1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2016 6WIND S.A.
3 Copyright 2016 Mellanox Technologies, Ltd
5 Generic flow API (rte_flow)
6 ===========================
11 This API provides a generic means to configure hardware to match specific
12 ingress or egress traffic, alter its fate and query related counters
13 according to any number of user-defined rules.
15 It is named *rte_flow* after the prefix used for all its symbols, and is
16 defined in ``rte_flow.h``.
18 - Matching can be performed on packet data (protocol headers, payload) and
19 properties (e.g. associated physical port, virtual device function ID).
21 - Possible operations include dropping traffic, diverting it to specific
22 queues, to virtual/physical device functions or ports, performing tunnel
23 offloads, adding marks and so on.
31 A flow rule is the combination of attributes with a matching pattern and a
32 list of actions. Flow rules form the basis of this API.
34 Flow rules can have several distinct actions (such as counting,
35 encapsulating, decapsulating before redirecting packets to a particular
36 queue, etc.), instead of relying on several rules to achieve this and having
37 applications deal with hardware implementation details regarding their
40 Support for different priority levels on a rule basis is provided, for
41 example in order to force a more specific rule to come before a more generic
42 one for packets matched by both. However hardware support for more than a
43 single priority level cannot be guaranteed. When supported, the number of
44 available priority levels is usually low, which is why they can also be
45 implemented in software by PMDs (e.g. missing priority levels may be
46 emulated by reordering rules).
48 In order to remain as hardware-agnostic as possible, by default all rules
49 are considered to have the same priority, which means that the order between
50 overlapping rules (when a packet is matched by several filters) is
53 PMDs may refuse to create overlapping rules at a given priority level when
54 they can be detected (e.g. if a pattern matches an existing filter).
56 Thus predictable results for a given priority level can only be achieved
57 with non-overlapping rules, using perfect matching on all protocol layers.
59 Flow rules can also be grouped, the flow rule priority is specific to the
60 group they belong to. All flow rules in a given group are thus processed within
61 the context of that group. Groups are not linked by default, so the logical
62 hierarchy of groups must be explicitly defined by flow rules themselves in each
63 group using the JUMP action to define the next group to redirect too. Only flow
64 rules defined in the default group 0 are guarantee to be matched against, this
65 makes group 0 the origin of any group hierarchy defined by an application.
67 Support for multiple actions per rule may be implemented internally on top
68 of non-default hardware priorities, as a result both features may not be
69 simultaneously available to applications.
71 Considering that allowed pattern/actions combinations cannot be known in
72 advance and would result in an impractically large number of capabilities to
73 expose, a method is provided to validate a given rule from the current
74 device configuration state.
76 This enables applications to check if the rule types they need is supported
77 at initialization time, before starting their data path. This method can be
78 used anytime, its only requirement being that the resources needed by a rule
79 should exist (e.g. a target RX queue should be configured first).
81 Each defined rule is associated with an opaque handle managed by the PMD,
82 applications are responsible for keeping it. These can be used for queries
83 and rules management, such as retrieving counters or other data and
86 To avoid resource leaks on the PMD side, handles must be explicitly
87 destroyed by the application before releasing associated resources such as
90 The following sections cover:
92 - **Attributes** (represented by ``struct rte_flow_attr``): properties of a
93 flow rule such as its direction (ingress or egress) and priority.
95 - **Pattern item** (represented by ``struct rte_flow_item``): part of a
96 matching pattern that either matches specific packet data or traffic
97 properties. It can also describe properties of the pattern itself, such as
100 - **Matching pattern**: traffic properties to look for, a combination of any
103 - **Actions** (represented by ``struct rte_flow_action``): operations to
104 perform whenever a packet is matched by a pattern.
112 Flow rules can be grouped by assigning them a common group number. Groups
113 allow a logical hierarchy of flow rule groups (tables) to be defined. These
114 groups can be supported virtually in the PMD or in the physical device.
115 Group 0 is the default group and this is the only group which flows are
116 guarantee to matched against, all subsequent groups can only be reached by
117 way of the JUMP action from a matched flow rule.
119 Although optional, applications are encouraged to group similar rules as
120 much as possible to fully take advantage of hardware capabilities
121 (e.g. optimized matching) and work around limitations (e.g. a single pattern
122 type possibly allowed in a given group), while being aware that the groups
123 hierarchies must be programmed explicitly.
125 Note that support for more than a single group is not guaranteed.
130 A priority level can be assigned to a flow rule, lower values
131 denote higher priority, with 0 as the maximum.
133 Priority levels are arbitrary and up to the application, they do
134 not need to be contiguous nor start from 0, however the maximum number
135 varies between devices and may be affected by existing flow rules.
137 A flow which matches multiple rules in the same group will always matched by
138 the rule with the highest priority in that group.
140 If a packet is matched by several rules of a given group for a given
141 priority level, the outcome is undefined. It can take any path, may be
142 duplicated or even cause unrecoverable errors.
144 Note that support for more than a single priority level is not guaranteed.
146 Attribute: Traffic direction
147 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
149 Flow rule patterns apply to inbound and/or outbound traffic.
151 In the context of this API, **ingress** and **egress** respectively stand
152 for **inbound** and **outbound** based on the standpoint of the application
153 creating a flow rule.
155 There are no exceptions to this definition.
157 Several pattern items and actions are valid and can be used in both
158 directions. At least one direction must be specified.
160 Specifying both directions at once for a given rule is not recommended but
161 may be valid in a few cases.
166 Instead of simply matching the properties of traffic as it would appear on a
167 given DPDK port ID, enabling this attribute transfers a flow rule to the
168 lowest possible level of any device endpoints found in the pattern.
170 When supported, this effectively enables an application to reroute traffic
171 not necessarily intended for it (e.g. coming from or addressed to different
172 physical ports, VFs or applications) at the device level.
174 It complements the behavior of some pattern items such as `Item: PHY_PORT`_
175 and is meaningless without them.
177 When transferring flow rules, **ingress** and **egress** attributes
178 (`Attribute: Traffic direction`_) keep their original meaning, as if
179 processing traffic emitted or received by the application.
184 Pattern items fall in two categories:
186 - Matching protocol headers and packet data, usually associated with a
187 specification structure. These must be stacked in the same order as the
188 protocol layers to match inside packets, starting from the lowest.
190 - Matching meta-data or affecting pattern processing, often without a
191 specification structure. Since they do not match packet contents, their
192 position in the list is usually not relevant.
194 Item specification structures are used to match specific values among
195 protocol fields (or item properties). Documentation describes for each item
196 whether they are associated with one and their type name if so.
198 Up to three structures of the same type can be set for a given item:
200 - ``spec``: values to match (e.g. a given IPv4 address).
202 - ``last``: upper bound for an inclusive range with corresponding fields in
205 - ``mask``: bit-mask applied to both ``spec`` and ``last`` whose purpose is
206 to distinguish the values to take into account and/or partially mask them
207 out (e.g. in order to match an IPv4 address prefix).
209 Usage restrictions and expected behavior:
211 - Setting either ``mask`` or ``last`` without ``spec`` is an error.
213 - Field values in ``last`` which are either 0 or equal to the corresponding
214 values in ``spec`` are ignored; they do not generate a range. Nonzero
215 values lower than those in ``spec`` are not supported.
217 - Setting ``spec`` and optionally ``last`` without ``mask`` causes the PMD
218 to use the default mask defined for that item (defined as
219 ``rte_flow_item_{name}_mask`` constants).
221 - Not setting any of them (assuming item type allows it) is equivalent to
222 providing an empty (zeroed) ``mask`` for broad (nonspecific) matching.
224 - ``mask`` is a simple bit-mask applied before interpreting the contents of
225 ``spec`` and ``last``, which may yield unexpected results if not used
226 carefully. For example, if for an IPv4 address field, ``spec`` provides
227 *10.1.2.3*, ``last`` provides *10.3.4.5* and ``mask`` provides
228 *255.255.0.0*, the effective range becomes *10.1.0.0* to *10.3.255.255*.
230 Example of an item specification matching an Ethernet header:
232 .. _table_rte_flow_pattern_item_example:
234 .. table:: Ethernet item
236 +----------+----------+-----------------------+
237 | Field | Subfield | Value |
238 +==========+==========+=======================+
239 | ``spec`` | ``src`` | ``00:00:01:02:03:04`` |
240 | +----------+-----------------------+
241 | | ``dst`` | ``00:00:2a:66:00:01`` |
242 | +----------+-----------------------+
243 | | ``type`` | ``0x22aa`` |
244 +----------+----------+-----------------------+
245 | ``last`` | unspecified |
246 +----------+----------+-----------------------+
247 | ``mask`` | ``src`` | ``00:00:ff:ff:ff:00`` |
248 | +----------+-----------------------+
249 | | ``dst`` | ``00:00:00:00:00:ff`` |
250 | +----------+-----------------------+
251 | | ``type`` | ``0x0000`` |
252 +----------+----------+-----------------------+
254 Non-masked bits stand for any value (shown as ``?`` below), Ethernet headers
255 with the following properties are thus matched:
257 - ``src``: ``??:??:01:02:03:??``
258 - ``dst``: ``??:??:??:??:??:01``
259 - ``type``: ``0x????``
264 A pattern is formed by stacking items starting from the lowest protocol
265 layer to match. This stacking restriction does not apply to meta items which
266 can be placed anywhere in the stack without affecting the meaning of the
269 Patterns are terminated by END items.
273 .. _table_rte_flow_tcpv4_as_l4:
275 .. table:: TCPv4 as L4
291 .. _table_rte_flow_tcpv6_in_vxlan:
293 .. table:: TCPv6 in VXLAN
295 +-------+------------+
297 +=======+============+
299 +-------+------------+
301 +-------+------------+
303 +-------+------------+
305 +-------+------------+
307 +-------+------------+
309 +-------+------------+
311 +-------+------------+
313 +-------+------------+
317 .. _table_rte_flow_tcpv4_as_l4_meta:
319 .. table:: TCPv4 as L4 with meta items
341 The above example shows how meta items do not affect packet data matching
342 items, as long as those remain stacked properly. The resulting matching
343 pattern is identical to "TCPv4 as L4".
345 .. _table_rte_flow_udpv6_anywhere:
347 .. table:: UDPv6 anywhere
359 If supported by the PMD, omitting one or several protocol layers at the
360 bottom of the stack as in the above example (missing an Ethernet
361 specification) enables looking up anywhere in packets.
363 It is unspecified whether the payload of supported encapsulations
364 (e.g. VXLAN payload) is matched by such a pattern, which may apply to inner,
365 outer or both packets.
367 .. _table_rte_flow_invalid_l3:
369 .. table:: Invalid, missing L3
381 The above pattern is invalid due to a missing L3 specification between L2
382 (Ethernet) and L4 (UDP). Doing so is only allowed at the bottom and at the
388 They match meta-data or affect pattern processing instead of matching packet
389 data directly, most of them do not need a specification structure. This
390 particularity allows them to be specified anywhere in the stack without
391 causing any side effect.
396 End marker for item lists. Prevents further processing of items, thereby
399 - Its numeric value is 0 for convenience.
400 - PMD support is mandatory.
401 - ``spec``, ``last`` and ``mask`` are ignored.
403 .. _table_rte_flow_item_end:
407 +----------+---------+
409 +==========+=========+
410 | ``spec`` | ignored |
411 +----------+---------+
412 | ``last`` | ignored |
413 +----------+---------+
414 | ``mask`` | ignored |
415 +----------+---------+
420 Used as a placeholder for convenience. It is ignored and simply discarded by
423 - PMD support is mandatory.
424 - ``spec``, ``last`` and ``mask`` are ignored.
426 .. _table_rte_flow_item_void:
430 +----------+---------+
432 +==========+=========+
433 | ``spec`` | ignored |
434 +----------+---------+
435 | ``last`` | ignored |
436 +----------+---------+
437 | ``mask`` | ignored |
438 +----------+---------+
440 One usage example for this type is generating rules that share a common
441 prefix quickly without reallocating memory, only by updating item types:
443 .. _table_rte_flow_item_void_example:
445 .. table:: TCP, UDP or ICMP as L4
447 +-------+--------------------+
449 +=======+====================+
451 +-------+--------------------+
453 +-------+------+------+------+
454 | 2 | UDP | VOID | VOID |
455 +-------+------+------+------+
456 | 3 | VOID | TCP | VOID |
457 +-------+------+------+------+
458 | 4 | VOID | VOID | ICMP |
459 +-------+------+------+------+
461 +-------+--------------------+
466 Inverted matching, i.e. process packets that do not match the pattern.
468 - ``spec``, ``last`` and ``mask`` are ignored.
470 .. _table_rte_flow_item_invert:
474 +----------+---------+
476 +==========+=========+
477 | ``spec`` | ignored |
478 +----------+---------+
479 | ``last`` | ignored |
480 +----------+---------+
481 | ``mask`` | ignored |
482 +----------+---------+
484 Usage example, matching non-TCPv4 packets only:
486 .. _table_rte_flow_item_invert_example:
488 .. table:: Anything but TCPv4
507 This item is deprecated. Consider:
508 - `Item: PORT_REPRESENTOR`_
509 - `Item: REPRESENTED_PORT`_
511 Matches traffic originating from (ingress) or going to (egress) the physical
512 function of the current device.
514 If supported, should work even if the physical function is not managed by
515 the application and thus not associated with a DPDK port ID.
517 - Can be combined with any number of `Item: VF`_ to match both PF and VF
519 - ``spec``, ``last`` and ``mask`` must not be set.
521 .. _table_rte_flow_item_pf:
538 This item is deprecated. Consider:
539 - `Item: PORT_REPRESENTOR`_
540 - `Item: REPRESENTED_PORT`_
542 Matches traffic originating from (ingress) or going to (egress) a given
543 virtual function of the current device.
545 If supported, should work even if the virtual function is not managed by the
546 application and thus not associated with a DPDK port ID.
548 Note this pattern item does not match VF representors traffic which, as
549 separate entities, should be addressed through their own DPDK port IDs.
551 - Can be specified multiple times to match traffic addressed to several VF
553 - Can be combined with a PF item to match both PF and VF traffic.
554 - Default ``mask`` matches any VF ID.
