src/swanctl/swanctl.opt

   1 connections { # }
   2         Section defining IKE connection configurations.
   3
   4         Section defining IKE connection configurations.
   5
   6         The connections section defines IKE connection configurations, each in
   7         its own subsections. In the keyword description below, the connection
   8         is named _<conn>_, but an arbitrary yet unique connection name can be
   9         chosen for each connection subsection.
  10
  11 connections.<conn> { # }
  12         Section for an IKE connection named <conn>.
  13
  14 connections.<conn>.version = 0
  15         IKE major version to use for connection.
  16
  17         IKE major version to use for connection. _1_ uses IKEv1 aka ISAKMP, _2_
  18         uses IKEv2. A connection using the default of _0_ accepts both IKEv1
  19         and IKEv2 as responder, and initiates the connection actively with IKEv2.
  20
  21 connections.<conn>.local_addrs = %any
  22         Local address(es) to use for IKE communication, comma separated.
  23
  24         Local address(es) to use for IKE communication, comma separated. Takes
  25         single IPv4/IPv6 addresses, DNS names, CIDR subnets or IP address ranges.
  26
  27         As initiator, the first non-range/non-subnet is used to initiate the
  28         connection from. As responder, the local destination address must match at
  29         least to one of the specified addresses, subnets or ranges.
  30
  31 connections.<conn>.remote_addrs = %any
  32         Remote address(es) to use for IKE communication, comma separated.
  33
  34         Remote address(es) to use for IKE communication, comma separated. Takes
  35         single IPv4/IPv6 addresses, DNS names, CIDR subnets or IP address ranges.
  36
  37         As initiator, the first non-range/non-subnet is used to initiate the
  38         connection to. As responder, the initiator source address must match at
  39         least to one of the specified addresses, subnets or ranges.
  40
  41         To initiate a connection, at least one specific address or DNS name must
  42         be specified.
  43
  44 connections.<conn>.local_port = 500
  45         Local UPD port for IKE communication.
  46
  47         Local UPD port for IKE communication. By default the port of the socket
  48         backend is used, which is usually _500_. If port _500_ is used, automatic
  49         IKE port floating to port 4500 is used to work around NAT issues.
  50
  51         Using a non-default local IKE port requires support from the socket backend
  52         in use (socket-dynamic).
  53
  54 connections.<conn>.remote_port = 500
  55         Remote UDP port for IKE communication.
  56
  57         Remote UPD port for IKE communication. If the default of port _500_ is used,
  58         automatic IKE port floating to port 4500 is used to work around NAT issues.
  59
  60 connections.<conn>.proposals = default
  61         Comma separated proposals to accept for IKE.
  62
  63         A proposal is a set of algorithms. For non-AEAD algorithms, this includes
  64         for IKE an encryption algorithm, an integrity algorithm, a pseudo random
  65         function and a Diffie-Hellman group. For AEAD algorithms, instead of
  66         encryption and integrity algorithms, a combined algorithm is used.
  67
  68         In IKEv2, multiple algorithms of the same kind can be specified in a single
  69         proposal, from which one gets selected. In IKEv1, only one algorithm per
  70         kind is allowed per proposal, more algorithms get implicitly stripped. Use
  71         multiple proposals to offer different algorithms combinations in IKEv1.
  72
  73         Algorithm keywords get separated using dashes. Multiple proposals may be
  74         separated by commas. The special value _default_ forms a default proposal
  75         of supported algorithms considered safe, and is usually a good choice
  76         for interoperability.
  77
  78 connections.<conn>.vips =
  79         Virtual IPs to request in configuration payload / Mode Config.
  80
  81         Comma separated list of virtual IPs to request in IKEv2 configuration
  82         payloads or IKEv1 Mode Config. The wildcard addresses _0.0.0.0_ and _::_
  83         request an arbitrary address, specific addresses may be defined. The
  84         responder may return a different address, though, or none at all.
  85
  86 connections.<conn>.aggressive = no
  87         Use Aggressive Mode in IKEv1.
  88
  89         Enables Aggressive Mode instead of Main Mode with Identity Protection.
  90         Aggressive Mode is considered less secure, because the ID and HASH
  91         payloads are exchanged unprotected. This allows a passive attacker to
  92         snoop peer identities, and even worse, start dictionary attacks on the
  93         Preshared Key.
  94
  95 connections.<conn>.pull = yes
  96         Set the Mode Config mode to use.
  97
  98         If the default of _yes_ is used, Mode Config works in pull mode, where
  99         the initiator actively requests a virtual IP. With _no_, push mode is used,
 100         where the responder pushes down a virtual IP to the initiating peer.
