2 Copyright 2016 6WIND S.A.
3 Copyright 2016 Mellanox.
5 Redistribution and use in source and binary forms, with or without
6 modification, are permitted provided that the following conditions
9 * Redistributions of source code must retain the above copyright
10 notice, this list of conditions and the following disclaimer.
11 * Redistributions in binary form must reproduce the above copyright
12 notice, this list of conditions and the following disclaimer in
13 the documentation and/or other materials provided with the
15 * Neither the name of 6WIND S.A. nor the names of its
16 contributors may be used to endorse or promote products derived
17 from this software without specific prior written permission.
19 THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
20 "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
21 LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
22 A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
23 OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
24 SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
25 LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
26 DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
27 THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
28 (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
29 OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
33 Generic flow API (rte_flow)
34 ===========================
39 This API provides a generic means to configure hardware to match specific
40 ingress or egress traffic, alter its fate and query related counters
41 according to any number of user-defined rules.
43 It is named *rte_flow* after the prefix used for all its symbols, and is
44 defined in ``rte_flow.h``.
46 - Matching can be performed on packet data (protocol headers, payload) and
47 properties (e.g. associated physical port, virtual device function ID).
49 - Possible operations include dropping traffic, diverting it to specific
50 queues, to virtual/physical device functions or ports, performing tunnel
51 offloads, adding marks and so on.
53 It is slightly higher-level than the legacy filtering framework which it
54 encompasses and supersedes (including all functions and filter types) in
55 order to expose a single interface with an unambiguous behavior that is
56 common to all poll-mode drivers (PMDs).
58 Several methods to migrate existing applications are described in `API
67 A flow rule is the combination of attributes with a matching pattern and a
68 list of actions. Flow rules form the basis of this API.
70 Flow rules can have several distinct actions (such as counting,
71 encapsulating, decapsulating before redirecting packets to a particular
72 queue, etc.), instead of relying on several rules to achieve this and having
73 applications deal with hardware implementation details regarding their
76 Support for different priority levels on a rule basis is provided, for
77 example in order to force a more specific rule to come before a more generic
78 one for packets matched by both. However hardware support for more than a
79 single priority level cannot be guaranteed. When supported, the number of
80 available priority levels is usually low, which is why they can also be
81 implemented in software by PMDs (e.g. missing priority levels may be
82 emulated by reordering rules).
84 In order to remain as hardware-agnostic as possible, by default all rules
85 are considered to have the same priority, which means that the order between
86 overlapping rules (when a packet is matched by several filters) is
89 PMDs may refuse to create overlapping rules at a given priority level when
90 they can be detected (e.g. if a pattern matches an existing filter).
92 Thus predictable results for a given priority level can only be achieved
93 with non-overlapping rules, using perfect matching on all protocol layers.
95 Flow rules can also be grouped, the flow rule priority is specific to the
96 group they belong to. All flow rules in a given group are thus processed
97 either before or after another group.
99 Support for multiple actions per rule may be implemented internally on top
100 of non-default hardware priorities, as a result both features may not be
101 simultaneously available to applications.
103 Considering that allowed pattern/actions combinations cannot be known in
104 advance and would result in an impractically large number of capabilities to
105 expose, a method is provided to validate a given rule from the current
106 device configuration state.
108 This enables applications to check if the rule types they need is supported
109 at initialization time, before starting their data path. This method can be
110 used anytime, its only requirement being that the resources needed by a rule
111 should exist (e.g. a target RX queue should be configured first).
113 Each defined rule is associated with an opaque handle managed by the PMD,
114 applications are responsible for keeping it. These can be used for queries
115 and rules management, such as retrieving counters or other data and
118 To avoid resource leaks on the PMD side, handles must be explicitly
119 destroyed by the application before releasing associated resources such as
122 The following sections cover:
124 - **Attributes** (represented by ``struct rte_flow_attr``): properties of a
125 flow rule such as its direction (ingress or egress) and priority.
127 - **Pattern item** (represented by ``struct rte_flow_item``): part of a
128 matching pattern that either matches specific packet data or traffic
129 properties. It can also describe properties of the pattern itself, such as
132 - **Matching pattern**: traffic properties to look for, a combination of any
135 - **Actions** (represented by ``struct rte_flow_action``): operations to
136 perform whenever a packet is matched by a pattern.
144 Flow rules can be grouped by assigning them a common group number. Lower
145 values have higher priority. Group 0 has the highest priority.
147 Although optional, applications are encouraged to group similar rules as
148 much as possible to fully take advantage of hardware capabilities
149 (e.g. optimized matching) and work around limitations (e.g. a single pattern
150 type possibly allowed in a given group).
152 Note that support for more than a single group is not guaranteed.
157 A priority level can be assigned to a flow rule. Like groups, lower values
158 denote higher priority, with 0 as the maximum.
160 A rule with priority 0 in group 8 is always matched after a rule with
161 priority 8 in group 0.
163 Group and priority levels are arbitrary and up to the application, they do
164 not need to be contiguous nor start from 0, however the maximum number
165 varies between devices and may be affected by existing flow rules.
167 If a packet is matched by several rules of a given group for a given
168 priority level, the outcome is undefined. It can take any path, may be
169 duplicated or even cause unrecoverable errors.
171 Note that support for more than a single priority level is not guaranteed.
173 Attribute: Traffic direction
174 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
176 Flow rules can apply to inbound and/or outbound traffic (ingress/egress).
178 Several pattern items and actions are valid and can be used in both
179 directions. At least one direction must be specified.
181 Specifying both directions at once for a given rule is not recommended but
182 may be valid in a few cases (e.g. shared counters).
187 Pattern items fall in two categories:
189 - Matching protocol headers and packet data (ANY, RAW, ETH, VLAN, IPV4,
190 IPV6, ICMP, UDP, TCP, SCTP, VXLAN and so on), usually associated with a
191 specification structure.
193 - Matching meta-data or affecting pattern processing (END, VOID, INVERT, PF,
194 VF, PORT and so on), often without a specification structure.
196 Item specification structures are used to match specific values among
197 protocol fields (or item properties). Documentation describes for each item
198 whether they are associated with one and their type name if so.
200 Up to three structures of the same type can be set for a given item:
202 - ``spec``: values to match (e.g. a given IPv4 address).
204 - ``last``: upper bound for an inclusive range with corresponding fields in
207 - ``mask``: bit-mask applied to both ``spec`` and ``last`` whose purpose is
208 to distinguish the values to take into account and/or partially mask them
209 out (e.g. in order to match an IPv4 address prefix).
211 Usage restrictions and expected behavior:
213 - Setting either ``mask`` or ``last`` without ``spec`` is an error.
215 - Field values in ``last`` which are either 0 or equal to the corresponding
216 values in ``spec`` are ignored; they do not generate a range. Nonzero
217 values lower than those in ``spec`` are not supported.
219 - Setting ``spec`` and optionally ``last`` without ``mask`` causes the PMD
220 to only take the fields it can recognize into account. There is no error
221 checking for unsupported fields.
223 - Not setting any of them (assuming item type allows it) uses default
224 parameters that depend on the item type. Most of the time, particularly
225 for protocol header items, it is equivalent to providing an empty (zeroed)
228 - ``mask`` is a simple bit-mask applied before interpreting the contents of
229 ``spec`` and ``last``, which may yield unexpected results if not used
230 carefully. For example, if for an IPv4 address field, ``spec`` provides
231 *10.1.2.3*, ``last`` provides *10.3.4.5* and ``mask`` provides
232 *255.255.0.0*, the effective range becomes *10.1.0.0* to *10.3.255.255*.