556 .. _table_rte_flow_item_vf:
560 +----------+----------+---------------------------+
561 | Field | Subfield | Value |
562 +==========+==========+===========================+
563 | ``spec`` | ``id`` | destination VF ID |
564 +----------+----------+---------------------------+
565 | ``last`` | ``id`` | upper range value |
566 +----------+----------+---------------------------+
567 | ``mask`` | ``id`` | zeroed to match any VF ID |
568 +----------+----------+---------------------------+
573 This item is deprecated. Consider:
574 - `Item: PORT_REPRESENTOR`_
575 - `Item: REPRESENTED_PORT`_
577 Matches traffic originating from (ingress) or going to (egress) a physical
578 port of the underlying device.
580 The first PHY_PORT item overrides the physical port normally associated with
581 the specified DPDK input port (port_id). This item can be provided several
582 times to match additional physical ports.
584 Note that physical ports are not necessarily tied to DPDK input ports
585 (port_id) when those are not under DPDK control. Possible values are
586 specific to each device, they are not necessarily indexed from zero and may
589 As a device property, the list of allowed values as well as the value
590 associated with a port_id should be retrieved by other means.
592 - Default ``mask`` matches any port index.
594 .. _table_rte_flow_item_phy_port:
598 +----------+-----------+--------------------------------+
599 | Field | Subfield | Value |
600 +==========+===========+================================+
601 | ``spec`` | ``index`` | physical port index |
602 +----------+-----------+--------------------------------+
603 | ``last`` | ``index`` | upper range value |
604 +----------+-----------+--------------------------------+
605 | ``mask`` | ``index`` | zeroed to match any port index |
606 +----------+-----------+--------------------------------+
611 This item is deprecated. Consider:
612 - `Item: PORT_REPRESENTOR`_
613 - `Item: REPRESENTED_PORT`_
615 Matches traffic originating from (ingress) or going to (egress) a given DPDK
618 Normally only supported if the port ID in question is known by the
619 underlying PMD and related to the device the flow rule is created against.
621 This must not be confused with `Item: PHY_PORT`_ which refers to the
622 physical port of a device, whereas `Item: PORT_ID`_ refers to a ``struct
623 rte_eth_dev`` object on the application side (also known as "port
624 representor" depending on the kind of underlying device).
626 - Default ``mask`` matches the specified DPDK port ID.
628 .. _table_rte_flow_item_port_id:
632 +----------+----------+-----------------------------+
633 | Field | Subfield | Value |
634 +==========+==========+=============================+
635 | ``spec`` | ``id`` | DPDK port ID |
636 +----------+----------+-----------------------------+
637 | ``last`` | ``id`` | upper range value |
638 +----------+----------+-----------------------------+
639 | ``mask`` | ``id`` | zeroed to match any port ID |
640 +----------+----------+-----------------------------+
645 Matches an arbitrary integer value which was set using the ``MARK`` action in
646 a previously matched rule.
648 This item can only specified once as a match criteria as the ``MARK`` action can
649 only be specified once in a flow action.
651 Note the value of MARK field is arbitrary and application defined.
653 Depending on the underlying implementation the MARK item may be supported on
654 the physical device, with virtual groups in the PMD or not at all.
656 - Default ``mask`` matches any integer value.
658 .. _table_rte_flow_item_mark:
662 +----------+----------+---------------------------+
663 | Field | Subfield | Value |
664 +==========+==========+===========================+
665 | ``spec`` | ``id`` | integer value |
666 +----------+--------------------------------------+
667 | ``last`` | ``id`` | upper range value |
668 +----------+----------+---------------------------+
669 | ``mask`` | ``id`` | zeroed to match any value |
670 +----------+----------+---------------------------+
675 Matches tag item set by other flows. Multiple tags are supported by specifying
678 - Default ``mask`` matches the specified tag value and index.
680 .. _table_rte_flow_item_tag:
684 +----------+----------+----------------------------------------+
685 | Field | Subfield | Value |
686 +==========+===========+=======================================+
687 | ``spec`` | ``data`` | 32 bit flow tag value |
688 | +-----------+---------------------------------------+
689 | | ``index`` | index of flow tag |
690 +----------+-----------+---------------------------------------+
691 | ``last`` | ``data`` | upper range value |
692 | +-----------+---------------------------------------+
693 | | ``index`` | field is ignored |
694 +----------+-----------+---------------------------------------+
695 | ``mask`` | ``data`` | bit-mask applies to "spec" and "last" |
696 | +-----------+---------------------------------------+
697 | | ``index`` | field is ignored |
698 +----------+-----------+---------------------------------------+
703 Matches 32 bit metadata item set.
705 On egress, metadata can be set either by mbuf metadata field with
706 PKT_TX_DYNF_METADATA flag or ``SET_META`` action. On ingress, ``SET_META``
707 action sets metadata for a packet and the metadata will be reported via
708 ``metadata`` dynamic field of ``rte_mbuf`` with PKT_RX_DYNF_METADATA flag.
710 - Default ``mask`` matches the specified Rx metadata value.
712 .. _table_rte_flow_item_meta:
716 +----------+----------+---------------------------------------+
717 | Field | Subfield | Value |
718 +==========+==========+=======================================+
719 | ``spec`` | ``data`` | 32 bit metadata value |
720 +----------+----------+---------------------------------------+
721 | ``last`` | ``data`` | upper range value |
722 +----------+----------+---------------------------------------+
723 | ``mask`` | ``data`` | bit-mask applies to "spec" and "last" |
724 +----------+----------+---------------------------------------+
726 Data matching item types
727 ~~~~~~~~~~~~~~~~~~~~~~~~
729 Most of these are basically protocol header definitions with associated
730 bit-masks. They must be specified (stacked) from lowest to highest protocol
731 layer to form a matching pattern.
736 Matches any protocol in place of the current layer, a single ANY may also
737 stand for several protocol layers.
739 This is usually specified as the first pattern item when looking for a
740 protocol anywhere in a packet.
742 - Default ``mask`` stands for any number of layers.
744 .. _table_rte_flow_item_any:
748 +----------+----------+--------------------------------------+
749 | Field | Subfield | Value |
750 +==========+==========+======================================+
751 | ``spec`` | ``num`` | number of layers covered |
752 +----------+----------+--------------------------------------+
753 | ``last`` | ``num`` | upper range value |
754 +----------+----------+--------------------------------------+
755 | ``mask`` | ``num`` | zeroed to cover any number of layers |
756 +----------+----------+--------------------------------------+
758 Example for VXLAN TCP payload matching regardless of outer L3 (IPv4 or IPv6)
759 and L4 (UDP) both matched by the first ANY specification, and inner L3 (IPv4
760 or IPv6) matched by the second ANY specification:
762 .. _table_rte_flow_item_any_example:
764 .. table:: TCP in VXLAN with wildcards
766 +-------+------+----------+----------+-------+
767 | Index | Item | Field | Subfield | Value |
768 +=======+======+==========+==========+=======+
770 +-------+------+----------+----------+-------+
771 | 1 | ANY | ``spec`` | ``num`` | 2 |
772 +-------+------+----------+----------+-------+
774 +-------+------------------------------------+
776 +-------+------+----------+----------+-------+
777 | 4 | ANY | ``spec`` | ``num`` | 1 |
778 +-------+------+----------+----------+-------+
780 +-------+------------------------------------+
782 +-------+------------------------------------+
787 Matches a byte string of a given length at a given offset.
789 Offset is either absolute (using the start of the packet) or relative to the
790 end of the previous matched item in the stack, in which case negative values
793 If search is enabled, offset is used as the starting point. The search area
794 can be delimited by setting limit to a nonzero value, which is the maximum
795 number of bytes after offset where the pattern may start.
797 Matching a zero-length pattern is allowed, doing so resets the relative
798 offset for subsequent items.
800 - This type does not support ranges (``last`` field).
801 - Default ``mask`` matches all fields exactly.
803 .. _table_rte_flow_item_raw:
807 +----------+--------------+-------------------------------------------------+
808 | Field | Subfield | Value |
809 +==========+==============+=================================================+
810 | ``spec`` | ``relative`` | look for pattern after the previous item |
811 | +--------------+-------------------------------------------------+
812 | | ``search`` | search pattern from offset (see also ``limit``) |
813 | +--------------+-------------------------------------------------+
814 | | ``reserved`` | reserved, must be set to zero |
815 | +--------------+-------------------------------------------------+
816 | | ``offset`` | absolute or relative offset for ``pattern`` |
817 | +--------------+-------------------------------------------------+
818 | | ``limit`` | search area limit for start of ``pattern`` |
819 | +--------------+-------------------------------------------------+
820 | | ``length`` | ``pattern`` length |
821 | +--------------+-------------------------------------------------+
822 | | ``pattern`` | byte string to look for |
823 +----------+--------------+-------------------------------------------------+
824 | ``last`` | if specified, either all 0 or with the same values as ``spec`` |
825 +----------+----------------------------------------------------------------+
826 | ``mask`` | bit-mask applied to ``spec`` values with usual behavior |
827 +----------+----------------------------------------------------------------+
829 Example pattern looking for several strings at various offsets of a UDP
830 payload, using combined RAW items:
832 .. _table_rte_flow_item_raw_example:
834 .. table:: UDP payload matching
836 +-------+------+----------+--------------+-------+
837 | Index | Item | Field | Subfield | Value |
838 +=======+======+==========+==============+=======+
840 +-------+----------------------------------------+
842 +-------+----------------------------------------+
844 +-------+------+----------+--------------+-------+
845 | 3 | RAW | ``spec`` | ``relative`` | 1 |
846 | | | +--------------+-------+
847 | | | | ``search`` | 1 |
848 | | | +--------------+-------+
849 | | | | ``offset`` | 10 |
850 | | | +--------------+-------+
851 | | | | ``limit`` | 0 |
852 | | | +--------------+-------+
853 | | | | ``length`` | 3 |
854 | | | +--------------+-------+
855 | | | | ``pattern`` | "foo" |
856 +-------+------+----------+--------------+-------+
857 | 4 | RAW | ``spec`` | ``relative`` | 1 |
858 | | | +--------------+-------+
859 | | | | ``search`` | 0 |
860 | | | +--------------+-------+
861 | | | | ``offset`` | 20 |
862 | | | +--------------+-------+
863 | | | | ``limit`` | 0 |
864 | | | +--------------+-------+
865 | | | | ``length`` | 3 |
866 | | | +--------------+-------+
867 | | | | ``pattern`` | "bar" |
868 +-------+------+----------+--------------+-------+
869 | 5 | RAW | ``spec`` | ``relative`` | 1 |
870 | | | +--------------+-------+
871 | | | | ``search`` | 0 |
872 | | | +--------------+-------+
873 | | | | ``offset`` | -29 |
874 | | | +--------------+-------+
875 | | | | ``limit`` | 0 |
876 | | | +--------------+-------+
877 | | | | ``length`` | 3 |
878 | | | +--------------+-------+
879 | | | | ``pattern`` | "baz" |
880 +-------+------+----------+--------------+-------+
882 +-------+----------------------------------------+
886 - Locate "foo" at least 10 bytes deep inside UDP payload.
887 - Locate "bar" after "foo" plus 20 bytes.
888 - Locate "baz" after "bar" minus 29 bytes.
890 Such a packet may be represented as follows (not to scale)::
893 | |<--------->| |<--------->|
895 |-----|------|-----|-----|-----|-----|-----------|-----|------|
896 | ETH | IPv4 | UDP | ... | baz | foo | ......... | bar | .... |
897 |-----|------|-----|-----|-----|-----|-----------|-----|------|
899 |<--------------------------->|
902 Note that matching subsequent pattern items would resume after "baz", not
903 "bar" since matching is always performed after the previous item of the
909 Matches an Ethernet header.
911 The ``type`` field either stands for "EtherType" or "TPID" when followed by
912 so-called layer 2.5 pattern items such as ``RTE_FLOW_ITEM_TYPE_VLAN``. In
913 the latter case, ``type`` refers to that of the outer header, with the inner
914 EtherType/TPID provided by the subsequent pattern item. This is the same
915 order as on the wire.
916 If the ``type`` field contains a TPID value, then only tagged packets with the
917 specified TPID will match the pattern.
918 The field ``has_vlan`` can be used to match any type of tagged packets,
919 instead of using the ``type`` field.
920 If the ``type`` and ``has_vlan`` fields are not specified, then both tagged
921 and untagged packets will match the pattern.
923 - ``dst``: destination MAC.
924 - ``src``: source MAC.
925 - ``type``: EtherType or TPID.
926 - ``has_vlan``: packet header contains at least one VLAN.
927 - Default ``mask`` matches destination and source addresses only.
932 Matches an 802.1Q/ad VLAN tag.
934 The corresponding standard outer EtherType (TPID) values are
935 ``RTE_ETHER_TYPE_VLAN`` or ``RTE_ETHER_TYPE_QINQ``. It can be overridden by the
936 preceding pattern item.
937 If a ``VLAN`` item is present in the pattern, then only tagged packets will
939 The field ``has_more_vlan`` can be used to match any type of tagged packets,
940 instead of using the ``inner_type field``.
941 If the ``inner_type`` and ``has_more_vlan`` fields are not specified,
942 then any tagged packets will match the pattern.
944 - ``tci``: tag control information.
945 - ``inner_type``: inner EtherType or TPID.
946 - ``has_more_vlan``: packet header contains at least one more VLAN, after this VLAN.
947 - Default ``mask`` matches the VID part of TCI only (lower 12 bits).
952 Matches an IPv4 header.
954 Note: IPv4 options are handled by dedicated pattern items.
956 - ``hdr``: IPv4 header definition (``rte_ip.h``).
957 - Default ``mask`` matches source and destination addresses only.
962 Matches an IPv6 header.
964 Dedicated flags indicate if header contains specific extension headers.
965 To match on packets containing a specific extension header, an application
966 should match on the dedicated flag set to 1.
967 To match on packets not containing a specific extension header, an application
968 should match on the dedicated flag clear to 0.
969 In case application doesn't care about the existence of a specific extension
970 header, it should not specify the dedicated flag for matching.
972 - ``hdr``: IPv6 header definition (``rte_ip.h``).
973 - ``has_hop_ext``: header contains Hop-by-Hop Options extension header.
974 - ``has_route_ext``: header contains Routing extension header.
975 - ``has_frag_ext``: header contains Fragment extension header.
976 - ``has_auth_ext``: header contains Authentication extension header.
977 - ``has_esp_ext``: header contains Encapsulation Security Payload extension header.
978 - ``has_dest_ext``: header contains Destination Options extension header.
979 - ``has_mobil_ext``: header contains Mobility extension header.
980 - ``has_hip_ext``: header contains Host Identity Protocol extension header.