 101
 102         Push mode is currently supported for IKEv1, but not in IKEv2. It is used
 103         by a few implementations only, pull mode is recommended.
 104
 105 connections.<conn>.encap = no
 106         Enforce UDP encapsulation by faking NAT-D payloads.
 107
 108         To enforce UDP encapsulation of ESP packets, the IKE daemon can fake the
 109         NAT detection payloads. This makes the peer believe that NAT takes
 110         place on the path, forcing it to encapsulate ESP packets in UDP.
 111
 112         Usually this is not required, but it can help to work around connectivity
 113         issues with too restrictive intermediary firewalls.
 114
 115 connections.<conn>.mobike = yes
 116         Enables MOBIKE on IKEv2 connections.
 117
 118         Enables MOBIKE on IKEv2 connections. MOBIKE is enabled by default on IKEv2
 119         connections, and allows mobility of clients and multi-homing on servers by
 120         migrating active IPsec tunnels.
 121
 122         Usually keeping MOBIKE enabled is unproblematic, as it is not used if the
 123         peer does not indicate support for it. However, due to the design of MOBIKE,
 124         IKEv2 always floats to port 4500 starting from the second exchange. Some
 125         implementations don't like this behavior, hence it can be disabled.
 126
 127 connections.<conn>.dpd_delay = 0s
 128         Interval of liveness checks (DPD).
 129
 130         Interval to check the liveness of a peer actively using IKEv2 INFORMATIONAL
 131         exchanges or IKEv1 R_U_THERE messages. Active DPD checking is only enforced
 132         if no IKE or ESP/AH packet has been received for the configured DPD delay.
 133
 134 connections.<conn>.dpd_timeout = 0s
 135         Timeout for DPD checks (IKEV1 only).
 136
 137         Charon by default uses the normal retransmission mechanism and timeouts to
 138         check the liveness of a peer, as all messages are used for liveness
 139         checking. For compatibility reasons, with IKEv1 a custom interval may be
 140         specified; this option has no effect on connections using IKE2.
 141
 142 connections.<conn>.fragmentation = no
 143         Use IKEv1 UDP packet fragmentation (_yes_, _no_ or _force_).
 144
 145         The default of _no_ disables IKEv1 fragmentation mechanism, _yes_ enables
 146         it if support has been indicated by the peer. _force_ enforces
 147         fragmentation if required even before the peer had a chance to indicate
 148         support for it.
 149
 150         IKE fragmentation is currently not supported with IKEv2.
 151
 152 connections.<conn>.send_certreq = yes
 153         Send certificate requests payloads (_yes_ or _no_).
 154
 155         Send certificate request payloads to offer trusted root CA certificates
 156         to the peer. Certificate requests help the peer to choose an appropriate
 157         certificate/private key for authentication and are enabled by default.
 158
 159         Disabling certificate requests can be useful if too many trusted root CA
 160         certificates are installed, as each certificate request increases the size
 161         of the initial IKE packets.
 162
 163 connections.<conn>.send_cert = ifasked
 164         Send certificate payloads (_yes_, _no_ or _ifasked_).
 165
 166         Send certificate payloads when using certificate authentication. With the
 167         default of _ifasked_ the daemon sends certificate payloads only if
 168         certificate requests have been received. _no_ disables sending of
 169         certificate payloads, _yes_ always sends certificate payloads whenever
 170         certificate authentication is used.
 171
 172 connections.<conn>.keyingtries = 1
 173         Number of retransmission sequences to perform during initial connect.
 174
 175         Number of retransmission sequences to perform during initial connect.
 176         Instead of giving up initiation after the first retransmission sequence with
 177         the default value of _1_, additional sequences may be started according to
 178         the configured value. A value of _0_ initiates a new sequence until the
 179         connection establishes or fails with a permanent error.
 180
 181 connections.<conn>.unique = no
 182         Connection uniqueness policy (_never_, _no_, _keep_ or _replace_).
 183
 184         Connection uniqueness policy to enforce. To avoid multiple connections
 185         from the same user, a uniqueness policy can be enforced. The value _never_
 186         does never enforce such a policy, even if a peer included INITIAL_CONTACT
 187         notification messages, whereas _no_ replaces existing connections for the
 188         same identity if a new one has the INITIAL_CONTACT notify. _keep_ rejects
 189         new connection attempts if the same user already has an active connection,
 190         _replace_ deletes any existing connection if a new one for the same user
 191         gets established.
