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 traffic, alter its fate and query related counters according to any
13 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
92 The following description of rule persistence is an experimental behavior
93 that may change without a prior notice.
95 When the device is stopped, its rules do not process the traffic.
96 In particular, transfer rules created using some device
97 stop affecting the traffic even if they refer to different ports.
99 If ``RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP`` is not advertised,
100 rules cannot be created until the device is started for the first time
101 and cannot be kept when the device is stopped.
102 However, PMD also does not flush them automatically on stop,
103 so the application must call ``rte_flow_flush()`` or ``rte_flow_destroy()``
104 before stopping the device to ensure no rules remain.
106 If ``RTE_ETH_DEV_CAPA_FLOW_RULE_KEEP`` is advertised, this means
107 the PMD can keep at least some rules across the device stop and start.
108 However, ``rte_eth_dev_configure()`` may fail if any rules remain,
109 so the application must flush them before attempting a reconfiguration.
110 Keeping may be unsupported for some types of rule items and actions,
111 as well as depending on the value of flow attributes transfer bit.
112 A combination of a single an item or action type
113 and a value of the transfer bit is called a rule feature.
114 For example: a COUNT action with the transfer bit set.
115 To test if rules with a particular feature are kept, the application must try
116 to create a valid rule using this feature when the device is not started
117 (either before the first start or after a stop).
118 If it fails with an error of type ``RTE_FLOW_ERROR_TYPE_STATE``,
119 all rules using this feature must be flushed by the application
120 before stopping the device.
121 If it succeeds, such rules will be kept when the device is stopped,
122 provided they do not use other features that are not supported.
123 Rules that are created when the device is stopped, including the rules
124 created for the test, will be kept after the device is started.
126 The following sections cover:
128 - **Attributes** (represented by ``struct rte_flow_attr``): properties of a
129 flow rule such as its direction (ingress or egress) and priority.
131 - **Pattern item** (represented by ``struct rte_flow_item``): part of a
132 matching pattern that either matches specific packet data or traffic
133 properties. It can also describe properties of the pattern itself, such as
136 - **Matching pattern**: traffic properties to look for, a combination of any
139 - **Actions** (represented by ``struct rte_flow_action``): operations to
140 perform whenever a packet is matched by a pattern.
148 Flow rules can be grouped by assigning them a common group number. Groups
149 allow a logical hierarchy of flow rule groups (tables) to be defined. These
150 groups can be supported virtually in the PMD or in the physical device.
151 Group 0 is the default group and this is the only group which flows are
152 guarantee to matched against, all subsequent groups can only be reached by
153 way of the JUMP action from a matched flow rule.
155 Although optional, applications are encouraged to group similar rules as
156 much as possible to fully take advantage of hardware capabilities
157 (e.g. optimized matching) and work around limitations (e.g. a single pattern
158 type possibly allowed in a given group), while being aware that the groups
159 hierarchies must be programmed explicitly.
161 Note that support for more than a single group is not guaranteed.
166 A priority level can be assigned to a flow rule, lower values
167 denote higher priority, with 0 as the maximum.
169 Priority levels are arbitrary and up to the application, they do
170 not need to be contiguous nor start from 0, however the maximum number
171 varies between devices and may be affected by existing flow rules.
173 A flow which matches multiple rules in the same group will always matched by
174 the rule with the highest priority in that group.
176 If a packet is matched by several rules of a given group for a given
177 priority level, the outcome is undefined. It can take any path, may be
178 duplicated or even cause unrecoverable errors.
180 Note that support for more than a single priority level is not guaranteed.
182 Attribute: Traffic direction
183 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
185 Unless `Attribute: Transfer`_ is specified, flow rule patterns apply
186 to inbound and / or outbound traffic. With this respect, ``ingress``
187 and ``egress`` respectively stand for **inbound** and **outbound**
188 based on the standpoint of the application creating a flow rule.
190 Several pattern items and actions are valid and can be used in both
191 directions. At least one direction must be specified.
193 Specifying both directions at once for a given rule is not recommended but
194 may be valid in a few cases.
199 Instead of simply matching the properties of traffic as it would appear on a
200 given DPDK port ID, enabling this attribute transfers a flow rule to the
201 lowest possible level of any device endpoints found in the pattern.
203 When supported, this effectively enables an application to reroute traffic
204 not necessarily intended for it (e.g. coming from or addressed to different
205 physical ports, VFs or applications) at the device level.
207 In "transfer" flows, the use of `Attribute: Traffic direction`_ in the sense of
208 implicitly matching packets going to or going from the ethdev used to create
209 flow rules is **deprecated**. `Attribute: Transfer`_ shifts the viewpoint to
210 the embedded switch. In it, `Attribute: Traffic direction`_ is ambiguous as
211 the switch serves many different endpoints. The application should match
212 traffic originating from precise locations. To do so, it should
213 use `Item: PORT_REPRESENTOR`_ and `Item: REPRESENTED_PORT`_.
218 Pattern items fall in two categories:
220 - Matching protocol headers and packet data, usually associated with a
221 specification structure. These must be stacked in the same order as the
222 protocol layers to match inside packets, starting from the lowest.
224 - Matching meta-data or affecting pattern processing, often without a
225 specification structure. Since they do not match packet contents, their
226 position in the list is usually not relevant.
228 Item specification structures are used to match specific values among
229 protocol fields (or item properties). Documentation describes for each item
230 whether they are associated with one and their type name if so.
232 Up to three structures of the same type can be set for a given item:
234 - ``spec``: values to match (e.g. a given IPv4 address).
236 - ``last``: upper bound for an inclusive range with corresponding fields in
239 - ``mask``: bit-mask applied to both ``spec`` and ``last`` whose purpose is
240 to distinguish the values to take into account and/or partially mask them
241 out (e.g. in order to match an IPv4 address prefix).
243 Usage restrictions and expected behavior:
245 - Setting either ``mask`` or ``last`` without ``spec`` is an error.
247 - Field values in ``last`` which are either 0 or equal to the corresponding
248 values in ``spec`` are ignored; they do not generate a range. Nonzero
249 values lower than those in ``spec`` are not supported.
251 - Setting ``spec`` and optionally ``last`` without ``mask`` causes the PMD
252 to use the default mask defined for that item (defined as
253 ``rte_flow_item_{name}_mask`` constants).
255 - Not setting any of them (assuming item type allows it) is equivalent to
256 providing an empty (zeroed) ``mask`` for broad (nonspecific) matching.
258 - ``mask`` is a simple bit-mask applied before interpreting the contents of
259 ``spec`` and ``last``, which may yield unexpected results if not used
260 carefully. For example, if for an IPv4 address field, ``spec`` provides
261 *10.1.2.3*, ``last`` provides *10.3.4.5* and ``mask`` provides
262 *255.255.0.0*, the effective range becomes *10.1.0.0* to *10.3.255.255*.
264 Example of an item specification matching an Ethernet header:
266 .. _table_rte_flow_pattern_item_example:
268 .. table:: Ethernet item
270 +----------+----------+-----------------------+
271 | Field | Subfield | Value |
272 +==========+==========+=======================+
273 | ``spec`` | ``src`` | ``00:00:01:02:03:04`` |
274 | +----------+-----------------------+
275 | | ``dst`` | ``00:00:2a:66:00:01`` |
276 | +----------+-----------------------+
277 | | ``type`` | ``0x22aa`` |
278 +----------+----------+-----------------------+
279 | ``last`` | unspecified |
280 +----------+----------+-----------------------+
281 | ``mask`` | ``src`` | ``00:00:ff:ff:ff:00`` |
282 | +----------+-----------------------+
283 | | ``dst`` | ``00:00:00:00:00:ff`` |
284 | +----------+-----------------------+
285 | | ``type`` | ``0x0000`` |
286 +----------+----------+-----------------------+
288 Non-masked bits stand for any value (shown as ``?`` below), Ethernet headers
289 with the following properties are thus matched:
291 - ``src``: ``??:??:01:02:03:??``
292 - ``dst``: ``??:??:??:??:??:01``
293 - ``type``: ``0x????``
298 A pattern is formed by stacking items starting from the lowest protocol
299 layer to match. This stacking restriction does not apply to meta items which
300 can be placed anywhere in the stack without affecting the meaning of the
303 Patterns are terminated by END items.
307 .. _table_rte_flow_tcpv4_as_l4:
309 .. table:: TCPv4 as L4
325 .. _table_rte_flow_tcpv6_in_vxlan:
327 .. table:: TCPv6 in VXLAN
329 +-------+------------+
331 +=======+============+
333 +-------+------------+
335 +-------+------------+
337 +-------+------------+
339 +-------+------------+
341 +-------+------------+
343 +-------+------------+
345 +-------+------------+
347 +-------+------------+
351 .. _table_rte_flow_tcpv4_as_l4_meta:
353 .. table:: TCPv4 as L4 with meta items
375 The above example shows how meta items do not affect packet data matching
376 items, as long as those remain stacked properly. The resulting matching
377 pattern is identical to "TCPv4 as L4".
379 .. _table_rte_flow_udpv6_anywhere:
381 .. table:: UDPv6 anywhere
393 If supported by the PMD, omitting one or several protocol layers at the
394 bottom of the stack as in the above example (missing an Ethernet
395 specification) enables looking up anywhere in packets.
397 It is unspecified whether the payload of supported encapsulations
398 (e.g. VXLAN payload) is matched by such a pattern, which may apply to inner,
399 outer or both packets.
401 .. _table_rte_flow_invalid_l3:
403 .. table:: Invalid, missing L3
415 The above pattern is invalid due to a missing L3 specification between L2
416 (Ethernet) and L4 (UDP). Doing so is only allowed at the bottom and at the
422 They match meta-data or affect pattern processing instead of matching packet
423 data directly, most of them do not need a specification structure. This
424 particularity allows them to be specified anywhere in the stack without
425 causing any side effect.
430 End marker for item lists. Prevents further processing of items, thereby
433 - Its numeric value is 0 for convenience.
434 - PMD support is mandatory.
435 - ``spec``, ``last`` and ``mask`` are ignored.
437 .. _table_rte_flow_item_end:
441 +----------+---------+
443 +==========+=========+
444 | ``spec`` | ignored |
445 +----------+---------+
446 | ``last`` | ignored |
447 +----------+---------+
448 | ``mask`` | ignored |
449 +----------+---------+
454 Used as a placeholder for convenience. It is ignored and simply discarded by
457 - PMD support is mandatory.
458 - ``spec``, ``last`` and ``mask`` are ignored.
460 .. _table_rte_flow_item_void:
464 +----------+---------+
466 +==========+=========+
467 | ``spec`` | ignored |
468 +----------+---------+
469 | ``last`` | ignored |
470 +----------+---------+
471 | ``mask`` | ignored |
472 +----------+---------+
474 One usage example for this type is generating rules that share a common
475 prefix quickly without reallocating memory, only by updating item types:
477 .. _table_rte_flow_item_void_example:
479 .. table:: TCP, UDP or ICMP as L4
481 +-------+--------------------+
483 +=======+====================+
485 +-------+--------------------+
487 +-------+------+------+------+
488 | 2 | UDP | VOID | VOID |
489 +-------+------+------+------+
490 | 3 | VOID | TCP | VOID |
491 +-------+------+------+------+
492 | 4 | VOID | VOID | ICMP |
493 +-------+------+------+------+
495 +-------+--------------------+
500 Inverted matching, i.e. process packets that do not match the pattern.
502 - ``spec``, ``last`` and ``mask`` are ignored.
504 .. _table_rte_flow_item_invert:
508 +----------+---------+
510 +==========+=========+
511 | ``spec`` | ignored |
512 +----------+---------+
513 | ``last`` | ignored |
514 +----------+---------+
515 | ``mask`` | ignored |
516 +----------+---------+
518 Usage example, matching non-TCPv4 packets only:
520 .. _table_rte_flow_item_invert_example:
522 .. table:: Anything but TCPv4
541 This item is deprecated. Consider:
542 - `Item: PORT_REPRESENTOR`_
543 - `Item: REPRESENTED_PORT`_
545 Matches traffic originating from (ingress) or going to (egress) the physical
546 function of the current device.
548 If supported, should work even if the physical function is not managed by
549 the application and thus not associated with a DPDK port ID.
551 - Can be combined with any number of `Item: VF`_ to match both PF and VF
553 - ``spec``, ``last`` and ``mask`` must not be set.
555 .. _table_rte_flow_item_pf:
572 This item is deprecated. Consider:
573 - `Item: PORT_REPRESENTOR`_
574 - `Item: REPRESENTED_PORT`_
576 Matches traffic originating from (ingress) or going to (egress) a given
577 virtual function of the current device.
579 If supported, should work even if the virtual function is not managed by the
580 application and thus not associated with a DPDK port ID.
582 Note this pattern item does not match VF representors traffic which, as
583 separate entities, should be addressed through their own DPDK port IDs.
585 - Can be specified multiple times to match traffic addressed to several VF
587 - Can be combined with a PF item to match both PF and VF traffic.
588 - Default ``mask`` matches any VF ID.
590 .. _table_rte_flow_item_vf:
594 +----------+----------+---------------------------+
595 | Field | Subfield | Value |
596 +==========+==========+===========================+
597 | ``spec`` | ``id`` | destination VF ID |
598 +----------+----------+---------------------------+
599 | ``last`` | ``id`` | upper range value |
600 +----------+----------+---------------------------+
601 | ``mask`` | ``id`` | zeroed to match any VF ID |
602 +----------+----------+---------------------------+
607 This item is deprecated. Consider:
608 - `Item: PORT_REPRESENTOR`_
609 - `Item: REPRESENTED_PORT`_
611 Matches traffic originating from (ingress) or going to (egress) a physical
612 port of the underlying device.
614 The first PHY_PORT item overrides the physical port normally associated with
615 the specified DPDK input port (port_id). This item can be provided several
616 times to match additional physical ports.
618 Note that physical ports are not necessarily tied to DPDK input ports
619 (port_id) when those are not under DPDK control. Possible values are
620 specific to each device, they are not necessarily indexed from zero and may
623 As a device property, the list of allowed values as well as the value
624 associated with a port_id should be retrieved by other means.
626 - Default ``mask`` matches any port index.
628 .. _table_rte_flow_item_phy_port:
632 +----------+-----------+--------------------------------+
633 | Field | Subfield | Value |
634 +==========+===========+================================+
635 | ``spec`` | ``index`` | physical port index |
636 +----------+-----------+--------------------------------+
637 | ``last`` | ``index`` | upper range value |
638 +----------+-----------+--------------------------------+
639 | ``mask`` | ``index`` | zeroed to match any port index |
640 +----------+-----------+--------------------------------+
645 This item is deprecated. Consider:
646 - `Item: PORT_REPRESENTOR`_
647 - `Item: REPRESENTED_PORT`_
649 Matches traffic originating from (ingress) or going to (egress) a given DPDK
652 Normally only supported if the port ID in question is known by the
653 underlying PMD and related to the device the flow rule is created against.
655 This must not be confused with `Item: PHY_PORT`_ which refers to the
656 physical port of a device, whereas `Item: PORT_ID`_ refers to a ``struct
657 rte_eth_dev`` object on the application side (also known as "port
658 representor" depending on the kind of underlying device).