981 - ``has_shim6_ext``: header contains Shim6 Protocol extension header.
982 - Default ``mask`` matches ``hdr`` source and destination addresses only.
987 Matches an ICMP header.
989 - ``hdr``: ICMP header definition (``rte_icmp.h``).
990 - Default ``mask`` matches ICMP type and code only.
995 Matches a UDP header.
997 - ``hdr``: UDP header definition (``rte_udp.h``).
998 - Default ``mask`` matches source and destination ports only.
1003 Matches a TCP header.
1005 - ``hdr``: TCP header definition (``rte_tcp.h``).
1006 - Default ``mask`` matches source and destination ports only.
1011 Matches a SCTP header.
1013 - ``hdr``: SCTP header definition (``rte_sctp.h``).
1014 - Default ``mask`` matches source and destination ports only.
1019 Matches a VXLAN header (RFC 7348).
1021 - ``flags``: normally 0x08 (I flag).
1022 - ``rsvd0``: reserved, normally 0x000000.
1023 - ``vni``: VXLAN network identifier.
1024 - ``rsvd1``: reserved, normally 0x00.
1025 - Default ``mask`` matches VNI only.
1030 Matches an IEEE 802.1BR E-Tag header.
1032 The corresponding standard outer EtherType (TPID) value is
1033 ``RTE_ETHER_TYPE_ETAG``. It can be overridden by the preceding pattern item.
1035 - ``epcp_edei_in_ecid_b``: E-Tag control information (E-TCI), E-PCP (3b),
1036 E-DEI (1b), ingress E-CID base (12b).
1037 - ``rsvd_grp_ecid_b``: reserved (2b), GRP (2b), E-CID base (12b).
1038 - ``in_ecid_e``: ingress E-CID ext.
1039 - ``ecid_e``: E-CID ext.
1040 - ``inner_type``: inner EtherType or TPID.
1041 - Default ``mask`` simultaneously matches GRP and E-CID base.
1046 Matches a NVGRE header (RFC 7637).
1048 - ``c_k_s_rsvd0_ver``: checksum (1b), undefined (1b), key bit (1b),
1049 sequence number (1b), reserved 0 (9b), version (3b). This field must have
1050 value 0x2000 according to RFC 7637.
1051 - ``protocol``: protocol type (0x6558).
1052 - ``tni``: virtual subnet ID.
1053 - ``flow_id``: flow ID.
1054 - Default ``mask`` matches TNI only.
1059 Matches a MPLS header.
1061 - ``label_tc_s_ttl``: label, TC, Bottom of Stack and TTL.
1062 - Default ``mask`` matches label only.
1067 Matches a GRE header.
1069 - ``c_rsvd0_ver``: checksum, reserved 0 and version.
1070 - ``protocol``: protocol type.
1071 - Default ``mask`` matches protocol only.
1076 Matches a GRE key field.
1077 This should be preceded by item ``GRE``.
1079 - Value to be matched is a big-endian 32 bit integer.
1080 - When this item present it implicitly match K bit in default mask as "1"
1085 Fuzzy pattern match, expect faster than default.
1087 This is for device that support fuzzy match option. Usually a fuzzy match is
1088 fast but the cost is accuracy. i.e. Signature Match only match pattern's hash
1089 value, but it is possible two different patterns have the same hash value.
1091 Matching accuracy level can be configured by threshold. Driver can divide the
1092 range of threshold and map to different accuracy levels that device support.
1094 Threshold 0 means perfect match (no fuzziness), while threshold 0xffffffff
1095 means fuzziest match.
1097 .. _table_rte_flow_item_fuzzy:
1101 +----------+---------------+--------------------------------------------------+
1102 | Field | Subfield | Value |
1103 +==========+===============+==================================================+
1104 | ``spec`` | ``threshold`` | 0 as perfect match, 0xffffffff as fuzziest match |
1105 +----------+---------------+--------------------------------------------------+
1106 | ``last`` | ``threshold`` | upper range value |
1107 +----------+---------------+--------------------------------------------------+
1108 | ``mask`` | ``threshold`` | bit-mask apply to "spec" and "last" |
1109 +----------+---------------+--------------------------------------------------+
1111 Usage example, fuzzy match a TCPv4 packets:
1113 .. _table_rte_flow_item_fuzzy_example:
1115 .. table:: Fuzzy matching
1117 +-------+----------+
1119 +=======+==========+
1121 +-------+----------+
1123 +-------+----------+
1125 +-------+----------+
1127 +-------+----------+
1129 +-------+----------+
1131 Item: ``GTP``, ``GTPC``, ``GTPU``
1132 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1134 Matches a GTPv1 header.
1136 Note: GTP, GTPC and GTPU use the same structure. GTPC and GTPU item
1137 are defined for a user-friendly API when creating GTP-C and GTP-U
1140 - ``v_pt_rsv_flags``: version (3b), protocol type (1b), reserved (1b),
1141 extension header flag (1b), sequence number flag (1b), N-PDU number
1143 - ``msg_type``: message type.
1144 - ``msg_len``: message length.
1145 - ``teid``: tunnel endpoint identifier.
1146 - Default ``mask`` matches teid only.
1151 Matches an ESP header.
1153 - ``hdr``: ESP header definition (``rte_esp.h``).
1154 - Default ``mask`` matches SPI only.
1159 Matches a GENEVE header.
1161 - ``ver_opt_len_o_c_rsvd0``: version (2b), length of the options fields (6b),
1162 OAM packet (1b), critical options present (1b), reserved 0 (6b).
1163 - ``protocol``: protocol type.
1164 - ``vni``: virtual network identifier.
1165 - ``rsvd1``: reserved, normally 0x00.
1166 - Default ``mask`` matches VNI only.
1171 Matches a VXLAN-GPE header (draft-ietf-nvo3-vxlan-gpe-05).
1173 - ``flags``: normally 0x0C (I and P flags).
1174 - ``rsvd0``: reserved, normally 0x0000.
1175 - ``protocol``: protocol type.
1176 - ``vni``: VXLAN network identifier.
1177 - ``rsvd1``: reserved, normally 0x00.
1178 - Default ``mask`` matches VNI only.
1180 Item: ``ARP_ETH_IPV4``
1181 ^^^^^^^^^^^^^^^^^^^^^^
1183 Matches an ARP header for Ethernet/IPv4.
1185 - ``hdr``: hardware type, normally 1.
1186 - ``pro``: protocol type, normally 0x0800.
1187 - ``hln``: hardware address length, normally 6.
1188 - ``pln``: protocol address length, normally 4.
1189 - ``op``: opcode (1 for request, 2 for reply).
1190 - ``sha``: sender hardware address.
1191 - ``spa``: sender IPv4 address.
1192 - ``tha``: target hardware address.
1193 - ``tpa``: target IPv4 address.
1194 - Default ``mask`` matches SHA, SPA, THA and TPA.
1199 Matches the presence of any IPv6 extension header.
1201 - ``next_hdr``: next header.
1202 - Default ``mask`` matches ``next_hdr``.
1204 Normally preceded by any of:
1209 Item: ``IPV6_FRAG_EXT``
1210 ^^^^^^^^^^^^^^^^^^^^^^^
1212 Matches the presence of IPv6 fragment extension header.
1214 - ``hdr``: IPv6 fragment extension header definition (``rte_ip.h``).
1216 Normally preceded by any of:
1224 Matches any ICMPv6 header.
1226 - ``type``: ICMPv6 type.
1227 - ``code``: ICMPv6 code.
1228 - ``checksum``: ICMPv6 checksum.
1229 - Default ``mask`` matches ``type`` and ``code``.
1231 Item: ``ICMP6_ND_NS``
1232 ^^^^^^^^^^^^^^^^^^^^^
1234 Matches an ICMPv6 neighbor discovery solicitation.
1236 - ``type``: ICMPv6 type, normally 135.
1237 - ``code``: ICMPv6 code, normally 0.
1238 - ``checksum``: ICMPv6 checksum.
1239 - ``reserved``: reserved, normally 0.
1240 - ``target_addr``: target address.
1241 - Default ``mask`` matches target address only.
1243 Item: ``ICMP6_ND_NA``
1244 ^^^^^^^^^^^^^^^^^^^^^
1246 Matches an ICMPv6 neighbor discovery advertisement.
1248 - ``type``: ICMPv6 type, normally 136.
1249 - ``code``: ICMPv6 code, normally 0.
1250 - ``checksum``: ICMPv6 checksum.
1251 - ``rso_reserved``: route flag (1b), solicited flag (1b), override flag
1252 (1b), reserved (29b).
1253 - ``target_addr``: target address.
1254 - Default ``mask`` matches target address only.
1256 Item: ``ICMP6_ND_OPT``
1257 ^^^^^^^^^^^^^^^^^^^^^^
1259 Matches the presence of any ICMPv6 neighbor discovery option.
1261 - ``type``: ND option type.
1262 - ``length``: ND option length.
1263 - Default ``mask`` matches type only.
1265 Normally preceded by any of:
1267 - `Item: ICMP6_ND_NA`_
1268 - `Item: ICMP6_ND_NS`_
1269 - `Item: ICMP6_ND_OPT`_
1271 Item: ``ICMP6_ND_OPT_SLA_ETH``
1272 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1274 Matches an ICMPv6 neighbor discovery source Ethernet link-layer address
1277 - ``type``: ND option type, normally 1.
1278 - ``length``: ND option length, normally 1.
1279 - ``sla``: source Ethernet LLA.
1280 - Default ``mask`` matches source link-layer address only.
1282 Normally preceded by any of:
1284 - `Item: ICMP6_ND_NA`_
1285 - `Item: ICMP6_ND_OPT`_
1287 Item: ``ICMP6_ND_OPT_TLA_ETH``
1288 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1290 Matches an ICMPv6 neighbor discovery target Ethernet link-layer address
1293 - ``type``: ND option type, normally 2.
1294 - ``length``: ND option length, normally 1.
1295 - ``tla``: target Ethernet LLA.
1296 - Default ``mask`` matches target link-layer address only.
1298 Normally preceded by any of:
1300 - `Item: ICMP6_ND_NS`_
1301 - `Item: ICMP6_ND_OPT`_
1306 Matches an application specific 32 bit metadata item.
1308 - Default ``mask`` matches the specified metadata value.
1313 Matches a GTP PDU extension header with type 0x85.
1315 - ``pdu_type``: PDU type.
1316 - ``qfi``: QoS flow identifier.
1317 - Default ``mask`` matches QFI only.
1319 Item: ``PPPOES``, ``PPPOED``
1320 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1322 Matches a PPPoE header.
1324 - ``version_type``: version (4b), type (4b).
1325 - ``code``: message type.
1326 - ``session_id``: session identifier.
1327 - ``length``: payload length.
1329 Item: ``PPPOE_PROTO_ID``
1330 ^^^^^^^^^^^^^^^^^^^^^^^^
1332 Matches a PPPoE session protocol identifier.
1334 - ``proto_id``: PPP protocol identifier.
1335 - Default ``mask`` matches proto_id only.
1340 Matches a network service header (RFC 8300).
1342 - ``version``: normally 0x0 (2 bits).
1343 - ``oam_pkt``: indicate oam packet (1 bit).
1344 - ``reserved``: reserved bit (1 bit).
1345 - ``ttl``: maximum SFF hopes (6 bits).
1346 - ``length``: total length in 4 bytes words (6 bits).
1347 - ``reserved1``: reserved1 bits (4 bits).
1348 - ``mdtype``: ndicates format of NSH header (4 bits).
1349 - ``next_proto``: indicates protocol type of encap data (8 bits).
1350 - ``spi``: service path identifier (3 bytes).
1351 - ``sindex``: service index (1 byte).
1352 - Default ``mask`` matches mdtype, next_proto, spi, sindex.
1358 Matches a Internet Group Management Protocol (RFC 2236).
1360 - ``type``: IGMP message type (Query/Report).
1361 - ``max_resp_time``: max time allowed before sending report.
1362 - ``checksum``: checksum, 1s complement of whole IGMP message.
1363 - ``group_addr``: group address, for Query value will be 0.
1364 - Default ``mask`` matches group_addr.
1370 Matches a IP Authentication Header (RFC 4302).
1372 - ``next_hdr``: next payload after AH.
1373 - ``payload_len``: total length of AH in 4B words.
1374 - ``reserved``: reserved bits.
1375 - ``spi``: security parameters index.
1376 - ``seq_num``: counter value increased by 1 on each packet sent.
1377 - Default ``mask`` matches spi.
1382 Matches a HIGIG2 header field. It is layer 2.5 protocol and used in
1385 - Default ``mask`` matches classification and vlan.
1390 Matches a L2TPv3 over IP header.
1392 - ``session_id``: L2TPv3 over IP session identifier.
1393 - Default ``mask`` matches session_id only.
1398 Matches a PFCP Header.
1400 - ``s_field``: S field.
1401 - ``msg_type``: message type.
1402 - ``msg_len``: message length.
1403 - ``seid``: session endpoint identifier.
1404 - Default ``mask`` matches s_field and seid.
1409 Matches a eCPRI header.
1411 - ``hdr``: eCPRI header definition (``rte_ecpri.h``).
1412 - Default ``mask`` matches nothing, for all eCPRI messages.
1414 Item: ``PACKET_INTEGRITY_CHECKS``
1415 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1417 Matches packet integrity.
1418 For some devices application needs to enable integration checks in HW
1419 before using this item.
1421 - ``level``: the encapsulation level that should be checked:
1422 - ``level == 0`` means the default PMD mode (can be inner most / outermost).
1423 - ``level == 1`` means outermost header.
1424 - ``level > 1`` means inner header. See also RSS level.
1425 - ``packet_ok``: All HW packet integrity checks have passed based on the
1426 topmost network layer. For example, for ICMP packet the topmost network
1427 layer is L3 and for TCP or UDP packet the topmost network layer is L4.
1428 - ``l2_ok``: all layer 2 HW integrity checks passed.
1429 - ``l3_ok``: all layer 3 HW integrity checks passed.
1430 - ``l4_ok``: all layer 4 HW integrity checks passed.
1431 - ``l2_crc_ok``: layer 2 CRC check passed.
1432 - ``ipv4_csum_ok``: IPv4 checksum check passed.
1433 - ``l4_csum_ok``: layer 4 checksum check passed.
1434 - ``l3_len_ok``: the layer 3 length is smaller than the frame length.
1439 Matches a conntrack state after conntrack action.
1441 - ``flags``: conntrack packet state flags.