 192
 193         To compare connections for uniqueness, the remote IKE identity is used. If
 194         EAP or XAuth authentication is involved, the EAP-Identity or XAuth username
 195         is used to enforce the uniqueness policy instead.
 196
 197 connections.<conn>.reauth_time = 0s
 198         Time to schedule IKE reauthentication.
 199
 200         Time to schedule IKE reauthentication. IKE reauthentication recreates the
 201         IKE/ISAKMP SA from scratch and re-evaluates the credentials. In asymmetric
 202         configurations (with EAP or configuration payloads) it might not be possible
 203         to actively reauthenticate as responder. The IKEv2 reauthentication lifetime
 204         negotiation can instruct the client to perform reauthentication.
 205
 206         Reauthentication is disabled by default. Enabling it usually may lead
 207         to small connection interruptions, as strongSwan uses a break-before-make
 208         policy with IKEv2 to avoid any conflicts with associated tunnel resources.
 209
 210 connections.<conn>.rekey_time = 4h
 211         Time to schedule IKE rekeying.
 212
 213         IKE rekeying refreshes key material using a Diffie-Hellman exchange, but
 214         does not re-check associated credentials. It is supported in IKEv2 only,
 215         IKEv1 performs a reauthentication procedure instead.
 216
 217         With the default value IKE rekeying is scheduled every 4 hours, minus the
 218         configured **rand_time**.
 219
 220 connections.<conn>.over_time = 10% of rekey_time/reauth_time
 221         Hard IKE_SA lifetime if rekey/reauth does not complete, as time.
 222
 223         Hard IKE_SA lifetime if rekey/reauth does not complete, as time.
 224         To avoid having an IKE/ISAKMP kept alive if IKE reauthentication or rekeying
 225         fails perpetually, a maximum hard lifetime may be specified. If the
 226         IKE_SA fails to rekey or reauthenticate within the specified time, the
 227         IKE_SA gets closed.
 228
 229         In contrast to CHILD_SA rekeying, **over_time** is relative in time to the
 230         **rekey_time** _and_ **reauth_time** values, as it applies to both.
 231
 232         The default is 10% of the longer of **rekey_time** and **reauth_time**.
 233
 234 connections.<conn>.rand_time = over_time
 235         Range of random time to subtract from rekey/reauth times.
 236
 237         Time range from which to choose a random value to subtract from
 238         rekey/reauth times. To avoid having both peers initiating the rekey/reauth
 239         procedure simultaneously, a random time gets subtracted from the
 240         rekey/reauth times.
 241
 242         The default is equal to the configured **over_time**.
 243
 244 connections.<conn>.pools =
 245         Comma separated list of named IP pools.
 246
 247         Comma separated list of named IP pools to allocate virtual IP addresses and
 248         other configuration attributes from. Each name references a pool by name
 249         from either the **pools** section or an external pool.
 250
 251 connections.<conn>.local<suffix> {}
 252         Section for a local authentication round.
 253
 254         Section for a local authentication round. A local authentication round
 255         defines the rules how authentication is performed for the local peer.
 256         Multiple rounds may be defined to use IKEv2 RFC 4739 Multiple Authentication
 257         or IKEv1 XAuth.
 258
 259         Each round is defined in a section having _local_ as prefix, and an optional
 260         unique suffix. To define a single authentication round, the suffix may be
 261         omitted.
 262
 263 connections.<conn>.local<suffix>.certs =
 264         Comma separated list of certificate candidates to use for authentication.
 265
 266         Comma separated list of certificate candidates to use for authentication.
 267         The certificates may use a relative path from the **swanctl** _x509_
 268         directory, or an absolute path.
 269
 270         The certificate used for authentication is selected based on the received
 271         certificate request payloads. If no appropriate CA can be located, the
 272         first certificate is used.
 273
 274 connections.<conn>.local<suffix>.auth = pubkey
 275         Authentication to perform locally (_pubkey_, _psk_, _xauth[-backend]_ or
 276         _eap[-method]_).
 277
 278         Authentication to perform locally. _pubkey_ uses public key authentication
 279         using a private key associated to a usable certificate. _psk_ uses
 280         pre-shared key authentication. The IKEv1 specific _xauth_ is used for
 281         XAuth or Hybrid authentication, while the IKEv2 specific _eap_ keyword
 282         defines EAP authentication.
