xref: /dpdk/doc/guides/sample_app_ug/ipsec_secgw.rst (revision 68a03efeed657e6e05f281479b33b51102797e15)
1..  SPDX-License-Identifier: BSD-3-Clause
2    Copyright(c) 2016-2017 Intel Corporation.
3    Copyright (C) 2020 Marvell International Ltd.
4
5IPsec Security Gateway Sample Application
6=========================================
7
8The IPsec Security Gateway application is an example of a "real world"
9application using DPDK cryptodev framework.
10
11Overview
12--------
13
14The application demonstrates the implementation of a Security Gateway
15(not IPsec compliant, see the Constraints section below) using DPDK based on RFC4301,
16RFC4303, RFC3602 and RFC2404.
17
18Internet Key Exchange (IKE) is not implemented, so only manual setting of
19Security Policies and Security Associations is supported.
20
21The Security Policies (SP) are implemented as ACL rules, the Security
22Associations (SA) are stored in a table and the routing is implemented
23using LPM.
24
25The application classifies the ports as *Protected* and *Unprotected*.
26Thus, traffic received on an Unprotected or Protected port is consider
27Inbound or Outbound respectively.
28
29The application also supports complete IPsec protocol offload to hardware
30(Look aside crypto accelerator or using ethernet device). It also support
31inline ipsec processing by the supported ethernet device during transmission.
32These modes can be selected during the SA creation configuration.
33
34In case of complete protocol offload, the processing of headers(ESP and outer
35IP header) is done by the hardware and the application does not need to
36add/remove them during outbound/inbound processing.
37
38For inline offloaded outbound traffic, the application will not do the LPM
39lookup for routing, as the port on which the packet has to be forwarded will be
40part of the SA. Security parameters will be configured on that port only, and
41sending the packet on other ports could result in unencrypted packets being
42sent out.
43
44The Path for IPsec Inbound traffic is:
45
46*  Read packets from the port.
47*  Classify packets between IPv4 and ESP.
48*  Perform Inbound SA lookup for ESP packets based on their SPI.
49*  Perform Verification/Decryption (Not needed in case of inline ipsec).
50*  Remove ESP and outer IP header (Not needed in case of protocol offload).
51*  Inbound SP check using ACL of decrypted packets and any other IPv4 packets.
52*  Routing.
53*  Write packet to port.
54
55The Path for the IPsec Outbound traffic is:
56
57*  Read packets from the port.
58*  Perform Outbound SP check using ACL of all IPv4 traffic.
59*  Perform Outbound SA lookup for packets that need IPsec protection.
60*  Add ESP and outer IP header (Not needed in case protocol offload).
61*  Perform Encryption/Digest (Not needed in case of inline ipsec).
62*  Routing.
63*  Write packet to port.
64
65The application supports two modes of operation: poll mode and event mode.
66
67* In the poll mode a core receives packets from statically configured list
68  of eth ports and eth ports' queues.
69
70* In the event mode a core receives packets as events. After packet processing
71  is done core submits them back as events to an event device. This enables
72  multicore scaling and HW assisted scheduling by making use of the event device
73  capabilities. The event mode configuration is predefined. All packets reaching
74  given eth port will arrive at the same event queue. All event queues are mapped
75  to all event ports. This allows all cores to receive traffic from all ports.
76  Since the underlying event device might have varying capabilities, the worker
77  threads can be drafted differently to maximize performance. For example, if an
78  event device - eth device pair has Tx internal port, then application can call
79  rte_event_eth_tx_adapter_enqueue() instead of regular rte_event_enqueue_burst().
80  So a thread which assumes that the device pair has internal port will not be the
81  right solution for another pair. The infrastructure added for the event mode aims
82  to help application to have multiple worker threads by maximizing performance from
83  every type of event device without affecting existing paths/use cases. The worker
84  to be used will be determined by the operating conditions and the underlying device
85  capabilities. **Currently the application provides non-burst, internal port worker
86  threads and supports inline protocol only.** It also provides infrastructure for
87  non-internal port however does not define any worker threads.
88
89Additionally the event mode introduces two submodes of processing packets:
90
91* Driver submode: This submode has bare minimum changes in the application to support
92  IPsec. There are no lookups, no routing done in the application. And for inline
93  protocol use case, the worker thread resembles l2fwd worker thread as the IPsec
94  processing is done entirely in HW. This mode can be used to benchmark the raw
95  performance of the HW. The driver submode is selected with --single-sa option
96  (used also by poll mode). When --single-sa option is used in conjution with event
97  mode then index passed to --single-sa is ignored.