1442 - Default ``mask`` matches all state bits.
1444 Item: ``PORT_REPRESENTOR``
1445 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1447 Matches traffic entering the embedded switch from the given ethdev.
1449 Term **ethdev** and the concept of **port representor** are synonymous.
1450 The **represented port** is an *entity* plugged to the embedded switch
1451 at the opposite end of the "wire" leading to the ethdev.
1455 .--------------------.
1456 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
1457 '--------------------'
1479 .--------------------.
1480 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
1481 '--------------------'
1484 - Incompatible with `Attribute: Traffic direction`_.
1485 - Requires `Attribute: Transfer`_.
1487 .. _table_rte_flow_item_ethdev:
1489 .. table:: ``struct rte_flow_item_ethdev``
1491 +----------+-------------+---------------------------+
1492 | Field | Subfield | Value |
1493 +==========+=============+===========================+
1494 | ``spec`` | ``port_id`` | ethdev port ID |
1495 +----------+-------------+---------------------------+
1496 | ``last`` | ``port_id`` | upper range value |
1497 +----------+-------------+---------------------------+
1498 | ``mask`` | ``port_id`` | zeroed for wildcard match |
1499 +----------+-------------+---------------------------+
1501 - Default ``mask`` provides exact match behaviour.
1503 See also `Action: PORT_REPRESENTOR`_.
1505 Item: ``REPRESENTED_PORT``
1506 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1508 Matches traffic entering the embedded switch from
1509 the entity represented by the given ethdev.
1511 Term **ethdev** and the concept of **port representor** are synonymous.
1512 The **represented port** is an *entity* plugged to the embedded switch
1513 at the opposite end of the "wire" leading to the ethdev.
1517 .--------------------.
1518 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
1519 '--------------------'
1541 .--------------------.
1542 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
1543 '--------------------'
1546 - Incompatible with `Attribute: Traffic direction`_.
1547 - Requires `Attribute: Transfer`_.
1549 This item is meant to use the same structure as `Item: PORT_REPRESENTOR`_.
1551 See also `Action: REPRESENTED_PORT`_.
1556 Each possible action is represented by a type.
1557 An action can have an associated configuration object.
1558 Several actions combined in a list can be assigned
1559 to a flow rule and are performed in order.
1561 They fall in three categories:
1563 - Actions that modify the fate of matching traffic, for instance by dropping
1564 or assigning it a specific destination.
1566 - Actions that modify matching traffic contents or its properties. This
1567 includes adding/removing encapsulation, encryption, compression and marks.
1569 - Actions related to the flow rule itself, such as updating counters or
1570 making it non-terminating.
1572 Flow rules being terminating by default, not specifying any action of the
1573 fate kind results in undefined behavior. This applies to both ingress and
1576 PASSTHRU, when supported, makes a flow rule non-terminating.
1578 Like matching patterns, action lists are terminated by END items.
1580 Example of action that redirects packets to queue index 10:
1582 .. _table_rte_flow_action_example:
1584 .. table:: Queue action
1586 +-----------+-------+
1588 +===========+=======+
1590 +-----------+-------+
1592 Actions are performed in list order:
1594 .. _table_rte_flow_count_then_drop:
1596 .. table:: Count then drop
1610 .. _table_rte_flow_mark_count_redirect:
1612 .. table:: Mark, count then redirect
1614 +-------+--------+------------+-------+
1615 | Index | Action | Field | Value |
1616 +=======+========+============+=======+
1617 | 0 | MARK | ``mark`` | 0x2a |
1618 +-------+--------+------------+-------+
1619 | 1 | COUNT | ``id`` | 0 |
1620 +-------+--------+------------+-------+
1621 | 2 | QUEUE | ``queue`` | 10 |
1622 +-------+--------+------------+-------+
1624 +-------+-----------------------------+
1628 .. _table_rte_flow_redirect_queue_5:
1630 .. table:: Redirect to queue 5
1632 +-------+--------+-----------+-------+
1633 | Index | Action | Field | Value |
1634 +=======+========+===========+=======+
1636 +-------+--------+-----------+-------+
1637 | 1 | QUEUE | ``queue`` | 5 |
1638 +-------+--------+-----------+-------+
1640 +-------+----------------------------+
1642 In the above example, while DROP and QUEUE must be performed in order, both
1643 have to happen before reaching END. Only QUEUE has a visible effect.
1645 Note that such a list may be thought as ambiguous and rejected on that
1648 .. _table_rte_flow_redirect_queue_5_3:
1650 .. table:: Redirect to queues 5 and 3
1652 +-------+--------+-----------+-------+
1653 | Index | Action | Field | Value |
1654 +=======+========+===========+=======+
1655 | 0 | QUEUE | ``queue`` | 5 |
1656 +-------+--------+-----------+-------+
1658 +-------+--------+-----------+-------+
1659 | 2 | QUEUE | ``queue`` | 3 |
1660 +-------+--------+-----------+-------+
1662 +-------+----------------------------+
1664 As previously described, all actions must be taken into account. This
1665 effectively duplicates traffic to both queues. The above example also shows
1666 that VOID is ignored.
1671 Common action types are described in this section.
1676 End marker for action lists. Prevents further processing of actions, thereby
1679 - Its numeric value is 0 for convenience.
1680 - PMD support is mandatory.
1681 - No configurable properties.
1683 .. _table_rte_flow_action_end:
1696 Used as a placeholder for convenience. It is ignored and simply discarded by
1699 - PMD support is mandatory.
1700 - No configurable properties.
1702 .. _table_rte_flow_action_void:
1712 Action: ``PASSTHRU``
1713 ^^^^^^^^^^^^^^^^^^^^
1715 Leaves traffic up for additional processing by subsequent flow rules; makes
1716 a flow rule non-terminating.
1718 - No configurable properties.
1720 .. _table_rte_flow_action_passthru:
1730 Example to copy a packet to a queue and continue processing by subsequent
1733 .. _table_rte_flow_action_passthru_example:
1735 .. table:: Copy to queue 8
1737 +-------+--------+-----------+-------+
1738 | Index | Action | Field | Value |
1739 +=======+========+===========+=======+
1741 +-------+--------+-----------+-------+
1742 | 1 | QUEUE | ``queue`` | 8 |
1743 +-------+--------+-----------+-------+
1745 +-------+----------------------------+
1750 Redirects packets to a group on the current device.
1752 In a hierarchy of groups, which can be used to represent physical or logical
1753 flow group/tables on the device, this action redirects the matched flow to
1754 the specified group on that device.
1756 If a matched flow is redirected to a table which doesn't contain a matching
1757 rule for that flow then the behavior is undefined and the resulting behavior
1758 is up to the specific device. Best practice when using groups would be define
1759 a default flow rule for each group which a defines the default actions in that
1760 group so a consistent behavior is defined.
1762 Defining an action for matched flow in a group to jump to a group which is
1763 higher in the group hierarchy may not be supported by physical devices,
1764 depending on how groups are mapped to the physical devices. In the
1765 definitions of jump actions, applications should be aware that it may be
1766 possible to define flow rules which trigger an undefined behavior causing
1767 flows to loop between groups.
1769 .. _table_rte_flow_action_jump:
1773 +-----------+------------------------------+
1775 +===========+==============================+
1776 | ``group`` | Group to redirect packets to |
1777 +-----------+------------------------------+
1782 Attaches an integer value to packets and sets ``PKT_RX_FDIR`` and
1783 ``PKT_RX_FDIR_ID`` mbuf flags.
1785 This value is arbitrary and application-defined. Maximum allowed value
1786 depends on the underlying implementation. It is returned in the
1787 ``hash.fdir.hi`` mbuf field.
1789 .. _table_rte_flow_action_mark:
1793 +--------+--------------------------------------+
1795 +========+======================================+
1796 | ``id`` | integer value to return with packets |
1797 +--------+--------------------------------------+
1802 Flags packets. Similar to `Action: MARK`_ without a specific value; only
1803 sets the ``PKT_RX_FDIR`` mbuf flag.
1805 - No configurable properties.
1807 .. _table_rte_flow_action_flag:
1820 Assigns packets to a given queue index.
1822 .. _table_rte_flow_action_queue:
1826 +-----------+--------------------+
1828 +===========+====================+
1829 | ``index`` | queue index to use |
1830 +-----------+--------------------+
1837 - No configurable properties.
1839 .. _table_rte_flow_action_drop:
1852 Adds a counter action to a matched flow.
1854 If more than one count action is specified in a single flow rule, then each
1855 action must specify a unique id.
1857 Counters can be retrieved and reset through ``rte_flow_query()``, see
1858 ``struct rte_flow_query_count``.
1860 For ports within the same switch domain then the counter id namespace extends
1861 to all ports within that switch domain.
1863 .. _table_rte_flow_action_count:
1867 +------------+---------------------------------+
1869 +============+=================================+
1870 | ``id`` | counter id |
1871 +------------+---------------------------------+
1873 Query structure to retrieve and reset flow rule counters:
1875 .. _table_rte_flow_query_count:
1877 .. table:: COUNT query
1879 +---------------+-----+-----------------------------------+
1880 | Field | I/O | Value |
1881 +===============+=====+===================================+
1882 | ``reset`` | in | reset counter after query |
1883 +---------------+-----+-----------------------------------+
1884 | ``hits_set`` | out | ``hits`` field is set |
1885 +---------------+-----+-----------------------------------+
1886 | ``bytes_set`` | out | ``bytes`` field is set |
1887 +---------------+-----+-----------------------------------+
1888 | ``hits`` | out | number of hits for this rule |
1889 +---------------+-----+-----------------------------------+
1890 | ``bytes`` | out | number of bytes through this rule |
1891 +---------------+-----+-----------------------------------+
1896 Similar to QUEUE, except RSS is additionally performed on packets to spread
1897 them among several queues according to the provided parameters.
1899 Unlike global RSS settings used by other DPDK APIs, unsetting the ``types``
1900 field does not disable RSS in a flow rule. Doing so instead requests safe
1901 unspecified "best-effort" settings from the underlying PMD, which depending
1902 on the flow rule, may result in anything ranging from empty (single queue)
1903 to all-inclusive RSS.
1905 If non-applicable for matching packets RSS types are requested,
1906 these RSS types are simply ignored. For example, it happens if:
1908 - Hashing of both TCP and UDP ports is requested
1909 (only one can be present in a packet).
1911 - Requested RSS types contradict to flow rule pattern
1912 (e.g. pattern has UDP item, but RSS types contain TCP).
1914 If requested RSS hash types are not supported by the Ethernet device at all
1915 (not reported in ``dev_info.flow_type_rss_offloads``),
1916 the flow creation will fail.
1918 Note: RSS hash result is stored in the ``hash.rss`` mbuf field which
1919 overlaps ``hash.fdir.lo``. Since `Action: MARK`_ sets the ``hash.fdir.hi``
1920 field only, both can be requested simultaneously.
1922 Also, regarding packet encapsulation ``level``:
1924 - ``0`` requests the default behavior. Depending on the packet type, it can
1925 mean outermost, innermost, anything in between or even no RSS.
1927 It basically stands for the innermost encapsulation level RSS can be
1928 performed on according to PMD and device capabilities.
1930 - ``1`` requests RSS to be performed on the outermost packet encapsulation
1933 - ``2`` and subsequent values request RSS to be performed on the specified
1934 inner packet encapsulation level, from outermost to innermost (lower to
1937 Values other than ``0`` are not necessarily supported.
1939 Requesting a specific RSS level on unrecognized traffic results in undefined
1940 behavior. For predictable results, it is recommended to make the flow rule
1941 pattern match packet headers up to the requested encapsulation level so that
1942 only matching traffic goes through.
1944 .. _table_rte_flow_action_rss:
1948 +---------------+---------------------------------------------+
1950 +===============+=============================================+
1951 | ``func`` | RSS hash function to apply |
1952 +---------------+---------------------------------------------+
1953 | ``level`` | encapsulation level for ``types`` |
1954 +---------------+---------------------------------------------+
1955 | ``types`` | specific RSS hash types (see ``ETH_RSS_*``) |
1956 +---------------+---------------------------------------------+
1957 | ``key_len`` | hash key length in bytes |
1958 +---------------+---------------------------------------------+
1959 | ``queue_num`` | number of entries in ``queue`` |
1960 +---------------+---------------------------------------------+
1961 | ``key`` | hash key |
1962 +---------------+---------------------------------------------+
1963 | ``queue`` | queue indices to use |
1964 +---------------+---------------------------------------------+
1969 This action is deprecated. Consider:
1970 - `Action: PORT_REPRESENTOR`_
1971 - `Action: REPRESENTED_PORT`_
1973 Directs matching traffic to the physical function (PF) of the current
1978 - No configurable properties.
1980 .. _table_rte_flow_action_pf:
1993 This action is deprecated. Consider:
1994 - `Action: PORT_REPRESENTOR`_
1995 - `Action: REPRESENTED_PORT`_
1997 Directs matching traffic to a given virtual function of the current device.
1999 Packets matched by a VF pattern item can be redirected to their original VF
2000 ID instead of the specified one. This parameter may not be available and is
2001 not guaranteed to work properly if the VF part is matched by a prior flow
2002 rule or if packets are not addressed to a VF in the first place.
2006 .. _table_rte_flow_action_vf:
2010 +--------------+--------------------------------+
2012 +==============+================================+
2013 | ``original`` | use original VF ID if possible |
2014 +--------------+--------------------------------+
2016 +--------------+--------------------------------+
2018 Action: ``PHY_PORT``
2019 ^^^^^^^^^^^^^^^^^^^^
2021 This action is deprecated. Consider:
2022 - `Action: PORT_REPRESENTOR`_
2023 - `Action: REPRESENTED_PORT`_
2025 Directs matching traffic to a given physical port index of the underlying
2028 See `Item: PHY_PORT`_.
2030 .. _table_rte_flow_action_phy_port:
2034 +--------------+-------------------------------------+
2036 +==============+=====================================+
2037 | ``original`` | use original port index if possible |
2038 +--------------+-------------------------------------+
2039 | ``index`` | physical port index |
2040 +--------------+-------------------------------------+
2044 This action is deprecated. Consider:
2045 - `Action: PORT_REPRESENTOR`_
2046 - `Action: REPRESENTED_PORT`_
2048 Directs matching traffic to a given DPDK port ID.
2050 See `Item: PORT_ID`_.
2052 .. _table_rte_flow_action_port_id:
2056 +--------------+---------------------------------------+
2058 +==============+=======================================+
2059 | ``original`` | use original DPDK port ID if possible |
2060 +--------------+---------------------------------------+
2061 | ``id`` | DPDK port ID |
2062 +--------------+---------------------------------------+
2067 Applies a stage of metering and policing.
2069 The metering and policing (MTR) object has to be first created using the
2070 rte_mtr_create() API function. The ID of the MTR object is specified as
2071 action parameter. More than one flow can use the same MTR object through
2072 the meter action. The MTR object can be further updated or queried using
2075 .. _table_rte_flow_action_meter:
2079 +--------------+---------------+
2081 +==============+===============+
2082 | ``mtr_id`` | MTR object ID |
2083 +--------------+---------------+
2085 Action: ``SECURITY``
2086 ^^^^^^^^^^^^^^^^^^^^
2088 Perform the security action on flows matched by the pattern items
2089 according to the configuration of the security session.