660 - Default ``mask`` matches the specified DPDK port ID.
662 .. _table_rte_flow_item_port_id:
666 +----------+----------+-----------------------------+
667 | Field | Subfield | Value |
668 +==========+==========+=============================+
669 | ``spec`` | ``id`` | DPDK port ID |
670 +----------+----------+-----------------------------+
671 | ``last`` | ``id`` | upper range value |
672 +----------+----------+-----------------------------+
673 | ``mask`` | ``id`` | zeroed to match any port ID |
674 +----------+----------+-----------------------------+
679 Matches an arbitrary integer value which was set using the ``MARK`` action in
680 a previously matched rule.
682 This item can only specified once as a match criteria as the ``MARK`` action can
683 only be specified once in a flow action.
685 Note the value of MARK field is arbitrary and application defined.
687 Depending on the underlying implementation the MARK item may be supported on
688 the physical device, with virtual groups in the PMD or not at all.
690 - Default ``mask`` matches any integer value.
692 .. _table_rte_flow_item_mark:
696 +----------+----------+---------------------------+
697 | Field | Subfield | Value |
698 +==========+==========+===========================+
699 | ``spec`` | ``id`` | integer value |
700 +----------+--------------------------------------+
701 | ``last`` | ``id`` | upper range value |
702 +----------+----------+---------------------------+
703 | ``mask`` | ``id`` | zeroed to match any value |
704 +----------+----------+---------------------------+
709 Matches tag item set by other flows. Multiple tags are supported by specifying
712 - Default ``mask`` matches the specified tag value and index.
714 .. _table_rte_flow_item_tag:
718 +----------+----------+----------------------------------------+
719 | Field | Subfield | Value |
720 +==========+===========+=======================================+
721 | ``spec`` | ``data`` | 32 bit flow tag value |
722 | +-----------+---------------------------------------+
723 | | ``index`` | index of flow tag |
724 +----------+-----------+---------------------------------------+
725 | ``last`` | ``data`` | upper range value |
726 | +-----------+---------------------------------------+
727 | | ``index`` | field is ignored |
728 +----------+-----------+---------------------------------------+
729 | ``mask`` | ``data`` | bit-mask applies to "spec" and "last" |
730 | +-----------+---------------------------------------+
731 | | ``index`` | field is ignored |
732 +----------+-----------+---------------------------------------+
737 Matches 32 bit metadata item set.
739 On egress, metadata can be set either by mbuf metadata field with
740 RTE_MBUF_DYNFLAG_TX_METADATA flag or ``SET_META`` action. On ingress, ``SET_META``
741 action sets metadata for a packet and the metadata will be reported via
742 ``metadata`` dynamic field of ``rte_mbuf`` with RTE_MBUF_DYNFLAG_RX_METADATA flag.
744 - Default ``mask`` matches the specified Rx metadata value.
746 .. _table_rte_flow_item_meta:
750 +----------+----------+---------------------------------------+
751 | Field | Subfield | Value |
752 +==========+==========+=======================================+
753 | ``spec`` | ``data`` | 32 bit metadata value |
754 +----------+----------+---------------------------------------+
755 | ``last`` | ``data`` | upper range value |
756 +----------+----------+---------------------------------------+
757 | ``mask`` | ``data`` | bit-mask applies to "spec" and "last" |
758 +----------+----------+---------------------------------------+
760 Data matching item types
761 ~~~~~~~~~~~~~~~~~~~~~~~~
763 Most of these are basically protocol header definitions with associated
764 bit-masks. They must be specified (stacked) from lowest to highest protocol
765 layer to form a matching pattern.
770 Matches any protocol in place of the current layer, a single ANY may also
771 stand for several protocol layers.
773 This is usually specified as the first pattern item when looking for a
774 protocol anywhere in a packet.
776 - Default ``mask`` stands for any number of layers.
778 .. _table_rte_flow_item_any:
782 +----------+----------+--------------------------------------+
783 | Field | Subfield | Value |
784 +==========+==========+======================================+
785 | ``spec`` | ``num`` | number of layers covered |
786 +----------+----------+--------------------------------------+
787 | ``last`` | ``num`` | upper range value |
788 +----------+----------+--------------------------------------+
789 | ``mask`` | ``num`` | zeroed to cover any number of layers |
790 +----------+----------+--------------------------------------+
792 Example for VXLAN TCP payload matching regardless of outer L3 (IPv4 or IPv6)
793 and L4 (UDP) both matched by the first ANY specification, and inner L3 (IPv4
794 or IPv6) matched by the second ANY specification:
796 .. _table_rte_flow_item_any_example:
798 .. table:: TCP in VXLAN with wildcards
800 +-------+------+----------+----------+-------+
801 | Index | Item | Field | Subfield | Value |
802 +=======+======+==========+==========+=======+
804 +-------+------+----------+----------+-------+
805 | 1 | ANY | ``spec`` | ``num`` | 2 |
806 +-------+------+----------+----------+-------+
808 +-------+------------------------------------+
810 +-------+------+----------+----------+-------+
811 | 4 | ANY | ``spec`` | ``num`` | 1 |
812 +-------+------+----------+----------+-------+
814 +-------+------------------------------------+
816 +-------+------------------------------------+
821 Matches a byte string of a given length at a given offset.
823 Offset is either absolute (using the start of the packet) or relative to the
824 end of the previous matched item in the stack, in which case negative values
827 If search is enabled, offset is used as the starting point. The search area
828 can be delimited by setting limit to a nonzero value, which is the maximum
829 number of bytes after offset where the pattern may start.
831 Matching a zero-length pattern is allowed, doing so resets the relative
832 offset for subsequent items.
834 - This type does not support ranges (``last`` field).
835 - Default ``mask`` matches all fields exactly.
837 .. _table_rte_flow_item_raw:
841 +----------+--------------+-------------------------------------------------+
842 | Field | Subfield | Value |
843 +==========+==============+=================================================+
844 | ``spec`` | ``relative`` | look for pattern after the previous item |
845 | +--------------+-------------------------------------------------+
846 | | ``search`` | search pattern from offset (see also ``limit``) |
847 | +--------------+-------------------------------------------------+
848 | | ``reserved`` | reserved, must be set to zero |
849 | +--------------+-------------------------------------------------+
850 | | ``offset`` | absolute or relative offset for ``pattern`` |
851 | +--------------+-------------------------------------------------+
852 | | ``limit`` | search area limit for start of ``pattern`` |
853 | +--------------+-------------------------------------------------+
854 | | ``length`` | ``pattern`` length |
855 | +--------------+-------------------------------------------------+
856 | | ``pattern`` | byte string to look for |
857 +----------+--------------+-------------------------------------------------+
858 | ``last`` | if specified, either all 0 or with the same values as ``spec`` |
859 +----------+----------------------------------------------------------------+
860 | ``mask`` | bit-mask applied to ``spec`` values with usual behavior |
861 +----------+----------------------------------------------------------------+
863 Example pattern looking for several strings at various offsets of a UDP
864 payload, using combined RAW items:
866 .. _table_rte_flow_item_raw_example:
868 .. table:: UDP payload matching
870 +-------+------+----------+--------------+-------+
871 | Index | Item | Field | Subfield | Value |
872 +=======+======+==========+==============+=======+
874 +-------+----------------------------------------+
876 +-------+----------------------------------------+
878 +-------+------+----------+--------------+-------+
879 | 3 | RAW | ``spec`` | ``relative`` | 1 |
880 | | | +--------------+-------+
881 | | | | ``search`` | 1 |
882 | | | +--------------+-------+
883 | | | | ``offset`` | 10 |
884 | | | +--------------+-------+
885 | | | | ``limit`` | 0 |
886 | | | +--------------+-------+
887 | | | | ``length`` | 3 |
888 | | | +--------------+-------+
889 | | | | ``pattern`` | "foo" |
890 +-------+------+----------+--------------+-------+
891 | 4 | RAW | ``spec`` | ``relative`` | 1 |
892 | | | +--------------+-------+
893 | | | | ``search`` | 0 |
894 | | | +--------------+-------+
895 | | | | ``offset`` | 20 |
896 | | | +--------------+-------+
897 | | | | ``limit`` | 0 |
898 | | | +--------------+-------+
899 | | | | ``length`` | 3 |
900 | | | +--------------+-------+
901 | | | | ``pattern`` | "bar" |
902 +-------+------+----------+--------------+-------+
903 | 5 | RAW | ``spec`` | ``relative`` | 1 |
904 | | | +--------------+-------+
905 | | | | ``search`` | 0 |
906 | | | +--------------+-------+
907 | | | | ``offset`` | -29 |
908 | | | +--------------+-------+
909 | | | | ``limit`` | 0 |
910 | | | +--------------+-------+
911 | | | | ``length`` | 3 |
912 | | | +--------------+-------+
913 | | | | ``pattern`` | "baz" |
914 +-------+------+----------+--------------+-------+
916 +-------+----------------------------------------+
920 - Locate "foo" at least 10 bytes deep inside UDP payload.
921 - Locate "bar" after "foo" plus 20 bytes.
922 - Locate "baz" after "bar" minus 29 bytes.
924 Such a packet may be represented as follows (not to scale)::
927 | |<--------->| |<--------->|
929 |-----|------|-----|-----|-----|-----|-----------|-----|------|
930 | ETH | IPv4 | UDP | ... | baz | foo | ......... | bar | .... |
931 |-----|------|-----|-----|-----|-----|-----------|-----|------|
933 |<--------------------------->|
936 Note that matching subsequent pattern items would resume after "baz", not
937 "bar" since matching is always performed after the previous item of the
943 Matches an Ethernet header.
945 The ``type`` field either stands for "EtherType" or "TPID" when followed by
946 so-called layer 2.5 pattern items such as ``RTE_FLOW_ITEM_TYPE_VLAN``. In
947 the latter case, ``type`` refers to that of the outer header, with the inner
948 EtherType/TPID provided by the subsequent pattern item. This is the same
949 order as on the wire.
950 If the ``type`` field contains a TPID value, then only tagged packets with the
951 specified TPID will match the pattern.
952 The field ``has_vlan`` can be used to match any type of tagged packets,
953 instead of using the ``type`` field.
954 If the ``type`` and ``has_vlan`` fields are not specified, then both tagged
955 and untagged packets will match the pattern.
957 - ``dst``: destination MAC.
958 - ``src``: source MAC.
959 - ``type``: EtherType or TPID.
960 - ``has_vlan``: packet header contains at least one VLAN.
961 - Default ``mask`` matches destination and source addresses only.
966 Matches an 802.1Q/ad VLAN tag.
968 The corresponding standard outer EtherType (TPID) values are
969 ``RTE_ETHER_TYPE_VLAN`` or ``RTE_ETHER_TYPE_QINQ``. It can be overridden by the
970 preceding pattern item.
971 If a ``VLAN`` item is present in the pattern, then only tagged packets will
973 The field ``has_more_vlan`` can be used to match any type of tagged packets,
974 instead of using the ``inner_type field``.
975 If the ``inner_type`` and ``has_more_vlan`` fields are not specified,
976 then any tagged packets will match the pattern.
978 - ``tci``: tag control information.
979 - ``inner_type``: inner EtherType or TPID.
980 - ``has_more_vlan``: packet header contains at least one more VLAN, after this VLAN.
981 - Default ``mask`` matches the VID part of TCI only (lower 12 bits).
986 Matches an IPv4 header.
988 Note: IPv4 options are handled by dedicated pattern items.
990 - ``hdr``: IPv4 header definition (``rte_ip.h``).
991 - Default ``mask`` matches source and destination addresses only.
996 Matches an IPv6 header.
998 Dedicated flags indicate if header contains specific extension headers.
999 To match on packets containing a specific extension header, an application
1000 should match on the dedicated flag set to 1.
1001 To match on packets not containing a specific extension header, an application
1002 should match on the dedicated flag clear to 0.
1003 In case application doesn't care about the existence of a specific extension
1004 header, it should not specify the dedicated flag for matching.
1006 - ``hdr``: IPv6 header definition (``rte_ip.h``).
1007 - ``has_hop_ext``: header contains Hop-by-Hop Options extension header.
1008 - ``has_route_ext``: header contains Routing extension header.
1009 - ``has_frag_ext``: header contains Fragment extension header.
1010 - ``has_auth_ext``: header contains Authentication extension header.
1011 - ``has_esp_ext``: header contains Encapsulation Security Payload extension header.
1012 - ``has_dest_ext``: header contains Destination Options extension header.
1013 - ``has_mobil_ext``: header contains Mobility extension header.
1014 - ``has_hip_ext``: header contains Host Identity Protocol extension header.
1015 - ``has_shim6_ext``: header contains Shim6 Protocol extension header.
1016 - Default ``mask`` matches ``hdr`` source and destination addresses only.
1021 Matches an ICMP header.
1023 - ``hdr``: ICMP header definition (``rte_icmp.h``).
1024 - Default ``mask`` matches ICMP type and code only.
1029 Matches a UDP header.
1031 - ``hdr``: UDP header definition (``rte_udp.h``).
1032 - Default ``mask`` matches source and destination ports only.
1037 Matches a TCP header.
1039 - ``hdr``: TCP header definition (``rte_tcp.h``).
1040 - Default ``mask`` matches source and destination ports only.
1045 Matches a SCTP header.
1047 - ``hdr``: SCTP header definition (``rte_sctp.h``).
1048 - Default ``mask`` matches source and destination ports only.
1053 Matches a VXLAN header (RFC 7348).
1055 - ``flags``: normally 0x08 (I flag).
1056 - ``rsvd0``: reserved, normally 0x000000.
1057 - ``vni``: VXLAN network identifier.
1058 - ``rsvd1``: reserved, normally 0x00.
1059 - Default ``mask`` matches VNI only.
1064 Matches an IEEE 802.1BR E-Tag header.
1066 The corresponding standard outer EtherType (TPID) value is
1067 ``RTE_ETHER_TYPE_ETAG``. It can be overridden by the preceding pattern item.
1069 - ``epcp_edei_in_ecid_b``: E-Tag control information (E-TCI), E-PCP (3b),
1070 E-DEI (1b), ingress E-CID base (12b).
1071 - ``rsvd_grp_ecid_b``: reserved (2b), GRP (2b), E-CID base (12b).
1072 - ``in_ecid_e``: ingress E-CID ext.
1073 - ``ecid_e``: E-CID ext.
1074 - ``inner_type``: inner EtherType or TPID.
1075 - Default ``mask`` simultaneously matches GRP and E-CID base.
1080 Matches a NVGRE header (RFC 7637).
1082 - ``c_k_s_rsvd0_ver``: checksum (1b), undefined (1b), key bit (1b),
1083 sequence number (1b), reserved 0 (9b), version (3b). This field must have
1084 value 0x2000 according to RFC 7637.
1085 - ``protocol``: protocol type (0x6558).
1086 - ``tni``: virtual subnet ID.
1087 - ``flow_id``: flow ID.
1088 - Default ``mask`` matches TNI only.
1093 Matches a MPLS header.
1095 - ``label_tc_s_ttl``: label, TC, Bottom of Stack and TTL.
1096 - Default ``mask`` matches label only.