98
99* App submode: This submode has all the features currently implemented with the
100  application (non librte_ipsec path). All the lookups, routing follows existing
101  methods and report numbers that can be compared against regular poll mode
102  benchmark numbers.
103
104Constraints
105-----------
106
107*  No IPv6 options headers.
108*  No AH mode.
109*  Supported algorithms: AES-CBC, AES-CTR, AES-GCM, 3DES-CBC, HMAC-SHA1 and NULL.
110*  Each SA must be handle by a unique lcore (*1 RX queue per port*).
111
112Compiling the Application
113-------------------------
114
115To compile the sample application see :doc:`compiling`.
116
117The application is located in the ``ipsec-secgw`` sub-directory.
118
119
120Running the Application
121-----------------------
122
123The application has a number of command line options::
124
125
126   ./<build_dir>/examples/dpdk-ipsec-secgw [EAL options] --
127                        -p PORTMASK -P -u PORTMASK -j FRAMESIZE
128                        -l -w REPLAY_WINDOW_SIZE -e -a
129                        -c SAD_CACHE_SIZE
130                        -s NUMBER_OF_MBUFS_IN_PACKET_POOL
131                        -f CONFIG_FILE_PATH
132                        --config (port,queue,lcore)[,(port,queue,lcore)]
133                        --single-sa SAIDX
134                        --cryptodev_mask MASK
135                        --transfer-mode MODE
136                        --event-schedule-type TYPE
137                        --rxoffload MASK
138                        --txoffload MASK
139                        --reassemble NUM
140                        --mtu MTU
141                        --frag-ttl FRAG_TTL_NS
142
143Where:
144
145*   ``-p PORTMASK``: Hexadecimal bitmask of ports to configure.
146
147*   ``-P``: *optional*. Sets all ports to promiscuous mode so that packets are
148    accepted regardless of the packet's Ethernet MAC destination address.
149    Without this option, only packets with the Ethernet MAC destination address
150    set to the Ethernet address of the port are accepted (default is enabled).
151
152*   ``-u PORTMASK``: hexadecimal bitmask of unprotected ports
153
154*   ``-j FRAMESIZE``: *optional*. data buffer size (in bytes),
155    in other words maximum data size for one segment.
156    Packets with length bigger then FRAMESIZE still can be received,
157    but will be segmented.
158    Default value: RTE_MBUF_DEFAULT_BUF_SIZE (2176)
159    Minimum value: RTE_MBUF_DEFAULT_BUF_SIZE (2176)
160    Maximum value: UINT16_MAX (65535).
161
162*   ``-l``: enables code-path that uses librte_ipsec.
163
164*   ``-w REPLAY_WINDOW_SIZE``: specifies the IPsec sequence number replay window
165    size for each Security Association (available only with librte_ipsec
166    code path).
167
168*   ``-e``: enables Security Association extended sequence number processing
169    (available only with librte_ipsec code path).
170
171*   ``-a``: enables Security Association sequence number atomic behavior
172    (available only with librte_ipsec code path).
173
174*   ``-c``: specifies the SAD cache size. Stores the most recent SA in a per
175    lcore cache. Cache represents flat array containing SA's indexed by SPI.
176    Zero value disables cache.
177    Default value: 128.
178
179*   ``-s``: sets number of mbufs in packet pool, if not provided number of mbufs
180    will be calculated based on number of cores, eth ports and crypto queues.
181
182*   ``-f CONFIG_FILE_PATH``: the full path of text-based file containing all
183    configuration items for running the application (See Configuration file
184    syntax section below). ``-f CONFIG_FILE_PATH`` **must** be specified.
185    **ONLY** the UNIX format configuration file is accepted.
186
187*   ``--config (port,queue,lcore)[,(port,queue,lcore)]``: in poll mode determines
188    which queues from which ports are mapped to which cores. In event mode this
189    option is not used as packets are dynamically scheduled to cores by HW.
190
191*   ``--single-sa SAIDX``: in poll mode use a single SA for outbound traffic,
192    bypassing the SP on both Inbound and Outbound. This option is meant for
193    debugging/performance purposes. In event mode selects driver submode, SA index
194    value is ignored.
195
196*   ``--cryptodev_mask MASK``: hexadecimal bitmask of the crypto devices
197    to configure.
198
199*   ``--transfer-mode MODE``: sets operating mode of the application
200    "poll"  : packet transfer via polling (default)
201    "event" : Packet transfer via event device
202
203*   ``--event-schedule-type TYPE``: queue schedule type, applies only when
204    --transfer-mode is set to event.
