::
957 testpmd> csum show (port_id)
962 Enable TCP Segmentation Offload (TSO) in the ``csum`` forwarding engine::
964 testpmd> tso set (segsize) (port_id)
968 Check the NIC datasheet for hardware limits.
973 Display the status of TCP Segmentation Offload::
975 testpmd> tso show (port_id)
980 Enable or disable GRO in ``csum`` forwarding engine::
982 testpmd> set port <port_id> gro on|off
984 If enabled, the csum forwarding engine will perform GRO on the TCP/IPv4
985 packets received from the given port.
987 If disabled, packets received from the given port won't be performed
988 GRO. By default, GRO is disabled for all ports.
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
1003 Display GRO configuration for a given port::
1005 testpmd> show port <port_id> gro
1010 Set the cycle to flush the GROed packets from reassembly tables::
1012 testpmd> set gro flush <cycles>
1014 When enable GRO, the csum forwarding engine performs GRO on received
1015 packets, and the GROed packets are stored in reassembly tables. Users
1016 can use this command to determine when the GROed packets are flushed
1017 from the reassembly tables.
1019 The ``cycles`` is measured in GRO operation times. The csum forwarding
1020 engine flushes the GROed packets from the tables every ``cycles`` GRO
1023 By default, the value of ``cycles`` is 1, which means flush GROed packets
1024 from the reassembly tables as soon as one GRO operation finishes. The value
1025 of ``cycles`` should be in the range of 1 to ``GRO_MAX_FLUSH_CYCLES``.
1027 Please note that the large value of ``cycles`` may cause the poor TCP/IP
1028 stack performance. Because the GROed packets are delayed to arrive the
1029 stack, thus causing more duplicated ACKs and TCP retransmissions.
1034 Toggle per-port GSO support in ``csum`` forwarding engine::
1036 testpmd> set port <port_id> gso on|off
1038 If enabled, the csum forwarding engine will perform GSO on supported IPv4
1039 packets, transmitted on the given port.
1041 If disabled, packets transmitted on the given port will not undergo GSO.
1042 By default, GSO is disabled for all ports.
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
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:
1056 testpmd> csum set outer_ip hw <port_id>
1058 testpmd> csum set ip hw <port_id>
1060 testpmd> csum set tcp hw <port_id>
1065 Set the maximum GSO segment size (measured in bytes), which includes the
1066 packet header and the packet payload for GSO-enabled ports (global)::
1068 testpmd> set gso segsz <length>
1073 Display the status of Generic Segmentation Offload for a given port::
1075 testpmd> show port <port_id> gso
1080 Add an alternative MAC address to a port::
1082 testpmd> mac_addr add (port_id) (XX:XX:XX:XX:XX:XX)
1087 Remove a MAC address from a port::
1089 testpmd> mac_addr remove (port_id) (XX:XX:XX:XX:XX:XX)
1091 mac_addr add (for VF)
1092 ~~~~~~~~~~~~~~~~~~~~~
1094 Add an alternative MAC address for a VF to a port::
1096 testpmd> mac_add add port (port_id) vf (vf_id) (XX:XX:XX:XX:XX:XX)
1101 Set the default MAC address for a port::
1103 testpmd> mac_addr set (port_id) (XX:XX:XX:XX:XX:XX)
1105 mac_addr set (for VF)
1106 ~~~~~~~~~~~~~~~~~~~~~
1108 Set the MAC address for a VF from the PF::
1110 testpmd> set vf mac addr (port_id) (vf_id) (XX:XX:XX:XX:XX:XX)
1115 Set the forwarding peer address for certain port::
1117 testpmd> set eth-peer (port_id) (perr_addr)
1119 This is equivalent to the ``--eth-peer`` command-line option.
1124 Set the unicast hash filter(s) on/off for a port::
1126 testpmd> set port (port_id) uta (XX:XX:XX:XX:XX:XX|all) (on|off)
1131 Set the promiscuous mode on for a port or for all ports.
1132 In promiscuous mode packets are not dropped if they aren't for the specified MAC address::
1134 testpmd> set promisc (port_id|all) (on|off)
1139 Set the allmulti mode for a port or for all ports::
1141 testpmd> set allmulti (port_id|all) (on|off)
1143 Same as the ifconfig (8) option. Controls how multicast packets are handled.
1145 set promisc (for VF)
1146 ~~~~~~~~~~~~~~~~~~~~
1148 Set the unicast promiscuous mode for a VF from PF.
1149 It's supported by Intel i40e NICs now.
1150 In promiscuous mode packets are not dropped if they aren't for the specified MAC address::
1152 testpmd> set vf promisc (port_id) (vf_id) (on|off)
1154 set allmulticast (for VF)
1155 ~~~~~~~~~~~~~~~~~~~~~~~~~
1157 Set the multicast promiscuous mode for a VF from PF.
1158 It's supported by Intel i40e NICs now.
1159 In promiscuous mode packets are not dropped if they aren't for the specified MAC address::
1161 testpmd> set vf allmulti (port_id) (vf_id) (on|off)
1163 set tx max bandwidth (for VF)
1164 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1166 Set TX max absolute bandwidth (Mbps) for a VF from PF::
1168 testpmd> set vf tx max-bandwidth (port_id) (vf_id) (max_bandwidth)
1170 set tc tx min bandwidth (for VF)
1171 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1173 Set all TCs' TX min relative bandwidth (%) for a VF from PF::
1175 testpmd> set vf tc tx min-bandwidth (port_id) (vf_id) (bw1, bw2, ...)
1177 set tc tx max bandwidth (for VF)
1178 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1180 Set a TC's TX max absolute bandwidth (Mbps) for a VF from PF::
1182 testpmd> set vf tc tx max-bandwidth (port_id) (vf_id) (tc_no) (max_bandwidth)
1184 set tc strict link priority mode
1185 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1187 Set some TCs' strict link priority mode on a physical port::
1189 testpmd> set tx strict-link-priority (port_id) (tc_bitmap)
1191 set tc tx min bandwidth
1192 ~~~~~~~~~~~~~~~~~~~~~~~
1194 Set all TCs' TX min relative bandwidth (%) globally for all PF and VFs::
1196 testpmd> set tc tx min-bandwidth (port_id) (bw1, bw2, ...)
1201 Set the link flow control parameter on a port::
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)
1209 * ``high_water`` (integer): High threshold value to trigger XOFF.
1211 * ``low_water`` (integer): Low threshold value to trigger XON.
1213 * ``pause_time`` (integer): Pause quota in the Pause frame.
1215 * ``send_xon`` (0/1): Send XON frame.
1217 * ``mac_ctrl_frame_fwd``: Enable receiving MAC control frames.
1219 * ``autoneg``: Change the auto-negotiation parameter.
1224 Set the priority flow control parameter on a port::
1226 testpmd> set pfc_ctrl rx (on|off) tx (on|off) (high_water) (low_water) \
1227 (pause_time) (priority) (port_id)
1231 * ``high_water`` (integer): High threshold value.
1233 * ``low_water`` (integer): Low threshold value.
1235 * ``pause_time`` (integer): Pause quota in the Pause frame.
1237 * ``priority`` (0-7): VLAN User Priority.
1242 Set statistics mapping (qmapping 0..15) for RX/TX queue on port::
1244 testpmd> set stat_qmap (tx|rx) (port_id) (queue_id) (qmapping)
1246 For example, to set rx queue 2 on port 0 to mapping 5::
1248 testpmd>set stat_qmap rx 0 2 5
1250 set xstats-hide-zero
1251 ~~~~~~~~~~~~~~~~~~~~
1253 Set the option to hide zero values for xstats display::
1255 testpmd> set xstats-hide-zero on|off
1259 By default, the zero values are displayed for xstats.
1261 set port - rx/tx (for VF)
1262 ~~~~~~~~~~~~~~~~~~~~~~~~~
1264 Set VF receive/transmit from a port::
1266 testpmd> set port (port_id) vf (vf_id) (rx|tx) (on|off)
1268 set port - mac address filter (for VF)
1269 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1271 Add/Remove unicast or multicast MAC addr filter for a VF::
1273 testpmd> set port (port_id) vf (vf_id) (mac_addr) \
1274 (exact-mac|exact-mac-vlan|hashmac|hashmac-vlan) (on|off)
1276 set port - rx mode(for VF)
1277 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1279 Set the VF receive mode of a port::
1281 testpmd> set port (port_id) vf (vf_id) \
1282 rxmode (AUPE|ROPE|BAM|MPE) (on|off)
1284 The available receive modes are:
1286 * ``AUPE``: Accepts untagged VLAN.
1288 * ``ROPE``: Accepts unicast hash.
1290 * ``BAM``: Accepts broadcast packets.
1292 * ``MPE``: Accepts all multicast packets.