 283
 284         For _xauth_, a specific backend name may be appended, separated by a dash.
 285         The appropriate _xauth_ backend is selected to perform the XAuth exchange.
 286         For traditional XAuth, the _xauth_ method is usually defined in the second
 287         authentication round following an initial _pubkey_ (or _psk_) round. Using
 288         _xauth_ in the first round performs Hybrid Mode client authentication.
 289
 290         For _eap_, a specific EAP method name may be appended, separated by a dash.
 291         An EAP module implementing the appropriate method is selected to perform
 292         the EAP conversation.
 293
 294 connections.<conn>.local<suffix>.id =
 295         IKE identity to use for authentication round.
 296
 297         IKE identity to use for authentication round. When using certificate
 298         authentication, the IKE identity must be contained in the certificate,
 299         either as subject or as subjectAltName.
 300
 301 connections.<conn>.local<suffix>.eap_id = id
 302         Client EAP-Identity to use in EAP-Identity exchange and the EAP method.
 303
 304 connections.<conn>.local<suffix>.aaa_id = remote-id
 305         Server side EAP-Identity to expect in the EAP method.
 306
 307         Server side EAP-Identity to expect in the EAP method. Some EAP methods, such
 308         as EAP-TLS, use an identity for the server to perform mutual authentication.
 309         This identity may differ from the IKE identity, especially when EAP
 310         authentication is delegated from the IKE responder to an AAA backend.
 311
 312         For EAP-(T)TLS, this defines the identity for wich the server must provide
 313         a certificate in the TLS exchange.
 314
 315 connections.<conn>.local<suffix>.xauth_id = id
 316         Client XAuth username used in the XAuth exchange.
 317
 318 connections.<conn>.remote<suffix> {}
 319         Section for a remote authentication round.
 320
 321         Section for a remote authentication round. A remote authentication round
 322         defines the constraints how the peers must authenticate to use this
 323         connection. Multiple rounds may be defined to use IKEv2 RFC 4739 Multiple
 324         Authentication or IKEv1 XAuth.
 325
 326         Each round is defined in a section having _remote_ as prefix, and an
 327         optional unique suffix. To define a single authentication round, the suffix
 328         may be omitted.
 329
 330 connections.<conn>.remote<suffix>.id = %any
 331         IKE identity to expect for authentication round.
 332
 333         IKE identity to expect for authentication round. When using certificate
 334         authentication, the IKE identity must be contained in the certificate,
 335         either as subject or as subjectAltName.
 336
 337 connections.<conn>.remote<suffix>.groups =
 338         Authorization group memberships to require.
 339
 340         Comma separated authorization group memberships to require. The peer must
 341         prove membership to at least one of the specified groups. Group membership
 342         can be certified by different means, for example by appropriate Attribute
 343         Certificates or by an AAA backend involved in the authentication.
 344
 345 connections.<conn>.remote<suffix>.certs =
 346         Comma separated list of certificate to accept for authentication.
 347
 348         Comma separated list of certificates to accept for authentication.
 349         The certificates may use a relative path from the **swanctl** _x509_
 350         directory, or an absolute path.
 351
 352 connections.<conn>.remote<suffix>.cacert =
 353         Comma separated list of CA certificates to accept for authentication.
 354
 355         Comma separated list of CA certificates to accept for authentication.
 356         The certificates may use a relative path from the **swanctl** _x509ca_
 357         directory, or an absolute path.
 358
 359 connections.<conn>.remote<suffix>.revocation = relaxed
 360         Certificate revocation policy, (_strict_, _ifuri_ or _relaxed_).
 361
 362         Certificate revocation policy for CRL or OCSP revocation.
 363
 364         A _strict_ revocation policy fails if no revocation information is
 365         available, i.e. the certificate is not known to be unrevoked.
 366
 367         _ifuri_ fails only if a CRL/OCSP URI is available, but certificate
 368         revocation checking fails, i.e. there should be revocation information
 369         available, but it could not be obtained.
 370
 371         The default revocation policy _relaxed_ fails only if a certificate
 372         is revoked, i.e. it is explicitly known that it is bad.
 373
 374 connections.<conn>.remote<suffix>.auth = pubkey
 375         Authentication to expect from remote (_pubkey_, _psk_, _xauth[-backend]_ or
 376         _eap[-method]_).