234 Example of an item specification matching an Ethernet header:
236 .. _table_rte_flow_pattern_item_example:
238 .. table:: Ethernet item
240 +----------+----------+--------------------+
241 | Field | Subfield | Value |
242 +==========+==========+====================+
243 | ``spec`` | ``src`` | ``00:01:02:03:04`` |
244 | +----------+--------------------+
245 | | ``dst`` | ``00:2a:66:00:01`` |
246 | +----------+--------------------+
247 | | ``type`` | ``0x22aa`` |
248 +----------+----------+--------------------+
249 | ``last`` | unspecified |
250 +----------+----------+--------------------+
251 | ``mask`` | ``src`` | ``00:ff:ff:ff:00`` |
252 | +----------+--------------------+
253 | | ``dst`` | ``00:00:00:00:ff`` |
254 | +----------+--------------------+
255 | | ``type`` | ``0x0000`` |
256 +----------+----------+--------------------+
258 Non-masked bits stand for any value (shown as ``?`` below), Ethernet headers
259 with the following properties are thus matched:
261 - ``src``: ``??:01:02:03:??``
262 - ``dst``: ``??:??:??:??:01``
263 - ``type``: ``0x????``
268 A pattern is formed by stacking items starting from the lowest protocol
269 layer to match. This stacking restriction does not apply to meta items which
270 can be placed anywhere in the stack without affecting the meaning of the
273 Patterns are terminated by END items.
277 .. _table_rte_flow_tcpv4_as_l4:
279 .. table:: TCPv4 as L4
295 .. _table_rte_flow_tcpv6_in_vxlan:
297 .. table:: TCPv6 in VXLAN
299 +-------+------------+
301 +=======+============+
303 +-------+------------+
305 +-------+------------+
307 +-------+------------+
309 +-------+------------+
311 +-------+------------+
313 +-------+------------+
315 +-------+------------+
317 +-------+------------+
321 .. _table_rte_flow_tcpv4_as_l4_meta:
323 .. table:: TCPv4 as L4 with meta items
345 The above example shows how meta items do not affect packet data matching
346 items, as long as those remain stacked properly. The resulting matching
347 pattern is identical to "TCPv4 as L4".
349 .. _table_rte_flow_udpv6_anywhere:
351 .. table:: UDPv6 anywhere
363 If supported by the PMD, omitting one or several protocol layers at the
364 bottom of the stack as in the above example (missing an Ethernet
365 specification) enables looking up anywhere in packets.
367 It is unspecified whether the payload of supported encapsulations
368 (e.g. VXLAN payload) is matched by such a pattern, which may apply to inner,
369 outer or both packets.
371 .. _table_rte_flow_invalid_l3:
373 .. table:: Invalid, missing L3
385 The above pattern is invalid due to a missing L3 specification between L2
386 (Ethernet) and L4 (UDP). Doing so is only allowed at the bottom and at the
392 They match meta-data or affect pattern processing instead of matching packet
393 data directly, most of them do not need a specification structure. This
394 particularity allows them to be specified anywhere in the stack without
395 causing any side effect.
400 End marker for item lists. Prevents further processing of items, thereby
403 - Its numeric value is 0 for convenience.
404 - PMD support is mandatory.
405 - ``spec``, ``last`` and ``mask`` are ignored.
407 .. _table_rte_flow_item_end:
411 +----------+---------+
413 +==========+=========+
414 | ``spec`` | ignored |
415 +----------+---------+
416 | ``last`` | ignored |
417 +----------+---------+
418 | ``mask`` | ignored |
419 +----------+---------+
424 Used as a placeholder for convenience. It is ignored and simply discarded by
427 - PMD support is mandatory.
428 - ``spec``, ``last`` and ``mask`` are ignored.
430 .. _table_rte_flow_item_void:
434 +----------+---------+
436 +==========+=========+
437 | ``spec`` | ignored |
438 +----------+---------+
439 | ``last`` | ignored |
440 +----------+---------+
441 | ``mask`` | ignored |
442 +----------+---------+
444 One usage example for this type is generating rules that share a common
445 prefix quickly without reallocating memory, only by updating item types:
447 .. _table_rte_flow_item_void_example:
449 .. table:: TCP, UDP or ICMP as L4
451 +-------+--------------------+
453 +=======+====================+
455 +-------+--------------------+
457 +-------+------+------+------+
458 | 2 | UDP | VOID | VOID |
459 +-------+------+------+------+
460 | 3 | VOID | TCP | VOID |
461 +-------+------+------+------+
462 | 4 | VOID | VOID | ICMP |
463 +-------+------+------+------+
465 +-------+--------------------+
470 Inverted matching, i.e. process packets that do not match the pattern.
472 - ``spec``, ``last`` and ``mask`` are ignored.
474 .. _table_rte_flow_item_invert:
478 +----------+---------+
480 +==========+=========+
481 | ``spec`` | ignored |
482 +----------+---------+
483 | ``last`` | ignored |
484 +----------+---------+
485 | ``mask`` | ignored |
486 +----------+---------+
488 Usage example, matching non-TCPv4 packets only:
490 .. _table_rte_flow_item_invert_example:
492 .. table:: Anything but TCPv4
511 Matches packets addressed to the physical function of the device.
513 If the underlying device function differs from the one that would normally
514 receive the matched traffic, specifying this item prevents it from reaching
515 that device unless the flow rule contains a `Action: PF`_. Packets are not
516 duplicated between device instances by default.
518 - Likely to return an error or never match any traffic if applied to a VF
520 - Can be combined with any number of `Item: VF`_ to match both PF and VF
522 - ``spec``, ``last`` and ``mask`` must not be set.
524 .. _table_rte_flow_item_pf:
541 Matches packets addressed to a virtual function ID of the device.
543 If the underlying device function differs from the one that would normally
544 receive the matched traffic, specifying this item prevents it from reaching
545 that device unless the flow rule contains a `Action: VF`_. Packets are not
546 duplicated between device instances by default.
548 - Likely to return an error or never match any traffic if this causes a VF
549 device to match traffic addressed to a different VF.
550 - Can be specified multiple times to match traffic addressed to several VF
552 - Can be combined with a PF item to match both PF and VF traffic.
554 .. _table_rte_flow_item_vf:
558 +----------+----------+---------------------------+
559 | Field | Subfield | Value |
560 +==========+==========+===========================+
561 | ``spec`` | ``id`` | destination VF ID |
562 +----------+----------+---------------------------+
563 | ``last`` | ``id`` | upper range value |
564 +----------+----------+---------------------------+
565 | ``mask`` | ``id`` | zeroed to match any VF ID |
566 +----------+----------+---------------------------+
571 Matches packets coming from the specified physical port of the underlying
574 The first PORT item overrides the physical port normally associated with the
575 specified DPDK input port (port_id). This item can be provided several times
576 to match additional physical ports.
578 Note that physical ports are not necessarily tied to DPDK input ports
579 (port_id) when those are not under DPDK control. Possible values are
580 specific to each device, they are not necessarily indexed from zero and may
583 As a device property, the list of allowed values as well as the value
584 associated with a port_id should be retrieved by other means.