1294 set port - tx_rate (for Queue)
1295 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1297 Set TX rate limitation for a queue on a port::
1299 testpmd> set port (port_id) queue (queue_id) rate (rate_value)
1301 set port - tx_rate (for VF)
1302 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
1304 Set TX rate limitation for queues in VF on a port::
1306 testpmd> set port (port_id) vf (vf_id) rate (rate_value) queue_mask (queue_mask)
1308 set port - mirror rule
1309 ~~~~~~~~~~~~~~~~~~~~~~
1311 Set pool or vlan type mirror rule for a port::
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)
1317 Set link mirror rule for a port::
1319 testpmd> set port (port_id) mirror-rule (rule_id) \
1320 (uplink-mirror|downlink-mirror) dst-pool (pool_id) (on|off)
1322 For example to enable mirror traffic with vlan 0,1 to pool 0::
1324 set port 0 mirror-rule 0 vlan-mirror 0,1 dst-pool 0 on
1326 reset port - mirror rule
1327 ~~~~~~~~~~~~~~~~~~~~~~~~
1329 Reset a mirror rule for a port::
1331 testpmd> reset port (port_id) mirror-rule (rule_id)
1336 Set the flush on RX streams before forwarding.
1337 The default is flush ``on``.
1338 Mainly used with PCAP drivers to turn off the default behavior of flushing the first 512 packets on RX streams::
1340 testpmd> set flush_rx off
1345 Set the bypass mode for the lowest port on bypass enabled NIC::
1347 testpmd> set bypass mode (normal|bypass|isolate) (port_id)
1352 Set the event required to initiate specified bypass mode for the lowest port on a bypass enabled::
1354 testpmd> set bypass event (timeout|os_on|os_off|power_on|power_off) \
1355 mode (normal|bypass|isolate) (port_id)
1359 * ``timeout``: Enable bypass after watchdog timeout.
1361 * ``os_on``: Enable bypass when OS/board is powered on.
1363 * ``os_off``: Enable bypass when OS/board is powered off.
1365 * ``power_on``: Enable bypass when power supply is turned on.
1367 * ``power_off``: Enable bypass when power supply is turned off.
1373 Set the bypass watchdog timeout to ``n`` seconds where 0 = instant::
1375 testpmd> set bypass timeout (0|1.5|2|3|4|8|16|32)
1380 Show the bypass configuration for a bypass enabled NIC using the lowest port on the NIC::
1382 testpmd> show bypass config (port_id)
1387 Set link up for a port::
1389 testpmd> set link-up port (port id)
1394 Set link down for a port::
1396 testpmd> set link-down port (port id)
1401 Enable E-tag insertion for a VF on a port::
1403 testpmd> E-tag set insertion on port-tag-id (value) port (port_id) vf (vf_id)
1405 Disable E-tag insertion for a VF on a port::
1407 testpmd> E-tag set insertion off port (port_id) vf (vf_id)
1409 Enable/disable E-tag stripping on a port::
1411 testpmd> E-tag set stripping (on|off) port (port_id)
1413 Enable/disable E-tag based forwarding on a port::
1415 testpmd> E-tag set forwarding (on|off) port (port_id)
1417 Add an E-tag forwarding filter on a port::
1419 testpmd> E-tag set filter add e-tag-id (value) dst-pool (pool_id) port (port_id)
1421 Delete an E-tag forwarding filter on a port::
1422 testpmd> E-tag set filter del e-tag-id (value) port (port_id)
1427 Load a dynamic device personalization (DDP) profile and store backup profile::
1429 testpmd> ddp add (port_id) (profile_path[,backup_profile_path])
1434 Delete a dynamic device personalization profile and restore backup profile::
1436 testpmd> ddp del (port_id) (backup_profile_path)
1441 List all items from the ptype mapping table::
1443 testpmd> ptype mapping get (port_id) (valid_only)
1447 * ``valid_only``: A flag indicates if only list valid items(=1) or all itemss(=0).
1449 Replace a specific or a group of software defined ptype with a new one::
1451 testpmd> ptype mapping replace (port_id) (target) (mask) (pkt_type)
1455 * ``target``: A specific software ptype or a mask to represent a group of software ptypes.
1457 * ``mask``: A flag indicate if "target" is a specific software ptype(=0) or a ptype mask(=1).
1459 * ``pkt_type``: The new software ptype to replace the old ones.
1461 Update hardware defined ptype to software defined packet type mapping table::
1463 testpmd> ptype mapping update (port_id) (hw_ptype) (sw_ptype)
1467 * ``hw_ptype``: hardware ptype as the index of the ptype mapping table.
1469 * ``sw_ptype``: software ptype as the value of the ptype mapping table.
1471 Reset ptype mapping table::
1473 testpmd> ptype mapping reset (port_id)
1475 config per port Rx offloading
1476 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1478 Enable or disable a per port Rx offloading on all Rx queues of a port::
1480 testpmd> port config (port_id) rx_offload (offloading) on|off
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
1488 This command should be run when the port is stopped, or else it will fail.
1490 config per queue Rx offloading
1491 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1493 Enable or disable a per queue Rx offloading only on a specific Rx queue::
1495 testpmd> port (port_id) rxq (queue_id) rx_offload (offloading) on|off
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
1503 This command should be run when the port is stopped, or else it will fail.
1505 config per port Tx offloading
1506 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1508 Enable or disable a per port Tx offloading on all Tx queues of a port::
1510 testpmd> port config (port_id) tx_offload (offloading) on|off
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
1519 This command should be run when the port is stopped, or else it will fail.
1521 config per queue Tx offloading
1522 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1524 Enable or disable a per queue Tx offloading only on a specific Tx queue::
1526 testpmd> port (port_id) txq (queue_id) tx_offload (offloading) on|off
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
1535 This command should be run when the port is stopped, or else it will fail.
1541 The following sections show functions for configuring ports.
1545 Port configuration changes only become active when forwarding is started/restarted.
1550 Attach a port specified by pci address or virtual device args::
1552 testpmd> port attach (identifier)
1554 To attach a new pci device, the device should be recognized by kernel first.
1555 Then it should be moved under DPDK management.
1556 Finally the port can be attached to testpmd.
1558 For example, to move a pci device using ixgbe under DPDK management:
1560 .. code-block:: console
1562 # Check the status of the available devices.
1563 ./usertools/dpdk-devbind.py --status
1565 Network devices using DPDK-compatible driver
1566 ============================================
1569 Network devices using kernel driver
1570 ===================================
1571 0000:0a:00.0 '82599ES 10-Gigabit' if=eth2 drv=ixgbe unused=
1574 # Bind the device to igb_uio.
1575 sudo ./usertools/dpdk-devbind.py -b igb_uio 0000:0a:00.0
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=
1584 To attach a port created by virtual device, above steps are not needed.
1586 For example, to attach a port whose pci address is 0000:0a:00.0.
1588 .. code-block:: console
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
1601 For example, to attach a port created by pcap PMD.
1603 .. code-block:: console
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
1612 In this case, identifier is ``net_pcap0``.
1613 This identifier format is the same as ``--vdev`` format of DPDK applications.
1615 For example, to re-attach a bonded port which has been previously detached,
1616 the mode and slave parameters must be given.
1618 .. code-block:: console
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
1631 Detach a specific port::
1633 testpmd> port detach (port_id)
1635 Before detaching a port, the port should be stopped and closed.
1637 For example, to detach a pci device port 0.
1639 .. code-block:: console
1641 testpmd> port stop 0
1644 testpmd> port close 0
1648 testpmd> port detach 0
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
1657 For example, to detach a virtual device port 0.
1659 .. code-block:: console
1661 testpmd> port stop 0
1664 testpmd> port close 0
1668 testpmd> port detach 0
1670 PMD: Closing pcap ethdev on numa socket 0
1671 Port 'net_pcap0' is detached. Now total ports is 0
1674 To remove a pci device completely from the system, first detach the port from testpmd.
1675 Then the device should be moved under kernel management.
1676 Finally the device can be removed using kernel pci hotplug functionality.
1678 For example, to move a pci device under kernel management:
1680 .. code-block:: console
1682 sudo ./usertools/dpdk-devbind.py -b ixgbe 0000:0a:00.0
1684 ./usertools/dpdk-devbind.py --status
1686 Network devices using DPDK-compatible driver
1687 ============================================
1690 Network devices using kernel driver
1691 ===================================
1692 0000:0a:00.0 '82599ES 10-Gigabit' if=eth2 drv=ixgbe unused=igb_uio
1694 To remove a port created by a virtual device, above steps are not needed.
1699 Start all ports or a specific port::
1701 testpmd> port start (port_id|all)
1706 Stop all ports or a specific port::
1708 testpmd> port stop (port_id|all)
1713 Close all ports or a specific port::
1715 testpmd> port close (port_id|all)
1717 port config - queue ring size
1718 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1720 Configure a rx/tx queue ring size::
1722 testpmd> port (port_id) (rxq|txq) (queue_id) ring_size (value)