1101 Matches a GRE header.
1103 - ``c_rsvd0_ver``: checksum, reserved 0 and version.
1104 - ``protocol``: protocol type.
1105 - Default ``mask`` matches protocol only.
1109 This action is deprecated. Consider `Item: GRE_OPTION`.
1111 Matches a GRE key field.
1112 This should be preceded by item ``GRE``.
1114 - Value to be matched is a big-endian 32 bit integer.
1115 - When this item present it implicitly match K bit in default mask as "1"
1117 Item: ``GRE_OPTION``
1118 ^^^^^^^^^^^^^^^^^^^^
1120 Matches a GRE optional fields (checksum/key/sequence).
1121 This should be preceded by item ``GRE``.
1123 - ``checksum``: checksum.
1125 - ``sequence``: sequence.
1126 - The items in GRE_OPTION do not change bit flags(c_bit/k_bit/s_bit) in GRE
1127 item. The bit flags need be set with GRE item by application. When the items
1128 present, the corresponding bits in GRE spec and mask should be set "1" by
1129 application, it means to match specified value of the fields. When the items
1130 no present, but the corresponding bits in GRE spec and mask is "1", it means
1131 to match any value of the fields.
1136 Fuzzy pattern match, expect faster than default.
1138 This is for device that support fuzzy match option. Usually a fuzzy match is
1139 fast but the cost is accuracy. i.e. Signature Match only match pattern's hash
1140 value, but it is possible two different patterns have the same hash value.
1142 Matching accuracy level can be configured by threshold. Driver can divide the
1143 range of threshold and map to different accuracy levels that device support.
1145 Threshold 0 means perfect match (no fuzziness), while threshold 0xffffffff
1146 means fuzziest match.
1148 .. _table_rte_flow_item_fuzzy:
1152 +----------+---------------+--------------------------------------------------+
1153 | Field | Subfield | Value |
1154 +==========+===============+==================================================+
1155 | ``spec`` | ``threshold`` | 0 as perfect match, 0xffffffff as fuzziest match |
1156 +----------+---------------+--------------------------------------------------+
1157 | ``last`` | ``threshold`` | upper range value |
1158 +----------+---------------+--------------------------------------------------+
1159 | ``mask`` | ``threshold`` | bit-mask apply to "spec" and "last" |
1160 +----------+---------------+--------------------------------------------------+
1162 Usage example, fuzzy match a TCPv4 packets:
1164 .. _table_rte_flow_item_fuzzy_example:
1166 .. table:: Fuzzy matching
1168 +-------+----------+
1170 +=======+==========+
1172 +-------+----------+
1174 +-------+----------+
1176 +-------+----------+
1178 +-------+----------+
1180 +-------+----------+
1182 Item: ``GTP``, ``GTPC``, ``GTPU``
1183 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1185 Matches a GTPv1 header.
1187 Note: GTP, GTPC and GTPU use the same structure. GTPC and GTPU item
1188 are defined for a user-friendly API when creating GTP-C and GTP-U
1191 - ``v_pt_rsv_flags``: version (3b), protocol type (1b), reserved (1b),
1192 extension header flag (1b), sequence number flag (1b), N-PDU number
1194 - ``msg_type``: message type.
1195 - ``msg_len``: message length.
1196 - ``teid``: tunnel endpoint identifier.
1197 - Default ``mask`` matches teid only.
1202 Matches an ESP header.
1204 - ``hdr``: ESP header definition (``rte_esp.h``).
1205 - Default ``mask`` matches SPI only.
1210 Matches a GENEVE header.
1212 - ``ver_opt_len_o_c_rsvd0``: version (2b), length of the options fields (6b),
1213 OAM packet (1b), critical options present (1b), reserved 0 (6b).
1214 - ``protocol``: protocol type.
1215 - ``vni``: virtual network identifier.
1216 - ``rsvd1``: reserved, normally 0x00.
1217 - Default ``mask`` matches VNI only.
1222 Matches a VXLAN-GPE header (draft-ietf-nvo3-vxlan-gpe-05).
1224 - ``flags``: normally 0x0C (I and P flags).
1225 - ``rsvd0``: reserved, normally 0x0000.
1226 - ``protocol``: protocol type.
1227 - ``vni``: VXLAN network identifier.
1228 - ``rsvd1``: reserved, normally 0x00.
1229 - Default ``mask`` matches VNI only.
1231 Item: ``ARP_ETH_IPV4``
1232 ^^^^^^^^^^^^^^^^^^^^^^
1234 Matches an ARP header for Ethernet/IPv4.
1236 - ``hdr``: hardware type, normally 1.
1237 - ``pro``: protocol type, normally 0x0800.
1238 - ``hln``: hardware address length, normally 6.
1239 - ``pln``: protocol address length, normally 4.
1240 - ``op``: opcode (1 for request, 2 for reply).
1241 - ``sha``: sender hardware address.
1242 - ``spa``: sender IPv4 address.
1243 - ``tha``: target hardware address.
1244 - ``tpa``: target IPv4 address.
1245 - Default ``mask`` matches SHA, SPA, THA and TPA.
1250 Matches the presence of any IPv6 extension header.
1252 - ``next_hdr``: next header.
1253 - Default ``mask`` matches ``next_hdr``.
1255 Normally preceded by any of:
1260 Item: ``IPV6_FRAG_EXT``
1261 ^^^^^^^^^^^^^^^^^^^^^^^
1263 Matches the presence of IPv6 fragment extension header.
1265 - ``hdr``: IPv6 fragment extension header definition (``rte_ip.h``).
1267 Normally preceded by any of:
1275 Matches any ICMPv6 header.
1277 - ``type``: ICMPv6 type.
1278 - ``code``: ICMPv6 code.
1279 - ``checksum``: ICMPv6 checksum.
1280 - Default ``mask`` matches ``type`` and ``code``.
1282 Item: ``ICMP6_ND_NS``
1283 ^^^^^^^^^^^^^^^^^^^^^
1285 Matches an ICMPv6 neighbor discovery solicitation.
1287 - ``type``: ICMPv6 type, normally 135.
1288 - ``code``: ICMPv6 code, normally 0.
1289 - ``checksum``: ICMPv6 checksum.
1290 - ``reserved``: reserved, normally 0.
1291 - ``target_addr``: target address.
1292 - Default ``mask`` matches target address only.
1294 Item: ``ICMP6_ND_NA``
1295 ^^^^^^^^^^^^^^^^^^^^^
1297 Matches an ICMPv6 neighbor discovery advertisement.
1299 - ``type``: ICMPv6 type, normally 136.
1300 - ``code``: ICMPv6 code, normally 0.
1301 - ``checksum``: ICMPv6 checksum.
1302 - ``rso_reserved``: route flag (1b), solicited flag (1b), override flag
1303 (1b), reserved (29b).
1304 - ``target_addr``: target address.
1305 - Default ``mask`` matches target address only.
1307 Item: ``ICMP6_ND_OPT``
1308 ^^^^^^^^^^^^^^^^^^^^^^
1310 Matches the presence of any ICMPv6 neighbor discovery option.
1312 - ``type``: ND option type.
1313 - ``length``: ND option length.
1314 - Default ``mask`` matches type only.
1316 Normally preceded by any of:
1318 - `Item: ICMP6_ND_NA`_
1319 - `Item: ICMP6_ND_NS`_
1320 - `Item: ICMP6_ND_OPT`_
1322 Item: ``ICMP6_ND_OPT_SLA_ETH``
1323 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1325 Matches an ICMPv6 neighbor discovery source Ethernet link-layer address
1328 - ``type``: ND option type, normally 1.
1329 - ``length``: ND option length, normally 1.
1330 - ``sla``: source Ethernet LLA.
1331 - Default ``mask`` matches source link-layer address only.
1333 Normally preceded by any of:
1335 - `Item: ICMP6_ND_NA`_
1336 - `Item: ICMP6_ND_OPT`_
1338 Item: ``ICMP6_ND_OPT_TLA_ETH``
1339 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1341 Matches an ICMPv6 neighbor discovery target Ethernet link-layer address
1344 - ``type``: ND option type, normally 2.
1345 - ``length``: ND option length, normally 1.
1346 - ``tla``: target Ethernet LLA.
1347 - Default ``mask`` matches target link-layer address only.
1349 Normally preceded by any of:
1351 - `Item: ICMP6_ND_NS`_
1352 - `Item: ICMP6_ND_OPT`_
1357 Matches an application specific 32 bit metadata item.
1359 - Default ``mask`` matches the specified metadata value.
1364 Matches a GTP PDU extension header with type 0x85.
1366 - ``pdu_type``: PDU type.
1367 - ``qfi``: QoS flow identifier.
1368 - Default ``mask`` matches QFI only.
1370 Item: ``PPPOES``, ``PPPOED``
1371 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1373 Matches a PPPoE header.
1375 - ``version_type``: version (4b), type (4b).
1376 - ``code``: message type.
1377 - ``session_id``: session identifier.
1378 - ``length``: payload length.
1380 Item: ``PPPOE_PROTO_ID``
1381 ^^^^^^^^^^^^^^^^^^^^^^^^
1383 Matches a PPPoE session protocol identifier.
1385 - ``proto_id``: PPP protocol identifier.
1386 - Default ``mask`` matches proto_id only.
1391 Matches a network service header (RFC 8300).
1393 - ``version``: normally 0x0 (2 bits).
1394 - ``oam_pkt``: indicate oam packet (1 bit).
1395 - ``reserved``: reserved bit (1 bit).
1396 - ``ttl``: maximum SFF hopes (6 bits).
1397 - ``length``: total length in 4 bytes words (6 bits).
1398 - ``reserved1``: reserved1 bits (4 bits).
1399 - ``mdtype``: indicates format of NSH header (4 bits).
1400 - ``next_proto``: indicates protocol type of encap data (8 bits).
1401 - ``spi``: service path identifier (3 bytes).
1402 - ``sindex``: service index (1 byte).
1403 - Default ``mask`` matches mdtype, next_proto, spi, sindex.
1409 Matches a Internet Group Management Protocol (RFC 2236).
1411 - ``type``: IGMP message type (Query/Report).
1412 - ``max_resp_time``: max time allowed before sending report.
1413 - ``checksum``: checksum, 1s complement of whole IGMP message.
1414 - ``group_addr``: group address, for Query value will be 0.
1415 - Default ``mask`` matches group_addr.
1421 Matches a IP Authentication Header (RFC 4302).
1423 - ``next_hdr``: next payload after AH.
1424 - ``payload_len``: total length of AH in 4B words.
1425 - ``reserved``: reserved bits.
1426 - ``spi``: security parameters index.
1427 - ``seq_num``: counter value increased by 1 on each packet sent.
1428 - Default ``mask`` matches spi.
1433 Matches a HIGIG2 header field. It is layer 2.5 protocol and used in
1436 - Default ``mask`` matches classification and vlan.
1441 Matches a L2TPv3 over IP header.
1443 - ``session_id``: L2TPv3 over IP session identifier.
1444 - Default ``mask`` matches session_id only.
1449 Matches a PFCP Header.
1451 - ``s_field``: S field.
1452 - ``msg_type``: message type.
1453 - ``msg_len``: message length.
1454 - ``seid``: session endpoint identifier.
1455 - Default ``mask`` matches s_field and seid.
1460 Matches a eCPRI header.
1462 - ``hdr``: eCPRI header definition (``rte_ecpri.h``).
1463 - Default ``mask`` matches nothing, for all eCPRI messages.
1465 Item: ``PACKET_INTEGRITY_CHECKS``
1466 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
1468 Matches packet integrity.
1469 For some devices application needs to enable integration checks in HW
1470 before using this item.
1472 - ``level``: the encapsulation level that should be checked:
1473 - ``level == 0`` means the default PMD mode (can be inner most / outermost).
1474 - ``level == 1`` means outermost header.
1475 - ``level > 1`` means inner header. See also RSS level.
1476 - ``packet_ok``: All HW packet integrity checks have passed based on the
1477 topmost network layer. For example, for ICMP packet the topmost network
1478 layer is L3 and for TCP or UDP packet the topmost network layer is L4.
1479 - ``l2_ok``: all layer 2 HW integrity checks passed.
1480 - ``l3_ok``: all layer 3 HW integrity checks passed.
1481 - ``l4_ok``: all layer 4 HW integrity checks passed.
1482 - ``l2_crc_ok``: layer 2 CRC check passed.
1483 - ``ipv4_csum_ok``: IPv4 checksum check passed.
1484 - ``l4_csum_ok``: layer 4 checksum check passed.
1485 - ``l3_len_ok``: the layer 3 length is smaller than the frame length.
1490 Matches a conntrack state after conntrack action.
1492 - ``flags``: conntrack packet state flags.
1493 - Default ``mask`` matches all state bits.
1495 Item: ``PORT_REPRESENTOR``
1496 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1498 Matches traffic entering the embedded switch from the given ethdev.
1500 Term **ethdev** and the concept of **port representor** are synonymous.
1501 The **represented port** is an *entity* plugged to the embedded switch
1502 at the opposite end of the "wire" leading to the ethdev.
1506 .--------------------.
1507 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
1508 '--------------------'
1530 .--------------------.
1531 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
1532 '--------------------'
1535 - Incompatible with `Attribute: Traffic direction`_.
1536 - Requires `Attribute: Transfer`_.
1538 .. _table_rte_flow_item_ethdev:
1540 .. table:: ``struct rte_flow_item_ethdev``
1542 +----------+-------------+---------------------------+
1543 | Field | Subfield | Value |
1544 +==========+=============+===========================+
1545 | ``spec`` | ``port_id`` | ethdev port ID |
1546 +----------+-------------+---------------------------+
1547 | ``last`` | ``port_id`` | upper range value |
1548 +----------+-------------+---------------------------+
1549 | ``mask`` | ``port_id`` | zeroed for wildcard match |
1550 +----------+-------------+---------------------------+
1552 - Default ``mask`` provides exact match behaviour.
1554 See also `Action: PORT_REPRESENTOR`_.
1556 Item: ``REPRESENTED_PORT``
1557 ^^^^^^^^^^^^^^^^^^^^^^^^^^
1559 Matches traffic entering the embedded switch from
1560 the entity represented by the given ethdev.
1562 Term **ethdev** and the concept of **port representor** are synonymous.
1563 The **represented port** is an *entity* plugged to the embedded switch
1564 at the opposite end of the "wire" leading to the ethdev.
1568 .--------------------.
1569 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
1570 '--------------------'
1592 .--------------------.
1593 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
1594 '--------------------'
1597 - Incompatible with `Attribute: Traffic direction`_.
1598 - Requires `Attribute: Transfer`_.
1600 This item is meant to use the same structure as `Item: PORT_REPRESENTOR`_.
1602 See also `Action: REPRESENTED_PORT`_.
1607 Matches with the custom network protocol header that was created
1608 using rte_flow_flex_item_create() API. The application describes
1609 the desired header structure, defines the header fields attributes
1610 and header relations with preceding and following protocols and
1611 configures the ethernet devices accordingly via
1612 rte_flow_flex_item_create() routine.
1614 - ``handle``: the flex item handle returned by the PMD on successful
1615 rte_flow_flex_item_create() call, mask for this field is ignored.
