linux/Documentation/networking/ip-sysctl.txt
<<
>>
Prefs
   1/proc/sys/net/ipv4/* Variables:
   2
   3ip_forward - BOOLEAN
   4        0 - disabled (default)
   5        not 0 - enabled
   6
   7        Forward Packets between interfaces.
   8
   9        This variable is special, its change resets all configuration
  10        parameters to their default state (RFC1122 for hosts, RFC1812
  11        for routers)
  12
  13ip_default_ttl - INTEGER
  14        Default value of TTL field (Time To Live) for outgoing (but not
  15        forwarded) IP packets. Should be between 1 and 255 inclusive.
  16        Default: 64 (as recommended by RFC1700)
  17
  18ip_no_pmtu_disc - INTEGER
  19        Disable Path MTU Discovery. If enabled in mode 1 and a
  20        fragmentation-required ICMP is received, the PMTU to this
  21        destination will be set to min_pmtu (see below). You will need
  22        to raise min_pmtu to the smallest interface MTU on your system
  23        manually if you want to avoid locally generated fragments.
  24
  25        In mode 2 incoming Path MTU Discovery messages will be
  26        discarded. Outgoing frames are handled the same as in mode 1,
  27        implicitly setting IP_PMTUDISC_DONT on every created socket.
  28
  29        Mode 3 is a hardend pmtu discover mode. The kernel will only
  30        accept fragmentation-needed errors if the underlying protocol
  31        can verify them besides a plain socket lookup. Current
  32        protocols for which pmtu events will be honored are TCP, SCTP
  33        and DCCP as they verify e.g. the sequence number or the
  34        association. This mode should not be enabled globally but is
  35        only intended to secure e.g. name servers in namespaces where
  36        TCP path mtu must still work but path MTU information of other
  37        protocols should be discarded. If enabled globally this mode
  38        could break other protocols.
  39
  40        Possible values: 0-3
  41        Default: FALSE
  42
  43min_pmtu - INTEGER
  44        default 552 - minimum discovered Path MTU
  45
  46ip_forward_use_pmtu - BOOLEAN
  47        By default we don't trust protocol path MTUs while forwarding
  48        because they could be easily forged and can lead to unwanted
  49        fragmentation by the router.
  50        You only need to enable this if you have user-space software
  51        which tries to discover path mtus by itself and depends on the
  52        kernel honoring this information. This is normally not the
  53        case.
  54        Default: 0 (disabled)
  55        Possible values:
  56        0 - disabled
  57        1 - enabled
  58
  59fwmark_reflect - BOOLEAN
  60        Controls the fwmark of kernel-generated IPv4 reply packets that are not
  61        associated with a socket for example, TCP RSTs or ICMP echo replies).
  62        If unset, these packets have a fwmark of zero. If set, they have the
  63        fwmark of the packet they are replying to.
  64        Default: 0
  65
  66fib_multipath_use_neigh - BOOLEAN
  67        Use status of existing neighbor entry when determining nexthop for
  68        multipath routes. If disabled, neighbor information is not used and
  69        packets could be directed to a failed nexthop. Only valid for kernels
  70        built with CONFIG_IP_ROUTE_MULTIPATH enabled.
  71        Default: 0 (disabled)
  72        Possible values:
  73        0 - disabled
  74        1 - enabled
  75
  76fib_multipath_hash_policy - INTEGER
  77        Controls which hash policy to use for multipath routes. Only valid
  78        for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
  79        Default: 0 (Layer 3)
  80        Possible values:
  81        0 - Layer 3
  82        1 - Layer 4
  83
  84route/max_size - INTEGER
  85        Maximum number of routes allowed in the kernel.  Increase
  86        this when using large numbers of interfaces and/or routes.
  87        From linux kernel 3.6 onwards, this is deprecated for ipv4
  88        as route cache is no longer used.
  89
  90neigh/default/gc_thresh1 - INTEGER
  91        Minimum number of entries to keep.  Garbage collector will not
  92        purge entries if there are fewer than this number.
  93        Default: 128
  94
  95neigh/default/gc_thresh2 - INTEGER
  96        Threshold when garbage collector becomes more aggressive about
  97        purging entries. Entries older than 5 seconds will be cleared
  98        when over this number.
  99        Default: 512
 100
 101neigh/default/gc_thresh3 - INTEGER
 102        Maximum number of neighbor entries allowed.  Increase this
 103        when using large numbers of interfaces and when communicating
 104        with large numbers of directly-connected peers.
 105        Default: 1024
 106
 107neigh/default/unres_qlen_bytes - INTEGER
 108        The maximum number of bytes which may be used by packets
 109        queued for each unresolved address by other network layers.
 110        (added in linux 3.3)
 111        Setting negative value is meaningless and will return error.
 112        Default: SK_WMEM_MAX, (same as net.core.wmem_default).
 113                Exact value depends on architecture and kernel options,
 114                but should be enough to allow queuing 256 packets
 115                of medium size.
 116
 117neigh/default/unres_qlen - INTEGER
 118        The maximum number of packets which may be queued for each
 119        unresolved address by other network layers.