586 .. _table_rte_flow_item_port:
590 +----------+-----------+--------------------------------+
591 | Field | Subfield | Value |
592 +==========+===========+================================+
593 | ``spec`` | ``index`` | physical port index |
594 +----------+-----------+--------------------------------+
595 | ``last`` | ``index`` | upper range value |
596 +----------+-----------+--------------------------------+
597 | ``mask`` | ``index`` | zeroed to match any port index |
598 +----------+-----------+--------------------------------+
600 Data matching item types
601 ~~~~~~~~~~~~~~~~~~~~~~~~
603 Most of these are basically protocol header definitions with associated
604 bit-masks. They must be specified (stacked) from lowest to highest protocol
605 layer to form a matching pattern.
607 The following list is not exhaustive, new protocols will be added in the
613 Matches any protocol in place of the current layer, a single ANY may also
614 stand for several protocol layers.
616 This is usually specified as the first pattern item when looking for a
617 protocol anywhere in a packet.
619 .. _table_rte_flow_item_any:
623 +----------+----------+--------------------------------------+
624 | Field | Subfield | Value |
625 +==========+==========+======================================+
626 | ``spec`` | ``num`` | number of layers covered |
627 +----------+----------+--------------------------------------+
628 | ``last`` | ``num`` | upper range value |
629 +----------+----------+--------------------------------------+
630 | ``mask`` | ``num`` | zeroed to cover any number of layers |
631 +----------+----------+--------------------------------------+
633 Example for VXLAN TCP payload matching regardless of outer L3 (IPv4 or IPv6)
634 and L4 (UDP) both matched by the first ANY specification, and inner L3 (IPv4
635 or IPv6) matched by the second ANY specification:
637 .. _table_rte_flow_item_any_example:
639 .. table:: TCP in VXLAN with wildcards
641 +-------+------+----------+----------+-------+
642 | Index | Item | Field | Subfield | Value |
643 +=======+======+==========+==========+=======+
645 +-------+------+----------+----------+-------+
646 | 1 | ANY | ``spec`` | ``num`` | 2 |
647 +-------+------+----------+----------+-------+
649 +-------+------------------------------------+
651 +-------+------+----------+----------+-------+
652 | 4 | ANY | ``spec`` | ``num`` | 1 |
653 +-------+------+----------+----------+-------+
655 +-------+------------------------------------+
657 +-------+------------------------------------+
662 Matches a byte string of a given length at a given offset.
664 Offset is either absolute (using the start of the packet) or relative to the
665 end of the previous matched item in the stack, in which case negative values
668 If search is enabled, offset is used as the starting point. The search area
669 can be delimited by setting limit to a nonzero value, which is the maximum
670 number of bytes after offset where the pattern may start.
672 Matching a zero-length pattern is allowed, doing so resets the relative
673 offset for subsequent items.
675 - This type does not support ranges (``last`` field).
677 .. _table_rte_flow_item_raw:
681 +----------+--------------+-------------------------------------------------+
682 | Field | Subfield | Value |
683 +==========+==============+=================================================+
684 | ``spec`` | ``relative`` | look for pattern after the previous item |
685 | +--------------+-------------------------------------------------+
686 | | ``search`` | search pattern from offset (see also ``limit``) |
687 | +--------------+-------------------------------------------------+
688 | | ``reserved`` | reserved, must be set to zero |
689 | +--------------+-------------------------------------------------+
690 | | ``offset`` | absolute or relative offset for ``pattern`` |
691 | +--------------+-------------------------------------------------+
692 | | ``limit`` | search area limit for start of ``pattern`` |
693 | +--------------+-------------------------------------------------+
694 | | ``length`` | ``pattern`` length |
695 | +--------------+-------------------------------------------------+
696 | | ``pattern`` | byte string to look for |
697 +----------+--------------+-------------------------------------------------+
698 | ``last`` | if specified, either all 0 or with the same values as ``spec`` |
699 +----------+----------------------------------------------------------------+
700 | ``mask`` | bit-mask applied to ``spec`` values with usual behavior |
701 +----------+----------------------------------------------------------------+
703 Example pattern looking for several strings at various offsets of a UDP
704 payload, using combined RAW items:
706 .. _table_rte_flow_item_raw_example:
708 .. table:: UDP payload matching
710 +-------+------+----------+--------------+-------+
711 | Index | Item | Field | Subfield | Value |
712 +=======+======+==========+==============+=======+
714 +-------+----------------------------------------+
716 +-------+----------------------------------------+
718 +-------+------+----------+--------------+-------+
719 | 3 | RAW | ``spec`` | ``relative`` | 1 |
720 | | | +--------------+-------+
721 | | | | ``search`` | 1 |
722 | | | +--------------+-------+
723 | | | | ``offset`` | 10 |
724 | | | +--------------+-------+
725 | | | | ``limit`` | 0 |
726 | | | +--------------+-------+
727 | | | | ``length`` | 3 |
728 | | | +--------------+-------+
729 | | | | ``pattern`` | "foo" |
730 +-------+------+----------+--------------+-------+
731 | 4 | RAW | ``spec`` | ``relative`` | 1 |
732 | | | +--------------+-------+
733 | | | | ``search`` | 0 |
734 | | | +--------------+-------+
735 | | | | ``offset`` | 20 |
736 | | | +--------------+-------+
737 | | | | ``limit`` | 0 |
738 | | | +--------------+-------+
739 | | | | ``length`` | 3 |
740 | | | +--------------+-------+
741 | | | | ``pattern`` | "bar" |
742 +-------+------+----------+--------------+-------+
743 | 5 | RAW | ``spec`` | ``relative`` | 1 |
744 | | | +--------------+-------+
745 | | | | ``search`` | 0 |
746 | | | +--------------+-------+
747 | | | | ``offset`` | -29 |
748 | | | +--------------+-------+
749 | | | | ``limit`` | 0 |
750 | | | +--------------+-------+
751 | | | | ``length`` | 3 |
752 | | | +--------------+-------+
753 | | | | ``pattern`` | "baz" |
754 +-------+------+----------+--------------+-------+
756 +-------+----------------------------------------+
760 - Locate "foo" at least 10 bytes deep inside UDP payload.
761 - Locate "bar" after "foo" plus 20 bytes.
762 - Locate "baz" after "bar" minus 29 bytes.
764 Such a packet may be represented as follows (not to scale)::
767 | |<--------->| |<--------->|
769 |-----|------|-----|-----|-----|-----|-----------|-----|------|
770 | ETH | IPv4 | UDP | ... | baz | foo | ......... | bar | .... |
771 |-----|------|-----|-----|-----|-----|-----------|-----|------|
773 |<--------------------------->|
776 Note that matching subsequent pattern items would resume after "baz", not
777 "bar" since matching is always performed after the previous item of the
783 Matches an Ethernet header.
785 - ``dst``: destination MAC.
786 - ``src``: source MAC.
787 - ``type``: EtherType.
792 Matches an 802.1Q/ad VLAN tag.
794 - ``tpid``: tag protocol identifier.
795 - ``tci``: tag control information.
800 Matches an IPv4 header.
802 Note: IPv4 options are handled by dedicated pattern items.
804 - ``hdr``: IPv4 header definition (``rte_ip.h``).
809 Matches an IPv6 header.
811 Note: IPv6 options are handled by dedicated pattern items.
813 - ``hdr``: IPv6 header definition (``rte_ip.h``).