1616 - ``length``: match pattern length in bytes. If the length does not cover
1617 all fields defined in item configuration, the pattern spec and mask are
1618 considered by the driver as padded with trailing zeroes till the full
1619 configured item pattern length.
1620 - ``pattern``: pattern to match. The pattern is concatenation of bit fields
1621 configured at item creation. At configuration the fields are presented
1622 by sample_data array. The order of the bitfields is defined by the order
1623 of sample_data elements. The width of each bitfield is defined by the width
1624 specified in the corresponding sample_data element as well. If pattern
1625 length is smaller than configured fields overall length it is considered
1626 as padded with trailing zeroes up to full configured length, both for
1632 Matches a L2TPv2 header.
1634 - ``flags_version``: flags(12b), version(4b).
1635 - ``length``: total length of the message.
1636 - ``tunnel_id``: identifier for the control connection.
1637 - ``session_id``: identifier for a session within a tunnel.
1638 - ``ns``: sequence number for this date or control message.
1639 - ``nr``: sequence number expected in the next control message to be received.
1640 - ``offset_size``: offset of payload data.
1641 - ``offset_padding``: offset padding, variable length.
1642 - Default ``mask`` matches flags_version only.
1647 Matches a PPP header.
1649 - ``addr``: PPP address.
1650 - ``ctrl``: PPP control.
1651 - ``proto_id``: PPP protocol identifier.
1652 - Default ``mask`` matches addr, ctrl, proto_id.
1657 Each possible action is represented by a type.
1658 An action can have an associated configuration object.
1659 Several actions combined in a list can be assigned
1660 to a flow rule and are performed in order.
1662 They fall in three categories:
1664 - Actions that modify the fate of matching traffic, for instance by dropping
1665 or assigning it a specific destination.
1667 - Actions that modify matching traffic contents or its properties. This
1668 includes adding/removing encapsulation, encryption, compression and marks.
1670 - Actions related to the flow rule itself, such as updating counters or
1671 making it non-terminating.
1673 Flow rules being terminating by default, not specifying any action of the
1674 fate kind results in undefined behavior. This applies to both ingress and
1677 PASSTHRU, when supported, makes a flow rule non-terminating.
1679 Like matching patterns, action lists are terminated by END items.
1681 Example of action that redirects packets to queue index 10:
1683 .. _table_rte_flow_action_example:
1685 .. table:: Queue action
1687 +-----------+-------+
1689 +===========+=======+
1691 +-----------+-------+
1693 Actions are performed in list order:
1695 .. _table_rte_flow_count_then_drop:
1697 .. table:: Count then drop
1711 .. _table_rte_flow_mark_count_redirect:
1713 .. table:: Mark, count then redirect
1715 +-------+--------+------------+-------+
1716 | Index | Action | Field | Value |
1717 +=======+========+============+=======+
1718 | 0 | MARK | ``mark`` | 0x2a |
1719 +-------+--------+------------+-------+
1720 | 1 | COUNT | ``id`` | 0 |
1721 +-------+--------+------------+-------+
1722 | 2 | QUEUE | ``queue`` | 10 |
1723 +-------+--------+------------+-------+
1725 +-------+-----------------------------+
1729 .. _table_rte_flow_redirect_queue_5:
1731 .. table:: Redirect to queue 5
1733 +-------+--------+-----------+-------+
1734 | Index | Action | Field | Value |
1735 +=======+========+===========+=======+
1737 +-------+--------+-----------+-------+
1738 | 1 | QUEUE | ``queue`` | 5 |
1739 +-------+--------+-----------+-------+
1741 +-------+----------------------------+
1743 In the above example, while DROP and QUEUE must be performed in order, both
1744 have to happen before reaching END. Only QUEUE has a visible effect.
1746 Note that such a list may be thought as ambiguous and rejected on that
1749 .. _table_rte_flow_redirect_queue_5_3:
1751 .. table:: Redirect to queues 5 and 3
1753 +-------+--------+-----------+-------+
1754 | Index | Action | Field | Value |
1755 +=======+========+===========+=======+
1756 | 0 | QUEUE | ``queue`` | 5 |
1757 +-------+--------+-----------+-------+
1759 +-------+--------+-----------+-------+
1760 | 2 | QUEUE | ``queue`` | 3 |
1761 +-------+--------+-----------+-------+
1763 +-------+----------------------------+
1765 As previously described, all actions must be taken into account. This
1766 effectively duplicates traffic to both queues. The above example also shows
1767 that VOID is ignored.
1772 Common action types are described in this section.
1777 End marker for action lists. Prevents further processing of actions, thereby
1780 - Its numeric value is 0 for convenience.
1781 - PMD support is mandatory.
1782 - No configurable properties.
1784 .. _table_rte_flow_action_end:
1797 Used as a placeholder for convenience. It is ignored and simply discarded by
1800 - PMD support is mandatory.
1801 - No configurable properties.
1803 .. _table_rte_flow_action_void:
1813 Action: ``PASSTHRU``
1814 ^^^^^^^^^^^^^^^^^^^^
1816 Leaves traffic up for additional processing by subsequent flow rules; makes
1817 a flow rule non-terminating.
1819 - No configurable properties.
1821 .. _table_rte_flow_action_passthru:
1831 Example to copy a packet to a queue and continue processing by subsequent
1834 .. _table_rte_flow_action_passthru_example:
1836 .. table:: Copy to queue 8
1838 +-------+--------+-----------+-------+
1839 | Index | Action | Field | Value |
1840 +=======+========+===========+=======+
1842 +-------+--------+-----------+-------+
1843 | 1 | QUEUE | ``queue`` | 8 |
1844 +-------+--------+-----------+-------+
1846 +-------+----------------------------+
1851 Redirects packets to a group on the current device.
1853 In a hierarchy of groups, which can be used to represent physical or logical
1854 flow group/tables on the device, this action redirects the matched flow to
1855 the specified group on that device.
1857 If a matched flow is redirected to a table which doesn't contain a matching
1858 rule for that flow then the behavior is undefined and the resulting behavior
1859 is up to the specific device. Best practice when using groups would be define
1860 a default flow rule for each group which a defines the default actions in that
1861 group so a consistent behavior is defined.
1863 Defining an action for matched flow in a group to jump to a group which is
1864 higher in the group hierarchy may not be supported by physical devices,
1865 depending on how groups are mapped to the physical devices. In the
1866 definitions of jump actions, applications should be aware that it may be
1867 possible to define flow rules which trigger an undefined behavior causing
1868 flows to loop between groups.
1870 .. _table_rte_flow_action_jump:
1874 +-----------+------------------------------+
1876 +===========+==============================+
1877 | ``group`` | Group to redirect packets to |
1878 +-----------+------------------------------+
1883 Attaches an integer value to packets and sets ``RTE_MBUF_F_RX_FDIR`` and
1884 ``RTE_MBUF_F_RX_FDIR_ID`` mbuf flags.
1886 This value is arbitrary and application-defined. Maximum allowed value
1887 depends on the underlying implementation. It is returned in the
1888 ``hash.fdir.hi`` mbuf field.
1890 .. _table_rte_flow_action_mark:
1894 +--------+--------------------------------------+
1896 +========+======================================+
1897 | ``id`` | integer value to return with packets |
1898 +--------+--------------------------------------+
1903 Flags packets. Similar to `Action: MARK`_ without a specific value; only
1904 sets the ``RTE_MBUF_F_RX_FDIR`` mbuf flag.
1906 - No configurable properties.
1908 .. _table_rte_flow_action_flag:
1921 Assigns packets to a given queue index.
1923 .. _table_rte_flow_action_queue:
1927 +-----------+--------------------+
1929 +===========+====================+
1930 | ``index`` | queue index to use |
1931 +-----------+--------------------+
1938 - No configurable properties.
1940 .. _table_rte_flow_action_drop:
1953 Adds a counter action to a matched flow.
1955 If more than one count action is specified in a single flow rule, then each
1956 action must specify a unique id.
1958 Counters can be retrieved and reset through ``rte_flow_query()``, see
1959 ``struct rte_flow_query_count``.
1961 For ports within the same switch domain then the counter id namespace extends
1962 to all ports within that switch domain.
1964 .. _table_rte_flow_action_count:
1968 +------------+---------------------------------+
1970 +============+=================================+
1971 | ``id`` | counter id |
1972 +------------+---------------------------------+
1974 Query structure to retrieve and reset flow rule counters:
1976 .. _table_rte_flow_query_count:
1978 .. table:: COUNT query
1980 +---------------+-----+-----------------------------------+
1981 | Field | I/O | Value |
1982 +===============+=====+===================================+
1983 | ``reset`` | in | reset counter after query |
1984 +---------------+-----+-----------------------------------+
1985 | ``hits_set`` | out | ``hits`` field is set |
1986 +---------------+-----+-----------------------------------+
1987 | ``bytes_set`` | out | ``bytes`` field is set |
1988 +---------------+-----+-----------------------------------+
1989 | ``hits`` | out | number of hits for this rule |
1990 +---------------+-----+-----------------------------------+
1991 | ``bytes`` | out | number of bytes through this rule |
1992 +---------------+-----+-----------------------------------+
1997 Similar to QUEUE, except RSS is additionally performed on packets to spread
1998 them among several queues according to the provided parameters.
2000 Unlike global RSS settings used by other DPDK APIs, unsetting the ``types``
2001 field does not disable RSS in a flow rule. Doing so instead requests safe
2002 unspecified "best-effort" settings from the underlying PMD, which depending
2003 on the flow rule, may result in anything ranging from empty (single queue)
2004 to all-inclusive RSS.
2006 If non-applicable for matching packets RSS types are requested,
2007 these RSS types are simply ignored. For example, it happens if:
2009 - Hashing of both TCP and UDP ports is requested
2010 (only one can be present in a packet).
2012 - Requested RSS types contradict to flow rule pattern
2013 (e.g. pattern has UDP item, but RSS types contain TCP).
2015 If requested RSS hash types are not supported by the Ethernet device at all
2016 (not reported in ``dev_info.flow_type_rss_offloads``),
2017 the flow creation will fail.
2019 Note: RSS hash result is stored in the ``hash.rss`` mbuf field which
2020 overlaps ``hash.fdir.lo``. Since `Action: MARK`_ sets the ``hash.fdir.hi``
2021 field only, both can be requested simultaneously.
2023 Also, regarding packet encapsulation ``level``:
2025 - ``0`` requests the default behavior. Depending on the packet type, it can
2026 mean outermost, innermost, anything in between or even no RSS.
2028 It basically stands for the innermost encapsulation level RSS can be
2029 performed on according to PMD and device capabilities.
2031 - ``1`` requests RSS to be performed on the outermost packet encapsulation
2034 - ``2`` and subsequent values request RSS to be performed on the specified
2035 inner packet encapsulation level, from outermost to innermost (lower to
2038 Values other than ``0`` are not necessarily supported.
2040 Requesting a specific RSS level on unrecognized traffic results in undefined
2041 behavior. For predictable results, it is recommended to make the flow rule
2042 pattern match packet headers up to the requested encapsulation level so that
2043 only matching traffic goes through.
2045 .. _table_rte_flow_action_rss:
2049 +---------------+-------------------------------------------------+
2051 +===============+=================================================+
2052 | ``func`` | RSS hash function to apply |
2053 +---------------+-------------------------------------------------+
2054 | ``level`` | encapsulation level for ``types`` |
2055 +---------------+-------------------------------------------------+
2056 | ``types`` | specific RSS hash types (see ``RTE_ETH_RSS_*``) |
2057 +---------------+-------------------------------------------------+
2058 | ``key_len`` | hash key length in bytes |
2059 +---------------+-------------------------------------------------+
2060 | ``queue_num`` | number of entries in ``queue`` |
2061 +---------------+-------------------------------------------------+
2062 | ``key`` | hash key |
2063 +---------------+-------------------------------------------------+
2064 | ``queue`` | queue indices to use |
2065 +---------------+-------------------------------------------------+
2070 This action is deprecated. Consider:
2071 - `Action: PORT_REPRESENTOR`_
2072 - `Action: REPRESENTED_PORT`_
2074 Directs matching traffic to the physical function (PF) of the current
2079 - No configurable properties.
2081 .. _table_rte_flow_action_pf:
2094 This action is deprecated. Consider:
2095 - `Action: PORT_REPRESENTOR`_
2096 - `Action: REPRESENTED_PORT`_
2098 Directs matching traffic to a given virtual function of the current device.
2100 Packets matched by a VF pattern item can be redirected to their original VF
2101 ID instead of the specified one. This parameter may not be available and is
2102 not guaranteed to work properly if the VF part is matched by a prior flow
2103 rule or if packets are not addressed to a VF in the first place.
2107 .. _table_rte_flow_action_vf:
2111 +--------------+--------------------------------+
2113 +==============+================================+
2114 | ``original`` | use original VF ID if possible |
2115 +--------------+--------------------------------+
2117 +--------------+--------------------------------+
2119 Action: ``PHY_PORT``
2120 ^^^^^^^^^^^^^^^^^^^^
2122 This action is deprecated. Consider:
2123 - `Action: PORT_REPRESENTOR`_
2124 - `Action: REPRESENTED_PORT`_
2126 Directs matching traffic to a given physical port index of the underlying
2129 See `Item: PHY_PORT`_.
2131 .. _table_rte_flow_action_phy_port:
2135 +--------------+-------------------------------------+
2137 +==============+=====================================+
2138 | ``original`` | use original port index if possible |
2139 +--------------+-------------------------------------+
2140 | ``index`` | physical port index |
2141 +--------------+-------------------------------------+
2145 This action is deprecated. Consider:
2146 - `Action: PORT_REPRESENTOR`_
2147 - `Action: REPRESENTED_PORT`_
2149 Directs matching traffic to a given DPDK port ID.
2151 See `Item: PORT_ID`_.
2153 .. _table_rte_flow_action_port_id:
2157 +--------------+---------------------------------------+
2159 +==============+=======================================+
2160 | ``original`` | use original DPDK port ID if possible |
2161 +--------------+---------------------------------------+
2162 | ``id`` | DPDK port ID |
2163 +--------------+---------------------------------------+
2168 Applies a stage of metering and policing.
2170 The metering and policing (MTR) object has to be first created using the
2171 rte_mtr_create() API function. The ID of the MTR object is specified as
2172 action parameter. More than one flow can use the same MTR object through
2173 the meter action. The MTR object can be further updated or queried using
2176 .. _table_rte_flow_action_meter:
2180 +--------------+---------------+
2182 +==============+===============+
2183 | ``mtr_id`` | MTR object ID |
2184 +--------------+---------------+
2186 Action: ``SECURITY``
2187 ^^^^^^^^^^^^^^^^^^^^
2189 Perform the security action on flows matched by the pattern items
2190 according to the configuration of the security session.
2192 This action modifies the payload of matched flows. For INLINE_CRYPTO, the
2193 security protocol headers and IV are fully provided by the application as
2194 specified in the flow pattern. The payload of matching packets is
2195 encrypted on egress, and decrypted and authenticated on ingress.