2091 This action modifies the payload of matched flows. For INLINE_CRYPTO, the
2092 security protocol headers and IV are fully provided by the application as
2093 specified in the flow pattern. The payload of matching packets is
2094 encrypted on egress, and decrypted and authenticated on ingress.
2095 For INLINE_PROTOCOL, the security protocol is fully offloaded to HW,
2096 providing full encapsulation and decapsulation of packets in security
2097 protocols. The flow pattern specifies both the outer security header fields
2098 and the inner packet fields. The security session specified in the action
2099 must match the pattern parameters.
2101 The security session specified in the action must be created on the same
2102 port as the flow action that is being specified.
2104 The ingress/egress flow attribute should match that specified in the
2105 security session if the security session supports the definition of the
2108 Multiple flows can be configured to use the same security session.
2110 .. _table_rte_flow_action_security:
2114 +----------------------+--------------------------------------+
2116 +======================+======================================+
2117 | ``security_session`` | security session to apply |
2118 +----------------------+--------------------------------------+
2120 The following is an example of configuring IPsec inline using the
2121 INLINE_CRYPTO security session:
2123 The encryption algorithm, keys and salt are part of the opaque
2124 ``rte_security_session``. The SA is identified according to the IP and ESP
2125 fields in the pattern items.
2127 .. _table_rte_flow_item_esp_inline_example:
2129 .. table:: IPsec inline crypto flow pattern items.
2131 +-------+----------+
2133 +=======+==========+
2135 +-------+----------+
2137 +-------+----------+
2139 +-------+----------+
2141 +-------+----------+
2143 .. _table_rte_flow_action_esp_inline_example:
2145 .. table:: IPsec inline flow actions.
2147 +-------+----------+
2149 +=======+==========+
2151 +-------+----------+
2153 +-------+----------+
2155 Action: ``OF_SET_MPLS_TTL``
2156 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2158 Implements ``OFPAT_SET_MPLS_TTL`` ("MPLS TTL") as defined by the `OpenFlow
2159 Switch Specification`_.
2161 .. _table_rte_flow_action_of_set_mpls_ttl:
2163 .. table:: OF_SET_MPLS_TTL
2165 +--------------+----------+
2167 +==============+==========+
2168 | ``mpls_ttl`` | MPLS TTL |
2169 +--------------+----------+
2171 Action: ``OF_DEC_MPLS_TTL``
2172 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2174 Implements ``OFPAT_DEC_MPLS_TTL`` ("decrement MPLS TTL") as defined by the
2175 `OpenFlow Switch Specification`_.
2177 .. _table_rte_flow_action_of_dec_mpls_ttl:
2179 .. table:: OF_DEC_MPLS_TTL
2187 Action: ``OF_SET_NW_TTL``
2188 ^^^^^^^^^^^^^^^^^^^^^^^^^
2190 Implements ``OFPAT_SET_NW_TTL`` ("IP TTL") as defined by the `OpenFlow
2191 Switch Specification`_.
2193 .. _table_rte_flow_action_of_set_nw_ttl:
2195 .. table:: OF_SET_NW_TTL
2197 +------------+--------+
2199 +============+========+
2200 | ``nw_ttl`` | IP TTL |
2201 +------------+--------+
2203 Action: ``OF_DEC_NW_TTL``
2204 ^^^^^^^^^^^^^^^^^^^^^^^^^
2206 Implements ``OFPAT_DEC_NW_TTL`` ("decrement IP TTL") as defined by the
2207 `OpenFlow Switch Specification`_.
2209 .. _table_rte_flow_action_of_dec_nw_ttl:
2211 .. table:: OF_DEC_NW_TTL
2219 Action: ``OF_COPY_TTL_OUT``
2220 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2222 Implements ``OFPAT_COPY_TTL_OUT`` ("copy TTL "outwards" -- from
2223 next-to-outermost to outermost") as defined by the `OpenFlow Switch
2226 .. _table_rte_flow_action_of_copy_ttl_out:
2228 .. table:: OF_COPY_TTL_OUT
2236 Action: ``OF_COPY_TTL_IN``
2237 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2239 Implements ``OFPAT_COPY_TTL_IN`` ("copy TTL "inwards" -- from outermost to
2240 next-to-outermost") as defined by the `OpenFlow Switch Specification`_.
2242 .. _table_rte_flow_action_of_copy_ttl_in:
2244 .. table:: OF_COPY_TTL_IN
2252 Action: ``OF_POP_VLAN``
2253 ^^^^^^^^^^^^^^^^^^^^^^^
2255 Implements ``OFPAT_POP_VLAN`` ("pop the outer VLAN tag") as defined
2256 by the `OpenFlow Switch Specification`_.
2258 .. _table_rte_flow_action_of_pop_vlan:
2260 .. table:: OF_POP_VLAN
2268 Action: ``OF_PUSH_VLAN``
2269 ^^^^^^^^^^^^^^^^^^^^^^^^
2271 Implements ``OFPAT_PUSH_VLAN`` ("push a new VLAN tag") as defined by the
2272 `OpenFlow Switch Specification`_.
2274 .. _table_rte_flow_action_of_push_vlan:
2276 .. table:: OF_PUSH_VLAN
2278 +---------------+-----------+
2280 +===============+===========+
2281 | ``ethertype`` | EtherType |
2282 +---------------+-----------+
2284 Action: ``OF_SET_VLAN_VID``
2285 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2287 Implements ``OFPAT_SET_VLAN_VID`` ("set the 802.1q VLAN id") as defined by
2288 the `OpenFlow Switch Specification`_.
2290 .. _table_rte_flow_action_of_set_vlan_vid:
2292 .. table:: OF_SET_VLAN_VID
2294 +--------------+---------+
2296 +==============+=========+
2297 | ``vlan_vid`` | VLAN id |
2298 +--------------+---------+
2300 Action: ``OF_SET_VLAN_PCP``
2301 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2303 Implements ``OFPAT_SET_LAN_PCP`` ("set the 802.1q priority") as defined by
2304 the `OpenFlow Switch Specification`_.
2306 .. _table_rte_flow_action_of_set_vlan_pcp:
2308 .. table:: OF_SET_VLAN_PCP
2310 +--------------+---------------+
2312 +==============+===============+
2313 | ``vlan_pcp`` | VLAN priority |
2314 +--------------+---------------+
2316 Action: ``OF_POP_MPLS``
2317 ^^^^^^^^^^^^^^^^^^^^^^^
2319 Implements ``OFPAT_POP_MPLS`` ("pop the outer MPLS tag") as defined by the
2320 `OpenFlow Switch Specification`_.
2322 .. _table_rte_flow_action_of_pop_mpls:
2324 .. table:: OF_POP_MPLS
2326 +---------------+-----------+
2328 +===============+===========+
2329 | ``ethertype`` | EtherType |
2330 +---------------+-----------+
2332 Action: ``OF_PUSH_MPLS``
2333 ^^^^^^^^^^^^^^^^^^^^^^^^
2335 Implements ``OFPAT_PUSH_MPLS`` ("push a new MPLS tag") as defined by the
2336 `OpenFlow Switch Specification`_.
2338 .. _table_rte_flow_action_of_push_mpls:
2340 .. table:: OF_PUSH_MPLS
2342 +---------------+-----------+
2344 +===============+===========+
2345 | ``ethertype`` | EtherType |
2346 +---------------+-----------+
2348 Action: ``VXLAN_ENCAP``
2349 ^^^^^^^^^^^^^^^^^^^^^^^
2351 Performs a VXLAN encapsulation action by encapsulating the matched flow in the
2352 VXLAN tunnel as defined in the``rte_flow_action_vxlan_encap`` flow items
2355 This action modifies the payload of matched flows. The flow definition specified
2356 in the ``rte_flow_action_tunnel_encap`` action structure must define a valid
2357 VLXAN network overlay which conforms with RFC 7348 (Virtual eXtensible Local
2358 Area Network (VXLAN): A Framework for Overlaying Virtualized Layer 2 Networks
2359 over Layer 3 Networks). The pattern must be terminated with the
2360 RTE_FLOW_ITEM_TYPE_END item type.
2362 .. _table_rte_flow_action_vxlan_encap:
2364 .. table:: VXLAN_ENCAP
2366 +----------------+-------------------------------------+
2368 +================+=====================================+
2369 | ``definition`` | Tunnel end-point overlay definition |
2370 +----------------+-------------------------------------+
2372 .. _table_rte_flow_action_vxlan_encap_example:
2374 .. table:: IPv4 VxLAN flow pattern example.
2376 +-------+----------+
2378 +=======+==========+
2380 +-------+----------+
2382 +-------+----------+
2384 +-------+----------+
2386 +-------+----------+
2388 +-------+----------+
2390 Action: ``VXLAN_DECAP``
2391 ^^^^^^^^^^^^^^^^^^^^^^^
2393 Performs a decapsulation action by stripping all headers of the VXLAN tunnel
2394 network overlay from the matched flow.
2396 The flow items pattern defined for the flow rule with which a ``VXLAN_DECAP``
2397 action is specified, must define a valid VXLAN tunnel as per RFC7348. If the
2398 flow pattern does not specify a valid VXLAN tunnel then a
2399 RTE_FLOW_ERROR_TYPE_ACTION error should be returned.
2401 This action modifies the payload of matched flows.
2403 Action: ``NVGRE_ENCAP``
2404 ^^^^^^^^^^^^^^^^^^^^^^^
2406 Performs a NVGRE encapsulation action by encapsulating the matched flow in the
2407 NVGRE tunnel as defined in the``rte_flow_action_tunnel_encap`` flow item
2410 This action modifies the payload of matched flows. The flow definition specified
2411 in the ``rte_flow_action_tunnel_encap`` action structure must defined a valid
2412 NVGRE network overlay which conforms with RFC 7637 (NVGRE: Network
2413 Virtualization Using Generic Routing Encapsulation). The pattern must be
2414 terminated with the RTE_FLOW_ITEM_TYPE_END item type.
2416 .. _table_rte_flow_action_nvgre_encap:
2418 .. table:: NVGRE_ENCAP
2420 +----------------+-------------------------------------+
2422 +================+=====================================+
2423 | ``definition`` | NVGRE end-point overlay definition |
2424 +----------------+-------------------------------------+
2426 .. _table_rte_flow_action_nvgre_encap_example:
2428 .. table:: IPv4 NVGRE flow pattern example.
2430 +-------+----------+
2432 +=======+==========+
2434 +-------+----------+
2436 +-------+----------+
2438 +-------+----------+
2440 +-------+----------+
2442 Action: ``NVGRE_DECAP``
2443 ^^^^^^^^^^^^^^^^^^^^^^^
2445 Performs a decapsulation action by stripping all headers of the NVGRE tunnel
2446 network overlay from the matched flow.
2448 The flow items pattern defined for the flow rule with which a ``NVGRE_DECAP``
2449 action is specified, must define a valid NVGRE tunnel as per RFC7637. If the
2450 flow pattern does not specify a valid NVGRE tunnel then a
2451 RTE_FLOW_ERROR_TYPE_ACTION error should be returned.
2453 This action modifies the payload of matched flows.
2455 Action: ``RAW_ENCAP``
2456 ^^^^^^^^^^^^^^^^^^^^^
2458 Adds outer header whose template is provided in its data buffer,
2459 as defined in the ``rte_flow_action_raw_encap`` definition.
2461 This action modifies the payload of matched flows. The data supplied must
2462 be a valid header, either holding layer 2 data in case of adding layer 2 after
2463 decap layer 3 tunnel (for example MPLSoGRE) or complete tunnel definition
2464 starting from layer 2 and moving to the tunnel item itself. When applied to
2465 the original packet the resulting packet must be a valid packet.
2467 .. _table_rte_flow_action_raw_encap:
2469 .. table:: RAW_ENCAP
2471 +----------------+----------------------------------------+
2473 +================+========================================+
2474 | ``data`` | Encapsulation data |
2475 +----------------+----------------------------------------+
2476 | ``preserve`` | Bit-mask of data to preserve on output |
2477 +----------------+----------------------------------------+
2478 | ``size`` | Size of data and preserve |
2479 +----------------+----------------------------------------+
2481 Action: ``RAW_DECAP``
2482 ^^^^^^^^^^^^^^^^^^^^^^^
2484 Remove outer header whose template is provided in its data buffer,
2485 as defined in the ``rte_flow_action_raw_decap``
2487 This action modifies the payload of matched flows. The data supplied must
2488 be a valid header, either holding layer 2 data in case of removing layer 2
2489 before encapsulation of layer 3 tunnel (for example MPLSoGRE) or complete
2490 tunnel definition starting from layer 2 and moving to the tunnel item itself.
2491 When applied to the original packet the resulting packet must be a
2494 .. _table_rte_flow_action_raw_decap:
2496 .. table:: RAW_DECAP
2498 +----------------+----------------------------------------+
2500 +================+========================================+
2501 | ``data`` | Decapsulation data |
2502 +----------------+----------------------------------------+
2503 | ``size`` | Size of data |
2504 +----------------+----------------------------------------+
2506 Action: ``SET_IPV4_SRC``
2507 ^^^^^^^^^^^^^^^^^^^^^^^^
2509 Set a new IPv4 source address in the outermost IPv4 header.
2511 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV4 flow pattern item.
2512 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2514 .. _table_rte_flow_action_set_ipv4_src:
2516 .. table:: SET_IPV4_SRC
2518 +-----------------------------------------+
2520 +===============+=========================+
2521 | ``ipv4_addr`` | new IPv4 source address |
2522 +---------------+-------------------------+
2524 Action: ``SET_IPV4_DST``
2525 ^^^^^^^^^^^^^^^^^^^^^^^^
2527 Set a new IPv4 destination address in the outermost IPv4 header.
2529 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV4 flow pattern item.