818 Matches an ICMP header.
820 - ``hdr``: ICMP header definition (``rte_icmp.h``).
825 Matches a UDP header.
827 - ``hdr``: UDP header definition (``rte_udp.h``).
832 Matches a TCP header.
834 - ``hdr``: TCP header definition (``rte_tcp.h``).
839 Matches a SCTP header.
841 - ``hdr``: SCTP header definition (``rte_sctp.h``).
846 Matches a VXLAN header (RFC 7348).
848 - ``flags``: normally 0x08 (I flag).
849 - ``rsvd0``: reserved, normally 0x000000.
850 - ``vni``: VXLAN network identifier.
851 - ``rsvd1``: reserved, normally 0x00.
856 Each possible action is represented by a type. Some have associated
857 configuration structures. Several actions combined in a list can be affected
858 to a flow rule. That list is not ordered.
860 They fall in three categories:
862 - Terminating actions (such as QUEUE, DROP, RSS, PF, VF) that prevent
863 processing matched packets by subsequent flow rules, unless overridden
866 - Non-terminating actions (PASSTHRU, DUP) that leave matched packets up for
867 additional processing by subsequent flow rules.
869 - Other non-terminating meta actions that do not affect the fate of packets
870 (END, VOID, MARK, FLAG, COUNT).
872 When several actions are combined in a flow rule, they should all have
873 different types (e.g. dropping a packet twice is not possible).
875 Only the last action of a given type is taken into account. PMDs still
876 perform error checking on the entire list.
878 Like matching patterns, action lists are terminated by END items.
880 *Note that PASSTHRU is the only action able to override a terminating rule.*
882 Example of action that redirects packets to queue index 10:
884 .. _table_rte_flow_action_example:
886 .. table:: Queue action
888 +-----------+-------+
890 +===========+=======+
892 +-----------+-------+
894 Action lists examples, their order is not significant, applications must
895 consider all actions to be performed simultaneously:
897 .. _table_rte_flow_count_and_drop:
899 .. table:: Count and drop
913 .. _table_rte_flow_mark_count_redirect:
915 .. table:: Mark, count and redirect
917 +-------+--------+-----------+-------+
918 | Index | Action | Field | Value |
919 +=======+========+===========+=======+
920 | 0 | MARK | ``mark`` | 0x2a |
921 +-------+--------+-----------+-------+
923 +-------+--------+-----------+-------+
924 | 2 | QUEUE | ``queue`` | 10 |
925 +-------+--------+-----------+-------+
927 +-------+----------------------------+
931 .. _table_rte_flow_redirect_queue_5:
933 .. table:: Redirect to queue 5
935 +-------+--------+-----------+-------+
936 | Index | Action | Field | Value |
937 +=======+========+===========+=======+
939 +-------+--------+-----------+-------+
940 | 1 | QUEUE | ``queue`` | 5 |
941 +-------+--------+-----------+-------+
943 +-------+----------------------------+
945 In the above example, considering both actions are performed simultaneously,
946 the end result is that only QUEUE has any effect.
948 .. _table_rte_flow_redirect_queue_3:
950 .. table:: Redirect to queue 3
952 +-------+--------+-----------+-------+
953 | Index | Action | Field | Value |
954 +=======+========+===========+=======+
955 | 0 | QUEUE | ``queue`` | 5 |
956 +-------+--------+-----------+-------+
958 +-------+--------+-----------+-------+
959 | 2 | QUEUE | ``queue`` | 3 |
960 +-------+--------+-----------+-------+
962 +-------+----------------------------+
964 As previously described, only the last action of a given type found in the
965 list is taken into account. The above example also shows that VOID is
971 Common action types are described in this section. Like pattern item types,
972 this list is not exhaustive as new actions will be added in the future.
977 End marker for action lists. Prevents further processing of actions, thereby
980 - Its numeric value is 0 for convenience.
981 - PMD support is mandatory.
982 - No configurable properties.
984 .. _table_rte_flow_action_end:
997 Used as a placeholder for convenience. It is ignored and simply discarded by
1000 - PMD support is mandatory.
1001 - No configurable properties.
1003 .. _table_rte_flow_action_void:
1013 Action: ``PASSTHRU``
1014 ^^^^^^^^^^^^^^^^^^^^
1016 Leaves packets up for additional processing by subsequent flow rules. This
1017 is the default when a rule does not contain a terminating action, but can be
1018 specified to force a rule to become non-terminating.
1020 - No configurable properties.
1022 .. _table_rte_flow_action_passthru:
1032 Example to copy a packet to a queue and continue processing by subsequent
1035 .. _table_rte_flow_action_passthru_example:
1037 .. table:: Copy to queue 8
1039 +-------+--------+-----------+-------+
1040 | Index | Action | Field | Value |
1041 +=======+========+===========+=======+
1043 +-------+--------+-----------+-------+
1044 | 1 | QUEUE | ``queue`` | 8 |
1045 +-------+--------+-----------+-------+
1047 +-------+----------------------------+
1052 Attaches an integer value to packets and sets ``PKT_RX_FDIR`` and
1053 ``PKT_RX_FDIR_ID`` mbuf flags.
1055 This value is arbitrary and application-defined. Maximum allowed value
1056 depends on the underlying implementation. It is returned in the
1057 ``hash.fdir.hi`` mbuf field.
1059 .. _table_rte_flow_action_mark:
1063 +--------+--------------------------------------+
1065 +========+======================================+
1066 | ``id`` | integer value to return with packets |
1067 +--------+--------------------------------------+
1072 Flags packets. Similar to `Action: MARK`_ without a specific value; only
1073 sets the ``PKT_RX_FDIR`` mbuf flag.
1075 - No configurable properties.
1077 .. _table_rte_flow_action_flag:
1090 Assigns packets to a given queue index.
1092 - Terminating by default.
1094 .. _table_rte_flow_action_queue:
1098 +-----------+--------------------+
1100 +===========+====================+
1101 | ``index`` | queue index to use |
1102 +-----------+--------------------+
1109 - No configurable properties.
1110 - Terminating by default.
1111 - PASSTHRU overrides this action if both are specified.
1113 .. _table_rte_flow_action_drop:
1126 Enables counters for this rule.
1128 These counters can be retrieved and reset through ``rte_flow_query()``, see
1129 ``struct rte_flow_query_count``.
1131 - Counters can be retrieved with ``rte_flow_query()``.
1132 - No configurable properties.
1134 .. _table_rte_flow_action_count:
1144 Query structure to retrieve and reset flow rule counters:
1146 .. _table_rte_flow_query_count:
1148 .. table:: COUNT query
1150 +---------------+-----+-----------------------------------+
1151 | Field | I/O | Value |
1152 +===============+=====+===================================+
1153 | ``reset`` | in | reset counter after query |
1154 +---------------+-----+-----------------------------------+
1155 | ``hits_set`` | out | ``hits`` field is set |
1156 +---------------+-----+-----------------------------------+
1157 | ``bytes_set`` | out | ``bytes`` field is set |
1158 +---------------+-----+-----------------------------------+
1159 | ``hits`` | out | number of hits for this rule |
1160 +---------------+-----+-----------------------------------+
1161 | ``bytes`` | out | number of bytes through this rule |
1162 +---------------+-----+-----------------------------------+
1167 Duplicates packets to a given queue index.
1169 This is normally combined with QUEUE, however when used alone, it is
1170 actually similar to QUEUE + PASSTHRU.