205    "ordered"  : Ordered (default)
206    "atomic"   : Atomic
207    "parallel" : Parallel
208    When --event-schedule-type is set as RTE_SCHED_TYPE_ORDERED/ATOMIC, event
209    device will ensure the ordering. Ordering will be lost when tried in PARALLEL.
210
211*   ``--rxoffload MASK``: RX HW offload capabilities to enable/use on this port
212    (bitmask of DEV_RX_OFFLOAD_* values). It is an optional parameter and
213    allows user to disable some of the RX HW offload capabilities.
214    By default all HW RX offloads are enabled.
215
216*   ``--txoffload MASK``: TX HW offload capabilities to enable/use on this port
217    (bitmask of DEV_TX_OFFLOAD_* values). It is an optional parameter and
218    allows user to disable some of the TX HW offload capabilities.
219    By default all HW TX offloads are enabled.
220
221*   ``--reassemble NUM``: max number of entries in reassemble fragment table.
222    Zero value disables reassembly functionality.
223    Default value: 0.
224
225*   ``--mtu MTU``: MTU value (in bytes) on all attached ethernet ports.
226    Outgoing packets with length bigger then MTU will be fragmented.
227    Incoming packets with length bigger then MTU will be discarded.
228    Default value: 1500.
229
230*   ``--frag-ttl FRAG_TTL_NS``: fragment lifetime (in nanoseconds).
231    If packet is not reassembled within this time, received fragments
232    will be discarded. Fragment lifetime should be decreased when
233    there is a high fragmented traffic loss in high bandwidth networks.
234    Should be lower for low number of reassembly buckets.
235    Valid values: from 1 ns to 10 s. Default value: 10000000 (10 s).
236
237
238The mapping of lcores to port/queues is similar to other l3fwd applications.
239
240For example, given the following command line to run application in poll mode::
241
242    ./<build_dir>/examples/dpdk-ipsec-secgw -l 20,21 -n 4 --socket-mem 0,2048       \
243           --vdev "crypto_null" -- -p 0xf -P -u 0x3             \
244           --config="(0,0,20),(1,0,20),(2,0,21),(3,0,21)"       \
245           -f /path/to/config_file --transfer-mode poll         \
246
247where each option means:
248
249*   The ``-l`` option enables cores 20 and 21.
250
251*   The ``-n`` option sets memory 4 channels.
252
253*   The ``--socket-mem`` to use 2GB on socket 1.
254
255*   The ``--vdev "crypto_null"`` option creates virtual NULL cryptodev PMD.
256
257*   The ``-p`` option enables ports (detected) 0, 1, 2 and 3.
258
259*   The ``-P`` option enables promiscuous mode.
260
261*   The ``-u`` option sets ports 0 and 1 as unprotected, leaving 2 and 3 as protected.
262
263*   The ``--config`` option enables one queue per port with the following mapping:
264
265    +----------+-----------+-----------+---------------------------------------+
266    | **Port** | **Queue** | **lcore** | **Description**                       |
267    |          |           |           |                                       |
268    +----------+-----------+-----------+---------------------------------------+
269    | 0        | 0         | 20        | Map queue 0 from port 0 to lcore 20.  |
270    |          |           |           |                                       |
271    +----------+-----------+-----------+---------------------------------------+
272    | 1        | 0         | 20        | Map queue 0 from port 1 to lcore 20.  |
273    |          |           |           |                                       |
274    +----------+-----------+-----------+---------------------------------------+
275    | 2        | 0         | 21        | Map queue 0 from port 2 to lcore 21.  |
276    |          |           |           |                                       |
277    +----------+-----------+-----------+---------------------------------------+
278    | 3        | 0         | 21        | Map queue 0 from port 3 to lcore 21.  |
279    |          |           |           |                                       |
280    +----------+-----------+-----------+---------------------------------------+
281
282*   The ``-f /path/to/config_file`` option enables the application read and
283    parse the configuration file specified, and configures the application
284    with a given set of SP, SA and Routing entries accordingly. The syntax of
285    the configuration file will be explained below in more detail. Please
286    **note** the parser only accepts UNIX format text file. Other formats
287    such as DOS/MAC format will cause a parse error.
288
289*   The ``--transfer-mode`` option selects poll mode for processing packets.