2530 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2532 .. _table_rte_flow_action_set_ipv4_dst:
2534 .. table:: SET_IPV4_DST
2536 +---------------+------------------------------+
2538 +===============+==============================+
2539 | ``ipv4_addr`` | new IPv4 destination address |
2540 +---------------+------------------------------+
2542 Action: ``SET_IPV6_SRC``
2543 ^^^^^^^^^^^^^^^^^^^^^^^^
2545 Set a new IPv6 source address in the outermost IPv6 header.
2547 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV6 flow pattern item.
2548 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2550 .. _table_rte_flow_action_set_ipv6_src:
2552 .. table:: SET_IPV6_SRC
2554 +---------------+-------------------------+
2556 +===============+=========================+
2557 | ``ipv6_addr`` | new IPv6 source address |
2558 +---------------+-------------------------+
2560 Action: ``SET_IPV6_DST``
2561 ^^^^^^^^^^^^^^^^^^^^^^^^
2563 Set a new IPv6 destination address in the outermost IPv6 header.
2565 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV6 flow pattern item.
2566 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2568 .. _table_rte_flow_action_set_ipv6_dst:
2570 .. table:: SET_IPV6_DST
2572 +---------------+------------------------------+
2574 +===============+==============================+
2575 | ``ipv6_addr`` | new IPv6 destination address |
2576 +---------------+------------------------------+
2578 Action: ``SET_TP_SRC``
2579 ^^^^^^^^^^^^^^^^^^^^^^^^^
2581 Set a new source port number in the outermost TCP/UDP header.
2583 It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP or RTE_FLOW_ITEM_TYPE_UDP
2584 flow pattern item. Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2586 .. _table_rte_flow_action_set_tp_src:
2588 .. table:: SET_TP_SRC
2590 +----------+-------------------------+
2592 +==========+=========================+
2593 | ``port`` | new TCP/UDP source port |
2594 +---------------+--------------------+
2596 Action: ``SET_TP_DST``
2597 ^^^^^^^^^^^^^^^^^^^^^^^^^
2599 Set a new destination port number in the outermost TCP/UDP header.
2601 It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP or RTE_FLOW_ITEM_TYPE_UDP
2602 flow pattern item. Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2604 .. _table_rte_flow_action_set_tp_dst:
2606 .. table:: SET_TP_DST
2608 +----------+------------------------------+
2610 +==========+==============================+
2611 | ``port`` | new TCP/UDP destination port |
2612 +---------------+-------------------------+
2614 Action: ``MAC_SWAP``
2615 ^^^^^^^^^^^^^^^^^^^^^^^^^
2617 Swap the source and destination MAC addresses in the outermost Ethernet
2620 It must be used with a valid RTE_FLOW_ITEM_TYPE_ETH flow pattern item.
2621 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2623 .. _table_rte_flow_action_mac_swap:
2638 If there is no valid RTE_FLOW_ITEM_TYPE_IPV4 or RTE_FLOW_ITEM_TYPE_IPV6
2639 in pattern, Some PMDs will reject rule because behavior will be undefined.
2641 .. _table_rte_flow_action_dec_ttl:
2654 Assigns a new TTL value.
2656 If there is no valid RTE_FLOW_ITEM_TYPE_IPV4 or RTE_FLOW_ITEM_TYPE_IPV6
2657 in pattern, Some PMDs will reject rule because behavior will be undefined.
2659 .. _table_rte_flow_action_set_ttl:
2663 +---------------+--------------------+
2665 +===============+====================+
2666 | ``ttl_value`` | new TTL value |
2667 +---------------+--------------------+
2669 Action: ``SET_MAC_SRC``
2670 ^^^^^^^^^^^^^^^^^^^^^^^
2672 Set source MAC address.
2674 It must be used with a valid RTE_FLOW_ITEM_TYPE_ETH flow pattern item.
2675 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2677 .. _table_rte_flow_action_set_mac_src:
2679 .. table:: SET_MAC_SRC
2681 +--------------+---------------+
2683 +==============+===============+
2684 | ``mac_addr`` | MAC address |
2685 +--------------+---------------+
2687 Action: ``SET_MAC_DST``
2688 ^^^^^^^^^^^^^^^^^^^^^^^
2690 Set destination MAC address.
2692 It must be used with a valid RTE_FLOW_ITEM_TYPE_ETH flow pattern item.
2693 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2695 .. _table_rte_flow_action_set_mac_dst:
2697 .. table:: SET_MAC_DST
2699 +--------------+---------------+
2701 +==============+===============+
2702 | ``mac_addr`` | MAC address |
2703 +--------------+---------------+
2705 Action: ``INC_TCP_SEQ``
2706 ^^^^^^^^^^^^^^^^^^^^^^^
2708 Increase sequence number in the outermost TCP header.
2709 Value to increase TCP sequence number by is a big-endian 32 bit integer.
2711 Using this action on non-matching traffic will result in undefined behavior.
2713 Action: ``DEC_TCP_SEQ``
2714 ^^^^^^^^^^^^^^^^^^^^^^^
2716 Decrease sequence number in the outermost TCP header.
2717 Value to decrease TCP sequence number by is a big-endian 32 bit integer.
2719 Using this action on non-matching traffic will result in undefined behavior.
2721 Action: ``INC_TCP_ACK``
2722 ^^^^^^^^^^^^^^^^^^^^^^^
2724 Increase acknowledgment number in the outermost TCP header.
2725 Value to increase TCP acknowledgment number by is a big-endian 32 bit integer.
2727 Using this action on non-matching traffic will result in undefined behavior.
2729 Action: ``DEC_TCP_ACK``
2730 ^^^^^^^^^^^^^^^^^^^^^^^
2732 Decrease acknowledgment number in the outermost TCP header.
2733 Value to decrease TCP acknowledgment number by is a big-endian 32 bit integer.
2735 Using this action on non-matching traffic will result in undefined behavior.
2742 Tag is a transient data used during flow matching. This is not delivered to
2743 application. Multiple tags are supported by specifying index.
2745 .. _table_rte_flow_action_set_tag:
2749 +-----------+----------------------------+
2751 +===========+============================+
2752 | ``data`` | 32 bit tag value |
2753 +-----------+----------------------------+
2754 | ``mask`` | bit-mask applies to "data" |
2755 +-----------+----------------------------+
2756 | ``index`` | index of tag to set |
2757 +-----------+----------------------------+
2759 Action: ``SET_META``
2760 ^^^^^^^^^^^^^^^^^^^^^^^
2762 Set metadata. Item ``META`` matches metadata.
2764 Metadata set by mbuf metadata field with PKT_TX_DYNF_METADATA flag on egress
2765 will be overridden by this action. On ingress, the metadata will be carried by
2766 ``metadata`` dynamic field of ``rte_mbuf`` which can be accessed by
2767 ``RTE_FLOW_DYNF_METADATA()``. PKT_RX_DYNF_METADATA flag will be set along
2770 The mbuf dynamic field must be registered by calling
2771 ``rte_flow_dynf_metadata_register()`` prior to use ``SET_META`` action.
2773 Altering partial bits is supported with ``mask``. For bits which have never been
2774 set, unpredictable value will be seen depending on driver implementation. For
2775 loopback/hairpin packet, metadata set on Rx/Tx may or may not be propagated to
2776 the other path depending on HW capability.
2778 In hairpin case with Tx explicit flow mode, metadata could (not mandatory) be
2779 used to connect the Rx and Tx flows if it can be propagated from Rx to Tx path.
2781 .. _table_rte_flow_action_set_meta:
2785 +----------+----------------------------+
2787 +==========+============================+
2788 | ``data`` | 32 bit metadata value |
2789 +----------+----------------------------+
2790 | ``mask`` | bit-mask applies to "data" |
2791 +----------+----------------------------+
2793 Action: ``SET_IPV4_DSCP``
2794 ^^^^^^^^^^^^^^^^^^^^^^^^^
2798 Modify DSCP in IPv4 header.
2800 It must be used with RTE_FLOW_ITEM_TYPE_IPV4 in pattern.
2801 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2803 .. _table_rte_flow_action_set_ipv4_dscp:
2805 .. table:: SET_IPV4_DSCP
2807 +-----------+---------------------------------+
2809 +===========+=================================+
2810 | ``dscp`` | DSCP in low 6 bits, rest ignore |
2811 +-----------+---------------------------------+
2813 Action: ``SET_IPV6_DSCP``
2814 ^^^^^^^^^^^^^^^^^^^^^^^^^
2818 Modify DSCP in IPv6 header.
2820 It must be used with RTE_FLOW_ITEM_TYPE_IPV6 in pattern.
2821 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2823 .. _table_rte_flow_action_set_ipv6_dscp:
2825 .. table:: SET_IPV6_DSCP
2827 +-----------+---------------------------------+
2829 +===========+=================================+
2830 | ``dscp`` | DSCP in low 6 bits, rest ignore |
2831 +-----------+---------------------------------+
2836 Set ageing timeout configuration to a flow.
2838 Event RTE_ETH_EVENT_FLOW_AGED will be reported if
2839 timeout passed without any matching on the flow.
2841 .. _table_rte_flow_action_age:
2845 +--------------+---------------------------------+
2847 +==============+=================================+
2848 | ``timeout`` | 24 bits timeout value |
2849 +--------------+---------------------------------+
2850 | ``reserved`` | 8 bits reserved, must be zero |
2851 +--------------+---------------------------------+
2852 | ``context`` | user input flow context |
2853 +--------------+---------------------------------+
2855 Query structure to retrieve ageing status information of a
2856 shared AGE action, or a flow rule using the AGE action:
2858 .. _table_rte_flow_query_age:
2860 .. table:: AGE query
2862 +------------------------------+-----+----------------------------------------+
2863 | Field | I/O | Value |
2864 +==============================+=====+========================================+
2865 | ``aged`` | out | Aging timeout expired |
2866 +------------------------------+-----+----------------------------------------+
2867 | ``sec_since_last_hit_valid`` | out | ``sec_since_last_hit`` value is valid |
2868 +------------------------------+-----+----------------------------------------+
2869 | ``sec_since_last_hit`` | out | Seconds since last traffic hit |
2870 +------------------------------+-----+----------------------------------------+
2875 Adds a sample action to a matched flow.
2877 The matching packets will be duplicated with the specified ``ratio`` and
2878 applied with own set of actions with a fate action, the packets sampled
2879 equals is '1/ratio'. All the packets continue to the target destination.
2881 When the ``ratio`` is set to 1 then the packets will be 100% mirrored.
2882 ``actions`` represent the different set of actions for the sampled or mirrored
2883 packets, and must have a fate action.
2885 .. _table_rte_flow_action_sample:
2889 +--------------+---------------------------------+
2891 +==============+=================================+
2892 | ``ratio`` | 32 bits sample ratio value |
2893 +--------------+---------------------------------+
2894 | ``actions`` | sub-action list for sampling |
2895 +--------------+---------------------------------+
2897 Action: ``INDIRECT``
2898 ^^^^^^^^^^^^^^^^^^^^
2900 Flow utilize indirect action by handle as returned from
2901 ``rte_flow_action_handle_create()``.
2903 The behaviour of the indirect action defined by ``action`` argument of type
2904 ``struct rte_flow_action`` passed to ``rte_flow_action_handle_create()``.
2906 The indirect action can be used by a single flow or shared among multiple flows.
2907 The indirect action can be in-place updated by ``rte_flow_action_handle_update()``
2908 without destroying flow and creating flow again. The fields that could be
2909 updated depend on the type of the ``action`` and different for every type.
2911 The indirect action specified data (e.g. counter) can be queried by
2912 ``rte_flow_action_handle_query()``.
2914 .. _table_rte_flow_action_handle:
2924 Action: ``MODIFY_FIELD``
2925 ^^^^^^^^^^^^^^^^^^^^^^^^
2927 Modify ``dst`` field according to ``op`` selected (set, addition,
2928 subtraction) with ``width`` bits of data from ``src`` field.
2930 Any arbitrary header field (as well as mark, metadata or tag values)
2931 can be used as both source and destination fields as set by ``field``.
2932 The immediate value ``RTE_FLOW_FIELD_VALUE`` (or a pointer to it
2933 ``RTE_FLOW_FIELD_POINTER``) is allowed as a source only.
2934 ``RTE_FLOW_FIELD_START`` is used to point to the beginning of a packet.
2935 See ``enum rte_flow_field_id`` for the list of supported fields.
2937 ``op`` selects the operation to perform on a destination field.
2938 - ``set`` copies the data from ``src`` field to ``dst`` field.
2939 - ``add`` adds together ``dst`` and ``src`` and stores the result into ``dst``.
2940 - ``sub`` subtracts ``src`` from ``dst`` and stores the result into ``dst``
2942 ``width`` defines a number of bits to use from ``src`` field.
2944 ``level`` is used to access any packet field on any encapsulation level
2945 as well as any tag element in the tag array.
2946 - ``0`` means the default behaviour. Depending on the packet type, it can
2947 mean outermost, innermost or anything in between.
2948 - ``1`` requests access to the outermost packet encapsulation level.
2949 - ``2`` and subsequent values requests access to the specified packet
2950 encapsulation level, from outermost to innermost (lower to higher values).
2951 For the tag array (in case of multiple tags are supported and present)
2952 ``level`` translates directly into the array index.
2954 ``offset`` specifies the number of bits to skip from a field's start.
2955 That allows performing a partial copy of the needed part or to divide a big
2956 packet field into multiple smaller fields. Alternatively, ``offset`` allows
2957 going past the specified packet field boundary to copy a field to an
2958 arbitrary place in a packet, essentially providing a way to copy any part of
2959 a packet to any other part of it.
2961 ``value`` sets an immediate value to be used as a source or points to a
2962 location of the value in memory. It is used instead of ``level`` and ``offset``
2963 for ``RTE_FLOW_FIELD_VALUE`` and ``RTE_FLOW_FIELD_POINTER`` respectively.