 120        (deprecated in linux 3.3) : use unres_qlen_bytes instead.
 121        Prior to linux 3.3, the default value is 3 which may cause
 122        unexpected packet loss. The current default value is calculated
 123        according to default value of unres_qlen_bytes and true size of
 124        packet.
 125        Default: 101
 126
 127mtu_expires - INTEGER
 128        Time, in seconds, that cached PMTU information is kept.
 129
 130min_adv_mss - INTEGER
 131        The advertised MSS depends on the first hop route MTU, but will
 132        never be lower than this setting.
 133
 134IP Fragmentation:
 135
 136ipfrag_high_thresh - INTEGER
 137        Maximum memory used to reassemble IP fragments. When
 138        ipfrag_high_thresh bytes of memory is allocated for this purpose,
 139        the fragment handler will toss packets until ipfrag_low_thresh
 140        is reached. This also serves as a maximum limit to namespaces
 141        different from the initial one.
 142
 143ipfrag_low_thresh - INTEGER
 144        Maximum memory used to reassemble IP fragments before the kernel
 145        begins to remove incomplete fragment queues to free up resources.
 146        The kernel still accepts new fragments for defragmentation.
 147
 148ipfrag_time - INTEGER
 149        Time in seconds to keep an IP fragment in memory.
 150
 151ipfrag_max_dist - INTEGER
 152        ipfrag_max_dist is a non-negative integer value which defines the
 153        maximum "disorder" which is allowed among fragments which share a
 154        common IP source address. Note that reordering of packets is
 155        not unusual, but if a large number of fragments arrive from a source
 156        IP address while a particular fragment queue remains incomplete, it
 157        probably indicates that one or more fragments belonging to that queue
 158        have been lost. When ipfrag_max_dist is positive, an additional check
 159        is done on fragments before they are added to a reassembly queue - if
 160        ipfrag_max_dist (or more) fragments have arrived from a particular IP
 161        address between additions to any IP fragment queue using that source
 162        address, it's presumed that one or more fragments in the queue are
 163        lost. The existing fragment queue will be dropped, and a new one
 164        started. An ipfrag_max_dist value of zero disables this check.
 165
 166        Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
 167        result in unnecessarily dropping fragment queues when normal
 168        reordering of packets occurs, which could lead to poor application
 169        performance. Using a very large value, e.g. 50000, increases the
 170        likelihood of incorrectly reassembling IP fragments that originate
 171        from different IP datagrams, which could result in data corruption.
 172        Default: 64
 173
 174INET peer storage:
 175
 176inet_peer_threshold - INTEGER
 177        The approximate size of the storage.  Starting from this threshold
 178        entries will be thrown aggressively.  This threshold also determines
 179        entries' time-to-live and time intervals between garbage collection
 180        passes.  More entries, less time-to-live, less GC interval.
 181
 182inet_peer_minttl - INTEGER
 183        Minimum time-to-live of entries.  Should be enough to cover fragment
 184        time-to-live on the reassembling side.  This minimum time-to-live  is
 185        guaranteed if the pool size is less than inet_peer_threshold.
 186        Measured in seconds.
 187
 188inet_peer_maxttl - INTEGER
 189        Maximum time-to-live of entries.  Unused entries will expire after
 190        this period of time if there is no memory pressure on the pool (i.e.
 191        when the number of entries in the pool is very small).
 192        Measured in seconds.
 193
 194TCP variables:
 195
 196somaxconn - INTEGER
 197        Limit of socket listen() backlog, known in userspace as SOMAXCONN.
 198        Defaults to 128.  See also tcp_max_syn_backlog for additional tuning
 199        for TCP sockets.
 200
 201tcp_abort_on_overflow - BOOLEAN
 202        If listening service is too slow to accept new connections,
 203        reset them. Default state is FALSE. It means that if overflow
 204        occurred due to a burst, connection will recover. Enable this
 205        option _only_ if you are really sure that listening daemon
 206        cannot be tuned to accept connections faster. Enabling this
 207        option can harm clients of your server.
 208
 209tcp_adv_win_scale - INTEGER
 210        Count buffering overhead as bytes/2^tcp_adv_win_scale
 211        (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
 212        if it is <= 0.
 213        Possible values are [-31, 31], inclusive.
 214        Default: 1
 215
 216tcp_allowed_congestion_control - STRING
 217        Show/set the congestion control choices available to non-privileged
 218        processes. The list is a subset of those listed in
 219        tcp_available_congestion_control.
 220        Default is "reno" and the default setting (tcp_congestion_control).
 221
 222tcp_app_win - INTEGER
 223        Reserve max(window/2^tcp_app_win, mss) of window for application
 224        buffer. Value 0 is special, it means that nothing is reserved.
 225        Default: 31
 226
 227tcp_autocorking - BOOLEAN
 228        Enable TCP auto corking :
 229        When applications do consecutive small write()/sendmsg() system calls,
 230        we try to coalesce these small writes as much as possible, to lower
 231        total amount of sent packets. This is done if at least one prior
 232        packet for the flow is waiting in Qdisc queues or device transmit
 233        queue. Applications can still use TCP_CORK for optimal behavior
 234        when they know how/when to uncork their sockets.