1172 - Non-terminating by default.
1174 .. _table_rte_flow_action_dup:
1178 +-----------+------------------------------------+
1180 +===========+====================================+
1181 | ``index`` | queue index to duplicate packet to |
1182 +-----------+------------------------------------+
1187 Similar to QUEUE, except RSS is additionally performed on packets to spread
1188 them among several queues according to the provided parameters.
1190 Note: RSS hash result is normally stored in the ``hash.rss`` mbuf field,
1191 however it conflicts with `Action: MARK`_ as they share the same space. When
1192 both actions are specified, the RSS hash is discarded and
1193 ``PKT_RX_RSS_HASH`` is not set in ``ol_flags``. MARK has priority. The mbuf
1194 structure should eventually evolve to store both.
1196 - Terminating by default.
1198 .. _table_rte_flow_action_rss:
1202 +--------------+------------------------------+
1204 +==============+==============================+
1205 | ``rss_conf`` | RSS parameters |
1206 +--------------+------------------------------+
1207 | ``num`` | number of entries in queue[] |
1208 +--------------+------------------------------+
1209 | ``queue[]`` | queue indices to use |
1210 +--------------+------------------------------+
1215 Redirects packets to the physical function (PF) of the current device.
1217 - No configurable properties.
1218 - Terminating by default.
1220 .. _table_rte_flow_action_pf:
1233 Redirects packets to a virtual function (VF) of the current device.
1235 Packets matched by a VF pattern item can be redirected to their original VF
1236 ID instead of the specified one. This parameter may not be available and is
1237 not guaranteed to work properly if the VF part is matched by a prior flow
1238 rule or if packets are not addressed to a VF in the first place.
1240 - Terminating by default.
1242 .. _table_rte_flow_action_vf:
1246 +--------------+--------------------------------+
1248 +==============+================================+
1249 | ``original`` | use original VF ID if possible |
1250 +--------------+--------------------------------+
1251 | ``vf`` | VF ID to redirect packets to |
1252 +--------------+--------------------------------+
1257 All specified pattern items (``enum rte_flow_item_type``) and actions
1258 (``enum rte_flow_action_type``) use positive identifiers.
1260 The negative space is reserved for dynamic types generated by PMDs during
1261 run-time. PMDs may encounter them as a result but must not accept negative
1262 identifiers they are not aware of.
1264 A method to generate them remains to be defined.
1269 Pattern item types will be added as new protocols are implemented.
1271 Variable headers support through dedicated pattern items, for example in
1272 order to match specific IPv4 options and IPv6 extension headers would be
1273 stacked after IPv4/IPv6 items.
1275 Other action types are planned but are not defined yet. These include the
1276 ability to alter packet data in several ways, such as performing
1277 encapsulation/decapsulation of tunnel headers.
1282 A rather simple API with few functions is provided to fully manage flow
1285 Each created flow rule is associated with an opaque, PMD-specific handle
1286 pointer. The application is responsible for keeping it until the rule is
1289 Flows rules are represented by ``struct rte_flow`` objects.
1294 Given that expressing a definite set of device capabilities is not
1295 practical, a dedicated function is provided to check if a flow rule is
1296 supported and can be created.
1301 rte_flow_validate(uint8_t port_id,
1302 const struct rte_flow_attr *attr,
1303 const struct rte_flow_item pattern[],
1304 const struct rte_flow_action actions[],
1305 struct rte_flow_error *error);
1307 While this function has no effect on the target device, the flow rule is
1308 validated against its current configuration state and the returned value
1309 should be considered valid by the caller for that state only.
1311 The returned value is guaranteed to remain valid only as long as no
1312 successful calls to ``rte_flow_create()`` or ``rte_flow_destroy()`` are made
1313 in the meantime and no device parameter affecting flow rules in any way are
1314 modified, due to possible collisions or resource limitations (although in
1315 such cases ``EINVAL`` should not be returned).
1319 - ``port_id``: port identifier of Ethernet device.
1320 - ``attr``: flow rule attributes.
1321 - ``pattern``: pattern specification (list terminated by the END pattern
1323 - ``actions``: associated actions (list terminated by the END action).
1324 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
1325 this structure in case of error only.
1329 - 0 if flow rule is valid and can be created. A negative errno value
1330 otherwise (``rte_errno`` is also set), the following errors are defined.
1331 - ``-ENOSYS``: underlying device does not support this functionality.
1332 - ``-EINVAL``: unknown or invalid rule specification.
1333 - ``-ENOTSUP``: valid but unsupported rule specification (e.g. partial
1334 bit-masks are unsupported).
1335 - ``-EEXIST``: collision with an existing rule.
1336 - ``-ENOMEM``: not enough resources.
1337 - ``-EBUSY``: action cannot be performed due to busy device resources, may
1338 succeed if the affected queues or even the entire port are in a stopped
1339 state (see ``rte_eth_dev_rx_queue_stop()`` and ``rte_eth_dev_stop()``).
1344 Creating a flow rule is similar to validating one, except the rule is
1345 actually created and a handle returned.
1350 rte_flow_create(uint8_t port_id,
1351 const struct rte_flow_attr *attr,
1352 const struct rte_flow_item pattern[],
1353 const struct rte_flow_action *actions[],
1354 struct rte_flow_error *error);
1358 - ``port_id``: port identifier of Ethernet device.
1359 - ``attr``: flow rule attributes.
1360 - ``pattern``: pattern specification (list terminated by the END pattern
1362 - ``actions``: associated actions (list terminated by the END action).
1363 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
1364 this structure in case of error only.
1368 A valid handle in case of success, NULL otherwise and ``rte_errno`` is set
1369 to the positive version of one of the error codes defined for
1370 ``rte_flow_validate()``.
1375 Flow rules destruction is not automatic, and a queue or a port should not be
1376 released if any are still attached to them. Applications must take care of
1377 performing this step before releasing resources.
1382 rte_flow_destroy(uint8_t port_id,
1383 struct rte_flow *flow,
1384 struct rte_flow_error *error);
1387 Failure to destroy a flow rule handle may occur when other flow rules depend
1388 on it, and destroying it would result in an inconsistent state.
1390 This function is only guaranteed to succeed if handles are destroyed in
1391 reverse order of their creation.
1395 - ``port_id``: port identifier of Ethernet device.
1396 - ``flow``: flow rule handle to destroy.
1397 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
1398 this structure in case of error only.
1402 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
1407 Convenience function to destroy all flow rule handles associated with a
1408 port. They are released as with successive calls to ``rte_flow_destroy()``.
1413 rte_flow_flush(uint8_t port_id,
1414 struct rte_flow_error *error);
1416 In the unlikely event of failure, handles are still considered destroyed and
1417 no longer valid but the port must be assumed to be in an inconsistent state.
1421 - ``port_id``: port identifier of Ethernet device.
1422 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
1423 this structure in case of error only.
1427 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
1432 Query an existing flow rule.
1434 This function allows retrieving flow-specific data such as counters. Data
1435 is gathered by special actions which must be present in the flow rule
1441 rte_flow_query(uint8_t port_id,
1442 struct rte_flow *flow,
1443 enum rte_flow_action_type action,
1445 struct rte_flow_error *error);
1449 - ``port_id``: port identifier of Ethernet device.