1724 Only take effect after command that (re-)start the port or command that setup specific queue.
1726 port start/stop queue
1727 ~~~~~~~~~~~~~~~~~~~~~
1729 Start/stop a rx/tx queue on a specific port::
1731 testpmd> port (port_id) (rxq|txq) (queue_id) (start|stop)
1734 ~~~~~~~~~~~~~~~~~~~~~
1736 Setup a rx/tx queue on a specific port::
1738 testpmd> port (port_id) (rxq|txq) (queue_id) setup
1740 Only take effect when port is started.
1745 Set the speed and duplex mode for all ports or a specific port::
1747 testpmd> port config (port_id|all) speed (10|100|1000|10000|25000|40000|50000|100000|auto) \
1748 duplex (half|full|auto)
1750 port config - queues/descriptors
1751 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1753 Set number of queues/descriptors for rxq, txq, rxd and txd::
1755 testpmd> port config all (rxq|txq|rxd|txd) (value)
1757 This is equivalent to the ``--rxq``, ``--txq``, ``--rxd`` and ``--txd`` command-line options.
1759 port config - max-pkt-len
1760 ~~~~~~~~~~~~~~~~~~~~~~~~~
1762 Set the maximum packet length::
1764 testpmd> port config all max-pkt-len (value)
1766 This is equivalent to the ``--max-pkt-len`` command-line option.
1768 port config - CRC Strip
1769 ~~~~~~~~~~~~~~~~~~~~~~~
1771 Set hardware CRC stripping on or off for all ports::
1773 testpmd> port config all crc-strip (on|off)
1775 CRC stripping is on by default.
1777 The ``off`` option is equivalent to the ``--disable-crc-strip`` command-line option.
1779 port config - scatter
1780 ~~~~~~~~~~~~~~~~~~~~~~~
1782 Set RX scatter mode on or off for all ports::
1784 testpmd> port config all scatter (on|off)
1786 RX scatter mode is off by default.
1788 The ``on`` option is equivalent to the ``--enable-scatter`` command-line option.
1790 port config - RX Checksum
1791 ~~~~~~~~~~~~~~~~~~~~~~~~~
1793 Set hardware RX checksum offload to on or off for all ports::
1795 testpmd> port config all rx-cksum (on|off)
1797 Checksum offload is off by default.
1799 The ``on`` option is equivalent to the ``--enable-rx-cksum`` command-line option.
1804 Set hardware VLAN on or off for all ports::
1806 testpmd> port config all hw-vlan (on|off)
1808 Hardware VLAN is off by default.
1810 The ``on`` option is equivalent to the ``--enable-hw-vlan`` command-line option.
1812 port config - VLAN filter
1813 ~~~~~~~~~~~~~~~~~~~~~~~~~
1815 Set hardware VLAN filter on or off for all ports::
1817 testpmd> port config all hw-vlan-filter (on|off)
1819 Hardware VLAN filter is off by default.
1821 The ``on`` option is equivalent to the ``--enable-hw-vlan-filter`` command-line option.
1823 port config - VLAN strip
1824 ~~~~~~~~~~~~~~~~~~~~~~~~
1826 Set hardware VLAN strip on or off for all ports::
1828 testpmd> port config all hw-vlan-strip (on|off)
1830 Hardware VLAN strip is off by default.
1832 The ``on`` option is equivalent to the ``--enable-hw-vlan-strip`` command-line option.
1834 port config - VLAN extend
1835 ~~~~~~~~~~~~~~~~~~~~~~~~~
1837 Set hardware VLAN extend on or off for all ports::
1839 testpmd> port config all hw-vlan-extend (on|off)
1841 Hardware VLAN extend is off by default.
1843 The ``on`` option is equivalent to the ``--enable-hw-vlan-extend`` command-line option.
1845 port config - Drop Packets
1846 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1848 Set packet drop for packets with no descriptors on or off for all ports::
1850 testpmd> port config all drop-en (on|off)
1852 Packet dropping for packets with no descriptors is off by default.
1854 The ``on`` option is equivalent to the ``--enable-drop-en`` command-line option.
1859 Set the RSS (Receive Side Scaling) mode on or off::
1861 testpmd> port config all rss (all|default|ip|tcp|udp|sctp|ether|port|vxlan|geneve|nvgre|none)
1863 RSS is on by default.
1865 The ``all`` option is equivalent to ip|tcp|udp|sctp|ether.
1866 The ``default`` option enables all supported RSS types reported by device info.
1867 The ``none`` option is equivalent to the ``--disable-rss`` command-line option.
1869 port config - RSS Reta
1870 ~~~~~~~~~~~~~~~~~~~~~~
1872 Set the RSS (Receive Side Scaling) redirection table::
1874 testpmd> port config all rss reta (hash,queue)[,(hash,queue)]
1879 Set the DCB mode for an individual port::
1881 testpmd> port config (port_id) dcb vt (on|off) (traffic_class) pfc (on|off)
1883 The traffic class should be 4 or 8.
1888 Set the number of packets per burst::
1890 testpmd> port config all burst (value)
1892 This is equivalent to the ``--burst`` command-line option.
1894 port config - Threshold
1895 ~~~~~~~~~~~~~~~~~~~~~~~
1897 Set thresholds for TX/RX queues::
1899 testpmd> port config all (threshold) (value)
1901 Where the threshold type can be:
1903 * ``txpt:`` Set the prefetch threshold register of the TX rings, 0 <= value <= 255.
1905 * ``txht:`` Set the host threshold register of the TX rings, 0 <= value <= 255.
1907 * ``txwt:`` Set the write-back threshold register of the TX rings, 0 <= value <= 255.
1909 * ``rxpt:`` Set the prefetch threshold register of the RX rings, 0 <= value <= 255.
1911 * ``rxht:`` Set the host threshold register of the RX rings, 0 <= value <= 255.
1913 * ``rxwt:`` Set the write-back threshold register of the RX rings, 0 <= value <= 255.
1915 * ``txfreet:`` Set the transmit free threshold of the TX rings, 0 <= value <= txd.
1917 * ``rxfreet:`` Set the transmit free threshold of the RX rings, 0 <= value <= rxd.
1919 * ``txrst:`` Set the transmit RS bit threshold of TX rings, 0 <= value <= txd.
1921 These threshold options are also available from the command-line.
1926 Set the value of ether-type for E-tag::
1928 testpmd> port config (port_id|all) l2-tunnel E-tag ether-type (value)
1930 Enable/disable the E-tag support::
1932 testpmd> port config (port_id|all) l2-tunnel E-tag (enable|disable)
1934 port config pctype mapping
1935 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1937 Reset pctype mapping table::
1939 testpmd> port config (port_id) pctype mapping reset
1941 Update hardware defined pctype to software defined flow type mapping table::
1943 testpmd> port config (port_id) pctype mapping update (pctype_id_0[,pctype_id_1]*) (flow_type_id)
1947 * ``pctype_id_x``: hardware pctype id as index of bit in bitmask value of the pctype mapping table.
1949 * ``flow_type_id``: software flow type id as the index of the pctype mapping table.
1951 port config input set
1952 ~~~~~~~~~~~~~~~~~~~~~
1954 Config 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)
1959 Clear 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
1965 * ``pctype_id``: hardware packet classification types.
1966 * ``field_idx``: hardware field index.
1968 port config udp_tunnel_port
1969 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
1971 Add/remove UDP tunnel port for VXLAN/GENEVE tunneling protocols::
1972 testpmd> port config (port_id) udp_tunnel_port add|rm vxlan|geneve (udp_port)
1974 Link Bonding Functions
1975 ----------------------
1977 The Link Bonding functions make it possible to dynamically create and
1978 manage link bonding devices from within testpmd interactive prompt.