290
291Similarly for example, given the following command line to run application in
292event app mode::
293
294    ./<build_dir>/examples/dpdk-ipsec-secgw -c 0x3 -- -P -p 0x3 -u 0x1       \
295           -f /path/to/config_file --transfer-mode event \
296           --event-schedule-type parallel                \
297
298where each option means:
299
300*   The ``-c`` option selects cores 0 and 1 to run on.
301
302*   The ``-P`` option enables promiscuous mode.
303
304*   The ``-p`` option enables ports (detected) 0 and 1.
305
306*   The ``-u`` option sets ports 0 as unprotected, leaving 1 as protected.
307
308*   The ``-f /path/to/config_file`` option has the same behavior as in poll
309    mode example.
310
311*   The ``--transfer-mode`` option selects event mode for processing packets.
312
313*   The ``--event-schedule-type`` option selects parallel ordering of event queues.
314
315
316Refer to the *DPDK Getting Started Guide* for general information on running
317applications and the Environment Abstraction Layer (EAL) options.
318
319The application would do a best effort to "map" crypto devices to cores, with
320hardware devices having priority. Basically, hardware devices if present would
321be assigned to a core before software ones.
322This means that if the application is using a single core and both hardware
323and software crypto devices are detected, hardware devices will be used.
324
325A way to achieve the case where you want to force the use of virtual crypto
326devices is to only use the Ethernet devices needed (via the allow flag)
327and therefore implicitly blocking all hardware crypto devices.
328
329For example, something like the following command line:
330
331.. code-block:: console
332
333    ./<build_dir>/examples/dpdk-ipsec-secgw -l 20,21 -n 4 --socket-mem 0,2048 \
334            -a 81:00.0 -a 81:00.1 -a 81:00.2 -a 81:00.3 \
335            --vdev "crypto_aesni_mb" --vdev "crypto_null" \
336	    -- \
337            -p 0xf -P -u 0x3 --config="(0,0,20),(1,0,20),(2,0,21),(3,0,21)" \
338            -f sample.cfg
339
340
341Configurations
342--------------
343
344The following sections provide the syntax of configurations to initialize
345your SP, SA, Routing, Flow and Neighbour tables.
346Configurations shall be specified in the configuration file to be passed to
347the application. The file is then parsed by the application. The successful
348parsing will result in the appropriate rules being applied to the tables
349accordingly.
350
351
352Configuration File Syntax
353~~~~~~~~~~~~~~~~~~~~~~~~~
354
355As mention in the overview, the Security Policies are ACL rules.
356The application parsers the rules specified in the configuration file and
357passes them to the ACL table, and replicates them per socket in use.
358
359Following are the configuration file syntax.
360
361General rule syntax
362^^^^^^^^^^^^^^^^^^^
363
364The parse treats one line in the configuration file as one configuration
365item (unless the line concatenation symbol exists). Every configuration
366item shall follow the syntax of either SP, SA, Routing, Flow or Neighbour
367rules specified below.
368
369The configuration parser supports the following special symbols:
370
371 * Comment symbol **#**. Any character from this symbol to the end of
372   line is treated as comment and will not be parsed.
373
374 * Line concatenation symbol **\\**. This symbol shall be placed in the end
375   of the line to be concatenated to the line below. Multiple lines'
376   concatenation is supported.