1450 - ``flow``: flow rule handle to query.
1451 - ``action``: action type to query.
1452 - ``data``: pointer to storage for the associated query data type.
1453 - ``error``: perform verbose error reporting if not NULL. PMDs initialize
1454 this structure in case of error only.
1458 - 0 on success, a negative errno value otherwise and ``rte_errno`` is set.
1460 Verbose error reporting
1461 -----------------------
1463 The defined *errno* values may not be accurate enough for users or
1464 application developers who want to investigate issues related to flow rules
1465 management. A dedicated error object is defined for this purpose:
1469 enum rte_flow_error_type {
1470 RTE_FLOW_ERROR_TYPE_NONE, /**< No error. */
1471 RTE_FLOW_ERROR_TYPE_UNSPECIFIED, /**< Cause unspecified. */
1472 RTE_FLOW_ERROR_TYPE_HANDLE, /**< Flow rule (handle). */
1473 RTE_FLOW_ERROR_TYPE_ATTR_GROUP, /**< Group field. */
1474 RTE_FLOW_ERROR_TYPE_ATTR_PRIORITY, /**< Priority field. */
1475 RTE_FLOW_ERROR_TYPE_ATTR_INGRESS, /**< Ingress field. */
1476 RTE_FLOW_ERROR_TYPE_ATTR_EGRESS, /**< Egress field. */
1477 RTE_FLOW_ERROR_TYPE_ATTR, /**< Attributes structure. */
1478 RTE_FLOW_ERROR_TYPE_ITEM_NUM, /**< Pattern length. */
1479 RTE_FLOW_ERROR_TYPE_ITEM, /**< Specific pattern item. */
1480 RTE_FLOW_ERROR_TYPE_ACTION_NUM, /**< Number of actions. */
1481 RTE_FLOW_ERROR_TYPE_ACTION, /**< Specific action. */
1484 struct rte_flow_error {
1485 enum rte_flow_error_type type; /**< Cause field and error types. */
1486 const void *cause; /**< Object responsible for the error. */
1487 const char *message; /**< Human-readable error message. */
1490 Error type ``RTE_FLOW_ERROR_TYPE_NONE`` stands for no error, in which case
1491 remaining fields can be ignored. Other error types describe the type of the
1492 object pointed by ``cause``.
1494 If non-NULL, ``cause`` points to the object responsible for the error. For a
1495 flow rule, this may be a pattern item or an individual action.
1497 If non-NULL, ``message`` provides a human-readable error message.
1499 This object is normally allocated by applications and set by PMDs in case of
1500 error, the message points to a constant string which does not need to be
1501 freed by the application, however its pointer can be considered valid only
1502 as long as its associated DPDK port remains configured. Closing the
1503 underlying device or unloading the PMD invalidates it.
1508 - DPDK does not keep track of flow rules definitions or flow rule objects
1509 automatically. Applications may keep track of the former and must keep
1510 track of the latter. PMDs may also do it for internal needs, however this
1511 must not be relied on by applications.
1513 - Flow rules are not maintained between successive port initializations. An
1514 application exiting without releasing them and restarting must re-create
1517 - API operations are synchronous and blocking (``EAGAIN`` cannot be
1520 - There is no provision for reentrancy/multi-thread safety, although nothing
1521 should prevent different devices from being configured at the same
1522 time. PMDs may protect their control path functions accordingly.
1524 - Stopping the data path (TX/RX) should not be necessary when managing flow
1525 rules. If this cannot be achieved naturally or with workarounds (such as
1526 temporarily replacing the burst function pointers), an appropriate error
1527 code must be returned (``EBUSY``).
1529 - PMDs, not applications, are responsible for maintaining flow rules
1530 configuration when stopping and restarting a port or performing other
1531 actions which may affect them. They can only be destroyed explicitly by
1534 For devices exposing multiple ports sharing global settings affected by flow
1537 - All ports under DPDK control must behave consistently, PMDs are
1538 responsible for making sure that existing flow rules on a port are not
1539 affected by other ports.
1541 - Ports not under DPDK control (unaffected or handled by other applications)
1542 are user's responsibility. They may affect existing flow rules and cause
1543 undefined behavior. PMDs aware of this may prevent flow rules creation
1544 altogether in such cases.
1549 The PMD interface is defined in ``rte_flow_driver.h``. It is not subject to
1550 API/ABI versioning constraints as it is not exposed to applications and may
1551 evolve independently.
1553 It is currently implemented on top of the legacy filtering framework through
1554 filter type *RTE_ETH_FILTER_GENERIC* that accepts the single operation
1555 *RTE_ETH_FILTER_GET* to return PMD-specific *rte_flow* callbacks wrapped
1556 inside ``struct rte_flow_ops``.
1558 This overhead is temporarily necessary in order to keep compatibility with
1559 the legacy filtering framework, which should eventually disappear.
1561 - PMD callbacks implement exactly the interface described in `Rules
1562 management`_, except for the port ID argument which has already been
1563 converted to a pointer to the underlying ``struct rte_eth_dev``.
1565 - Public API functions do not process flow rules definitions at all before
1566 calling PMD functions (no basic error checking, no validation
1567 whatsoever). They only make sure these callbacks are non-NULL or return
1568 the ``ENOSYS`` (function not supported) error.
1570 This interface additionally defines the following helper functions:
1572 - ``rte_flow_ops_get()``: get generic flow operations structure from a
1575 - ``rte_flow_error_set()``: initialize generic flow error structure.
1577 More will be added over time.
1579 Device compatibility
1580 --------------------
1582 No known implementation supports all the described features.
1584 Unsupported features or combinations are not expected to be fully emulated
1585 in software by PMDs for performance reasons. Partially supported features
1586 may be completed in software as long as hardware performs most of the work
1587 (such as queue redirection and packet recognition).
1589 However PMDs are expected to do their best to satisfy application requests
1590 by working around hardware limitations as long as doing so does not affect
1591 the behavior of existing flow rules.
1593 The following sections provide a few examples of such cases and describe how
1594 PMDs should handle them, they are based on limitations built into the
1600 Each flow rule comes with its own, per-layer bit-masks, while hardware may
1601 support only a single, device-wide bit-mask for a given layer type, so that
1602 two IPv4 rules cannot use different bit-masks.
1604 The expected behavior in this case is that PMDs automatically configure
1605 global bit-masks according to the needs of the first flow rule created.
1607 Subsequent rules are allowed only if their bit-masks match those, the
1608 ``EEXIST`` error code should be returned otherwise.
1610 Unsupported layer types
1611 ~~~~~~~~~~~~~~~~~~~~~~~
1613 Many protocols can be simulated by crafting patterns with the `Item: RAW`_
1616 PMDs can rely on this capability to simulate support for protocols with
1617 headers not directly recognized by hardware.