1980 create bonded device
1981 ~~~~~~~~~~~~~~~~~~~~
1983 Create a new bonding device::
1985 testpmd> create bonded device (mode) (socket)
1987 For example, to create a bonded device in mode 1 on socket 0::
1989 testpmd> create bonded 1 0
1990 created new bonded device (port X)
1995 Adds Ethernet device to a Link Bonding device::
1997 testpmd> add bonding slave (slave id) (port id)
1999 For example, to add Ethernet device (port 6) to a Link Bonding device (port 10)::
2001 testpmd> add bonding slave 6 10
2004 remove bonding slave
2005 ~~~~~~~~~~~~~~~~~~~~
2007 Removes an Ethernet slave device from a Link Bonding device::
2009 testpmd> remove bonding slave (slave id) (port id)
2011 For example, to remove Ethernet slave device (port 6) to a Link Bonding device (port 10)::
2013 testpmd> remove bonding slave 6 10
2018 Set the Link Bonding mode of a Link Bonding device::
2020 testpmd> set bonding mode (value) (port id)
2022 For example, to set the bonding mode of a Link Bonding device (port 10) to broadcast (mode 3)::
2024 testpmd> set bonding mode 3 10
2029 Set an Ethernet slave device as the primary device on a Link Bonding device::
2031 testpmd> set bonding primary (slave id) (port id)
2033 For example, to set the Ethernet slave device (port 6) as the primary port of a Link Bonding device (port 10)::
2035 testpmd> set bonding primary 6 10
2040 Set the MAC address of a Link Bonding device::
2042 testpmd> set bonding mac (port id) (mac)
2044 For example, to set the MAC address of a Link Bonding device (port 10) to 00:00:00:00:00:01::
2046 testpmd> set bonding mac 10 00:00:00:00:00:01
2048 set bonding xmit_balance_policy
2049 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2051 Set the transmission policy for a Link Bonding device when it is in Balance XOR mode::
2053 testpmd> set bonding xmit_balance_policy (port_id) (l2|l23|l34)
2055 For example, set a Link Bonding device (port 10) to use a balance policy of layer 3+4 (IP addresses & UDP ports)::
2057 testpmd> set bonding xmit_balance_policy 10 l34
2060 set bonding mon_period
2061 ~~~~~~~~~~~~~~~~~~~~~~
2063 Set the link status monitoring polling period in milliseconds for a bonding device.
2065 This adds support for PMD slave devices which do not support link status interrupts.
2066 When the mon_period is set to a value greater than 0 then all PMD's which do not support
2067 link status ISR will be queried every polling interval to check if their link status has changed::
2069 testpmd> set bonding mon_period (port_id) (value)
2071 For example, to set the link status monitoring polling period of bonded device (port 5) to 150ms::
2073 testpmd> set bonding mon_period 5 150
2076 set bonding lacp dedicated_queue
2077 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2079 Enable dedicated tx/rx queues on bonding devices slaves to handle LACP control plane traffic
2080 when in mode 4 (link-aggregration-802.3ad)::
2082 testpmd> set bonding lacp dedicated_queues (port_id) (enable|disable)
2085 set bonding agg_mode
2086 ~~~~~~~~~~~~~~~~~~~~
2088 Enable one of the specific aggregators mode when in mode 4 (link-aggregration-802.3ad)::
2090 testpmd> set bonding agg_mode (port_id) (bandwidth|count|stable)
2096 Show the current configuration of a Link Bonding device::
2098 testpmd> show bonding config (port id)
2101 to show the configuration a Link Bonding device (port 9) with 3 slave devices (1, 3, 4)
2102 in balance mode with a transmission policy of layer 2+3::
2104 testpmd> show bonding config 9
2106 Balance Xmit Policy: BALANCE_XMIT_POLICY_LAYER23
2108 Active Slaves (3): [1 3 4]
2115 The Register Functions can be used to read from and write to registers on the network card referenced by a port number.
2116 This is mainly useful for debugging purposes.
2117 Reference should be made to the appropriate datasheet for the network card for details on the register addresses
2118 and fields that can be accessed.
2123 Display the value of a port register::
2125 testpmd> read reg (port_id) (address)
2127 For example, to examine the Flow Director control register (FDIRCTL, 0x0000EE000) on an Intel 82599 10 GbE Controller::
2129 testpmd> read reg 0 0xEE00
2130 port 0 PCI register at offset 0xEE00: 0x4A060029 (1241907241)
2135 Display a port register bit field::
2137 testpmd> read regfield (port_id) (address) (bit_x) (bit_y)
2139 For example, reading the lowest two bits from the register in the example above::
2141 testpmd> read regfield 0 0xEE00 0 1
2142 port 0 PCI register at offset 0xEE00: bits[0, 1]=0x1 (1)
2147 Display a single port register bit::
2149 testpmd> read regbit (port_id) (address) (bit_x)
2151 For example, reading the lowest bit from the register in the example above::
2153 testpmd> read regbit 0 0xEE00 0
2154 port 0 PCI register at offset 0xEE00: bit 0=1
2159 Set the value of a port register::
2161 testpmd> write reg (port_id) (address) (value)
2163 For example, to clear a register::
2165 testpmd> write reg 0 0xEE00 0x0
2166 port 0 PCI register at offset 0xEE00: 0x00000000 (0)
2171 Set bit field of a port register::
2173 testpmd> write regfield (port_id) (address) (bit_x) (bit_y) (value)
2175 For example, writing to the register cleared in the example above::
2177 testpmd> write regfield 0 0xEE00 0 1 2
2178 port 0 PCI register at offset 0xEE00: 0x00000002 (2)
2183 Set single bit value of a port register::
2185 testpmd> write regbit (port_id) (address) (bit_x) (value)
2187 For example, to set the high bit in the register from the example above::
2189 testpmd> write regbit 0 0xEE00 31 1
2190 port 0 PCI register at offset 0xEE00: 0x8000000A (2147483658)
2192 Traffic Metering and Policing
2193 -----------------------------
2195 The following section shows functions for configuring traffic metering and
2196 policing on the ethernet device through the use of generic ethdev API.
2198 show port traffic management capability
2199 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2201 Show traffic metering and policing capability of the port::
2203 testpmd> show port meter cap (port_id)
2205 add port meter profile (srTCM rfc2967)
2206 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2208 Add meter profile (srTCM rfc2697) to the ethernet device::
2210 testpmd> add port meter profile srtcm_rfc2697 (port_id) (profile_id) \
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).
2220 add port meter profile (trTCM rfc2968)
2221 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2223 Add meter profile (srTCM rfc2698) to the ethernet device::
2225 testpmd> add port meter profile trtcm_rfc2698 (port_id) (profile_id) \
2226 (cir) (pir) (cbs) (pbs)
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).
2236 add port meter profile (trTCM rfc4115)
2237 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2239 Add meter profile (trTCM rfc4115) to the ethernet device::
2241 testpmd> add port meter profile trtcm_rfc4115 (port_id) (profile_id) \
2242 (cir) (eir) (cbs) (ebs)
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).
2252 delete port meter profile
2253 ~~~~~~~~~~~~~~~~~~~~~~~~~
2255 Delete meter profile from the ethernet device::
2257 testpmd> del port meter profile (port_id) (profile_id)
2262 Create new meter object for the ethernet device::
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)...\
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
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.
2292 Enable meter for the ethernet device::
2294 testpmd> enable port meter (port_id) (mtr_id)
2299 Disable meter for the ethernet device::
2301 testpmd> disable port meter (port_id) (mtr_id)
2306 Delete meter for the ethernet device::
2308 testpmd> del port meter (port_id) (mtr_id)
2310 Set port meter profile
2311 ~~~~~~~~~~~~~~~~~~~~~~
2313 Set meter profile for the ethernet device::
2315 testpmd> set port meter profile (port_id) (mtr_id) (profile_id)
2317 set port meter dscp table
2318 ~~~~~~~~~~~~~~~~~~~~~~~~~
2320 Set meter dscp table for the ethernet device::
2322 testpmd> set port meter dscp table (port_id) (mtr_id) [(dscp_tbl_entry0) \
2323 (dscp_tbl_entry1)...(dscp_tbl_entry63)]
2325 set port meter policer action
2326 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2328 Set meter policer action for the ethernet device::
2330 testpmd> set port meter policer action (port_id) (mtr_id) (action_mask) \
2331 (action0) [(action1) (action1)]
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
2343 set port meter stats mask
2344 ~~~~~~~~~~~~~~~~~~~~~~~~~
2346 Set meter stats mask for the ethernet device::
2348 testpmd> set port meter stats mask (port_id) (mtr_id) (stats_mask)
2352 * ``stats_mask``: Bit mask indicating statistics counter types to be enabled.
2354 show port meter stats
2355 ~~~~~~~~~~~~~~~~~~~~~
2357 Show meter stats of the ethernet device::
2359 testpmd> show port meter stats (port_id) (mtr_id) (clear)
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.
2369 The following section shows functions for configuring traffic management on
2370 on the ethernet device through the use of generic TM API.
2372 show port traffic management capability
2373 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2375 Show traffic management capability of the port::
2377 testpmd> show port tm cap (port_id)
2379 show port traffic management capability (hierarchy level)
2380 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2382 Show traffic management hierarchy level capability of the port::
2384 testpmd> show port tm level cap (port_id) (level_id)
2386 show port traffic management capability (hierarchy node level)
2387 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2389 Show the traffic management hierarchy node capability of the port::
2391 testpmd> show port tm node cap (port_id) (node_id)
2393 show port traffic management hierarchy node type
2394 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2396 Show the port traffic management hierarchy node type::
2398 testpmd> show port tm node type (port_id) (node_id)
2400 show port traffic management hierarchy node stats
2401 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2403 Show the port traffic management hierarchy node statistics::
2405 testpmd> show port tm node stats (port_id) (node_id) (clear)
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.