377
378
379SP rule syntax
380^^^^^^^^^^^^^^
381
382The SP rule syntax is shown as follows:
383
384.. code-block:: console
385
386    sp <ip_ver> <dir> esp <action> <priority> <src_ip> <dst_ip>
387    <proto> <sport> <dport>
388
389
390where each options means:
391
392``<ip_ver>``
393
394 * IP protocol version
395
396 * Optional: No
397
398 * Available options:
399
400   * *ipv4*: IP protocol version 4
401   * *ipv6*: IP protocol version 6
402
403``<dir>``
404
405 * The traffic direction
406
407 * Optional: No
408
409 * Available options:
410
411   * *in*: inbound traffic
412   * *out*: outbound traffic
413
414``<action>``
415
416 * IPsec action
417
418 * Optional: No
419
420 * Available options:
421
422   * *protect <SA_idx>*: the specified traffic is protected by SA rule
423     with id SA_idx
424   * *bypass*: the specified traffic traffic is bypassed
425   * *discard*: the specified traffic is discarded
426
427``<priority>``
428
429 * Rule priority
430
431 * Optional: Yes, default priority 0 will be used
432
433 * Syntax: *pri <id>*
434
435``<src_ip>``
436
437 * The source IP address and mask
438
439 * Optional: Yes, default address 0.0.0.0 and mask of 0 will be used
440
441 * Syntax:
442
443   * *src X.X.X.X/Y* for IPv4
444   * *src XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX/Y* for IPv6
445
446``<dst_ip>``
447
448 * The destination IP address and mask
449
450 * Optional: Yes, default address 0.0.0.0 and mask of 0 will be used
451
452 * Syntax:
453
454   * *dst X.X.X.X/Y* for IPv4
455   * *dst XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX/Y* for IPv6
456
457``<proto>``
458
459 * The protocol start and end range
460
461 * Optional: yes, default range of 0 to 0 will be used
462
463 * Syntax: *proto X:Y*
464
465``<sport>``
466
467 * The source port start and end range
468
469 * Optional: yes, default range of 0 to 0 will be used
470
471 * Syntax: *sport X:Y*
472
473``<dport>``
474
475 * The destination port start and end range
476
477 * Optional: yes, default range of 0 to 0 will be used
478
479 * Syntax: *dport X:Y*
480
481Example SP rules:
482
483.. code-block:: console
484
485    sp ipv4 out esp protect 105 pri 1 dst 192.168.115.0/24 sport 0:65535 \
486    dport 0:65535
487
488    sp ipv6 in esp bypass pri 1 dst 0000:0000:0000:0000:5555:5555:\
489    0000:0000/96 sport 0:65535 dport 0:65535
490
491
492SA rule syntax
493^^^^^^^^^^^^^^
494
495The successfully parsed SA rules will be stored in an array table.
496
497The SA rule syntax is shown as follows:
498
499.. code-block:: console
500
501    sa <dir> <spi> <cipher_algo> <cipher_key> <auth_algo> <auth_key>
502    <mode> <src_ip> <dst_ip> <action_type> <port_id> <fallback>
503    <flow-direction> <port_id> <queue_id>
504
505where each options means:
506
507``<dir>``
508
509 * The traffic direction
510
511 * Optional: No
512
513 * Available options:
514
515   * *in*: inbound traffic
516   * *out*: outbound traffic
517
518``<spi>``
519
520 * The SPI number
521
522 * Optional: No
523
524 * Syntax: unsigned integer number
525
526``<cipher_algo>``
527
528 * Cipher algorithm
529
530 * Optional: Yes, unless <aead_algo> is not used
531
532 * Available options:
533
534   * *null*: NULL algorithm
535   * *aes-128-cbc*: AES-CBC 128-bit algorithm
536   * *aes-192-cbc*: AES-CBC 192-bit algorithm
537   * *aes-256-cbc*: AES-CBC 256-bit algorithm
538   * *aes-128-ctr*: AES-CTR 128-bit algorithm
539   * *3des-cbc*: 3DES-CBC 192-bit algorithm
540
541 * Syntax: *cipher_algo <your algorithm>*
542
543``<cipher_key>``
544
545 * Cipher key, NOT available when 'null' algorithm is used
546
547 * Optional: Yes, unless <aead_algo> is not used.
548   Must be followed by <cipher_algo> option
549
550 * Syntax: Hexadecimal bytes (0x0-0xFF) concatenate by colon symbol ':'.
551   The number of bytes should be as same as the specified cipher algorithm
552   key size.
553
554   For example: *cipher_key A1:B2:C3:D4:A1:B2:C3:D4:A1:B2:C3:D4:
555   A1:B2:C3:D4*
556
557``<auth_algo>``
558
559 * Authentication algorithm
560
561 * Optional: Yes, unless <aead_algo> is not used
562
563 * Available options:
564
565    * *null*: NULL algorithm
566    * *sha1-hmac*: HMAC SHA1 algorithm
567
568``<auth_key>``
569
570 * Authentication key, NOT available when 'null' or 'aes-128-gcm' algorithm
571   is used.
572
573 * Optional: Yes, unless <aead_algo> is not used.
574   Must be followed by <auth_algo> option
575
576 * Syntax: Hexadecimal bytes (0x0-0xFF) concatenate by colon symbol ':'.
577   The number of bytes should be as same as the specified authentication
578   algorithm key size.
579
580   For example: *auth_key A1:B2:C3:D4:A1:B2:C3:D4:A1:B2:C3:D4:A1:B2:C3:D4:
581   A1:B2:C3:D4*
582
583``<aead_algo>``
584
585 * AEAD algorithm
586
587 * Optional: Yes, unless <cipher_algo> and <auth_algo> are not used
588
589 * Available options:
590
591   * *aes-128-gcm*: AES-GCM 128-bit algorithm
592   * *aes-192-gcm*: AES-GCM 192-bit algorithm
593   * *aes-256-gcm*: AES-GCM 256-bit algorithm
594
595 * Syntax: *cipher_algo <your algorithm>*
596
597``<aead_key>``
598
599 * Cipher key, NOT available when 'null' algorithm is used
600
601 * Optional: Yes, unless <cipher_algo> and <auth_algo> are not used.