2196 For INLINE_PROTOCOL, the security protocol is fully offloaded to HW,
2197 providing full encapsulation and decapsulation of packets in security
2198 protocols. The flow pattern specifies both the outer security header fields
2199 and the inner packet fields. The security session specified in the action
2200 must match the pattern parameters.
2202 The security session specified in the action must be created on the same
2203 port as the flow action that is being specified.
2205 The ingress/egress flow attribute should match that specified in the
2206 security session if the security session supports the definition of the
2209 Multiple flows can be configured to use the same security session.
2211 .. _table_rte_flow_action_security:
2215 +----------------------+--------------------------------------+
2217 +======================+======================================+
2218 | ``security_session`` | security session to apply |
2219 +----------------------+--------------------------------------+
2221 The following is an example of configuring IPsec inline using the
2222 INLINE_CRYPTO security session:
2224 The encryption algorithm, keys and salt are part of the opaque
2225 ``rte_security_session``. The SA is identified according to the IP and ESP
2226 fields in the pattern items.
2228 .. _table_rte_flow_item_esp_inline_example:
2230 .. table:: IPsec inline crypto flow pattern items.
2232 +-------+----------+
2234 +=======+==========+
2236 +-------+----------+
2238 +-------+----------+
2240 +-------+----------+
2242 +-------+----------+
2244 .. _table_rte_flow_action_esp_inline_example:
2246 .. table:: IPsec inline flow actions.
2248 +-------+----------+
2250 +=======+==========+
2252 +-------+----------+
2254 +-------+----------+
2256 Action: ``OF_SET_MPLS_TTL``
2257 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2258 This action is deprecated. Consider `Action: MODIFY_FIELD`_.
2260 Implements ``OFPAT_SET_MPLS_TTL`` ("MPLS TTL") as defined by the `OpenFlow
2261 Switch Specification`_.
2263 .. _table_rte_flow_action_of_set_mpls_ttl:
2265 .. table:: OF_SET_MPLS_TTL
2267 +--------------+----------+
2269 +==============+==========+
2270 | ``mpls_ttl`` | MPLS TTL |
2271 +--------------+----------+
2273 Action: ``OF_DEC_MPLS_TTL``
2274 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2275 This action is deprecated. Consider `Action: MODIFY_FIELD`_.
2277 Implements ``OFPAT_DEC_MPLS_TTL`` ("decrement MPLS TTL") as defined by the
2278 `OpenFlow Switch Specification`_.
2280 .. _table_rte_flow_action_of_dec_mpls_ttl:
2282 .. table:: OF_DEC_MPLS_TTL
2290 Action: ``OF_SET_NW_TTL``
2291 ^^^^^^^^^^^^^^^^^^^^^^^^^
2292 This action is deprecated. Consider `Action: MODIFY_FIELD`_.
2294 Implements ``OFPAT_SET_NW_TTL`` ("IP TTL") as defined by the `OpenFlow
2295 Switch Specification`_.
2297 .. _table_rte_flow_action_of_set_nw_ttl:
2299 .. table:: OF_SET_NW_TTL
2301 +------------+--------+
2303 +============+========+
2304 | ``nw_ttl`` | IP TTL |
2305 +------------+--------+
2307 Action: ``OF_DEC_NW_TTL``
2308 ^^^^^^^^^^^^^^^^^^^^^^^^^
2309 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2311 Implements ``OFPAT_DEC_NW_TTL`` ("decrement IP TTL") as defined by the
2312 `OpenFlow Switch Specification`_.
2314 .. _table_rte_flow_action_of_dec_nw_ttl:
2316 .. table:: OF_DEC_NW_TTL
2324 Action: ``OF_COPY_TTL_OUT``
2325 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2326 This action is deprecated. Consider `Action: MODIFY_FIELD`_.
2328 Implements ``OFPAT_COPY_TTL_OUT`` ("copy TTL "outwards" -- from
2329 next-to-outermost to outermost") as defined by the `OpenFlow Switch
2332 .. _table_rte_flow_action_of_copy_ttl_out:
2334 .. table:: OF_COPY_TTL_OUT
2342 Action: ``OF_COPY_TTL_IN``
2343 ^^^^^^^^^^^^^^^^^^^^^^^^^^
2344 This action is deprecated. Consider `Action: MODIFY_FIELD`_.
2346 Implements ``OFPAT_COPY_TTL_IN`` ("copy TTL "inwards" -- from outermost to
2347 next-to-outermost") as defined by the `OpenFlow Switch Specification`_.
2349 .. _table_rte_flow_action_of_copy_ttl_in:
2351 .. table:: OF_COPY_TTL_IN
2359 Action: ``OF_POP_VLAN``
2360 ^^^^^^^^^^^^^^^^^^^^^^^
2362 Implements ``OFPAT_POP_VLAN`` ("pop the outer VLAN tag") as defined
2363 by the `OpenFlow Switch Specification`_.
2365 .. _table_rte_flow_action_of_pop_vlan:
2367 .. table:: OF_POP_VLAN
2375 Action: ``OF_PUSH_VLAN``
2376 ^^^^^^^^^^^^^^^^^^^^^^^^
2378 Implements ``OFPAT_PUSH_VLAN`` ("push a new VLAN tag") as defined by the
2379 `OpenFlow Switch Specification`_.
2381 .. _table_rte_flow_action_of_push_vlan:
2383 .. table:: OF_PUSH_VLAN
2385 +---------------+-----------+
2387 +===============+===========+
2388 | ``ethertype`` | EtherType |
2389 +---------------+-----------+
2391 Action: ``OF_SET_VLAN_VID``
2392 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2394 Implements ``OFPAT_SET_VLAN_VID`` ("set the 802.1q VLAN id") as defined by
2395 the `OpenFlow Switch Specification`_.
2397 .. _table_rte_flow_action_of_set_vlan_vid:
2399 .. table:: OF_SET_VLAN_VID
2401 +--------------+---------+
2403 +==============+=========+
2404 | ``vlan_vid`` | VLAN id |
2405 +--------------+---------+
2407 Action: ``OF_SET_VLAN_PCP``
2408 ^^^^^^^^^^^^^^^^^^^^^^^^^^^
2410 Implements ``OFPAT_SET_LAN_PCP`` ("set the 802.1q priority") as defined by
2411 the `OpenFlow Switch Specification`_.
2413 .. _table_rte_flow_action_of_set_vlan_pcp:
2415 .. table:: OF_SET_VLAN_PCP
2417 +--------------+---------------+
2419 +==============+===============+
2420 | ``vlan_pcp`` | VLAN priority |
2421 +--------------+---------------+
2423 Action: ``OF_POP_MPLS``
2424 ^^^^^^^^^^^^^^^^^^^^^^^
2426 Implements ``OFPAT_POP_MPLS`` ("pop the outer MPLS tag") as defined by the
2427 `OpenFlow Switch Specification`_.
2429 .. _table_rte_flow_action_of_pop_mpls:
2431 .. table:: OF_POP_MPLS
2433 +---------------+-----------+
2435 +===============+===========+
2436 | ``ethertype`` | EtherType |
2437 +---------------+-----------+
2439 Action: ``OF_PUSH_MPLS``
2440 ^^^^^^^^^^^^^^^^^^^^^^^^
2442 Implements ``OFPAT_PUSH_MPLS`` ("push a new MPLS tag") as defined by the
2443 `OpenFlow Switch Specification`_.
2445 .. _table_rte_flow_action_of_push_mpls:
2447 .. table:: OF_PUSH_MPLS
2449 +---------------+-----------+
2451 +===============+===========+
2452 | ``ethertype`` | EtherType |
2453 +---------------+-----------+
2455 Action: ``VXLAN_ENCAP``
2456 ^^^^^^^^^^^^^^^^^^^^^^^
2458 Performs a VXLAN encapsulation action by encapsulating the matched flow in the
2459 VXLAN tunnel as defined in the``rte_flow_action_vxlan_encap`` flow items
2462 This action modifies the payload of matched flows. The flow definition specified
2463 in the ``rte_flow_action_tunnel_encap`` action structure must define a valid
2464 VLXAN network overlay which conforms with RFC 7348 (Virtual eXtensible Local
2465 Area Network (VXLAN): A Framework for Overlaying Virtualized Layer 2 Networks
2466 over Layer 3 Networks). The pattern must be terminated with the
2467 RTE_FLOW_ITEM_TYPE_END item type.
2469 .. _table_rte_flow_action_vxlan_encap:
2471 .. table:: VXLAN_ENCAP
2473 +----------------+-------------------------------------+
2475 +================+=====================================+
2476 | ``definition`` | Tunnel end-point overlay definition |
2477 +----------------+-------------------------------------+
2479 .. _table_rte_flow_action_vxlan_encap_example:
2481 .. table:: IPv4 VxLAN flow pattern example.
2483 +-------+----------+
2485 +=======+==========+
2487 +-------+----------+
2489 +-------+----------+
2491 +-------+----------+
2493 +-------+----------+
2495 +-------+----------+
2497 Action: ``VXLAN_DECAP``
2498 ^^^^^^^^^^^^^^^^^^^^^^^
2500 Performs a decapsulation action by stripping all headers of the VXLAN tunnel
2501 network overlay from the matched flow.
2503 The flow items pattern defined for the flow rule with which a ``VXLAN_DECAP``
2504 action is specified, must define a valid VXLAN tunnel as per RFC7348. If the
2505 flow pattern does not specify a valid VXLAN tunnel then a
2506 RTE_FLOW_ERROR_TYPE_ACTION error should be returned.
2508 This action modifies the payload of matched flows.
2510 Action: ``NVGRE_ENCAP``
2511 ^^^^^^^^^^^^^^^^^^^^^^^
2513 Performs a NVGRE encapsulation action by encapsulating the matched flow in the
2514 NVGRE tunnel as defined in the``rte_flow_action_tunnel_encap`` flow item
2517 This action modifies the payload of matched flows. The flow definition specified
2518 in the ``rte_flow_action_tunnel_encap`` action structure must defined a valid
2519 NVGRE network overlay which conforms with RFC 7637 (NVGRE: Network
2520 Virtualization Using Generic Routing Encapsulation). The pattern must be
2521 terminated with the RTE_FLOW_ITEM_TYPE_END item type.
2523 .. _table_rte_flow_action_nvgre_encap:
2525 .. table:: NVGRE_ENCAP
2527 +----------------+-------------------------------------+
2529 +================+=====================================+
2530 | ``definition`` | NVGRE end-point overlay definition |
2531 +----------------+-------------------------------------+
2533 .. _table_rte_flow_action_nvgre_encap_example:
2535 .. table:: IPv4 NVGRE flow pattern example.
2537 +-------+----------+
2539 +=======+==========+
2541 +-------+----------+
2543 +-------+----------+
2545 +-------+----------+
2547 +-------+----------+
2549 Action: ``NVGRE_DECAP``
2550 ^^^^^^^^^^^^^^^^^^^^^^^
2552 Performs a decapsulation action by stripping all headers of the NVGRE tunnel
2553 network overlay from the matched flow.
2555 The flow items pattern defined for the flow rule with which a ``NVGRE_DECAP``
2556 action is specified, must define a valid NVGRE tunnel as per RFC7637. If the
2557 flow pattern does not specify a valid NVGRE tunnel then a
2558 RTE_FLOW_ERROR_TYPE_ACTION error should be returned.
2560 This action modifies the payload of matched flows.
2562 Action: ``RAW_ENCAP``
2563 ^^^^^^^^^^^^^^^^^^^^^
2565 Adds outer header whose template is provided in its data buffer,
2566 as defined in the ``rte_flow_action_raw_encap`` definition.
2568 This action modifies the payload of matched flows. The data supplied must
2569 be a valid header, either holding layer 2 data in case of adding layer 2 after
2570 decap layer 3 tunnel (for example MPLSoGRE) or complete tunnel definition
2571 starting from layer 2 and moving to the tunnel item itself. When applied to
2572 the original packet the resulting packet must be a valid packet.
2574 .. _table_rte_flow_action_raw_encap:
2576 .. table:: RAW_ENCAP
2578 +----------------+----------------------------------------+
2580 +================+========================================+
2581 | ``data`` | Encapsulation data |
2582 +----------------+----------------------------------------+
2583 | ``preserve`` | Bit-mask of data to preserve on output |
2584 +----------------+----------------------------------------+
2585 | ``size`` | Size of data and preserve |
2586 +----------------+----------------------------------------+
2588 Action: ``RAW_DECAP``
2589 ^^^^^^^^^^^^^^^^^^^^^^^
2591 Remove outer header whose template is provided in its data buffer,
2592 as defined in the ``rte_flow_action_raw_decap``
2594 This action modifies the payload of matched flows. The data supplied must
2595 be a valid header, either holding layer 2 data in case of removing layer 2
2596 before encapsulation of layer 3 tunnel (for example MPLSoGRE) or complete
2597 tunnel definition starting from layer 2 and moving to the tunnel item itself.
2598 When applied to the original packet the resulting packet must be a
2601 .. _table_rte_flow_action_raw_decap:
2603 .. table:: RAW_DECAP
2605 +----------------+----------------------------------------+
2607 +================+========================================+
2608 | ``data`` | Decapsulation data |
2609 +----------------+----------------------------------------+
2610 | ``size`` | Size of data |
2611 +----------------+----------------------------------------+
2613 Action: ``SET_IPV4_SRC``
2614 ^^^^^^^^^^^^^^^^^^^^^^^^
2615 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2617 Set a new IPv4 source address in the outermost IPv4 header.
2619 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV4 flow pattern item.
2620 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2622 .. _table_rte_flow_action_set_ipv4_src:
2624 .. table:: SET_IPV4_SRC
2626 +-----------------------------------------+
2628 +===============+=========================+
2629 | ``ipv4_addr`` | new IPv4 source address |
2630 +---------------+-------------------------+
2632 Action: ``SET_IPV4_DST``
2633 ^^^^^^^^^^^^^^^^^^^^^^^^
2634 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2636 Set a new IPv4 destination address in the outermost IPv4 header.
2638 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV4 flow pattern item.
2639 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2641 .. _table_rte_flow_action_set_ipv4_dst:
2643 .. table:: SET_IPV4_DST
2645 +---------------+------------------------------+
2647 +===============+==============================+
2648 | ``ipv4_addr`` | new IPv4 destination address |
2649 +---------------+------------------------------+
2651 Action: ``SET_IPV6_SRC``
2652 ^^^^^^^^^^^^^^^^^^^^^^^^
2653 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2655 Set a new IPv6 source address in the outermost IPv6 header.
2657 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV6 flow pattern item.
2658 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2660 .. _table_rte_flow_action_set_ipv6_src:
2662 .. table:: SET_IPV6_SRC
2664 +---------------+-------------------------+
2666 +===============+=========================+
2667 | ``ipv6_addr`` | new IPv6 source address |
2668 +---------------+-------------------------+
2670 Action: ``SET_IPV6_DST``
2671 ^^^^^^^^^^^^^^^^^^^^^^^^
2672 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2674 Set a new IPv6 destination address in the outermost IPv6 header.