1619 ``ANY`` pattern item
1620 ~~~~~~~~~~~~~~~~~~~~
1622 This pattern item stands for anything, which can be difficult to translate
1623 to something hardware would understand, particularly if followed by more
1626 Consider the following pattern:
1628 .. _table_rte_flow_unsupported_any:
1630 .. table:: Pattern with ANY as L3
1632 +-------+-----------------------+
1634 +=======+=======================+
1636 +-------+-----+---------+-------+
1637 | 1 | ANY | ``num`` | ``1`` |
1638 +-------+-----+---------+-------+
1640 +-------+-----------------------+
1642 +-------+-----------------------+
1644 Knowing that TCP does not make sense with something other than IPv4 and IPv6
1645 as L3, such a pattern may be translated to two flow rules instead:
1647 .. _table_rte_flow_unsupported_any_ipv4:
1649 .. table:: ANY replaced with IPV4
1651 +-------+--------------------+
1653 +=======+====================+
1655 +-------+--------------------+
1656 | 1 | IPV4 (zeroed mask) |
1657 +-------+--------------------+
1659 +-------+--------------------+
1661 +-------+--------------------+
1665 .. _table_rte_flow_unsupported_any_ipv6:
1667 .. table:: ANY replaced with IPV6
1669 +-------+--------------------+
1671 +=======+====================+
1673 +-------+--------------------+
1674 | 1 | IPV6 (zeroed mask) |
1675 +-------+--------------------+
1677 +-------+--------------------+
1679 +-------+--------------------+
1681 Note that as soon as a ANY rule covers several layers, this approach may
1682 yield a large number of hidden flow rules. It is thus suggested to only
1683 support the most common scenarios (anything as L2 and/or L3).
1688 - When combined with `Action: QUEUE`_, packet counting (`Action: COUNT`_)
1689 and tagging (`Action: MARK`_ or `Action: FLAG`_) may be implemented in
1690 software as long as the target queue is used by a single rule.
1692 - A rule specifying both `Action: DUP`_ + `Action: QUEUE`_ may be translated
1693 to two hidden rules combining `Action: QUEUE`_ and `Action: PASSTHRU`_.
1695 - When a single target queue is provided, `Action: RSS`_ can also be
1696 implemented through `Action: QUEUE`_.
1701 While it would naturally make sense, flow rules cannot be assumed to be
1702 processed by hardware in the same order as their creation for several
1705 - They may be managed internally as a tree or a hash table instead of a
1707 - Removing a flow rule before adding another one can either put the new rule
1708 at the end of the list or reuse a freed entry.
1709 - Duplication may occur when packets are matched by several rules.
1711 For overlapping rules (particularly in order to use `Action: PASSTHRU`_)
1712 predictable behavior is only guaranteed by using different priority levels.
1714 Priority levels are not necessarily implemented in hardware, or may be
1715 severely limited (e.g. a single priority bit).
1717 For these reasons, priority levels may be implemented purely in software by
1720 - For devices expecting flow rules to be added in the correct order, PMDs
1721 may destroy and re-create existing rules after adding a new one with
1724 - A configurable number of dummy or empty rules can be created at
1725 initialization time to save high priority slots for later.
1727 - In order to save priority levels, PMDs may evaluate whether rules are
1728 likely to collide and adjust their priority accordingly.
1733 - A device profile selection function which could be used to force a
1734 permanent profile instead of relying on its automatic configuration based
1735 on existing flow rules.
1737 - A method to optimize *rte_flow* rules with specific pattern items and
1738 action types generated on the fly by PMDs. DPDK should assign negative
1739 numbers to these in order to not collide with the existing types. See
1742 - Adding specific egress pattern items and actions as described in
1743 `Attribute: Traffic direction`_.
1745 - Optional software fallback when PMDs are unable to handle requested flow
1746 rules so applications do not have to implement their own.
1751 Exhaustive list of deprecated filter types (normally prefixed with
1752 *RTE_ETH_FILTER_*) found in ``rte_eth_ctrl.h`` and methods to convert them
1753 to *rte_flow* rules.
1755 ``MACVLAN`` to ``ETH`` → ``VF``, ``PF``
1756 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1758 *MACVLAN* can be translated to a basic `Item: ETH`_ flow rule with a
1759 terminating `Action: VF`_ or `Action: PF`_.
1761 .. _table_rte_flow_migration_macvlan:
1763 .. table:: MACVLAN conversion
1765 +--------------------------+---------+
1766 | Pattern | Actions |
1767 +===+=====+==========+=====+=========+
1768 | 0 | ETH | ``spec`` | any | VF, |
1769 | | +----------+-----+ PF |
1770 | | | ``last`` | N/A | |
1771 | | +----------+-----+ |
1772 | | | ``mask`` | any | |
1773 +---+-----+----------+-----+---------+
1775 +---+----------------------+---------+
1777 ``ETHERTYPE`` to ``ETH`` → ``QUEUE``, ``DROP``
1778 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1780 *ETHERTYPE* is basically an `Item: ETH`_ flow rule with a terminating
1781 `Action: QUEUE`_ or `Action: DROP`_.
1783 .. _table_rte_flow_migration_ethertype:
1785 .. table:: ETHERTYPE conversion
1787 +--------------------------+---------+
1788 | Pattern | Actions |
1789 +===+=====+==========+=====+=========+
1790 | 0 | ETH | ``spec`` | any | QUEUE, |
1791 | | +----------+-----+ DROP |
1792 | | | ``last`` | N/A | |
1793 | | +----------+-----+ |
1794 | | | ``mask`` | any | |
1795 +---+-----+----------+-----+---------+
1797 +---+----------------------+---------+
1799 ``FLEXIBLE`` to ``RAW`` → ``QUEUE``
1800 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1802 *FLEXIBLE* can be translated to one `Item: RAW`_ pattern with a terminating
1803 `Action: QUEUE`_ and a defined priority level.
1805 .. _table_rte_flow_migration_flexible:
1807 .. table:: FLEXIBLE conversion
1809 +--------------------------+---------+
1810 | Pattern | Actions |
1811 +===+=====+==========+=====+=========+
1812 | 0 | RAW | ``spec`` | any | QUEUE |
1813 | | +----------+-----+ |
1814 | | | ``last`` | N/A | |
1815 | | +----------+-----+ |
1816 | | | ``mask`` | any | |
1817 +---+-----+----------+-----+---------+
1819 +---+----------------------+---------+
1821 ``SYN`` to ``TCP`` → ``QUEUE``
1822 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1824 *SYN* is a `Item: TCP`_ rule with only the ``syn`` bit enabled and masked,
1825 and a terminating `Action: QUEUE`_.
1827 Priority level can be set to simulate the high priority bit.
1829 .. _table_rte_flow_migration_syn:
1831 .. table:: SYN conversion
1833 +-----------------------------------+---------+
1834 | Pattern | Actions |
1835 +===+======+==========+=============+=========+
1836 | 0 | ETH | ``spec`` | unset | QUEUE |
1837 | | +----------+-------------+ |
1838 | | | ``last`` | unset | |
1839 | | +----------+-------------+ |
1840 | | | ``mask`` | unset | |
1841 +---+------+----------+-------------+---------+
1842 | 1 | IPV4 | ``spec`` | unset | END |
1843 | | +----------+-------------+ |
1844 | | | ``mask`` | unset | |
1845 | | +----------+-------------+ |
1846 | | | ``mask`` | unset | |
1847 +---+------+----------+---------+---+ |
1848 | 2 | TCP | ``spec`` | ``syn`` | 1 | |
1849 | | +----------+---------+---+ |
1850 | | | ``mask`` | ``syn`` | 1 | |
1851 +---+------+----------+---------+---+ |
1853 +---+-------------------------------+---------+
1855 ``NTUPLE`` to ``IPV4``, ``TCP``, ``UDP`` → ``QUEUE``
1856 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1858 *NTUPLE* is similar to specifying an empty L2, `Item: IPV4`_ as L3 with
1859 `Item: TCP`_ or `Item: UDP`_ as L4 and a terminating `Action: QUEUE`_.