602   Must be followed by <aead_algo> option
603
604 * Syntax: Hexadecimal bytes (0x0-0xFF) concatenate by colon symbol ':'.
605   Last 4 bytes of the provided key will be used as 'salt' and so, the
606   number of bytes should be same as the sum of specified AEAD algorithm
607   key size and salt size (4 bytes).
608
609   For example: *aead_key A1:B2:C3:D4:A1:B2:C3:D4:A1:B2:C3:D4:
610   A1:B2:C3:D4:A1:B2:C3:D4*
611
612``<mode>``
613
614 * The operation mode
615
616 * Optional: No
617
618 * Available options:
619
620   * *ipv4-tunnel*: Tunnel mode for IPv4 packets
621   * *ipv6-tunnel*: Tunnel mode for IPv6 packets
622   * *transport*: transport mode
623
624 * Syntax: mode XXX
625
626``<src_ip>``
627
628 * The source IP address. This option is not available when
629   transport mode is used
630
631 * Optional: Yes, default address 0.0.0.0 will be used
632
633 * Syntax:
634
635   * *src X.X.X.X* for IPv4
636   * *src XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX* for IPv6
637
638``<dst_ip>``
639
640 * The destination IP address. This option is not available when
641   transport mode is used
642
643 * Optional: Yes, default address 0.0.0.0 will be used
644
645 * Syntax:
646
647   * *dst X.X.X.X* for IPv4
648   * *dst XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX* for IPv6
649
650``<type>``
651
652 * Action type to specify the security action. This option specify
653   the SA to be performed with look aside protocol offload to HW
654   accelerator or protocol offload on ethernet device or inline
655   crypto processing on the ethernet device during transmission.
656
657 * Optional: Yes, default type *no-offload*
658
659 * Available options:
660
661   * *lookaside-protocol-offload*: look aside protocol offload to HW accelerator
662   * *inline-protocol-offload*: inline protocol offload on ethernet device
663   * *inline-crypto-offload*: inline crypto processing on ethernet device
664   * *no-offload*: no offloading to hardware
665
666 ``<port_id>``
667
668 * Port/device ID of the ethernet/crypto accelerator for which the SA is
669   configured. For *inline-crypto-offload* and *inline-protocol-offload*, this
670   port will be used for routing. The routing table will not be referred in
671   this case.
672
673 * Optional: No, if *type* is not *no-offload*
674
675 * Syntax:
676
677   * *port_id X* X is a valid device number in decimal
678
679 ``<fallback>``
680
681 * Action type for ingress IPsec packets that inline processor failed to
682   process. Only a combination of *inline-crypto-offload* as a primary
683   session and *lookaside-none* as a fall-back session is supported at the
684   moment.
685
686   If used in conjunction with IPsec window, its width needs be increased
687   due to different processing times of inline and lookaside modes which
688   results in packet reordering.
689
690 * Optional: Yes.
691
692 * Available options:
693
694   * *lookaside-none*: use automatically chosen cryptodev to process packets
695
696 * Syntax:
697
698   * *fallback lookaside-none*
699
700``<flow-direction>``
701
702 * Option for redirecting a specific inbound ipsec flow of a port to a specific
703   queue of that port.
704
705 * Optional: Yes.
706
707 * Available options:
708
709   * *port_id*: Port ID of the NIC for which the SA is configured.
710   * *queue_id*: Queue ID to which traffic should be redirected.