 377
 378         Authentication to expect from remote. See the **local** sections **auth**
 379         keyword description about the details of supported mechanisms.
 380
 381 connections.<conn>.children.<child> {}
 382         CHILD_SA configuration sub-section.
 383
 384         CHILD_SA configuration sub-section. Each connection definition may have
 385         one or more sections in its _children_ subsection. The section name
 386         defines the name of the CHILD_SA configuration, which must be unique within
 387         the connection.
 388
 389 connections.<conn>.children.<child>.ah_proposals =
 390         AH proposals to offer for the CHILD_SA.
 391
 392         AH proposals to offer for the CHILD_SA. A proposal is a set of algorithms.
 393         For AH, this includes an integrity algorithm and an optional Diffie-Hellman
 394         group. If a DH group is specified, CHILD_SA/Quick Mode rekeying and initial
 395         negotiation uses a separate Diffie-Hellman exchange using the specified
 396         group.
 397
 398         In IKEv2, multiple algorithms of the same kind can be specified in a single
 399         proposal, from which one gets selected. In IKEv1, only one algorithm per
 400         kind is allowed per proposal, more algorithms get implicitly stripped. Use
 401         multiple proposals to offer different algorithms combinations in IKEv1.
 402
 403         Algorithm keywords get separated using dashes. Multiple proposals may be
 404         separated by commas. The special value _default_ forms a default proposal
 405         of supported algorithms considered safe, and is usually a good choice
 406         for interoperability. By default no AH proposals are included, instead ESP
 407         is proposed.
 408
 409 connections.<conn>.children.<child>.esp_proposals = default
 410         ESP proposals to offer for the CHILD_SA.
 411
 412         ESP proposals to offer for the CHILD_SA. A proposal is a set of algorithms.
 413         For ESP non-AEAD proposals, this includes an integrity algorithm, an
 414         encryption algorithm, an optional Diffie-Hellman group and an optional
 415         Extended Sequence Number Mode indicator. For AEAD proposals, a combined
 416         mode algorithm is used instead of the separate encryption/integrity
 417         algorithms.
 418
 419         If a DH group is specified, CHILD_SA/Quick Mode rekeying and initial (non
 420         IKE_AUTH piggybacked) negotiation uses a separate Diffie-Hellman exchange
 421         using the specified group. Extended Sequence Number support may be indicated
 422         with the _esn_ and _noesn_ values, both may be included to indicate support
 423         for both modes. If omitted, _noesn_ is assumed.
 424
 425         In IKEv2, multiple algorithms of the same kind can be specified in a single
 426         proposal, from which one gets selected. In IKEv1, only one algorithm per
 427         kind is allowed per proposal, more algorithms get implicitly stripped. Use
 428         multiple proposals to offer different algorithms combinations in IKEv1.
 429
 430         Algorithm keywords get separated using dashes. Multiple proposals may be
 431         separated by commas. The special value _default_ forms a default proposal
 432         of supported algorithms considered safe, and is usually a good choice
 433         for interoperability. If no algorithms are specified for AH nor ESP,
 434         the _default_ set of algorithms for ESP is included.
 435
 436 connections.<conn>.children.<child>.local_ts = dynamic
 437         Local traffic selectors to include in CHILD_SA.
 438
 439         Comma separated list of local traffic selectors to include in CHILD_SA.
 440         Each selector is a CIDR subnet definition, followed by an optional
 441         proto/port selector. The special value _dynamic_ may be used instead of a
 442         subnet definition, which gets replaced by the tunnel outer address or the
 443         virtual IP, if negotiated. This is the default.
 444
 445         A protocol/port selector is surrounded by opening and closing square
 446         brackets. Between these brackets, a numeric or **getservent**(3) protocol
 447         name may be specified. After the optional protocol restriction, an optional
 448         port restriction may be specified, separated by a slash. The port
 449         restriction may be numeric, a **getservent**(3) service name, or the special
 450         value _opaque_ for RFC 4301 OPAQUE selectors. Port ranges may be specified
 451         as well, none of the kernel backends currently support port ranges, though.
 452
 453         Unless the Unity extension is used, IKEv1 supports the first specified
 454         selector only. IKEv1 uses very similar traffic selector narrowing as it is
 455         supported in the IKEv2 protocol.
 456
 457 connections.<conn>.children.<child>.remote_ts = dynamic
 458         Remote selectors to include in CHILD_SA.
 459
 460         Comma separated list of remote selectors to include in CHILD_SA. See
 461         **local_ts** for a description of the selector syntax.