2676 It must be used with a valid RTE_FLOW_ITEM_TYPE_IPV6 flow pattern item.
2677 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2679 .. _table_rte_flow_action_set_ipv6_dst:
2681 .. table:: SET_IPV6_DST
2683 +---------------+------------------------------+
2685 +===============+==============================+
2686 | ``ipv6_addr`` | new IPv6 destination address |
2687 +---------------+------------------------------+
2689 Action: ``SET_TP_SRC``
2690 ^^^^^^^^^^^^^^^^^^^^^^^^^
2691 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2693 Set a new source port number in the outermost TCP/UDP header.
2695 It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP or RTE_FLOW_ITEM_TYPE_UDP
2696 flow pattern item. Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2698 .. _table_rte_flow_action_set_tp_src:
2700 .. table:: SET_TP_SRC
2702 +----------+-------------------------+
2704 +==========+=========================+
2705 | ``port`` | new TCP/UDP source port |
2706 +---------------+--------------------+
2708 Action: ``SET_TP_DST``
2709 ^^^^^^^^^^^^^^^^^^^^^^^^^
2710 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2712 Set a new destination port number in the outermost TCP/UDP header.
2714 It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP or RTE_FLOW_ITEM_TYPE_UDP
2715 flow pattern item. Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2717 .. _table_rte_flow_action_set_tp_dst:
2719 .. table:: SET_TP_DST
2721 +----------+------------------------------+
2723 +==========+==============================+
2724 | ``port`` | new TCP/UDP destination port |
2725 +---------------+-------------------------+
2727 Action: ``MAC_SWAP``
2728 ^^^^^^^^^^^^^^^^^^^^^^^^^
2730 Swap the source and destination MAC addresses in the outermost Ethernet
2733 It must be used with a valid RTE_FLOW_ITEM_TYPE_ETH flow pattern item.
2734 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2736 .. _table_rte_flow_action_mac_swap:
2748 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2752 If there is no valid RTE_FLOW_ITEM_TYPE_IPV4 or RTE_FLOW_ITEM_TYPE_IPV6
2753 in pattern, Some PMDs will reject rule because behavior will be undefined.
2755 .. _table_rte_flow_action_dec_ttl:
2767 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2769 Assigns a new TTL value.
2771 If there is no valid RTE_FLOW_ITEM_TYPE_IPV4 or RTE_FLOW_ITEM_TYPE_IPV6
2772 in pattern, Some PMDs will reject rule because behavior will be undefined.
2774 .. _table_rte_flow_action_set_ttl:
2778 +---------------+--------------------+
2780 +===============+====================+
2781 | ``ttl_value`` | new TTL value |
2782 +---------------+--------------------+
2784 Action: ``SET_MAC_SRC``
2785 ^^^^^^^^^^^^^^^^^^^^^^^
2786 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2788 Set source MAC address.
2790 It must be used with a valid RTE_FLOW_ITEM_TYPE_ETH flow pattern item.
2791 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2793 .. _table_rte_flow_action_set_mac_src:
2795 .. table:: SET_MAC_SRC
2797 +--------------+---------------+
2799 +==============+===============+
2800 | ``mac_addr`` | MAC address |
2801 +--------------+---------------+
2803 Action: ``SET_MAC_DST``
2804 ^^^^^^^^^^^^^^^^^^^^^^^
2805 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2807 Set destination MAC address.
2809 It must be used with a valid RTE_FLOW_ITEM_TYPE_ETH flow pattern item.
2810 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2812 .. _table_rte_flow_action_set_mac_dst:
2814 .. table:: SET_MAC_DST
2816 +--------------+---------------+
2818 +==============+===============+
2819 | ``mac_addr`` | MAC address |
2820 +--------------+---------------+
2822 Action: ``INC_TCP_SEQ``
2823 ^^^^^^^^^^^^^^^^^^^^^^^
2824 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2826 Increase sequence number in the outermost TCP header.
2827 Value to increase TCP sequence number by is a big-endian 32 bit integer.
2829 Using this action on non-matching traffic will result in undefined behavior.
2831 Action: ``DEC_TCP_SEQ``
2832 ^^^^^^^^^^^^^^^^^^^^^^^
2833 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2835 Decrease sequence number in the outermost TCP header.
2836 Value to decrease TCP sequence number by is a big-endian 32 bit integer.
2838 Using this action on non-matching traffic will result in undefined behavior.
2840 Action: ``INC_TCP_ACK``
2841 ^^^^^^^^^^^^^^^^^^^^^^^
2842 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2844 Increase acknowledgment number in the outermost TCP header.
2845 Value to increase TCP acknowledgment number by is a big-endian 32 bit integer.
2847 Using this action on non-matching traffic will result in undefined behavior.
2849 Action: ``DEC_TCP_ACK``
2850 ^^^^^^^^^^^^^^^^^^^^^^^
2851 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2853 Decrease acknowledgment number in the outermost TCP header.
2854 Value to decrease TCP acknowledgment number by is a big-endian 32 bit integer.
2856 Using this action on non-matching traffic will result in undefined behavior.
2860 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2864 Tag is a transient data used during flow matching. This is not delivered to
2865 application. Multiple tags are supported by specifying index.
2867 .. _table_rte_flow_action_set_tag:
2871 +-----------+----------------------------+
2873 +===========+============================+
2874 | ``data`` | 32 bit tag value |
2875 +-----------+----------------------------+
2876 | ``mask`` | bit-mask applies to "data" |
2877 +-----------+----------------------------+
2878 | ``index`` | index of tag to set |
2879 +-----------+----------------------------+
2881 Action: ``SET_META``
2882 ^^^^^^^^^^^^^^^^^^^^^^^
2883 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2885 Set metadata. Item ``META`` matches metadata.
2887 Metadata set by mbuf metadata field with RTE_MBUF_DYNFLAG_TX_METADATA flag on egress
2888 will be overridden by this action. On ingress, the metadata will be carried by
2889 ``metadata`` dynamic field of ``rte_mbuf`` which can be accessed by
2890 ``RTE_FLOW_DYNF_METADATA()``. RTE_MBUF_DYNFLAG_RX_METADATA flag will be set along
2893 The mbuf dynamic field must be registered by calling
2894 ``rte_flow_dynf_metadata_register()`` prior to use ``SET_META`` action.
2896 Altering partial bits is supported with ``mask``. For bits which have never been
2897 set, unpredictable value will be seen depending on driver implementation. For
2898 loopback/hairpin packet, metadata set on Rx/Tx may or may not be propagated to
2899 the other path depending on HW capability.
2901 In hairpin case with Tx explicit flow mode, metadata could (not mandatory) be
2902 used to connect the Rx and Tx flows if it can be propagated from Rx to Tx path.
2904 .. _table_rte_flow_action_set_meta:
2908 +----------+----------------------------+
2910 +==========+============================+
2911 | ``data`` | 32 bit metadata value |
2912 +----------+----------------------------+
2913 | ``mask`` | bit-mask applies to "data" |
2914 +----------+----------------------------+
2916 Action: ``SET_IPV4_DSCP``
2917 ^^^^^^^^^^^^^^^^^^^^^^^^^
2918 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2922 Modify DSCP in IPv4 header.
2924 It must be used with RTE_FLOW_ITEM_TYPE_IPV4 in pattern.
2925 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2927 .. _table_rte_flow_action_set_ipv4_dscp:
2929 .. table:: SET_IPV4_DSCP
2931 +-----------+---------------------------------+
2933 +===========+=================================+
2934 | ``dscp`` | DSCP in low 6 bits, rest ignore |
2935 +-----------+---------------------------------+
2937 Action: ``SET_IPV6_DSCP``
2938 ^^^^^^^^^^^^^^^^^^^^^^^^^
2939 This is a legacy action. Consider `Action: MODIFY_FIELD`_ as alternative.
2943 Modify DSCP in IPv6 header.
2945 It must be used with RTE_FLOW_ITEM_TYPE_IPV6 in pattern.
2946 Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned.
2948 .. _table_rte_flow_action_set_ipv6_dscp:
2950 .. table:: SET_IPV6_DSCP
2952 +-----------+---------------------------------+
2954 +===========+=================================+
2955 | ``dscp`` | DSCP in low 6 bits, rest ignore |
2956 +-----------+---------------------------------+
2961 Set ageing timeout configuration to a flow.
2963 Event RTE_ETH_EVENT_FLOW_AGED will be reported if
2964 timeout passed without any matching on the flow.
2966 .. _table_rte_flow_action_age:
2970 +--------------+---------------------------------+
2972 +==============+=================================+
2973 | ``timeout`` | 24 bits timeout value |
2974 +--------------+---------------------------------+
2975 | ``reserved`` | 8 bits reserved, must be zero |
2976 +--------------+---------------------------------+
2977 | ``context`` | user input flow context |
2978 +--------------+---------------------------------+
2980 Query structure to retrieve ageing status information of a
2981 shared AGE action, or a flow rule using the AGE action:
2983 .. _table_rte_flow_query_age:
2985 .. table:: AGE query
2987 +------------------------------+-----+----------------------------------------+
2988 | Field | I/O | Value |
2989 +==============================+=====+========================================+
2990 | ``aged`` | out | Aging timeout expired |
2991 +------------------------------+-----+----------------------------------------+
2992 | ``sec_since_last_hit_valid`` | out | ``sec_since_last_hit`` value is valid |
2993 +------------------------------+-----+----------------------------------------+
2994 | ``sec_since_last_hit`` | out | Seconds since last traffic hit |
2995 +------------------------------+-----+----------------------------------------+
3000 Adds a sample action to a matched flow.
3002 The matching packets will be duplicated with the specified ``ratio`` and
3003 applied with own set of actions with a fate action, the packets sampled
3004 equals is '1/ratio'. All the packets continue to the target destination.
3006 When the ``ratio`` is set to 1 then the packets will be 100% mirrored.
3007 ``actions`` represent the different set of actions for the sampled or mirrored
3008 packets, and must have a fate action.
3010 .. _table_rte_flow_action_sample:
3014 +--------------+---------------------------------+
3016 +==============+=================================+
3017 | ``ratio`` | 32 bits sample ratio value |
3018 +--------------+---------------------------------+
3019 | ``actions`` | sub-action list for sampling |
3020 +--------------+---------------------------------+
3022 Action: ``INDIRECT``
3023 ^^^^^^^^^^^^^^^^^^^^
3025 Flow utilize indirect action by handle as returned from
3026 ``rte_flow_action_handle_create()``.
3028 The behaviour of the indirect action defined by ``action`` argument of type
3029 ``struct rte_flow_action`` passed to ``rte_flow_action_handle_create()``.
3031 The indirect action can be used by a single flow or shared among multiple flows.
3032 The indirect action can be in-place updated by ``rte_flow_action_handle_update()``
3033 without destroying flow and creating flow again. The fields that could be
3034 updated depend on the type of the ``action`` and different for every type.
3036 The indirect action specified data (e.g. counter) can be queried by
3037 ``rte_flow_action_handle_query()``.
3041 The following description of indirect action persistence
3042 is an experimental behavior that may change without a prior notice.
3044 If ``RTE_ETH_DEV_CAPA_FLOW_SHARED_OBJECT_KEEP`` is not advertised,
3045 indirect actions cannot be created until the device is started for the first time
3046 and cannot be kept when the device is stopped.
3047 However, PMD also does not flush them automatically on stop,
3048 so the application must call ``rte_flow_action_handle_destroy()``
3049 before stopping the device to ensure no indirect actions remain.
3051 If ``RTE_ETH_DEV_CAPA_FLOW_SHARED_OBJECT_KEEP`` is advertised,
3052 this means that the PMD can keep at least some indirect actions
3053 across device stop and start.
3054 However, ``rte_eth_dev_configure()`` may fail if any indirect actions remain,
3055 so the application must destroy them before attempting a reconfiguration.
3056 Keeping may be only supported for certain kinds of indirect actions.
3057 A kind is a combination of an action type and a value of its transfer bit.
3058 For example: an indirect counter with the transfer bit reset.
3059 To test if a particular kind of indirect actions is kept,
3060 the application must try to create a valid indirect action of that kind
3061 when the device is not started (either before the first start of after a stop).
3062 If it fails with an error of type ``RTE_FLOW_ERROR_TYPE_STATE``,
3063 application must destroy all indirect actions of this kind
3064 before stopping the device.
3065 If it succeeds, all indirect actions of the same kind are kept
3066 when the device is stopped.
3067 Indirect actions of a kept kind that are created when the device is stopped,
3068 including the ones created for the test, will be kept after the device start.
3070 .. _table_rte_flow_action_handle:
3080 Action: ``MODIFY_FIELD``
3081 ^^^^^^^^^^^^^^^^^^^^^^^^
3083 Modify ``dst`` field according to ``op`` selected (set, addition,
3084 subtraction) with ``width`` bits of data from ``src`` field.
3086 Any arbitrary header field (as well as mark, metadata or tag values)
3087 can be used as both source and destination fields as set by ``field``.
3088 The immediate value ``RTE_FLOW_FIELD_VALUE`` (or a pointer to it
3089 ``RTE_FLOW_FIELD_POINTER``) is allowed as a source only.
3090 ``RTE_FLOW_FIELD_START`` is used to point to the beginning of a packet.
3091 See ``enum rte_flow_field_id`` for the list of supported fields.
3093 ``op`` selects the operation to perform on a destination field.
3094 - ``set`` copies the data from ``src`` field to ``dst`` field.
3095 - ``add`` adds together ``dst`` and ``src`` and stores the result into ``dst``.
3096 - ``sub`` subtracts ``src`` from ``dst`` and stores the result into ``dst``
3098 ``width`` defines a number of bits to use from ``src`` field.
3100 ``level`` is used to access any packet field on any encapsulation level
3101 as well as any tag element in the tag array.
3102 - ``0`` means the default behaviour. Depending on the packet type, it can
3103 mean outermost, innermost or anything in between.
3104 - ``1`` requests access to the outermost packet encapsulation level.
3105 - ``2`` and subsequent values requests access to the specified packet
3106 encapsulation level, from outermost to innermost (lower to higher values).
3107 For the tag array (in case of multiple tags are supported and present)
3108 ``level`` translates directly into the array index.
3110 ``offset`` specifies the number of bits to skip from a field's start.
3111 That allows performing a partial copy of the needed part or to divide a big
3112 packet field into multiple smaller fields. Alternatively, ``offset`` allows
3113 going past the specified packet field boundary to copy a field to an
3114 arbitrary place in a packet, essentially providing a way to copy any part of
3115 a packet to any other part of it.
3117 ``value`` sets an immediate value to be used as a source or points to a
3118 location of the value in memory. It is used instead of ``level`` and ``offset``
3119 for ``RTE_FLOW_FIELD_VALUE`` and ``RTE_FLOW_FIELD_POINTER`` respectively.
3120 The data in memory should be presented exactly in the same byte order and
3121 length as in the relevant flow item, i.e. data for field with type
3122 ``RTE_FLOW_FIELD_MAC_DST`` should follow the conventions of ``dst`` field
3123 in ``rte_flow_item_eth`` structure, with type ``RTE_FLOW_FIELD_IPV6_SRC`` -
3124 ``rte_flow_item_ipv6`` conventions, and so on. If the field size is larger than
3125 16 bytes the pattern can be provided as pointer only.
3127 The bitfield extracted from the memory being applied as second operation
3128 parameter is defined by action width and by the destination field offset.