2965 .. _table_rte_flow_action_modify_field:
2967 .. table:: MODIFY_FIELD
2969 +---------------+-------------------------+
2971 +===============+=========================+
2972 | ``op`` | operation to perform |
2973 +---------------+-------------------------+
2974 | ``dst`` | destination field |
2975 +---------------+-------------------------+
2976 | ``src`` | source field |
2977 +---------------+-------------------------+
2978 | ``width`` | number of bits to use |
2979 +---------------+-------------------------+
2981 .. _table_rte_flow_action_modify_data:
2983 .. table:: destination/source field definition
2985 +---------------+----------------------------------------------------------+
2987 +===============+==========================================================+
2988 | ``field`` | ID: packet field, mark, meta, tag, immediate, pointer |
2989 +---------------+----------------------------------------------------------+
2990 | ``level`` | encapsulation level of a packet field or tag array index |
2991 +---------------+----------------------------------------------------------+
2992 | ``offset`` | number of bits to skip at the beginning |
2993 +---------------+----------------------------------------------------------+
2994 | ``value`` | immediate value or a pointer to this value |
2995 +---------------+----------------------------------------------------------+
2997 Action: ``CONNTRACK``
2998 ^^^^^^^^^^^^^^^^^^^^^
3000 Create a conntrack (connection tracking) context with the provided information.
3002 In stateful session like TCP, the conntrack action provides the ability to
3003 examine every packet of this connection and associate the state to every
3004 packet. It will help to realize the stateful offload of connections with little
3005 software participation. For example, the packets with invalid state may be
3006 handled by the software. The control packets could be handled in the hardware.
3007 The software just need to query the state of a connection when needed, and then
3008 decide how to handle the flow rules and conntrack context.
3010 A conntrack context should be created via ``rte_flow_action_handle_create()``
3011 before using. Then the handle with ``INDIRECT`` type is used for a flow rule
3012 creation. If a flow rule with an opposite direction needs to be created, the
3013 ``rte_flow_action_handle_update()`` should be used to modify the direction.
3015 Not all the fields of the ``struct rte_flow_action_conntrack`` will be used
3016 for a conntrack context creating, depending on the HW, and they should be
3017 in host byte order. PMD should convert them into network byte order when
3020 The ``struct rte_flow_modify_conntrack`` should be used for an updating.
3022 The current conntrack context information could be queried via the
3023 ``rte_flow_action_handle_query()`` interface.
3025 .. _table_rte_flow_action_conntrack:
3027 .. table:: CONNTRACK
3029 +--------------------------+-------------------------------------------------------------+
3031 +==========================+=============================================================+
3032 | ``peer_port`` | peer port number |
3033 +--------------------------+-------------------------------------------------------------+
3034 | ``is_original_dir`` | direction of this connection for creating flow rule |
3035 +--------------------------+-------------------------------------------------------------+
3036 | ``enable`` | enable the conntrack context |
3037 +--------------------------+-------------------------------------------------------------+
3038 | ``live_connection`` | one ack was seen for this connection |
3039 +--------------------------+-------------------------------------------------------------+
3040 | ``selective_ack`` | SACK enabled |
3041 +--------------------------+-------------------------------------------------------------+
3042 | ``challenge_ack_passed`` | a challenge ack has passed |
3043 +--------------------------+-------------------------------------------------------------+
3044 | ``last_direction`` | direction of the last passed packet |
3045 +--------------------------+-------------------------------------------------------------+
3046 | ``liberal_mode`` | only report state change |
3047 +--------------------------+-------------------------------------------------------------+
3048 | ``state`` | current state |
3049 +--------------------------+-------------------------------------------------------------+
3050 | ``max_ack_window`` | maximal window scaling factor |
3051 +--------------------------+-------------------------------------------------------------+
3052 | ``retransmission_limit`` | maximal retransmission times |
3053 +--------------------------+-------------------------------------------------------------+
3054 | ``original_dir`` | TCP parameters of the original direction |
3055 +--------------------------+-------------------------------------------------------------+
3056 | ``reply_dir`` | TCP parameters of the reply direction |
3057 +--------------------------+-------------------------------------------------------------+
3058 | ``last_window`` | window size of the last passed packet |
3059 +--------------------------+-------------------------------------------------------------+
3060 | ``last_seq`` | sequence number of the last passed packet |
3061 +--------------------------+-------------------------------------------------------------+
3062 | ``last_ack`` | acknowledgment number the last passed packet |
3063 +--------------------------+-------------------------------------------------------------+
3064 | ``last_end`` | sum of ack number and length of the last passed packet |
3065 +--------------------------+-------------------------------------------------------------+
3067 .. _table_rte_flow_tcp_dir_param:
3069 .. table:: configuration parameters for each direction
3071 +---------------------+---------------------------------------------------------+
3073 +=====================+=========================================================+
3074 | ``scale`` | TCP window scaling factor |
3075 +---------------------+---------------------------------------------------------+
3076 | ``close_initiated`` | FIN sent from this direction |
3077 +---------------------+---------------------------------------------------------+
3078 | ``last_ack_seen`` | an ACK packet received |
3079 +---------------------+---------------------------------------------------------+
3080 | ``data_unacked`` | unacknowledged data for packets from this direction |
3081 +---------------------+---------------------------------------------------------+
3082 | ``sent_end`` | max{seq + len} seen in sent packets |
3083 +---------------------+---------------------------------------------------------+
3084 | ``reply_end`` | max{sack + max{win, 1}} seen in reply packets |
3085 +---------------------+---------------------------------------------------------+
3086 | ``max_win`` | max{max{win, 1}} + {sack - ack} seen in sent packets |
3087 +---------------------+---------------------------------------------------------+
3088 | ``max_ack`` | max{ack} + seen in sent packets |
3089 +---------------------+---------------------------------------------------------+
3091 .. _table_rte_flow_modify_conntrack:
3093 .. table:: update a conntrack context
3095 +----------------+-------------------------------------------------+
3097 +================+=================================================+
3098 | ``new_ct`` | new conntrack information |
3099 +----------------+-------------------------------------------------+
3100 | ``direction`` | direction will be updated |
3101 +----------------+-------------------------------------------------+
3102 | ``state`` | other fields except direction will be updated |
3103 +----------------+-------------------------------------------------+
3104 | ``reserved`` | reserved bits |
3105 +----------------+-------------------------------------------------+
3107 Action: ``METER_COLOR``
3108 ^^^^^^^^^^^^^^^^^^^^^^^
3110 Color the packet to reflect the meter color result.
3112 The meter action must be configured before meter color action.
3113 Meter color action is set to a color to reflect the meter color result.
3114 Set the meter color in the mbuf to the selected color.
3115 The meter color action output color is the output color of the packet,
3116 which is set in the packet meta-data (i.e. struct ``rte_mbuf::sched::color``)
3118 .. _table_rte_flow_action_meter_color:
3120 .. table:: METER_COLOR
3122 +-----------------+--------------+
3124 +=================+==============+
3125 | ``meter_color`` | Packet color |
3126 +-----------------+--------------+
3128 Action: ``PORT_REPRESENTOR``
3129 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3131 At embedded switch level, send matching traffic to the given ethdev.
3133 Term **ethdev** and the concept of **port representor** are synonymous.
3134 The **represented port** is an *entity* plugged to the embedded switch
3135 at the opposite end of the "wire" leading to the ethdev.
3139 .--------------------.
3140 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
3141 '--------------------'
3151 .----------. .--------------------.
3152 | Switch | <== | Matching Traffic |
3153 '----------' '--------------------'
3163 .--------------------.
3164 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
3165 '--------------------'
3168 - Requires `Attribute: Transfer`_.
3170 .. _table_rte_flow_action_ethdev:
3172 .. table:: ``struct rte_flow_action_ethdev``
3174 +-------------+----------------+
3176 +=============+================+
3177 | ``port_id`` | ethdev port ID |
3178 +-------------+----------------+
3180 See also `Item: PORT_REPRESENTOR`_.
3182 Action: ``REPRESENTED_PORT``
3183 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3185 At embedded switch level, send matching traffic to
3186 the entity represented by the given ethdev.
3188 Term **ethdev** and the concept of **port representor** are synonymous.
3189 The **represented port** is an *entity* plugged to the embedded switch
3190 at the opposite end of the "wire" leading to the ethdev.
3194 .--------------------.
3195 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
3196 '--------------------'
3206 .----------. .--------------------.
3207 | Switch | <== | Matching Traffic |
3208 '----------' '--------------------'
3218 .--------------------.
3219 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
3220 '--------------------'
3223 - Requires `Attribute: Transfer`_.
3225 This action is meant to use the same structure as `Action: PORT_REPRESENTOR`_.
3227 See also `Item: REPRESENTED_PORT`_.
3232 All specified pattern items (``enum rte_flow_item_type``) and actions
3233 (``enum rte_flow_action_type``) use positive identifiers.
3235 The negative space is reserved for dynamic types generated by PMDs during
3236 run-time. PMDs may encounter them as a result but must not accept negative
3237 identifiers they are not aware of.
3239 A method to generate them remains to be defined.
3241 Application may use PMD dynamic items or actions in flow rules. In that case
3242 size of configuration object in dynamic element must be a pointer size.
3247 A rather simple API with few functions is provided to fully manage flow
3250 Each created flow rule is associated with an opaque, PMD-specific handle
3251 pointer. The application is responsible for keeping it until the rule is
3254 Flows rules are represented by ``struct rte_flow`` objects.
3259 Given that expressing a definite set of device capabilities is not
3260 practical, a dedicated function is provided to check if a flow rule is
3261 supported and can be created.
3266 rte_flow_validate(uint16_t port_id,
3267 const struct rte_flow_attr *attr,
3268 const struct rte_flow_item pattern[],
3269 const struct rte_flow_action actions[],
3270 struct rte_flow_error *error);
3272 The flow rule is validated for correctness and whether it could be accepted
3273 by the device given sufficient resources. The rule is checked against the
3274 current device mode and queue configuration. The flow rule may also
3275 optionally be validated against existing flow rules and device resources.
3276 This function has no effect on the target device.
3278 The returned value is guaranteed to remain valid only as long as no
3279 successful calls to ``rte_flow_create()`` or ``rte_flow_destroy()`` are made
3280 in the meantime and no device parameter affecting flow rules in any way are
3281 modified, due to possible collisions or resource limitations (although in
3282 such cases ``EINVAL`` should not be returned).
3286 - ``port_id``: port identifier of Ethernet device.
3287 - ``attr``: flow rule attributes.
3288 - ``pattern``: pattern specification (list terminated by the END pattern
3290 - ``actions``: associated actions (list terminated by the END action).
3291 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3292 this structure in case of error only.
3296 - 0 if flow rule is valid and can be created. A negative errno value
3297 otherwise (``rte_errno`` is also set), the following errors are defined.
3298 - ``-ENOSYS``: underlying device does not support this functionality.
3299 - ``-EINVAL``: unknown or invalid rule specification.
3300 - ``-ENOTSUP``: valid but unsupported rule specification (e.g. partial
3301 bit-masks are unsupported).
3302 - ``EEXIST``: collision with an existing rule. Only returned if device
3303 supports flow rule collision checking and there was a flow rule
3304 collision. Not receiving this return code is no guarantee that creating
3305 the rule will not fail due to a collision.
3306 - ``ENOMEM``: not enough memory to execute the function, or if the device
3307 supports resource validation, resource limitation on the device.
3308 - ``-EBUSY``: action cannot be performed due to busy device resources, may
3309 succeed if the affected queues or even the entire port are in a stopped
3310 state (see ``rte_eth_dev_rx_queue_stop()`` and ``rte_eth_dev_stop()``).
3315 Creating a flow rule is similar to validating one, except the rule is
3316 actually created and a handle returned.
3321 rte_flow_create(uint16_t port_id,
3322 const struct rte_flow_attr *attr,
3323 const struct rte_flow_item pattern[],
3324 const struct rte_flow_action *actions[],
3325 struct rte_flow_error *error);
3329 - ``port_id``: port identifier of Ethernet device.
3330 - ``attr``: flow rule attributes.
3331 - ``pattern``: pattern specification (list terminated by the END pattern
3333 - ``actions``: associated actions (list terminated by the END action).
3334 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3335 this structure in case of error only.
3339 A valid handle in case of success, NULL otherwise and ``rte_errno`` is set
3340 to the positive version of one of the error codes defined for
3341 ``rte_flow_validate()``.
3346 Flow rules destruction is not automatic, and a queue or a port should not be
3347 released if any are still attached to them. Applications must take care of
3348 performing this step before releasing resources.
3353 rte_flow_destroy(uint16_t port_id,
3354 struct rte_flow *flow,
3355 struct rte_flow_error *error);
3358 Failure to destroy a flow rule handle may occur when other flow rules depend
3359 on it, and destroying it would result in an inconsistent state.
3361 This function is only guaranteed to succeed if handles are destroyed in
3362 reverse order of their creation.
3366 - ``port_id``: port identifier of Ethernet device.
3367 - ``flow``: flow rule handle to destroy.
3368 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3369 this structure in case of error only.
3373 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3378 Convenience function to destroy all flow rule handles associated with a
3379 port. They are released as with successive calls to ``rte_flow_destroy()``.
3384 rte_flow_flush(uint16_t port_id,
3385 struct rte_flow_error *error);
3387 In the unlikely event of failure, handles are still considered destroyed and
3388 no longer valid but the port must be assumed to be in an inconsistent state.
3392 - ``port_id``: port identifier of Ethernet device.
3393 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3394 this structure in case of error only.
3398 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3403 Query an existing flow rule.
3405 This function allows retrieving flow-specific data such as counters. Data
3406 is gathered by special actions which must be present in the flow rule
3412 rte_flow_query(uint16_t port_id,
3413 struct rte_flow *flow,
3414 const struct rte_flow_action *action,
3416 struct rte_flow_error *error);
3420 - ``port_id``: port identifier of Ethernet device.
3421 - ``flow``: flow rule handle to query.
3422 - ``action``: action to query, this must match prototype from flow rule.
3423 - ``data``: pointer to storage for the associated query data type.
3424 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3425 this structure in case of error only.
3429 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3431 .. _flow_isolated_mode:
3436 The general expectation for ingress traffic is that flow rules process it
3437 first; the remaining unmatched or pass-through traffic usually ends up in a
3438 queue (with or without RSS, locally or in some sub-device instance)
3439 depending on the global configuration settings of a port.
3441 While fine from a compatibility standpoint, this approach makes drivers more
3442 complex as they have to check for possible side effects outside of this API
3443 when creating or destroying flow rules. It results in a more limited set of
3444 available rule types due to the way device resources are assigned (e.g. no
3445 support for the RSS action even on capable hardware).
3447 Given that nonspecific traffic can be handled by flow rules as well,
3448 isolated mode is a means for applications to tell a driver that ingress on
3449 the underlying port must be injected from the defined flow rules only; that
3450 no default traffic is expected outside those rules.