2413 Add port traffic management private shaper profile
2414 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2416 Add the port traffic management private shaper profile::
2418 testpmd> add port tm node shaper profile (port_id) (shaper_profile_id) \
2419 (tb_rate) (tb_size) (packet_length_adjust)
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
2431 Delete port traffic management private shaper profile
2432 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2434 Delete the port traffic management private shaper::
2436 testpmd> del port tm node shaper profile (port_id) (shaper_profile_id)
2440 * ``shaper_profile id``: Shaper profile ID that needs to be deleted.
2442 Add port traffic management shared shaper
2443 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2445 Create the port traffic management shared shaper::
2447 testpmd> add port tm node shared shaper (port_id) (shared_shaper_id) \
2452 * ``shared_shaper_id``: Shared shaper ID to be created.
2453 * ``shaper_profile id``: Shaper profile ID for shared shaper.
2455 Set port traffic management shared shaper
2456 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2458 Update the port traffic management shared shaper::
2460 testpmd> set port tm node shared shaper (port_id) (shared_shaper_id) \
2465 * ``shared_shaper_id``: Shared shaper ID to be update.
2466 * ``shaper_profile id``: Shaper profile ID for shared shaper.
2468 Delete port traffic management shared shaper
2469 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2471 Delete the port traffic management shared shaper::
2473 testpmd> del port tm node shared shaper (port_id) (shared_shaper_id)
2477 * ``shared_shaper_id``: Shared shaper ID to be deleted.
2479 Set port traffic management hiearchy node private shaper
2480 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2482 set the port traffic management hierarchy node private shaper::
2484 testpmd> set port tm node shaper profile (port_id) (node_id) \
2489 * ``shaper_profile id``: Private shaper profile ID to be enabled on the
2492 Add port traffic management WRED profile
2493 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2495 Create a new WRED profile::
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)
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)
2521 Delete port traffic management WRED profile
2522 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2524 Delete the WRED profile::
2526 testpmd> del port tm node wred profile (port_id) (wred_profile_id)
2528 Add port traffic management hierarchy nonleaf node
2529 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2531 Add nonleaf node to port traffic management hiearchy::
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) ...] \
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
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.
2554 Add port traffic management hierarchy leaf node
2555 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2557 Add leaf node to port traffic management hiearchy::
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) ...] \
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
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.
2581 Delete port traffic management hierarchy node
2582 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2584 Delete node from port traffic management hiearchy::
2586 testpmd> del port tm node (port_id) (node_id)
2588 Update port traffic management hierarchy parent node
2589 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2591 Update port traffic management hierarchy parent node::
2593 testpmd> set port tm node parent (port_id) (node_id) (parent_node_id) \
2596 This function can only be called after the hierarchy commit invocation. Its
2597 success depends on the port support for this operation, as advertised through
2598 the port capability set. This function is valid for all nodes of the traffic
2599 management hierarchy except root node.
2601 Suspend port traffic management hierarchy node
2602 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2604 testpmd> suspend port tm node (port_id) (node_id)
2606 Resume port traffic management hierarchy node
2607 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2609 testpmd> resume port tm node (port_id) (node_id)
2611 Commit port traffic management hierarchy
2612 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2614 Commit the traffic management hierarchy on the port::
2616 testpmd> port tm hierarchy commit (port_id) (clean_on_fail)
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
2624 Set port traffic management default hierarchy (tm forwarding mode)
2625 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2627 set the traffic management default hierarchy on the port::
2629 testpmd> set port tm hierarchy default (port_id)
2634 This section details the available filter functions that are available.
2636 Note these functions interface the deprecated legacy filtering framework,
2637 superseded by *rte_flow*. See `Flow rules management`_.
2640 ~~~~~~~~~~~~~~~~~~~~
2642 Add or delete a L2 Ethertype filter, which identify packets by their L2 Ethertype mainly assign them to a receive queue::
2644 ethertype_filter (port_id) (add|del) (mac_addr|mac_ignr) (mac_address) \
2645 ethertype (ether_type) (drop|fwd) queue (queue_id)
2647 The available information parameters are:
2649 * ``port_id``: The port which the Ethertype filter assigned on.
2651 * ``mac_addr``: Compare destination mac address.
2653 * ``mac_ignr``: Ignore destination mac address match.
2655 * ``mac_address``: Destination mac address to match.
2657 * ``ether_type``: The EtherType value want to match,
2658 for example 0x0806 for ARP packet. 0x0800 (IPv4) and 0x86DD (IPv6) are invalid.
2660 * ``queue_id``: The receive queue associated with this EtherType filter.
2661 It is meaningless when deleting or dropping.
2663 Example, to add/remove an ethertype filter rule::
2665 testpmd> ethertype_filter 0 add mac_ignr 00:11:22:33:44:55 \
2666 ethertype 0x0806 fwd queue 3
2668 testpmd> ethertype_filter 0 del mac_ignr 00:11:22:33:44:55 \
2669 ethertype 0x0806 fwd queue 3
2674 Add or delete a 2-tuple filter,
2675 which identifies packets by specific protocol and destination TCP/UDP port
2676 and forwards packets into one of the receive queues::
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) \
2683 The available information parameters are:
2685 * ``port_id``: The port which the 2-tuple filter assigned on.
2687 * ``dst_port_value``: Destination port in L4.
2689 * ``protocol_value``: IP L4 protocol.
2691 * ``mask_value``: Participates in the match or not by bit for field above, 1b means participate.
2693 * ``tcp_flags_value``: TCP control bits. The non-zero value is invalid, when the pro_value is not set to 0x06 (TCP).
2695 * ``prio_value``: Priority of this filter.
2697 * ``queue_id``: The receive queue associated with this 2-tuple filter.
2699 Example, to add/remove an 2tuple filter rule::
2701 testpmd> 2tuple_filter 0 add dst_port 32 protocol 0x06 mask 0x03 \
2702 tcp_flags 0x02 priority 3 queue 3
2704 testpmd> 2tuple_filter 0 del dst_port 32 protocol 0x06 mask 0x03 \
2705 tcp_flags 0x02 priority 3 queue 3
2710 Add or delete a 5-tuple filter,
2711 which consists of a 5-tuple (protocol, source and destination IP addresses, source and destination TCP/UDP/SCTP port)
2712 and routes packets into one of the receive queues::
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)
2720 The available information parameters are:
2722 * ``port_id``: The port which the 5-tuple filter assigned on.
2724 * ``dst_address``: Destination IP address.
2726 * ``src_address``: Source IP address.
2728 * ``dst_port_value``: TCP/UDP destination port.
2730 * ``src_port_value``: TCP/UDP source port.
2732 * ``protocol_value``: L4 protocol.
2734 * ``mask_value``: Participates in the match or not by bit for field above, 1b means participate
2736 * ``tcp_flags_value``: TCP control bits. The non-zero value is invalid, when the protocol_value is not set to 0x06 (TCP).
2738 * ``prio_value``: The priority of this filter.
2740 * ``queue_id``: The receive queue associated with this 5-tuple filter.
2742 Example, to add/remove an 5tuple filter rule::
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
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
2755 Using the SYN filter, TCP packets whose *SYN* flag is set can be forwarded to a separate queue::
2757 syn_filter (port_id) (add|del) priority (high|low) queue (queue_id)
2759 The available information parameters are:
2761 * ``port_id``: The port which the SYN filter assigned on.
2763 * ``high``: This SYN filter has higher priority than other filters.
2765 * ``low``: This SYN filter has lower priority than other filters.
2767 * ``queue_id``: The receive queue associated with this SYN filter
2771 testpmd> syn_filter 0 add priority high queue 3
2776 With flex filter, packets can be recognized by any arbitrary pattern within the first 128 bytes of the packet
2777 and routed into one of the receive queues::
2779 flex_filter (port_id) (add|del) len (len_value) bytes (bytes_value) \
2780 mask (mask_value) priority (prio_value) queue (queue_id)
2782 The available information parameters are:
2784 * ``port_id``: The port which the Flex filter is assigned on.
2786 * ``len_value``: Filter length in bytes, no greater than 128.
2788 * ``bytes_value``: A string in hexadecimal, means the value the flex filter needs to match.
2790 * ``mask_value``: A string in hexadecimal, bit 1 means corresponding byte participates in the match.
2792 * ``prio_value``: The priority of this filter.
2794 * ``queue_id``: The receive queue associated with this Flex filter.
2798 testpmd> flex_filter 0 add len 16 bytes 0x00000000000000000000000008060000 \
2799 mask 000C priority 3 queue 3
2801 testpmd> flex_filter 0 del len 16 bytes 0x00000000000000000000000008060000 \
2802 mask 000C priority 3 queue 3
2805 .. _testpmd_flow_director:
2807 flow_director_filter
2808 ~~~~~~~~~~~~~~~~~~~~
2810 The Flow Director works in receive mode to identify specific flows or sets of flows and route them to specific queues.
2812 Four types of filtering are supported which are referred to as Perfect Match, Signature, Perfect-mac-vlan and
2813 Perfect-tunnel filters, the match mode is set by the ``--pkt-filter-mode`` command-line parameter:
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.