 462
 463 connections.<conn>.children.<child>.rekey_time = 1h
 464         Time to schedule CHILD_SA rekeying.
 465
 466         Time to schedule CHILD_SA rekeying. CHILD_SA rekeying refreshes key
 467         material, optionally using a Diffie-Hellman exchange if a group is
 468         specified in the proposal.
 469
 470         To avoid rekey collisions initiated by both ends simultaneously, a value
 471         in the range of **rand_time** gets subtracted to form the effective soft
 472         lifetime.
 473
 474         By default CHILD_SA rekeying is scheduled every hour, minus **rand_time**.
 475
 476 connections.<conn>.children.<child>.life_time = rekey_time + 10%
 477         Maximum lifetime before CHILD_SA gets closed, as time.
 478
 479         Maximum lifetime before CHILD_SA gets closed. Usually this hard lifetime
 480         is never reached, because the CHILD_SA gets rekeyed before.
 481         If that fails for whatever reason, this limit closes the CHILD_SA.
 482
 483         The default is 10% more than the **rekey_time**.
 484
 485 connections.<conn>.children.<child>.rand_time = life_time - rekey_time
 486         Range of random time to subtract from **rekey_time**.
 487
 488         Time range from which to choose a random value to subtract from
 489         **rekey_time**. The default is the difference between **life_time** and
 490         **rekey_time**.
 491
 492 connections.<conn>.children.<child>.rekey_bytes = 0
 493         Number of bytes processed before initiating CHILD_SA rekeying.
 494
 495         Number of bytes processed before initiating CHILD_SA rekeying. CHILD_SA
 496         rekeying refreshes key material, optionally using a Diffie-Hellman exchange
 497         if a group is specified in the proposal.
 498
 499         To avoid rekey collisions initiated by both ends simultaneously, a value
 500         in the range of **rand_bytes** gets subtracted to form the effective soft
 501         volume limit.
 502
 503         Volume based CHILD_SA rekeying is disabled by default.
 504
 505 connections.<conn>.children.<child>.life_bytes = rekey_bytes + 10%
 506         Maximum bytes processed before CHILD_SA gets closed.
 507
 508         Maximum bytes processed before CHILD_SA gets closed. Usually this hard
 509         volume limit is never reached, because the CHILD_SA gets rekeyed before.
 510         If that fails for whatever reason, this limit closes the CHILD_SA.
 511
 512         The default is 10% more than **rekey_bytes**.
 513
 514 connections.<conn>.children.<child>.rand_bytes = life_bytes - rekey_bytes
 515         Range of random bytes to subtract from **rekey_bytes**.
 516
 517         Byte range from which to choose a random value to subtract from
 518         **rekey_bytes**. The default is the difference between **life_bytes** and
 519         **rekey_bytes**.
 520
 521 connections.<conn>.children.<child>.rekey_packets = 0
 522         Number of packets processed before initiating CHILD_SA rekeying.
 523
 524         Number of packets processed before initiating CHILD_SA rekeying. CHILD_SA
 525         rekeying refreshes key material, optionally using a Diffie-Hellman exchange
 526         if a group is specified in the proposal.
 527
 528         To avoid rekey collisions initiated by both ends simultaneously, a value
 529         in the range of **rand_packets** gets subtracted to form the effective soft
 530         packet count limit.
 531
 532         Packet count based CHILD_SA rekeying is disabled by default.
 533
 534 connections.<conn>.children.<child>.life_packets = rekey_packets + 10%
 535         Maximum number of packets processed before CHILD_SA gets closed.
 536
 537         Maximum number of packets processed before CHILD_SA gets closed. Usually
 538         this hard packets limit is never reached, because the CHILD_SA gets rekeyed
 539         before. If that fails for whatever reason, this limit closes the CHILD_SA.
 540
 541         The default is 10% more than **rekey_bytes**.
 542
 543 connections.<conn>.children.<child>.rand_packets = life_packets - rekey_packets
 544         Range of random packets to subtract from **packets_bytes**.
 545
 546         Packet range from which to choose a random value to subtract from
 547         **rekey_packets**. The default is the difference between **life_packets**
 548         and **rekey_packets**.