3129 Application should provide the data in immediate value memory (either as
3130 buffer or by pointer) exactly as item field without any applied explicit offset,
3131 and destination packet field (with specified width and bit offset) will be
3132 replaced by immediate source bits from the same bit offset. For example,
3133 to replace the third byte of MAC address with value 0x85, application should
3134 specify destination width as 8, destination offset as 16, and provide immediate
3135 value as sequence of bytes {xxx, xxx, 0x85, xxx, xxx, xxx}.
3137 .. _table_rte_flow_action_modify_field:
3139 .. table:: MODIFY_FIELD
3141 +---------------+-------------------------+
3143 +===============+=========================+
3144 | ``op`` | operation to perform |
3145 +---------------+-------------------------+
3146 | ``dst`` | destination field |
3147 +---------------+-------------------------+
3148 | ``src`` | source field |
3149 +---------------+-------------------------+
3150 | ``width`` | number of bits to use |
3151 +---------------+-------------------------+
3153 .. _table_rte_flow_action_modify_data:
3155 .. table:: destination/source field definition
3157 +---------------+----------------------------------------------------------+
3159 +===============+==========================================================+
3160 | ``field`` | ID: packet field, mark, meta, tag, immediate, pointer |
3161 +---------------+----------------------------------------------------------+
3162 | ``level`` | encapsulation level of a packet field or tag array index |
3163 +---------------+----------------------------------------------------------+
3164 | ``offset`` | number of bits to skip at the beginning |
3165 +---------------+----------------------------------------------------------+
3166 | ``value`` | immediate value buffer (source field only, not |
3167 | | applicable to destination) for RTE_FLOW_FIELD_VALUE |
3169 +---------------+----------------------------------------------------------+
3170 | ``pvalue`` | pointer to immediate value data (source field only, not |
3171 | | applicable to destination) for RTE_FLOW_FIELD_POINTER |
3173 +---------------+----------------------------------------------------------+
3175 Action: ``CONNTRACK``
3176 ^^^^^^^^^^^^^^^^^^^^^
3178 Create a conntrack (connection tracking) context with the provided information.
3180 In stateful session like TCP, the conntrack action provides the ability to
3181 examine every packet of this connection and associate the state to every
3182 packet. It will help to realize the stateful offload of connections with little
3183 software participation. For example, the packets with invalid state may be
3184 handled by the software. The control packets could be handled in the hardware.
3185 The software just need to query the state of a connection when needed, and then
3186 decide how to handle the flow rules and conntrack context.
3188 A conntrack context should be created via ``rte_flow_action_handle_create()``
3189 before using. Then the handle with ``INDIRECT`` type is used for a flow rule
3190 creation. If a flow rule with an opposite direction needs to be created, the
3191 ``rte_flow_action_handle_update()`` should be used to modify the direction.
3193 Not all the fields of the ``struct rte_flow_action_conntrack`` will be used
3194 for a conntrack context creating, depending on the HW, and they should be
3195 in host byte order. PMD should convert them into network byte order when
3198 The ``struct rte_flow_modify_conntrack`` should be used for an updating.
3200 The current conntrack context information could be queried via the
3201 ``rte_flow_action_handle_query()`` interface.
3203 .. _table_rte_flow_action_conntrack:
3205 .. table:: CONNTRACK
3207 +--------------------------+-------------------------------------------------------------+
3209 +==========================+=============================================================+
3210 | ``peer_port`` | peer port number |
3211 +--------------------------+-------------------------------------------------------------+
3212 | ``is_original_dir`` | direction of this connection for creating flow rule |
3213 +--------------------------+-------------------------------------------------------------+
3214 | ``enable`` | enable the conntrack context |
3215 +--------------------------+-------------------------------------------------------------+
3216 | ``live_connection`` | one ack was seen for this connection |
3217 +--------------------------+-------------------------------------------------------------+
3218 | ``selective_ack`` | SACK enabled |
3219 +--------------------------+-------------------------------------------------------------+
3220 | ``challenge_ack_passed`` | a challenge ack has passed |
3221 +--------------------------+-------------------------------------------------------------+
3222 | ``last_direction`` | direction of the last passed packet |
3223 +--------------------------+-------------------------------------------------------------+
3224 | ``liberal_mode`` | only report state change |
3225 +--------------------------+-------------------------------------------------------------+
3226 | ``state`` | current state |
3227 +--------------------------+-------------------------------------------------------------+
3228 | ``max_ack_window`` | maximal window scaling factor |
3229 +--------------------------+-------------------------------------------------------------+
3230 | ``retransmission_limit`` | maximal retransmission times |
3231 +--------------------------+-------------------------------------------------------------+
3232 | ``original_dir`` | TCP parameters of the original direction |
3233 +--------------------------+-------------------------------------------------------------+
3234 | ``reply_dir`` | TCP parameters of the reply direction |
3235 +--------------------------+-------------------------------------------------------------+
3236 | ``last_window`` | window size of the last passed packet |
3237 +--------------------------+-------------------------------------------------------------+
3238 | ``last_seq`` | sequence number of the last passed packet |
3239 +--------------------------+-------------------------------------------------------------+
3240 | ``last_ack`` | acknowledgment number the last passed packet |
3241 +--------------------------+-------------------------------------------------------------+
3242 | ``last_end`` | sum of ack number and length of the last passed packet |
3243 +--------------------------+-------------------------------------------------------------+
3245 .. _table_rte_flow_tcp_dir_param:
3247 .. table:: configuration parameters for each direction
3249 +---------------------+---------------------------------------------------------+
3251 +=====================+=========================================================+
3252 | ``scale`` | TCP window scaling factor |
3253 +---------------------+---------------------------------------------------------+
3254 | ``close_initiated`` | FIN sent from this direction |
3255 +---------------------+---------------------------------------------------------+
3256 | ``last_ack_seen`` | an ACK packet received |
3257 +---------------------+---------------------------------------------------------+
3258 | ``data_unacked`` | unacknowledged data for packets from this direction |
3259 +---------------------+---------------------------------------------------------+
3260 | ``sent_end`` | max{seq + len} seen in sent packets |
3261 +---------------------+---------------------------------------------------------+
3262 | ``reply_end`` | max{sack + max{win, 1}} seen in reply packets |
3263 +---------------------+---------------------------------------------------------+
3264 | ``max_win`` | max{max{win, 1}} + {sack - ack} seen in sent packets |
3265 +---------------------+---------------------------------------------------------+
3266 | ``max_ack`` | max{ack} + seen in sent packets |
3267 +---------------------+---------------------------------------------------------+
3269 .. _table_rte_flow_modify_conntrack:
3271 .. table:: update a conntrack context
3273 +----------------+-------------------------------------------------+
3275 +================+=================================================+
3276 | ``new_ct`` | new conntrack information |
3277 +----------------+-------------------------------------------------+
3278 | ``direction`` | direction will be updated |
3279 +----------------+-------------------------------------------------+
3280 | ``state`` | other fields except direction will be updated |
3281 +----------------+-------------------------------------------------+
3282 | ``reserved`` | reserved bits |
3283 +----------------+-------------------------------------------------+
3285 Action: ``METER_COLOR``
3286 ^^^^^^^^^^^^^^^^^^^^^^^
3288 Color the packet to reflect the meter color result.
3290 The meter action must be configured before meter color action.
3291 Meter color action is set to a color to reflect the meter color result.
3292 Set the meter color in the mbuf to the selected color.
3293 The meter color action output color is the output color of the packet,
3294 which is set in the packet meta-data (i.e. struct ``rte_mbuf::sched::color``)
3296 .. _table_rte_flow_action_meter_color:
3298 .. table:: METER_COLOR
3300 +-----------------+--------------+
3302 +=================+==============+
3303 | ``meter_color`` | Packet color |
3304 +-----------------+--------------+
3306 Action: ``PORT_REPRESENTOR``
3307 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3309 At embedded switch level, send matching traffic to the given ethdev.
3311 Term **ethdev** and the concept of **port representor** are synonymous.
3312 The **represented port** is an *entity* plugged to the embedded switch
3313 at the opposite end of the "wire" leading to the ethdev.
3317 .--------------------.
3318 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
3319 '--------------------'
3329 .----------. .--------------------.
3330 | Switch | <== | Matching Traffic |
3331 '----------' '--------------------'
3341 .--------------------.
3342 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
3343 '--------------------'
3346 - Requires `Attribute: Transfer`_.
3348 .. _table_rte_flow_action_ethdev:
3350 .. table:: ``struct rte_flow_action_ethdev``
3352 +-------------+----------------+
3354 +=============+================+
3355 | ``port_id`` | ethdev port ID |
3356 +-------------+----------------+
3358 See also `Item: PORT_REPRESENTOR`_.
3360 Action: ``REPRESENTED_PORT``
3361 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
3363 At embedded switch level, send matching traffic to
3364 the entity represented by the given ethdev.
3366 Term **ethdev** and the concept of **port representor** are synonymous.
3367 The **represented port** is an *entity* plugged to the embedded switch
3368 at the opposite end of the "wire" leading to the ethdev.
3372 .--------------------.
3373 | PORT_REPRESENTOR | Ethdev (Application Port Referred to by its ID)
3374 '--------------------'
3384 .----------. .--------------------.
3385 | Switch | <== | Matching Traffic |
3386 '----------' '--------------------'
3396 .--------------------.
3397 | REPRESENTED_PORT | Net / Guest / Another Ethdev (Same Application)
3398 '--------------------'
3401 - Requires `Attribute: Transfer`_.
3403 This action is meant to use the same structure as `Action: PORT_REPRESENTOR`_.
3405 See also `Item: REPRESENTED_PORT`_.
3410 All specified pattern items (``enum rte_flow_item_type``) and actions
3411 (``enum rte_flow_action_type``) use positive identifiers.
3413 The negative space is reserved for dynamic types generated by PMDs during
3414 run-time. PMDs may encounter them as a result but must not accept negative
3415 identifiers they are not aware of.
3417 A method to generate them remains to be defined.
3419 Application may use PMD dynamic items or actions in flow rules. In that case
3420 size of configuration object in dynamic element must be a pointer size.
3425 A rather simple API with few functions is provided to fully manage flow
3428 Each created flow rule is associated with an opaque, PMD-specific handle
3429 pointer. The application is responsible for keeping it until the rule is
3432 Flows rules are represented by ``struct rte_flow`` objects.
3437 Given that expressing a definite set of device capabilities is not
3438 practical, a dedicated function is provided to check if a flow rule is
3439 supported and can be created.
3444 rte_flow_validate(uint16_t port_id,
3445 const struct rte_flow_attr *attr,
3446 const struct rte_flow_item pattern[],
3447 const struct rte_flow_action actions[],
3448 struct rte_flow_error *error);
3450 The flow rule is validated for correctness and whether it could be accepted
3451 by the device given sufficient resources. The rule is checked against the
3452 current device mode and queue configuration. The flow rule may also
3453 optionally be validated against existing flow rules and device resources.
3454 This function has no effect on the target device.
3456 The returned value is guaranteed to remain valid only as long as no
3457 successful calls to ``rte_flow_create()`` or ``rte_flow_destroy()`` are made
3458 in the meantime and no device parameter affecting flow rules in any way are
3459 modified, due to possible collisions or resource limitations (although in
3460 such cases ``EINVAL`` should not be returned).
3464 - ``port_id``: port identifier of Ethernet device.
3465 - ``attr``: flow rule attributes.
3466 - ``pattern``: pattern specification (list terminated by the END pattern
3468 - ``actions``: associated actions (list terminated by the END action).
3469 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3470 this structure in case of error only.
3474 - 0 if flow rule is valid and can be created. A negative errno value
3475 otherwise (``rte_errno`` is also set), the following errors are defined.
3476 - ``-ENOSYS``: underlying device does not support this functionality.
3477 - ``-EINVAL``: unknown or invalid rule specification.
3478 - ``-ENOTSUP``: valid but unsupported rule specification (e.g. partial
3479 bit-masks are unsupported).
3480 - ``EEXIST``: collision with an existing rule. Only returned if device
3481 supports flow rule collision checking and there was a flow rule
3482 collision. Not receiving this return code is no guarantee that creating
3483 the rule will not fail due to a collision.
3484 - ``ENOMEM``: not enough memory to execute the function, or if the device
3485 supports resource validation, resource limitation on the device.
3486 - ``-EBUSY``: action cannot be performed due to busy device resources, may
3487 succeed if the affected queues or even the entire port are in a stopped
3488 state (see ``rte_eth_dev_rx_queue_stop()`` and ``rte_eth_dev_stop()``).
3493 Creating a flow rule is similar to validating one, except the rule is
3494 actually created and a handle returned.
3499 rte_flow_create(uint16_t port_id,
3500 const struct rte_flow_attr *attr,
3501 const struct rte_flow_item pattern[],
3502 const struct rte_flow_action *actions[],
3503 struct rte_flow_error *error);
3507 - ``port_id``: port identifier of Ethernet device.
3508 - ``attr``: flow rule attributes.
3509 - ``pattern``: pattern specification (list terminated by the END pattern
3511 - ``actions``: associated actions (list terminated by the END action).
3512 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3513 this structure in case of error only.
3517 A valid handle in case of success, NULL otherwise and ``rte_errno`` is set
3518 to the positive version of one of the error codes defined for
3519 ``rte_flow_validate()``.
3524 Flow rules destruction is not automatic, and a queue or a port should not be
3525 released if any are still attached to them. Applications must take care of
3526 performing this step before releasing resources.
3531 rte_flow_destroy(uint16_t port_id,
3532 struct rte_flow *flow,
3533 struct rte_flow_error *error);
3536 Failure to destroy a flow rule handle may occur when other flow rules depend
3537 on it, and destroying it would result in an inconsistent state.
3539 This function is only guaranteed to succeed if handles are destroyed in
3540 reverse order of their creation.
3544 - ``port_id``: port identifier of Ethernet device.
3545 - ``flow``: flow rule handle to destroy.
3546 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3547 this structure in case of error only.
3551 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3556 Convenience function to destroy all flow rule handles associated with a
3557 port. They are released as with successive calls to ``rte_flow_destroy()``.
3562 rte_flow_flush(uint16_t port_id,
3563 struct rte_flow_error *error);
3565 In the unlikely event of failure, handles are still considered destroyed and
3566 no longer valid but the port must be assumed to be in an inconsistent state.
3570 - ``port_id``: port identifier of Ethernet device.
3571 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3572 this structure in case of error only.
3576 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3581 Query an existing flow rule.
3583 This function allows retrieving flow-specific data such as counters. Data
3584 is gathered by special actions which must be present in the flow rule
3590 rte_flow_query(uint16_t port_id,
3591 struct rte_flow *flow,
3592 const struct rte_flow_action *action,
3594 struct rte_flow_error *error);
3598 - ``port_id``: port identifier of Ethernet device.
3599 - ``flow``: flow rule handle to query.
3600 - ``action``: action to query, this must match prototype from flow rule.
3601 - ``data``: pointer to storage for the associated query data type.
3602 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3603 this structure in case of error only.
3607 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3609 .. _flow_isolated_mode:
3614 The general expectation for ingress traffic is that flow rules process it
3615 first; the remaining unmatched or pass-through traffic usually ends up in a
3616 queue (with or without RSS, locally or in some sub-device instance)
3617 depending on the global configuration settings of a port.