2819 * Signature filters.
2820 The hardware checks a match between a hash-based signature of the masked fields of the received packet.
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.
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.
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.
2834 The Flow Director filters can match the different fields for different type of packet: flow type, specific input set
2835 per flow type and the flexible payload.
2837 The Flow Director can also mask out parts of all of these fields so that filters
2838 are only applied to certain fields or parts of the fields.
2840 Note that for raw flow type mode the source and destination fields in the
2841 raw packet buffer need to be presented in a reversed order with respect
2842 to the expected received packets.
2843 For example: IP source and destination addresses or TCP/UDP/SCTP
2844 source and destination ports
2846 Different NICs may have different capabilities, command show port fdir (port_id) can be used to acquire the information.
2848 # Commands to add flow director filters of different flow types::
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) \
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) \
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)
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)
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)
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)
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)
2896 For example, to add an ipv4-udp flow type filter::
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
2902 For example, add an ipv4-other flow type filter::
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
2911 Flush all flow director filters on a device::
2913 testpmd> flush_flow_director (port_id)
2915 Example, to flush all flow director filter on port 0::
2917 testpmd> flush_flow_director 0
2922 Set flow director's input masks::
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)
2928 flow_director_mask (port_id) mode MAC-VLAN vlan (vlan_value)
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)
2934 Example, to set flow director mask on port 0::
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
2942 flow_director_flex_mask
2943 ~~~~~~~~~~~~~~~~~~~~~~~
2945 set masks of flow director's flexible payload based on certain flow type::
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)
2952 Example, to set flow director's flex mask for all flow type on port 0::
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)
2958 flow_director_flex_payload
2959 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2961 Configure flexible payload selection::
2963 flow_director_flex_payload (port_id) (raw|l2|l3|l4) (config)
2965 For example, to select the first 16 bytes from the offset 4 (bytes) of packet's payload as flexible payload::
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)
2970 get_sym_hash_ena_per_port
2971 ~~~~~~~~~~~~~~~~~~~~~~~~~
2973 Get symmetric hash enable configuration per port::
2975 get_sym_hash_ena_per_port (port_id)
2977 For example, to get symmetric hash enable configuration of port 1::
2979 testpmd> get_sym_hash_ena_per_port 1
2981 set_sym_hash_ena_per_port
2982 ~~~~~~~~~~~~~~~~~~~~~~~~~
2984 Set symmetric hash enable configuration per port to enable or disable::
2986 set_sym_hash_ena_per_port (port_id) (enable|disable)
2988 For example, to set symmetric hash enable configuration of port 1 to enable::
2990 testpmd> set_sym_hash_ena_per_port 1 enable
2992 get_hash_global_config
2993 ~~~~~~~~~~~~~~~~~~~~~~
2995 Get the global configurations of hash filters::
2997 get_hash_global_config (port_id)
2999 For example, to get the global configurations of hash filters of port 1::
3001 testpmd> get_hash_global_config 1
3003 set_hash_global_config
3004 ~~~~~~~~~~~~~~~~~~~~~~
3006 Set the global configurations of hash filters::
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>) \
3013 For example, to enable simple_xor for flow type of ipv6 on port 2::
3015 testpmd> set_hash_global_config 2 simple_xor ipv6 enable
3020 Set the input set for hash::
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)
3030 For example, to add source IP to hash input set for flow type of ipv4-udp on port 0::
3032 testpmd> set_hash_input_set 0 ipv4-udp src-ipv4 add
3037 The Flow Director filters can match the different fields for different type of packet, i.e. specific input set
3038 on per flow type and the flexible payload. This command can be used to change input set for each flow type.
3040 Set the input set for flow director::
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)
3049 For example to add source IP to FD input set for flow type of ipv4-udp on port 0::
3051 testpmd> set_fdir_input_set 0 ipv4-udp src-ipv4 add
3056 Set different GRE key length for input set::
3058 global_config (port_id) gre-key-len (number in bytes)
3060 For example to set GRE key length for input set to 4 bytes on port 0::
3062 testpmd> global_config 0 gre-key-len 4
3065 .. _testpmd_rte_flow:
3067 Flow rules management
3068 ---------------------
3070 Control of the generic flow API (*rte_flow*) is fully exposed through the
3071 ``flow`` command (validation, creation, destruction, queries and operation
3074 Considering *rte_flow* overlaps with all `Filter Functions`_, using both
3075 features simultaneously may cause undefined side-effects and is therefore
3081 Because the ``flow`` command uses dynamic tokens to handle the large number
3082 of possible flow rules combinations, its behavior differs slightly from
3083 other commands, in particular:
3085 - Pressing *?* or the *<tab>* key displays contextual help for the current
3086 token, not that of the entire command.
3088 - Optional and repeated parameters are supported (provided they are listed
3089 in the contextual help).
3091 The first parameter stands for the operation mode. Possible operations and
3092 their general syntax are described below. They are covered in detail in the
3095 - Check whether a flow rule can be created::
3097 flow validate {port_id}
3098 [group {group_id}] [priority {level}] [ingress] [egress] [transfer]
3099 pattern {item} [/ {item} [...]] / end
3100 actions {action} [/ {action} [...]] / end
3102 - Create a flow rule::
3104 flow create {port_id}
3105 [group {group_id}] [priority {level}] [ingress] [egress] [transfer]
3106 pattern {item} [/ {item} [...]] / end
3107 actions {action} [/ {action} [...]] / end
3109 - Destroy specific flow rules::
3111 flow destroy {port_id} rule {rule_id} [...]
3113 - Destroy all flow rules::
3115 flow flush {port_id}
3117 - Query an existing flow rule::
3119 flow query {port_id} {rule_id} {action}
3121 - List existing flow rules sorted by priority, filtered by group
3124 flow list {port_id} [group {group_id}] [...]
3126 - Restrict ingress traffic to the defined flow rules::
3128 flow isolate {port_id} {boolean}
3130 Validating flow rules
3131 ~~~~~~~~~~~~~~~~~~~~~
3133 ``flow validate`` reports whether a flow rule would be accepted by the
3134 underlying device in its current state but stops short of creating it. It is
3135 bound to ``rte_flow_validate()``::
3137 flow validate {port_id}
3138 [group {group_id}] [priority {level}] [ingress] [egress] [transfer]
3139 pattern {item} [/ {item} [...]] / end
3140 actions {action} [/ {action} [...]] / end
3142 If successful, it will show::
3146 Otherwise it will show an error message of the form::
3148 Caught error type [...] ([...]): [...]
3150 This command uses the same parameters as ``flow create``, their format is
3151 described in `Creating flow rules`_.
3153 Check whether redirecting any Ethernet packet received on port 0 to RX queue
3154 index 6 is supported::
3156 testpmd> flow validate 0 ingress pattern eth / end
3157 actions queue index 6 / end
3161 Port 0 does not support TCPv6 rules::
3163 testpmd> flow validate 0 ingress pattern eth / ipv6 / tcp / end
3165 Caught error type 9 (specific pattern item): Invalid argument
3171 ``flow create`` validates and creates the specified flow rule. It is bound
3172 to ``rte_flow_create()``::
3174 flow create {port_id}
3175 [group {group_id}] [priority {level}] [ingress] [egress] [transfer]
3176 pattern {item} [/ {item} [...]] / end
3177 actions {action} [/ {action} [...]] / end
3179 If successful, it will return a flow rule ID usable with other commands::
3181 Flow rule #[...] created
3183 Otherwise it will show an error message of the form::
3185 Caught error type [...] ([...]): [...]
3187 Parameters describe in the following order:
3189 - Attributes (*group*, *priority*, *ingress*, *egress*, *transfer* tokens).
3190 - A matching pattern, starting with the *pattern* token and terminated by an
3192 - Actions, starting with the *actions* token and terminated by an *end*
3195 These translate directly to *rte_flow* objects provided as-is to the
3196 underlying functions.
3198 The shortest valid definition only comprises mandatory tokens::
3200 testpmd> flow create 0 pattern end actions end
3202 Note that PMDs may refuse rules that essentially do nothing such as this
3205 **All unspecified object values are automatically initialized to 0.**
3210 These tokens affect flow rule attributes (``struct rte_flow_attr``) and are
3211 specified before the ``pattern`` token.
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.
3219 Each instance of an attribute specified several times overrides the previous
3220 value as shown below (group 4 is used)::
3222 testpmd> flow create 0 group 42 group 24 group 4 [...]
3224 Note that once enabled, ``ingress`` and ``egress`` cannot be disabled.
3226 While not specifying a direction is an error, some rules may allow both
3229 Most rules affect RX therefore contain the ``ingress`` token::
3231 testpmd> flow create 0 ingress pattern [...]
3236 A matching pattern starts after the ``pattern`` token. It is made of pattern
3237 items and is terminated by a mandatory ``end`` item.