 549
 550 connections.<conn>.children.<child>.updown =
 551         Updown script to invoke on CHILD_SA up and down events.
 552
 553 connections.<conn>.children.<child>.hostaccess = yes
 554         Hostaccess variable to pass to **updown** script.
 555
 556 connections.<conn>.children.<child>.mode = tunnel
 557         IPsec Mode to establish (_tunnel_, _transport_, _beet_, _pass_ or _drop_).
 558
 559         IPsec Mode to establish CHILD_SA with. _tunnel_ negotiates the CHILD_SA
 560         in IPsec Tunnel Mode, whereas _transport_ uses IPsec Transport Mode. _beet_
 561         is the Bound End to End Tunnel mixture mode, working with fixed inner
 562         addresses without the need to include them in each packet.
 563
 564         Both _transport_ and _beet_ modes are subject to mode negotiation; _tunnel_
 565         mode is negotiated if the preferred mode is not available.
 566
 567         _pass_ and _drop_ are used to install shunt policies, which explicitly
 568         bypass the defined traffic from IPsec processing, or drop it, respectively.
 569
 570 connections.<conn>.children.<child>.dpd_action = clear
 571         Action to perform on DPD timeout (_clear_, _trap_ or _restart_).
 572
 573         Action to perform for this CHILD_SA on DPD timeout. The default _clear_
 574         closes the CHILD_SA and does not take further action. _trap_ installs
 575         a trap policy, which will catch matching traffic and tries to re-negotiate
 576         the tunnel on-demand. _restart_ immediately tries to re-negotiate the
 577         CHILD_SA under a fresh IKE_SA.
 578
 579 connections.<conn>.children.<child>.ipcomp = no
 580         Enable IPComp compression before encryption.
 581
 582         Enable IPComp compression before encryption. If enabled, IKE tries to
 583         negotiate IPComp compression to compress ESP payload data prior to
 584         encryption.
 585
 586 connections.<conn>.children.<child>.inactivity = 0s
 587         Timeout before closing CHILD_SA after inactivity.
 588
 589         Timeout before closing CHILD_SA after inactivity. If no traffic has
 590         been processed in either direction for the configured timeout, the CHILD_SA
 591         gets closed due to inactivity. The default value of _0_ disables inactivity
 592         checks.
 593
 594 connections.<conn>.children.<child>.reqid = 0
 595         Fixed reqid to use for this CHILD_SA.
 596
 597         Fixed reqid to use for this CHILD_SA. This might be helpful in some
 598         scenarios, but works only if each CHILD_SA configuration is instantiated
 599         not more than once. The default of _0_ uses dynamic reqids, allocated
 600         incrementally.
 601
 602 connections.<conn>.children.<child>.mark_in = 0/0x00000000
 603         Netfilter mark and mask for input traffic.
 604
 605         Netfilter mark and mask for input traffic. On Linux Netfilter may apply
 606         marks to each packet coming from a tunnel having that option set. The
 607         mark may then be used by Netfilter to match rules.
 608
 609         An additional mask may be appended to the mark, separated by _/_. The
 610         default mask if omitted is 0xffffffff.
 611
 612 connections.<conn>.children.<child>.mark_out = 0/0x00000000
 613         Netfilter mark and mask for output traffic.
 614
 615         Netfilter mark and mask for output traffic. On Linux Netfilter may require
 616         marks on each packet to match a policy having that option set. This allows
 617         Netfilter rules to select specific tunnels for outgoing traffic.
 618
 619         An additional mask may be appended to the mark, separated by _/_. The
 620         default mask if omitted is 0xffffffff.
 621
 622 connections.<conn>.children.<child>.tfc_padding = 0
 623         Traffic Flow Confidentiality padding.
 624
 625         Pads ESP packets with additional data to have a consistent ESP packet size
 626         for improved Traffic Flow Confidentiality. The padding defines the minimum
 627         size of all ESP packets sent.
 628
 629         The default value of 0 disables TFC padding, the special value _mtu_ adds
 630         TFC padding to create a packet size equal to the Path Maximum Transfer Unit.
 631
 632 connections.<conn>.children.<child>.replay_window = 32
 633         IPsec replay window to configure for this CHILD_SA.
 634
 635         IPsec replay window to configure for this CHILD_SA. Larger values than the
 636         default of 32 are supported using the Netlink backend only, a value of 0
 637         disables IPsec replay protection.
 638
 639 connections.<conn>.children.<child>.start_action = none
 640         Action to perform after loading the configuration (_none_, _trap_, _start_).
 641
 642         Action to perform after loading the configuration. The default of _none_
 643         loads the connection only, which then can be manually initiated or used as
 644         a responder configuration.