3619 While fine from a compatibility standpoint, this approach makes drivers more
3620 complex as they have to check for possible side effects outside of this API
3621 when creating or destroying flow rules. It results in a more limited set of
3622 available rule types due to the way device resources are assigned (e.g. no
3623 support for the RSS action even on capable hardware).
3625 Given that nonspecific traffic can be handled by flow rules as well,
3626 isolated mode is a means for applications to tell a driver that ingress on
3627 the underlying port must be injected from the defined flow rules only; that
3628 no default traffic is expected outside those rules.
3630 This has the following benefits:
3632 - Applications get finer-grained control over the kind of traffic they want
3633 to receive (no traffic by default).
3635 - More importantly they control at what point nonspecific traffic is handled
3636 relative to other flow rules, by adjusting priority levels.
3638 - Drivers can assign more hardware resources to flow rules and expand the
3639 set of supported rule types.
3641 Because toggling isolated mode may cause profound changes to the ingress
3642 processing path of a driver, it may not be possible to leave it once
3643 entered. Likewise, existing flow rules or global configuration settings may
3644 prevent a driver from entering isolated mode.
3646 Applications relying on this mode are therefore encouraged to toggle it as
3647 soon as possible after device initialization, ideally before the first call
3648 to ``rte_eth_dev_configure()`` to avoid possible failures due to conflicting
3651 Once effective, the following functionality has no effect on the underlying
3652 port and may return errors such as ``ENOTSUP`` ("not supported"):
3654 - Toggling promiscuous mode.
3655 - Toggling allmulticast mode.
3656 - Configuring MAC addresses.
3657 - Configuring multicast addresses.
3658 - Configuring VLAN filters.
3659 - Configuring global RSS settings.
3664 rte_flow_isolate(uint16_t port_id, int set, struct rte_flow_error *error);
3668 - ``port_id``: port identifier of Ethernet device.
3669 - ``set``: nonzero to enter isolated mode, attempt to leave it otherwise.
3670 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
3671 this structure in case of error only.
3675 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
3677 Verbose error reporting
3678 -----------------------
3680 The defined *errno* values may not be accurate enough for users or
3681 application developers who want to investigate issues related to flow rules
3682 management. A dedicated error object is defined for this purpose:
3686 enum rte_flow_error_type {
3687 RTE_FLOW_ERROR_TYPE_NONE, /**< No error. */
3688 RTE_FLOW_ERROR_TYPE_UNSPECIFIED, /**< Cause unspecified. */
3689 RTE_FLOW_ERROR_TYPE_HANDLE, /**< Flow rule (handle). */
3690 RTE_FLOW_ERROR_TYPE_ATTR_GROUP, /**< Group field. */
3691 RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY, /**< Priority field. */
3692 RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, /**< Ingress field. */
3693 RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, /**< Egress field. */
3694 RTE_FLOW_ERROR_TYPE_ATTR, /**< Attributes structure. */
3695 RTE_FLOW_ERROR_TYPE_ITEM_NUM, /**< Pattern length. */
3696 RTE_FLOW_ERROR_TYPE_ITEM, /**< Specific pattern item. */
3697 RTE_FLOW_ERROR_TYPE_ACTION_NUM, /**< Number of actions. */
3698 RTE_FLOW_ERROR_TYPE_ACTION, /**< Specific action. */
3701 struct rte_flow_error {
3702 enum rte_flow_error_type type; /**< Cause field and error types. */
3703 const void *cause; /**< Object responsible for the error. */
3704 const char *message; /**< Human-readable error message. */
3707 Error type ``RTE_FLOW_ERROR_TYPE_NONE`` stands for no error, in which case
3708 remaining fields can be ignored. Other error types describe the type of the
3709 object pointed by ``cause``.
3711 If non-NULL, ``cause`` points to the object responsible for the error. For a
3712 flow rule, this may be a pattern item or an individual action.
3714 If non-NULL, ``message`` provides a human-readable error message.
3716 This object is normally allocated by applications and set by PMDs in case of
3717 error, the message points to a constant string which does not need to be
3718 freed by the application, however its pointer can be considered valid only
3719 as long as its associated DPDK port remains configured. Closing the
3720 underlying device or unloading the PMD invalidates it.
3731 rte_flow_error_set(struct rte_flow_error *error,
3733 enum rte_flow_error_type type,
3735 const char *message);
3737 This function initializes ``error`` (if non-NULL) with the provided
3738 parameters and sets ``rte_errno`` to ``code``. A negative error ``code`` is
3747 rte_flow_conv(enum rte_flow_conv_op op,
3751 struct rte_flow_error *error);
3753 Convert ``src`` to ``dst`` according to operation ``op``. Possible
3756 - Attributes, pattern item or action duplication.
3757 - Duplication of an entire pattern or list of actions.
3758 - Duplication of a complete flow rule description.
3759 - Pattern item or action name retrieval.
3761 Tunneled traffic offload
3762 ~~~~~~~~~~~~~~~~~~~~~~~~
3764 rte_flow API provides the building blocks for vendor-agnostic flow
3765 classification offloads. The rte_flow "patterns" and "actions"
3766 primitives are fine-grained, thus enabling DPDK applications the
3767 flexibility to offload network stacks and complex pipelines.
3768 Applications wishing to offload tunneled traffic are required to use
3769 the rte_flow primitives, such as group, meta, mark, tag, and others to
3770 model their high-level objects. The hardware model design for
3771 high-level software objects is not trivial. Furthermore, an optimal
3772 design is often vendor-specific.
3774 When hardware offloads tunneled traffic in multi-group logic,
3775 partially offloaded packets may arrive to the application after they
3776 were modified in hardware. In this case, the application may need to
3777 restore the original packet headers. Consider the following sequence:
3778 The application decaps a packet in one group and jumps to a second
3779 group where it tries to match on a 5-tuple, that will miss and send
3780 the packet to the application. In this case, the application does not
3781 receive the original packet but a modified one. Also, in this case,
3782 the application cannot match on the outer header fields, such as VXLAN
3785 There are several possible ways to use rte_flow "patterns" and
3786 "actions" to resolve the issues above. For example:
3788 1 Mapping headers to a hardware registers using the
3789 rte_flow_action_mark/rte_flow_action_tag/rte_flow_set_meta objects.
3791 2 Apply the decap only at the last offload stage after all the
3792 "patterns" were matched and the packet will be fully offloaded.
3794 Every approach has its pros and cons and is highly dependent on the
3795 hardware vendor. For example, some hardware may have a limited number
3796 of registers while other hardware could not support inner actions and
3797 must decap before accessing inner headers.
3799 The tunnel offload model resolves these issues. The model goals are:
3801 1 Provide a unified application API to offload tunneled traffic that
3802 is capable to match on outer headers after decap.
3804 2 Allow the application to restore the outer header of partially
3807 The tunnel offload model does not introduce new elements to the
3808 existing RTE flow model and is implemented as a set of helper
3811 For the application to work with the tunnel offload API it
3812 has to adjust flow rules in multi-table tunnel offload in the
3815 1 Remove explicit call to decap action and replace it with PMD actions
3816 obtained from rte_flow_tunnel_decap_and_set() helper.
3818 2 Add PMD items obtained from rte_flow_tunnel_match() helper to all
3819 other rules in the tunnel offload sequence.
3821 The model requirements:
3823 Software application must initialize
3824 rte_tunnel object with tunnel parameters before calling
3825 rte_flow_tunnel_decap_set() & rte_flow_tunnel_match().
3827 PMD actions array obtained in rte_flow_tunnel_decap_set() must be
3828 released by application with rte_flow_action_release() call.
3830 PMD items array obtained with rte_flow_tunnel_match() must be released
3831 by application with rte_flow_item_release() call. Application can
3832 release PMD items and actions after rule was created. However, if the
3833 application needs to create additional rule for the same tunnel it
3834 will need to obtain PMD items again.
3836 Application cannot destroy rte_tunnel object before it releases all
3837 PMD actions & PMD items referencing that tunnel.
3842 - DPDK does not keep track of flow rules definitions or flow rule objects
3843 automatically. Applications may keep track of the former and must keep
3844 track of the latter. PMDs may also do it for internal needs, however this
3845 must not be relied on by applications.
3847 - Flow rules are not maintained between successive port initializations. An
3848 application exiting without releasing them and restarting must re-create
3851 - API operations are synchronous and blocking (``EAGAIN`` cannot be
3854 - Stopping the data path (TX/RX) should not be necessary when managing flow
3855 rules. If this cannot be achieved naturally or with workarounds (such as
3856 temporarily replacing the burst function pointers), an appropriate error
3857 code must be returned (``EBUSY``).
3859 - Applications, not PMDs, are responsible for maintaining flow rules
3860 configuration when closing, stopping or restarting a port or performing other
3861 actions which may affect them.
3862 Applications must assume that after port close, stop or restart all flows
3863 related to that port are not valid, hardware rules are destroyed and relevant
3864 PMD resources are released.
3866 For devices exposing multiple ports sharing global settings affected by flow
3869 - All ports under DPDK control must behave consistently, PMDs are
3870 responsible for making sure that existing flow rules on a port are not
3871 affected by other ports.
3873 - Ports not under DPDK control (unaffected or handled by other applications)
3874 are user's responsibility. They may affect existing flow rules and cause
3875 undefined behavior. PMDs aware of this may prevent flow rules creation
3876 altogether in such cases.
3881 The PMD interface is defined in ``rte_flow_driver.h``. It is not subject to
3882 API/ABI versioning constraints as it is not exposed to applications and may
3883 evolve independently.
3885 The PMD interface is based on callbacks pointed by the ``struct rte_flow_ops``.
3887 - PMD callbacks implement exactly the interface described in `Rules
3888 management`_, except for the port ID argument which has already been
3889 converted to a pointer to the underlying ``struct rte_eth_dev``.
3891 - Public API functions do not process flow rules definitions at all before
3892 calling PMD functions (no basic error checking, no validation
3893 whatsoever). They only make sure these callbacks are non-NULL or return
3894 the ``ENOSYS`` (function not supported) error.
3896 This interface additionally defines the following helper function:
3898 - ``rte_flow_ops_get()``: get generic flow operations structure from a
3901 If PMD interfaces don't support re-entrancy/multi-thread safety,
3902 the rte_flow API functions will protect threads by mutex per port.
3903 The application can check whether ``RTE_ETH_DEV_FLOW_OPS_THREAD_SAFE``
3904 is set in ``dev_flags``, meaning the PMD is thread-safe regarding rte_flow,
3905 so the API level protection is disabled.
3906 Please note that this API-level mutex protects only rte_flow functions,
3907 other control path functions are not in scope.
3909 Device compatibility
3910 --------------------
3912 No known implementation supports all the described features.
3914 Unsupported features or combinations are not expected to be fully emulated
3915 in software by PMDs for performance reasons. Partially supported features
3916 may be completed in software as long as hardware performs most of the work
3917 (such as queue redirection and packet recognition).
3919 However PMDs are expected to do their best to satisfy application requests
3920 by working around hardware limitations as long as doing so does not affect
3921 the behavior of existing flow rules.
3923 The following sections provide a few examples of such cases and describe how
3924 PMDs should handle them, they are based on limitations built into the
3930 Each flow rule comes with its own, per-layer bit-masks, while hardware may
3931 support only a single, device-wide bit-mask for a given layer type, so that
3932 two IPv4 rules cannot use different bit-masks.
3934 The expected behavior in this case is that PMDs automatically configure
3935 global bit-masks according to the needs of the first flow rule created.
3937 Subsequent rules are allowed only if their bit-masks match those, the
3938 ``EEXIST`` error code should be returned otherwise.
3940 Unsupported layer types
3941 ~~~~~~~~~~~~~~~~~~~~~~~
3943 Many protocols can be simulated by crafting patterns with the `Item: RAW`_
3946 PMDs can rely on this capability to simulate support for protocols with
3947 headers not directly recognized by hardware.
3949 ``ANY`` pattern item
3950 ~~~~~~~~~~~~~~~~~~~~
3952 This pattern item stands for anything, which can be difficult to translate
3953 to something hardware would understand, particularly if followed by more
3956 Consider the following pattern:
3958 .. _table_rte_flow_unsupported_any:
3960 .. table:: Pattern with ANY as L3
3962 +-------+-----------------------+
3964 +=======+=======================+
3966 +-------+-----+---------+-------+
3967 | 1 | ANY | ``num`` | ``1`` |
3968 +-------+-----+---------+-------+
3970 +-------+-----------------------+
3972 +-------+-----------------------+
3974 Knowing that TCP does not make sense with something other than IPv4 and IPv6
3975 as L3, such a pattern may be translated to two flow rules instead:
3977 .. _table_rte_flow_unsupported_any_ipv4:
3979 .. table:: ANY replaced with IPV4
3981 +-------+--------------------+
3983 +=======+====================+
3985 +-------+--------------------+
3986 | 1 | IPV4 (zeroed mask) |
3987 +-------+--------------------+
3989 +-------+--------------------+
3991 +-------+--------------------+
3995 .. _table_rte_flow_unsupported_any_ipv6:
3997 .. table:: ANY replaced with IPV6
3999 +-------+--------------------+
4001 +=======+====================+
4003 +-------+--------------------+
4004 | 1 | IPV6 (zeroed mask) |
4005 +-------+--------------------+
4007 +-------+--------------------+
4009 +-------+--------------------+
4011 Note that as soon as a ANY rule covers several layers, this approach may
4012 yield a large number of hidden flow rules. It is thus suggested to only
4013 support the most common scenarios (anything as L2 and/or L3).
4018 - When combined with `Action: QUEUE`_, packet counting (`Action: COUNT`_)
4019 and tagging (`Action: MARK`_ or `Action: FLAG`_) may be implemented in
4020 software as long as the target queue is used by a single rule.
4022 - When a single target queue is provided, `Action: RSS`_ can also be
4023 implemented through `Action: QUEUE`_.
4028 While it would naturally make sense, flow rules cannot be assumed to be
4029 processed by hardware in the same order as their creation for several
4032 - They may be managed internally as a tree or a hash table instead of a
4034 - Removing a flow rule before adding another one can either put the new rule
4035 at the end of the list or reuse a freed entry.
4036 - Duplication may occur when packets are matched by several rules.
4038 For overlapping rules (particularly in order to use `Action: PASSTHRU`_)
4039 predictable behavior is only guaranteed by using different priority levels.
4041 Priority levels are not necessarily implemented in hardware, or may be
4042 severely limited (e.g. a single priority bit).
4044 For these reasons, priority levels may be implemented purely in software by
4047 - For devices expecting flow rules to be added in the correct order, PMDs
4048 may destroy and re-create existing rules after adding a new one with
4051 - A configurable number of dummy or empty rules can be created at
4052 initialization time to save high priority slots for later.
4054 - In order to save priority levels, PMDs may evaluate whether rules are
4055 likely to collide and adjust their priority accordingly.
4058 .. _OpenFlow Switch Specification: https://www.opennetworking.org/software-defined-standards/specifications/
git.droids-corp.org / dpdk.git / blob