3239 Items are named after their type (*RTE_FLOW_ITEM_TYPE_* from ``enum
3240 rte_flow_item_type``).
3242 The ``/`` token is used as a separator between pattern items as shown
3245 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end [...]
3247 Note that protocol items like these must be stacked from lowest to highest
3248 layer to make sense. For instance, the following rule is either invalid or
3249 unlikely to match any packet::
3251 testpmd> flow create 0 ingress pattern eth / udp / ipv4 / end [...]
3253 More information on these restrictions can be found in the *rte_flow*
3256 Several items support additional specification structures, for example
3257 ``ipv4`` allows specifying source and destination addresses as follows::
3259 testpmd> flow create 0 ingress pattern eth / ipv4 src is 10.1.1.1
3260 dst is 10.2.0.0 / end [...]
3262 This rule matches all IPv4 traffic with the specified properties.
3264 In this example, ``src`` and ``dst`` are field names of the underlying
3265 ``struct rte_flow_item_ipv4`` object. All item properties can be specified
3266 in a similar fashion.
3268 The ``is`` token means that the subsequent value must be matched exactly,
3269 and assigns ``spec`` and ``mask`` fields in ``struct rte_flow_item``
3270 accordingly. Possible assignment tokens are:
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.
3278 These yield identical results::
3280 ipv4 src is 10.1.1.1
3284 ipv4 src spec 10.1.1.1 src mask 255.255.255.255
3288 ipv4 src spec 10.1.1.1 src prefix 32
3292 ipv4 src is 10.1.1.1 src last 10.1.1.1 # range with a single value
3296 ipv4 src is 10.1.1.1 src last 0 # 0 disables range
3298 Inclusive ranges can be defined with ``last``::
3300 ipv4 src is 10.1.1.1 src last 10.2.3.4 # 10.1.1.1 to 10.2.3.4
3302 Note that ``mask`` affects both ``spec`` and ``last``::
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
3307 Properties can be modified multiple times::
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
3313 ipv4 src is 10.1.1.1 src prefix 24 src prefix 16 # matches 10.1.0.0/16
3318 This section lists supported pattern items and their attributes, if any.
3320 - ``end``: end list of pattern items.
3322 - ``void``: no-op pattern item.
3324 - ``invert``: perform actions when pattern does not match.
3326 - ``any``: match any protocol for the current layer.
3328 - ``num {unsigned}``: number of layers covered.
3330 - ``pf``: match traffic from/to the physical function.
3332 - ``vf``: match traffic from/to a virtual function ID.
3334 - ``id {unsigned}``: VF ID.
3336 - ``phy_port``: match traffic from/to a specific physical port.
3338 - ``index {unsigned}``: physical port index.
3340 - ``port_id``: match traffic from/to a given DPDK port ID.
3342 - ``id {unsigned}``: DPDK port ID.
3344 - ``mark``: match value set in previously matched flow rule using the mark action.
3346 - ``id {unsigned}``: arbitrary integer value.
3348 - ``raw``: match an arbitrary byte string.
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.
3356 - ``eth``: match Ethernet header.
3358 - ``dst {MAC-48}``: destination MAC.
3359 - ``src {MAC-48}``: source MAC.
3360 - ``type {unsigned}``: EtherType or TPID.
3362 - ``vlan``: match 802.1Q/ad VLAN tag.
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.
3370 - ``ipv4``: match IPv4 header.
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.
3378 - ``ipv6``: match IPv6 header.
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.
3387 - ``icmp``: match ICMP header.
3389 - ``type {unsigned}``: ICMP packet type.
3390 - ``code {unsigned}``: ICMP packet code.
3392 - ``udp``: match UDP header.
3394 - ``src {unsigned}``: UDP source port.
3395 - ``dst {unsigned}``: UDP destination port.
3397 - ``tcp``: match TCP header.
3399 - ``src {unsigned}``: TCP source port.
3400 - ``dst {unsigned}``: TCP destination port.
3402 - ``sctp``: match SCTP header.
3404 - ``src {unsigned}``: SCTP source port.
3405 - ``dst {unsigned}``: SCTP destination port.
3406 - ``tag {unsigned}``: validation tag.
3407 - ``cksum {unsigned}``: checksum.
3409 - ``vxlan``: match VXLAN header.
3411 - ``vni {unsigned}``: VXLAN identifier.
3413 - ``e_tag``: match IEEE 802.1BR E-Tag header.
3415 - ``grp_ecid_b {unsigned}``: GRP and E-CID base.
3417 - ``nvgre``: match NVGRE header.
3419 - ``tni {unsigned}``: virtual subnet ID.
3421 - ``mpls``: match MPLS header.
3423 - ``label {unsigned}``: MPLS label.
3425 - ``gre``: match GRE header.
3427 - ``protocol {unsigned}``: protocol type.
3429 - ``fuzzy``: fuzzy pattern match, expect faster than default.
3431 - ``thresh {unsigned}``: accuracy threshold.
3433 - ``gtp``, ``gtpc``, ``gtpu``: match GTPv1 header.
3435 - ``teid {unsigned}``: tunnel endpoint identifier.
3437 - ``geneve``: match GENEVE header.
3439 - ``vni {unsigned}``: virtual network identifier.
3440 - ``protocol {unsigned}``: protocol type.
3442 - ``vxlan-gpe``: match VXLAN-GPE header.
3444 - ``vni {unsigned}``: VXLAN-GPE identifier.
3446 - ``arp_eth_ipv4``: match ARP header for Ethernet/IPv4.
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.
3453 - ``ipv6_ext``: match presence of any IPv6 extension header.
3455 - ``next_hdr {unsigned}``: next header.
3457 - ``icmp6``: match any ICMPv6 header.
3459 - ``type {unsigned}``: ICMPv6 type.
3460 - ``code {unsigned}``: ICMPv6 code.
3462 - ``icmp6_nd_ns``: match ICMPv6 neighbor discovery solicitation.
3464 - ``target_addr {ipv6 address}``: target address.
3466 - ``icmp6_nd_na``: match ICMPv6 neighbor discovery advertisement.
3468 - ``target_addr {ipv6 address}``: target address.
3470 - ``icmp6_nd_opt``: match presence of any ICMPv6 neighbor discovery option.
3472 - ``type {unsigned}``: ND option type.
3474 - ``icmp6_nd_opt_sla_eth``: match ICMPv6 neighbor discovery source Ethernet
3475 link-layer address option.
3477 - ``sla {MAC-48}``: source Ethernet LLA.
3479 - ``icmp6_nd_opt_sla_eth``: match ICMPv6 neighbor discovery target Ethernet
3480 link-layer address option.
3482 - ``tla {MAC-48}``: target Ethernet LLA.
3487 A 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
3489 terminated by a mandatory ``end`` action.
3491 Actions are named after their type (*RTE_FLOW_ACTION_TYPE_* from ``enum
3492 rte_flow_action_type``).
3494 Dropping all incoming UDPv4 packets can be expressed as follows::
3496 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end
3499 Several actions have configurable properties which must be specified when
3500 there is no valid default value. For example, ``queue`` requires a target
3503 This rule redirects incoming UDPv4 traffic to queue index 6::
3505 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end
3506 actions queue index 6 / end
3508 While this one could be rejected by PMDs (unspecified queue index)::
3510 testpmd> flow create 0 ingress pattern eth / ipv4 / udp / end
3513 As defined by *rte_flow*, the list is not ordered, all actions of a given
3514 rule are performed simultaneously. These are equivalent::
3516 queue index 6 / void / mark id 42 / end
3520 void / mark id 42 / queue index 6 / end
3522 All actions in a list should have different types, otherwise only the last
3523 action of a given type is taken into account::
3525 queue index 4 / queue index 5 / queue index 6 / end # will use queue 6
3529 drop / drop / drop / end # drop is performed only once
3533 mark id 42 / queue index 3 / mark id 24 / end # mark will be 24
3535 Considering they are performed simultaneously, opposite and overlapping
3536 actions can sometimes be combined when the end result is unambiguous::
3538 drop / queue index 6 / end # drop has no effect
3542 queue index 6 / rss queues 6 7 8 / end # queue has no effect
3546 drop / passthru / end # drop has no effect
3548 Note that PMDs may still refuse such combinations.
3553 This section lists supported actions and their attributes, if any.
3555 - ``end``: end list of actions.
3557 - ``void``: no-op action.
3559 - ``passthru``: let subsequent rule process matched packets.
3561 - ``jump``: redirect traffic to group on device.
3563 - ``group {unsigned}``: group to redirect to.
3565 - ``mark``: attach 32 bit value to packets.
3567 - ``id {unsigned}``: 32 bit value to return with packets.
3569 - ``flag``: flag packets.
3571 - ``queue``: assign packets to a given queue index.
3573 - ``index {unsigned}``: queue index to use.
3575 - ``drop``: drop packets (note: passthru has priority).