 645
 646         The value _trap_ installs a trap policy, which triggers the tunnel as soon
 647         as matching traffic has been detected. The value _start_ initiates
 648         the connection actively.
 649
 650         When unloading or replacing a CHILD_SA configuration having a
 651         **start_action** different from _none_, the inverse action is performed.
 652         Configurations with _start_ get closed, while such with _trap_ get
 653         uninstalled.
 654
 655 connections.<conn>.children.<child>.close_action = none
 656         Action to perform after a CHILD_SA gets closed (_none_, _trap_, _start_).
 657
 658         Action to perform after a CHILD_SA gets closed by the peer. The default of
 659         _none_ does not take any action, _trap_ installs a trap policy for the
 660         CHILD_SA. _start_ tries to re-create the CHILD_SA.
 661
 662         **close_action** does not provide any guarantee that the CHILD_SA is kept
 663         alive. It acts on explicit close messages only, but not on negotiation
 664         failures. Use trap policies to reliably re-create failed CHILD_SAs.
 665
 666 secrets { # }
 667         Section defining secrets for IKE and EAP/XAuth authentication.
 668
 669         Section defining secrets for IKE and EAP/XAuth authentication. The
 670         **secrets** section takes sub-sections having a specific prefix which
 671         defines the secret type.
 672
 673 secrets.eap<suffix> { # }
 674         EAP secret section for a specific secret.
 675
 676         EAP secret section for a specific secret. Each EAP secret is defined in
 677         a unique section having the _eap_ prefix. EAP secrets are used for XAuth
 678         authentication as well.
 679
 680 secrets.xauth<suffix> { # }
 681         XAuth secret section for a specific secret.
 682
 683         XAuth secret section for a specific secret. **xauth** is just an alias
 684         for **eap**, secrets under both section prefixes are used for both EAP and
 685         XAuth authentication.
 686
 687 secrets.eap<suffix>.secret =
 688         Value of the EAP/XAuth secret.
 689
 690         Value of the EAP/XAuth secret. It may either be an ASCII string, a hex
 691         encoded string if it has a _0x_ prefix, or a Base64 encoded string if it
 692         has a _0s_ prefix in its value.
 693
 694 secrets.eap<suffix>.id<suffix> =
 695         Identity the EAP/XAuth secret belongs to.
 696
 697         Identity the EAP/XAuth secret belongs to. Multiple unique identities may
 698         be specified, each having an _id_ prefix, if a secret is shared between
 699         multiple users.
 700
 701 secrets.ike<suffix> { # }
 702         IKE preshared secret section for a specific secret.
 703
 704         IKE preshared secret section for a specific secret. Each IKE PSK is defined
 705         in a unique section having the _ike_ prefix.
 706
 707 secrets.ike<suffix>.secret =
 708         Value of the IKE preshared secret.
 709
 710         Value of the IKE preshared secret. It may either be an ASCII string,
 711         a hex encoded string if it has a _0x_ prefix, or a Base64 encoded string if
 712         it has a _0s_ prefix in its value.
 713
 714 secrets.ike<suffix>.id<suffix> =
 715         IKE identity the IKE preshared secret belongs to.
 716
 717         IKE identity the IKE preshared secret belongs to. Multiple unique identities
 718         may be specified, each having an _id_ prefix, if a secret is shared between
 719         multiple peers.
 720
 721 pools { # }
 722         Section defining named pools.
 723
 724         Section defining named pools. Named pools may be referenced by connections
 725         with the **pools** option to assign virtual IPs and other configuration
 726         attributes.
 727
 728 pools.<name> { # }
 729         Section defining a single pool with a unique name.
 730
 731 pools.<name>.addrs =
 732         Subnet defining addresses allocated in pool.
 733
 734         Subnet defining addresses allocated in pool. Accepts a single CIDR subnet
 735         defining the pool to allocate addresses from. Pools must be unique and
 736         non-overlapping.
 737
 738 pools.<name>.<attr> =
 739         Comma separated list of additional attributes from type <attr>.
 740
 741         Comma separated list of additional attributes of type **<attr>**. The
 742         attribute type may be one of _dns_, _nbns_, _dhcp_, _netmask_, _server_,
 743         _subnet_, _split_include_ and _split_exclude_ to define addresses or CIDR
 744         subnets for the corresponding attribute types. Alternatively, **<attr>** can
 745         be a numerical identifier, for which string attribute values are accepted
 746         as well.