3452 This has the following benefits:
3454 - Applications get finer-grained control over the kind of traffic they want
3455 to receive (no traffic by default).
3457 - More importantly they control at what point nonspecific traffic is handled
3458 relative to other flow rules, by adjusting priority levels.
3460 - Drivers can assign more hardware resources to flow rules and expand the
3461 set of supported rule types.
3463 Because toggling isolated mode may cause profound changes to the ingress
3464 processing path of a driver, it may not be possible to leave it once
3465 entered. Likewise, existing flow rules or global configuration settings may
3466 prevent a driver from entering isolated mode.
3468 Applications relying on this mode are therefore encouraged to toggle it as
3469 soon as possible after device initialization, ideally before the first call
3470 to ``rte_eth_dev_configure()`` to avoid possible failures due to conflicting
3473 Once effective, the following functionality has no effect on the underlying
3474 port and may return errors such as ``ENOTSUP`` ("not supported"):
3476 - Toggling promiscuous mode.
3477 - Toggling allmulticast mode.
3478 - Configuring MAC addresses.
3479 - Configuring multicast addresses.
3480 - Configuring VLAN filters.
3481 - Configuring global RSS settings.
3486 rte_flow_isolate(uint16_t port_id, int set, struct rte_flow_error *error);
3490 - ``port_id``: port identifier of Ethernet device.
3491 - ``set``: nonzero to enter isolated mode, attempt to leave it otherwise.
3492 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3493 this structure in case of error only.
3497 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3499 Verbose error reporting
3500 -----------------------
3502 The defined *errno* values may not be accurate enough for users or
3503 application developers who want to investigate issues related to flow rules
3504 management. A dedicated error object is defined for this purpose:
3508 enum rte_flow_error_type {
3509 RTE_FLOW_ERROR_TYPE_NONE, /**< No error. */
3510 RTE_FLOW_ERROR_TYPE_UNSPECIFIED, /**< Cause unspecified. */
3511 RTE_FLOW_ERROR_TYPE_HANDLE, /**< Flow rule (handle). */
3512 RTE_FLOW_ERROR_TYPE_ATTR_GROUP, /**< Group field. */
3513 RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY, /**< Priority field. */
3514 RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, /**< Ingress field. */
3515 RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, /**< Egress field. */
3516 RTE_FLOW_ERROR_TYPE_ATTR, /**< Attributes structure. */
3517 RTE_FLOW_ERROR_TYPE_ITEM_NUM, /**< Pattern length. */
3518 RTE_FLOW_ERROR_TYPE_ITEM, /**< Specific pattern item. */
3519 RTE_FLOW_ERROR_TYPE_ACTION_NUM, /**< Number of actions. */
3520 RTE_FLOW_ERROR_TYPE_ACTION, /**< Specific action. */
3523 struct rte_flow_error {
3524 enum rte_flow_error_type type; /**< Cause field and error types. */
3525 const void *cause; /**< Object responsible for the error. */
3526 const char *message; /**< Human-readable error message. */
3529 Error type ``RTE_FLOW_ERROR_TYPE_NONE`` stands for no error, in which case
3530 remaining fields can be ignored. Other error types describe the type of the
3531 object pointed by ``cause``.
3533 If non-NULL, ``cause`` points to the object responsible for the error. For a
3534 flow rule, this may be a pattern item or an individual action.
3536 If non-NULL, ``message`` provides a human-readable error message.
3538 This object is normally allocated by applications and set by PMDs in case of
3539 error, the message points to a constant string which does not need to be
3540 freed by the application, however its pointer can be considered valid only
3541 as long as its associated DPDK port remains configured. Closing the
3542 underlying device or unloading the PMD invalidates it.
3553 rte_flow_error_set(struct rte_flow_error *error,
3555 enum rte_flow_error_type type,
3557 const char *message);
3559 This function initializes ``error`` (if non-NULL) with the provided
3560 parameters and sets ``rte_errno`` to ``code``. A negative error ``code`` is
3569 rte_flow_conv(enum rte_flow_conv_op op,
3573 struct rte_flow_error *error);
3575 Convert ``src`` to ``dst`` according to operation ``op``. Possible
3578 - Attributes, pattern item or action duplication.
3579 - Duplication of an entire pattern or list of actions.
3580 - Duplication of a complete flow rule description.
3581 - Pattern item or action name retrieval.
3583 Tunneled traffic offload
3584 ~~~~~~~~~~~~~~~~~~~~~~~~
3586 rte_flow API provides the building blocks for vendor-agnostic flow
3587 classification offloads. The rte_flow "patterns" and "actions"
3588 primitives are fine-grained, thus enabling DPDK applications the
3589 flexibility to offload network stacks and complex pipelines.
3590 Applications wishing to offload tunneled traffic are required to use
3591 the rte_flow primitives, such as group, meta, mark, tag, and others to
3592 model their high-level objects. The hardware model design for
3593 high-level software objects is not trivial. Furthermore, an optimal
3594 design is often vendor-specific.
3596 When hardware offloads tunneled traffic in multi-group logic,
3597 partially offloaded packets may arrive to the application after they
3598 were modified in hardware. In this case, the application may need to
3599 restore the original packet headers. Consider the following sequence:
3600 The application decaps a packet in one group and jumps to a second
3601 group where it tries to match on a 5-tuple, that will miss and send
3602 the packet to the application. In this case, the application does not
3603 receive the original packet but a modified one. Also, in this case,
3604 the application cannot match on the outer header fields, such as VXLAN
3607 There are several possible ways to use rte_flow "patterns" and
3608 "actions" to resolve the issues above. For example:
3610 1 Mapping headers to a hardware registers using the
3611 rte_flow_action_mark/rte_flow_action_tag/rte_flow_set_meta objects.
3613 2 Apply the decap only at the last offload stage after all the
3614 "patterns" were matched and the packet will be fully offloaded.
3616 Every approach has its pros and cons and is highly dependent on the
3617 hardware vendor. For example, some hardware may have a limited number
3618 of registers while other hardware could not support inner actions and
3619 must decap before accessing inner headers.
3621 The tunnel offload model resolves these issues. The model goals are:
3623 1 Provide a unified application API to offload tunneled traffic that
3624 is capable to match on outer headers after decap.
3626 2 Allow the application to restore the outer header of partially
3629 The tunnel offload model does not introduce new elements to the
3630 existing RTE flow model and is implemented as a set of helper
3633 For the application to work with the tunnel offload API it
3634 has to adjust flow rules in multi-table tunnel offload in the
3637 1 Remove explicit call to decap action and replace it with PMD actions
3638 obtained from rte_flow_tunnel_decap_and_set() helper.
3640 2 Add PMD items obtained from rte_flow_tunnel_match() helper to all
3641 other rules in the tunnel offload sequence.
3643 The model requirements:
3645 Software application must initialize
3646 rte_tunnel object with tunnel parameters before calling
3647 rte_flow_tunnel_decap_set() & rte_flow_tunnel_match().
3649 PMD actions array obtained in rte_flow_tunnel_decap_set() must be
3650 released by application with rte_flow_action_release() call.
3652 PMD items array obtained with rte_flow_tunnel_match() must be released
3653 by application with rte_flow_item_release() call. Application can
3654 release PMD items and actions after rule was created. However, if the
3655 application needs to create additional rule for the same tunnel it
3656 will need to obtain PMD items again.
3658 Application cannot destroy rte_tunnel object before it releases all
3659 PMD actions & PMD items referencing that tunnel.
3664 - DPDK does not keep track of flow rules definitions or flow rule objects
3665 automatically. Applications may keep track of the former and must keep
3666 track of the latter. PMDs may also do it for internal needs, however this
3667 must not be relied on by applications.
3669 - Flow rules are not maintained between successive port initializations. An
3670 application exiting without releasing them and restarting must re-create
3673 - API operations are synchronous and blocking (``EAGAIN`` cannot be
3676 - Stopping the data path (TX/RX) should not be necessary when managing flow
3677 rules. If this cannot be achieved naturally or with workarounds (such as
3678 temporarily replacing the burst function pointers), an appropriate error
3679 code must be returned (``EBUSY``).
3681 - Applications, not PMDs, are responsible for maintaining flow rules
3682 configuration when closing, stopping or restarting a port or performing other
3683 actions which may affect them.
3684 Applications must assume that after port close, stop or restart all flows
3685 related to that port are not valid, hardware rules are destroyed and relevant
3686 PMD resources are released.
3688 For devices exposing multiple ports sharing global settings affected by flow
3691 - All ports under DPDK control must behave consistently, PMDs are
3692 responsible for making sure that existing flow rules on a port are not
3693 affected by other ports.
3695 - Ports not under DPDK control (unaffected or handled by other applications)
3696 are user's responsibility. They may affect existing flow rules and cause
3697 undefined behavior. PMDs aware of this may prevent flow rules creation
3698 altogether in such cases.
3703 The PMD interface is defined in ``rte_flow_driver.h``. It is not subject to
3704 API/ABI versioning constraints as it is not exposed to applications and may
3705 evolve independently.
3707 The PMD interface is based on callbacks pointed by the ``struct rte_flow_ops``.
3709 - PMD callbacks implement exactly the interface described in `Rules
3710 management`_, except for the port ID argument which has already been
3711 converted to a pointer to the underlying ``struct rte_eth_dev``.
3713 - Public API functions do not process flow rules definitions at all before
3714 calling PMD functions (no basic error checking, no validation
3715 whatsoever). They only make sure these callbacks are non-NULL or return
3716 the ``ENOSYS`` (function not supported) error.
3718 This interface additionally defines the following helper function:
3720 - ``rte_flow_ops_get()``: get generic flow operations structure from a
3723 If PMD interfaces don't support re-entrancy/multi-thread safety,
3724 the rte_flow API functions will protect threads by mutex per port.
3725 The application can check whether ``RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE``
3726 is set in ``dev_flags``, meaning the PMD is thread-safe regarding rte_flow,
3727 so the API level protection is disabled.
3728 Please note that this API-level mutex protects only rte_flow functions,
3729 other control path functions are not in scope.
3731 Device compatibility
3732 --------------------
3734 No known implementation supports all the described features.
3736 Unsupported features or combinations are not expected to be fully emulated
3737 in software by PMDs for performance reasons. Partially supported features
3738 may be completed in software as long as hardware performs most of the work
3739 (such as queue redirection and packet recognition).
3741 However PMDs are expected to do their best to satisfy application requests
3742 by working around hardware limitations as long as doing so does not affect
3743 the behavior of existing flow rules.
3745 The following sections provide a few examples of such cases and describe how
3746 PMDs should handle them, they are based on limitations built into the
3752 Each flow rule comes with its own, per-layer bit-masks, while hardware may
3753 support only a single, device-wide bit-mask for a given layer type, so that
3754 two IPv4 rules cannot use different bit-masks.
3756 The expected behavior in this case is that PMDs automatically configure
3757 global bit-masks according to the needs of the first flow rule created.
3759 Subsequent rules are allowed only if their bit-masks match those, the
3760 ``EEXIST`` error code should be returned otherwise.
3762 Unsupported layer types
3763 ~~~~~~~~~~~~~~~~~~~~~~~
3765 Many protocols can be simulated by crafting patterns with the `Item: RAW`_
3768 PMDs can rely on this capability to simulate support for protocols with
3769 headers not directly recognized by hardware.
3771 ``ANY`` pattern item
3772 ~~~~~~~~~~~~~~~~~~~~
3774 This pattern item stands for anything, which can be difficult to translate
3775 to something hardware would understand, particularly if followed by more
3778 Consider the following pattern:
3780 .. _table_rte_flow_unsupported_any:
3782 .. table:: Pattern with ANY as L3
3784 +-------+-----------------------+
3786 +=======+=======================+
3788 +-------+-----+---------+-------+
3789 | 1 | ANY | ``num`` | ``1`` |
3790 +-------+-----+---------+-------+
3792 +-------+-----------------------+
3794 +-------+-----------------------+
3796 Knowing that TCP does not make sense with something other than IPv4 and IPv6
3797 as L3, such a pattern may be translated to two flow rules instead:
3799 .. _table_rte_flow_unsupported_any_ipv4:
3801 .. table:: ANY replaced with IPV4
3803 +-------+--------------------+
3805 +=======+====================+
3807 +-------+--------------------+
3808 | 1 | IPV4 (zeroed mask) |
3809 +-------+--------------------+
3811 +-------+--------------------+
3813 +-------+--------------------+
3817 .. _table_rte_flow_unsupported_any_ipv6:
3819 .. table:: ANY replaced with IPV6
3821 +-------+--------------------+
3823 +=======+====================+
3825 +-------+--------------------+
3826 | 1 | IPV6 (zeroed mask) |
3827 +-------+--------------------+
3829 +-------+--------------------+
3831 +-------+--------------------+
3833 Note that as soon as a ANY rule covers several layers, this approach may
3834 yield a large number of hidden flow rules. It is thus suggested to only
3835 support the most common scenarios (anything as L2 and/or L3).
3840 - When combined with `Action: QUEUE`_, packet counting (`Action: COUNT`_)
3841 and tagging (`Action: MARK`_ or `Action: FLAG`_) may be implemented in
3842 software as long as the target queue is used by a single rule.
3844 - When a single target queue is provided, `Action: RSS`_ can also be
3845 implemented through `Action: QUEUE`_.
3850 While it would naturally make sense, flow rules cannot be assumed to be
3851 processed by hardware in the same order as their creation for several
3854 - They may be managed internally as a tree or a hash table instead of a
3856 - Removing a flow rule before adding another one can either put the new rule
3857 at the end of the list or reuse a freed entry.
3858 - Duplication may occur when packets are matched by several rules.
3860 For overlapping rules (particularly in order to use `Action: PASSTHRU`_)
3861 predictable behavior is only guaranteed by using different priority levels.
3863 Priority levels are not necessarily implemented in hardware, or may be
3864 severely limited (e.g. a single priority bit).
3866 For these reasons, priority levels may be implemented purely in software by
3869 - For devices expecting flow rules to be added in the correct order, PMDs
3870 may destroy and re-create existing rules after adding a new one with
3873 - A configurable number of dummy or empty rules can be created at
3874 initialization time to save high priority slots for later.
3876 - In order to save priority levels, PMDs may evaluate whether rules are
3877 likely to collide and adjust their priority accordingly.
3880 .. _OpenFlow Switch Specification: https://www.opennetworking.org/software-defined-standards/specifications/