1861 A priority level can be specified as well.
1863 .. _table_rte_flow_migration_ntuple:
1865 .. table:: NTUPLE conversion
1867 +-----------------------------+---------+
1868 | Pattern | Actions |
1869 +===+======+==========+=======+=========+
1870 | 0 | ETH | ``spec`` | unset | QUEUE |
1871 | | +----------+-------+ |
1872 | | | ``last`` | unset | |
1873 | | +----------+-------+ |
1874 | | | ``mask`` | unset | |
1875 +---+------+----------+-------+---------+
1876 | 1 | IPV4 | ``spec`` | any | END |
1877 | | +----------+-------+ |
1878 | | | ``last`` | unset | |
1879 | | +----------+-------+ |
1880 | | | ``mask`` | any | |
1881 +---+------+----------+-------+ |
1882 | 2 | TCP, | ``spec`` | any | |
1883 | | UDP +----------+-------+ |
1884 | | | ``last`` | unset | |
1885 | | +----------+-------+ |
1886 | | | ``mask`` | any | |
1887 +---+------+----------+-------+ |
1889 +---+-------------------------+---------+
1891 ``TUNNEL`` to ``ETH``, ``IPV4``, ``IPV6``, ``VXLAN`` (or other) → ``QUEUE``
1892 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1894 *TUNNEL* matches common IPv4 and IPv6 L3/L4-based tunnel types.
1896 In the following table, `Item: ANY`_ is used to cover the optional L4.
1898 .. _table_rte_flow_migration_tunnel:
1900 .. table:: TUNNEL conversion
1902 +-------------------------------------------------------+---------+
1903 | Pattern | Actions |
1904 +===+==========================+==========+=============+=========+
1905 | 0 | ETH | ``spec`` | any | QUEUE |
1906 | | +----------+-------------+ |
1907 | | | ``last`` | unset | |
1908 | | +----------+-------------+ |
1909 | | | ``mask`` | any | |
1910 +---+--------------------------+----------+-------------+---------+
1911 | 1 | IPV4, IPV6 | ``spec`` | any | END |
1912 | | +----------+-------------+ |
1913 | | | ``last`` | unset | |
1914 | | +----------+-------------+ |
1915 | | | ``mask`` | any | |
1916 +---+--------------------------+----------+-------------+ |
1917 | 2 | ANY | ``spec`` | any | |
1918 | | +----------+-------------+ |
1919 | | | ``last`` | unset | |
1920 | | +----------+---------+---+ |
1921 | | | ``mask`` | ``num`` | 0 | |
1922 +---+--------------------------+----------+---------+---+ |
1923 | 3 | VXLAN, GENEVE, TEREDO, | ``spec`` | any | |
1924 | | NVGRE, GRE, ... +----------+-------------+ |
1925 | | | ``last`` | unset | |
1926 | | +----------+-------------+ |
1927 | | | ``mask`` | any | |
1928 +---+--------------------------+----------+-------------+ |
1930 +---+---------------------------------------------------+---------+
1932 ``FDIR`` to most item types → ``QUEUE``, ``DROP``, ``PASSTHRU``
1933 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1935 *FDIR* is more complex than any other type, there are several methods to
1936 emulate its functionality. It is summarized for the most part in the table
1939 A few features are intentionally not supported:
1941 - The ability to configure the matching input set and masks for the entire
1942 device, PMDs should take care of it automatically according to the
1943 requested flow rules.
1945 For example if a device supports only one bit-mask per protocol type,
1946 source/address IPv4 bit-masks can be made immutable by the first created
1947 rule. Subsequent IPv4 or TCPv4 rules can only be created if they are
1950 Note that only protocol bit-masks affected by existing flow rules are
1951 immutable, others can be changed later. They become mutable again after
1952 the related flow rules are destroyed.
1954 - Returning four or eight bytes of matched data when using flex bytes
1955 filtering. Although a specific action could implement it, it conflicts
1956 with the much more useful 32 bits tagging on devices that support it.
1958 - Side effects on RSS processing of the entire device. Flow rules that
1959 conflict with the current device configuration should not be
1960 allowed. Similarly, device configuration should not be allowed when it
1961 affects existing flow rules.
1963 - Device modes of operation. "none" is unsupported since filtering cannot be
1964 disabled as long as a flow rule is present.
1966 - "MAC VLAN" or "tunnel" perfect matching modes should be automatically set
1967 according to the created flow rules.
1969 - Signature mode of operation is not defined but could be handled through a
1970 specific item type if needed.
1972 .. _table_rte_flow_migration_fdir:
1974 .. table:: FDIR conversion
1976 +----------------------------------------+-----------------------+
1977 | Pattern | Actions |
1978 +===+===================+==========+=====+=======================+
1979 | 0 | ETH, RAW | ``spec`` | any | QUEUE, DROP, PASSTHRU |
1980 | | +----------+-----+ |
1981 | | | ``last`` | N/A | |
1982 | | +----------+-----+ |
1983 | | | ``mask`` | any | |
1984 +---+-------------------+----------+-----+-----------------------+
1985 | 1 | IPV4, IPv6 | ``spec`` | any | MARK |
1986 | | +----------+-----+ |
1987 | | | ``last`` | N/A | |
1988 | | +----------+-----+ |
1989 | | | ``mask`` | any | |
1990 +---+-------------------+----------+-----+-----------------------+
1991 | 2 | TCP, UDP, SCTP | ``spec`` | any | END |
1992 | | +----------+-----+ |
1993 | | | ``last`` | N/A | |
1994 | | +----------+-----+ |
1995 | | | ``mask`` | any | |
1996 +---+-------------------+----------+-----+ |
1997 | 3 | VF, PF (optional) | ``spec`` | any | |
1998 | | +----------+-----+ |
1999 | | | ``last`` | N/A | |
2000 | | +----------+-----+ |
2001 | | | ``mask`` | any | |
2002 +---+-------------------+----------+-----+ |
2004 +---+------------------------------------+-----------------------+
2009 There is no counterpart to this filter type because it translates to a
2010 global device setting instead of a pattern item. Device settings are
2011 automatically set according to the created flow rules.
2013 ``L2_TUNNEL`` to ``VOID`` → ``VXLAN`` (or others)
2014 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2016 All packets are matched. This type alters incoming packets to encapsulate
2017 them in a chosen tunnel type, optionally redirect them to a VF as well.
2019 The destination pool for tag based forwarding can be emulated with other
2020 flow rules using `Action: DUP`_.
2022 .. _table_rte_flow_migration_l2tunnel:
2024 .. table:: L2_TUNNEL conversion
2026 +---------------------------+--------------------+
2027 | Pattern | Actions |
2028 +===+======+==========+=====+====================+
2029 | 0 | VOID | ``spec`` | N/A | VXLAN, GENEVE, ... |
2032 | | +----------+-----+ |
2033 | | | ``last`` | N/A | |
2034 | | +----------+-----+ |
2035 | | | ``mask`` | N/A | |
2037 +---+------+----------+-----+--------------------+
2038 | 1 | END | VF (optional) |
2039 +---+ +--------------------+
2041 +---+-----------------------+--------------------+