711
712Example SA rules:
713
714.. code-block:: console
715
716    sa out 5 cipher_algo null auth_algo null mode ipv4-tunnel \
717    src 172.16.1.5 dst 172.16.2.5
718
719    sa out 25 cipher_algo aes-128-cbc \
720    cipher_key c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3 \
721    auth_algo sha1-hmac \
722    auth_key c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3:c3 \
723    mode ipv6-tunnel \
724    src 1111:1111:1111:1111:1111:1111:1111:5555 \
725    dst 2222:2222:2222:2222:2222:2222:2222:5555
726
727    sa in 105 aead_algo aes-128-gcm \
728    aead_key de:ad:be:ef:de:ad:be:ef:de:ad:be:ef:de:ad:be:ef:de:ad:be:ef \
729    mode ipv4-tunnel src 172.16.2.5 dst 172.16.1.5
730
731    sa out 5 cipher_algo aes-128-cbc cipher_key 0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0 \
732    auth_algo sha1-hmac auth_key 0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0:0 \
733    mode ipv4-tunnel src 172.16.1.5 dst 172.16.2.5 \
734    type lookaside-protocol-offload port_id 4
735
736    sa in 35 aead_algo aes-128-gcm \
737    aead_key de:ad:be:ef:de:ad:be:ef:de:ad:be:ef:de:ad:be:ef:de:ad:be:ef \
738    mode ipv4-tunnel src 172.16.2.5 dst 172.16.1.5 \
739    type inline-crypto-offload port_id 0
740
741    sa in 117 cipher_algo null auth_algo null mode ipv4-tunnel src 172.16.2.7 \
742    dst 172.16.1.7 flow-direction 0 2
743
744Routing rule syntax
745^^^^^^^^^^^^^^^^^^^
746
747The Routing rule syntax is shown as follows:
748
749.. code-block:: console
750
751    rt <ip_ver> <src_ip> <dst_ip> <port>
752
753
754where each options means:
755
756``<ip_ver>``
757
758 * IP protocol version
759
760 * Optional: No
761
762 * Available options:
763
764   * *ipv4*: IP protocol version 4
765   * *ipv6*: IP protocol version 6
766
767``<src_ip>``
768
769 * The source IP address and mask
770
771 * Optional: Yes, default address 0.0.0.0 and mask of 0 will be used
772
773 * Syntax:
774
775   * *src X.X.X.X/Y* for IPv4
776   * *src XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX/Y* for IPv6
777
778``<dst_ip>``
779
780 * The destination IP address and mask
781
782 * Optional: Yes, default address 0.0.0.0 and mask of 0 will be used
783
784 * Syntax:
785
786   * *dst X.X.X.X/Y* for IPv4
787   * *dst XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX/Y* for IPv6
788
789``<port>``
790
791 * The traffic output port id
792
793 * Optional: yes, default output port 0 will be used
794
795 * Syntax: *port X*
796
797Example SP rules:
798
799.. code-block:: console
800
801    rt ipv4 dst 172.16.1.5/32 port 0
802
803    rt ipv6 dst 1111:1111:1111:1111:1111:1111:1111:5555/116 port 0
804
805Flow rule syntax
806^^^^^^^^^^^^^^^^
807
808Flow rule enables the usage of hardware classification capabilities to match specific
809ingress traffic and redirect the packets to the specified queue. This feature is
810optional and relies on hardware ``rte_flow`` support.
811
812The flow rule syntax is shown as follows:
813
814.. code-block:: console
815
816    flow <ip_ver> <src_ip> <dst_ip> <port> <queue>
817
818
819where each options means:
820
821``<ip_ver>``
822
823 * IP protocol version
824
825 * Optional: No
826
827 * Available options:
828
829   * *ipv4*: IP protocol version 4
830   * *ipv6*: IP protocol version 6
831
832``<src_ip>``
833
834 * The source IP address and mask
835
836 * Optional: Yes, default address 0.0.0.0 and mask of 0 will be used
837
838 * Syntax:
839
840   * *src X.X.X.X/Y* for IPv4
841   * *src XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX/Y* for IPv6
842
843``<dst_ip>``
844
845 * The destination IP address and mask
846
847 * Optional: Yes, default address 0.0.0.0 and mask of 0 will be used
848
849 * Syntax:
850
851   * *dst X.X.X.X/Y* for IPv4
852   * *dst XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX:XXXX/Y* for IPv6
853
854``<port>``
855
856 * The traffic input port id
857
858 * Optional: yes, default input port 0 will be used
859
860 * Syntax: *port X*
861
862``<queue>``
863
864 * The traffic input queue id
865
866 * Optional: yes, default input queue 0 will be used
867
868 * Syntax: *queue X*
869
870Example flow rules:
871
872.. code-block:: console
873
874    flow ipv4 dst 172.16.1.5/32 port 0 queue 0
875
876    flow ipv6 dst 1111:1111:1111:1111:1111:1111:1111:5555/116 port 1 queue 0
877
878
879Neighbour rule syntax
880^^^^^^^^^^^^^^^^^^^^^
881
882The Neighbour rule syntax is shown as follows:
883
884.. code-block:: console
885
886    neigh <port> <dst_mac>
887
888
889where each options means:
890
891``<port>``
892
893 * The output port id
894
895 * Optional: No
896
897 * Syntax: *port X*
898
899``<dst_mac>``
900
901 * The destination ethernet address to use for that port
902
903 * Optional: No
904
905 * Syntax:
906
907   * XX:XX:XX:XX:XX:XX
908
909Example Neighbour rules:
910
911.. code-block:: console
912
913    neigh port 0 DE:AD:BE:EF:01:02
914
915Test directory
916--------------
917
918The test directory contains scripts for testing the various encryption
919algorithms.