3577 - ``count``: enable counters for this rule.
3579 - ``rss``: spread packets among several queues.
3581 - ``func {hash function}``: RSS hash function to apply, allowed tokens are
3582 the same as `set_hash_global_config`_.
3584 - ``level {unsigned}``: encapsulation level for ``types``.
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"
3591 - ``key {string}``: RSS hash key, overrides ``key_len``.
3593 - ``key_len {unsigned}``: RSS hash key length in bytes, can be used in
3594 conjunction with ``key`` to pad or truncate it.
3596 - ``queues [{unsigned} [...]] end``: queue indices to use.
3598 - ``pf``: direct traffic to physical function.
3600 - ``vf``: direct traffic to a virtual function ID.
3602 - ``original {boolean}``: use original VF ID if possible.
3603 - ``id {unsigned}``: VF ID.
3605 - ``phy_port``: direct packets to physical port index.
3607 - ``original {boolean}``: use original port index if possible.
3608 - ``index {unsigned}``: physical port index.
3610 - ``port_id``: direct matching traffic to a given DPDK port ID.
3612 - ``original {boolean}``: use original DPDK port ID if possible.
3613 - ``id {unsigned}``: DPDK port ID.
3615 - ``of_set_mpls_ttl``: OpenFlow's ``OFPAT_SET_MPLS_TTL``.
3617 - ``mpls_ttl``: MPLS TTL.
3619 - ``of_dec_mpls_ttl``: OpenFlow's ``OFPAT_DEC_MPLS_TTL``.
3621 - ``of_set_nw_ttl``: OpenFlow's ``OFPAT_SET_NW_TTL``.
3623 - ``nw_ttl``: IP TTL.
3625 - ``of_dec_nw_ttl``: OpenFlow's ``OFPAT_DEC_NW_TTL``.
3627 - ``of_copy_ttl_out``: OpenFlow's ``OFPAT_COPY_TTL_OUT``.
3629 - ``of_copy_ttl_in``: OpenFlow's ``OFPAT_COPY_TTL_IN``.
3631 - ``of_pop_vlan``: OpenFlow's ``OFPAT_POP_VLAN``.
3633 - ``of_push_vlan``: OpenFlow's ``OFPAT_PUSH_VLAN``.
3635 - ``ethertype``: Ethertype.
3637 - ``of_set_vlan_vid``: OpenFlow's ``OFPAT_SET_VLAN_VID``.
3639 - ``vlan_vid``: VLAN id.
3641 - ``of_set_vlan_pcp``: OpenFlow's ``OFPAT_SET_VLAN_PCP``.
3643 - ``vlan_pcp``: VLAN priority.
3645 - ``of_pop_mpls``: OpenFlow's ``OFPAT_POP_MPLS``.
3647 - ``ethertype``: Ethertype.
3649 - ``of_push_mpls``: OpenFlow's ``OFPAT_PUSH_MPLS``.
3651 - ``ethertype``: Ethertype.
3653 Destroying flow rules
3654 ~~~~~~~~~~~~~~~~~~~~~
3656 ``flow destroy`` destroys one or more rules from their rule ID (as returned
3657 by ``flow create``), this command calls ``rte_flow_destroy()`` as many
3658 times as necessary::
3660 flow destroy {port_id} rule {rule_id} [...]
3662 If successful, it will show::
3664 Flow rule #[...] destroyed
3666 It does not report anything for rule IDs that do not exist. The usual error
3667 message is shown when a rule cannot be destroyed::
3669 Caught error type [...] ([...]): [...]
3671 ``flow flush`` destroys all rules on a device and does not take extra
3672 arguments. It is bound to ``rte_flow_flush()``::
3674 flow flush {port_id}
3676 Any errors are reported as above.
3678 Creating several rules and destroying them::
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
3691 The same result can be achieved using ``flow flush``::
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
3702 Non-existent rule IDs are ignored::
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
3712 testpmd> flow destroy 0 rule 0
3713 Flow rule #0 destroyed
3719 ``flow query`` queries a specific action of a flow rule having that
3720 ability. Such actions collect information that can be reported using this
3721 command. It is bound to ``rte_flow_query()``::
3723 flow query {port_id} {rule_id} {action}
3725 If successful, it will display either the retrieved data for known actions
3726 or the following message::
3728 Cannot display result for action type [...] ([...])
3730 Otherwise, it will complain either that the rule does not exist or that some
3733 Flow rule #[...] not found
3737 Caught error type [...] ([...]): [...]
3739 Currently only the ``count`` action is supported. This action reports the
3740 number of packets that hit the flow rule and the total number of bytes. Its
3741 output has the following format::
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
3749 Querying counters for TCPv6 packets redirected to queue 6::
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
3765 ``flow list`` lists existing flow rules sorted by priority and optionally
3766 filtered by group identifiers::
3768 flow list {port_id} [group {group_id}] [...]
3770 This command only fails with the following message if the device does not
3775 Output consists of a header line followed by a short description of each
3776 flow rule, one per line. There is no output at all when no flow rules are
3777 configured on the device::
3779 ID Group Prio Attr Rule
3780 [...] [...] [...] [...] [...]
3782 ``Attr`` column flags:
3784 - ``i`` for ``ingress``.
3785 - ``e`` for ``egress``.
3787 Creating several flow rules and listing them::
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
3805 Rules are sorted by priority (i.e. group ID first, then priority level)::
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
3819 Output can be limited to specific groups::
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
3829 Toggling isolated mode
3830 ~~~~~~~~~~~~~~~~~~~~~~
3832 ``flow isolate`` can be used to tell the underlying PMD that ingress traffic
3833 must only be injected from the defined flow rules; that no default traffic
3834 is expected outside those rules and the driver is free to assign more
3835 resources to handle them. It is bound to ``rte_flow_isolate()``::
3837 flow isolate {port_id} {boolean}
3839 If successful, enabling or disabling isolated mode shows either::
3841 Ingress traffic on port [...]
3842 is now restricted to the defined flow rules
3846 Ingress traffic on port [...]
3847 is not restricted anymore to the defined flow rules
3849 Otherwise, in case of error::
3851 Caught error type [...] ([...]): [...]
3853 Mainly due to its side effects, PMDs supporting this mode may not have the
3854 ability to toggle it more than once without reinitializing affected ports
3855 first (e.g. by exiting testpmd).
3857 Enabling isolated mode::
3859 testpmd> flow isolate 0 true
3860 Ingress traffic on port 0 is now restricted to the defined flow rules
3863 Disabling isolated mode::
3865 testpmd> flow isolate 0 false
3866 Ingress traffic on port 0 is not restricted anymore to the defined flow rules
3869 Sample QinQ flow rules
3870 ~~~~~~~~~~~~~~~~~~~~~~
3872 Before creating QinQ rule(s) the following commands should be issued to enable QinQ::
3874 testpmd> port stop 0
3875 testpmd> vlan set qinq on 0
3877 The above command sets the inner and outer TPID's to 0x8100.
3879 To change the TPID's the following commands should be used::
3881 testpmd> vlan set outer tpid 0xa100 0
3882 testpmd> vlan set inner tpid 0x9100 0
3883 testpmd> port start 0
3885 Validate and create a QinQ rule on port 0 to steer traffic to a VF queue in a VM.
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
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
3897 testpmd> flow list 0
3898 ID Group Prio Attr Rule
3899 0 0 0 i- ETH VLAN VLAN=>VF QUEUE
3901 Validate and create a QinQ rule on port 0 to steer traffic to a queue on the host.
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
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
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
3921 The following sections show functions to load/unload eBPF based filters.
3926 Load an eBPF program as a callback for partciular RX/TX queue::
3928 testpmd> bpf-load rx|tx (portid) (queueid) (load-flags) (bpf-prog-filename)
3930 The available load-flags are:
3932 * ``J``: use JIT generated native code, otherwise BPF interpreter will be used.
3934 * ``M``: assume input parameter is a pointer to rte_mbuf, otherwise assume it is a pointer to first segment's data.
3940 You'll need clang v3.7 or above to build bpf program you'd like to load
3944 .. code-block:: console
3947 clang -O2 -target bpf -c t1.c
3949 Then to load (and JIT compile) t1.o at RX queue 0, port 1::
3951 .. code-block:: console
3953 testpmd> bpf-load rx 1 0 J ./dpdk.org/test/bpf/t1.o
3955 To load (not JITed) t1.o at TX queue 0, port 0::
3957 .. code-block:: console
3959 testpmd> bpf-load tx 0 0 - ./dpdk.org/test/bpf/t1.o
3964 Unload previously loaded eBPF program for partciular RX/TX queue::
3966 testpmd> bpf-unload rx|tx (portid) (queueid)
3968 For example to unload BPF filter from TX queue 0, port 0:
3970 .. code-block:: console
3972 testpmd> bpf-load tx 0 0 - ./dpdk.org/test/bpf/t1.o