920
921The purpose of the scripts is to automate ipsec-secgw testing
922using another system running linux as a DUT.
923
924The user must setup the following environment variables:
925
926*   ``SGW_PATH``: path to the ipsec-secgw binary to test.
927
928*   ``REMOTE_HOST``: IP address/hostname of the DUT.
929
930*   ``REMOTE_IFACE``: interface name for the test-port on the DUT.
931
932*   ``ETH_DEV``: ethernet device to be used on the SUT by DPDK ('-a <pci-id>')
933
934Also the user can optionally setup:
935
936*   ``SGW_LCORE``: lcore to run ipsec-secgw on (default value is 0)
937
938*   ``CRYPTO_DEV``: crypto device to be used ('-a <pci-id>'). If none specified
939    appropriate vdevs will be created by the script
940
941Scripts can be used for multiple test scenarios. To check all available
942options run:
943
944.. code-block:: console
945
946    /bin/bash run_test.sh -h
947
948Note that most of the tests require the appropriate crypto PMD/device to be
949available.
950
951Server configuration
952~~~~~~~~~~~~~~~~~~~~
953
954Two servers are required for the tests, SUT and DUT.
955
956Make sure the user from the SUT can ssh to the DUT without entering the password.
957To enable this feature keys must be setup on the DUT.
958
959``ssh-keygen`` will make a private & public key pair on the SUT.
960
961``ssh-copy-id`` <user name>@<target host name> on the SUT will copy the public
962key to the DUT. It will ask for credentials so that it can upload the public key.
963
964The SUT and DUT are connected through at least 2 NIC ports.
965
966One NIC port is expected to be managed by linux on both machines and will be
967used as a control path.
968
969The second NIC port (test-port) should be bound to DPDK on the SUT, and should
970be managed by linux on the DUT.
971
972The script starts ``ipsec-secgw`` with 2 NIC devices: ``test-port`` and
973``tap vdev``.
974
975It then configures the local tap interface and the remote interface and IPsec
976policies in the following way:
977
978Traffic going over the test-port in both directions has to be protected by IPsec.
979
980Traffic going over the TAP port in both directions does not have to be protected.
981
982i.e:
983
984DUT OS(NIC1)--(IPsec)-->(NIC1)ipsec-secgw(TAP)--(plain)-->(TAP)SUT OS
985
986SUT OS(TAP)--(plain)-->(TAP)psec-secgw(NIC1)--(IPsec)-->(NIC1)DUT OS
987
988It then tries to perform some data transfer using the scheme described above.
989
990Usage
991~~~~~
992
993In the ipsec-secgw/test directory run
994
995/bin/bash run_test.sh <options> <ipsec_mode>
996
997Available options:
998
999*   ``-4`` Perform tests with use of IPv4. One or both [-46] options needs to be
1000    selected.
1001
1002*   ``-6`` Perform tests with use of IPv6. One or both [-46] options needs to be
1003    selected.
1004
1005*   ``-m`` Add IPSec tunnel mixed IP version tests - outer IP version different
1006    than inner. Inner IP version will match selected option [-46].
1007
1008*   ``-i`` Run tests in inline mode. Regular tests will not be invoked.
1009
1010*   ``-f`` Run tests for fallback mechanism. Regular tests will not be invoked.
1011
1012*   ``-l`` Run tests in legacy mode only. It cannot be used with options [-fsc].
1013    On default library mode is used.
1014
1015*   ``-s`` Run all tests with reassembly support. On default only tests for
1016    fallback mechanism use reassembly support.
1017
1018*   ``-c`` Run tests with use of cpu-crypto. For inline tests it will not be
1019    applied. On default lookaside-none is used.
1020
1021*   ``-p`` Perform packet validation tests. Option [-46] is not required.
1022
1023*   ``-h`` Show usage.
1024
1025If <ipsec_mode> is specified, only tests for that mode will be invoked. For the
1026list of available modes please refer to run_test.sh.