 235        Default : 1
 236
 237tcp_available_congestion_control - STRING
 238        Shows the available congestion control choices that are registered.
 239        More congestion control algorithms may be available as modules,
 240        but not loaded.
 241
 242tcp_base_mss - INTEGER
 243        The initial value of search_low to be used by the packetization layer
 244        Path MTU discovery (MTU probing).  If MTU probing is enabled,
 245        this is the initial MSS used by the connection.
 246
 247tcp_congestion_control - STRING
 248        Set the congestion control algorithm to be used for new
 249        connections. The algorithm "reno" is always available, but
 250        additional choices may be available based on kernel configuration.
 251        Default is set as part of kernel configuration.
 252        For passive connections, the listener congestion control choice
 253        is inherited.
 254        [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
 255
 256tcp_dsack - BOOLEAN
 257        Allows TCP to send "duplicate" SACKs.
 258
 259tcp_early_retrans - INTEGER
 260        Tail loss probe (TLP) converts RTOs occurring due to tail
 261        losses into fast recovery (draft-ietf-tcpm-rack). Note that
 262        TLP requires RACK to function properly (see tcp_recovery below)
 263        Possible values:
 264                0 disables TLP
 265                3 or 4 enables TLP
 266        Default: 3
 267
 268tcp_ecn - INTEGER
 269        Control use of Explicit Congestion Notification (ECN) by TCP.
 270        ECN is used only when both ends of the TCP connection indicate
 271        support for it.  This feature is useful in avoiding losses due
 272        to congestion by allowing supporting routers to signal
 273        congestion before having to drop packets.
 274        Possible values are:
 275                0 Disable ECN.  Neither initiate nor accept ECN.
 276                1 Enable ECN when requested by incoming connections and
 277                  also request ECN on outgoing connection attempts.
 278                2 Enable ECN when requested by incoming connections
 279                  but do not request ECN on outgoing connections.
 280        Default: 2
 281
 282tcp_ecn_fallback - BOOLEAN
 283        If the kernel detects that ECN connection misbehaves, enable fall
 284        back to non-ECN. Currently, this knob implements the fallback
 285        from RFC3168, section 6.1.1.1., but we reserve that in future,
 286        additional detection mechanisms could be implemented under this
 287        knob. The value is not used, if tcp_ecn or per route (or congestion
 288        control) ECN settings are disabled.
 289        Default: 1 (fallback enabled)
 290
 291tcp_fack - BOOLEAN
 292        This is a legacy option, it has no effect anymore.
 293
 294tcp_fin_timeout - INTEGER
 295        The length of time an orphaned (no longer referenced by any
 296        application) connection will remain in the FIN_WAIT_2 state
 297        before it is aborted at the local end.  While a perfectly
 298        valid "receive only" state for an un-orphaned connection, an
 299        orphaned connection in FIN_WAIT_2 state could otherwise wait
 300        forever for the remote to close its end of the connection.
 301        Cf. tcp_max_orphans
 302        Default: 60 seconds
 303
 304tcp_frto - INTEGER
 305        Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
 306        F-RTO is an enhanced recovery algorithm for TCP retransmission
 307        timeouts.  It is particularly beneficial in networks where the
 308        RTT fluctuates (e.g., wireless). F-RTO is sender-side only
 309        modification. It does not require any support from the peer.
 310
 311        By default it's enabled with a non-zero value. 0 disables F-RTO.
 312
 313tcp_invalid_ratelimit - INTEGER
 314        Limit the maximal rate for sending duplicate acknowledgments
 315        in response to incoming TCP packets that are for an existing
 316        connection but that are invalid due to any of these reasons:
 317
 318          (a) out-of-window sequence number,
 319          (b) out-of-window acknowledgment number, or
 320          (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
 321
 322        This can help mitigate simple "ack loop" DoS attacks, wherein
 323        a buggy or malicious middlebox or man-in-the-middle can
 324        rewrite TCP header fields in manner that causes each endpoint
 325        to think that the other is sending invalid TCP segments, thus
 326        causing each side to send an unterminating stream of duplicate
 327        acknowledgments for invalid segments.
 328
 329        Using 0 disables rate-limiting of dupacks in response to
 330        invalid segments; otherwise this value specifies the minimal
 331        space between sending such dupacks, in milliseconds.
 332
 333        Default: 500 (milliseconds).
 334
 335tcp_keepalive_time - INTEGER
 336        How often TCP sends out keepalive messages when keepalive is enabled.
 337        Default: 2hours.
 338
 339tcp_keepalive_probes - INTEGER
 340        How many keepalive probes TCP sends out, until it decides that the
 341        connection is broken. Default value: 9.
 342
 343tcp_keepalive_intvl - INTEGER
 344        How frequently the probes are send out. Multiplied by
 345        tcp_keepalive_probes it is time to kill not responding connection,
 346        after probes started. Default value: 75sec i.e. connection
 347        will be aborted after ~11 minutes of retries.
 348
 349tcp_l3mdev_accept - BOOLEAN
 350        Enables child sockets to inherit the L3 master device index.
 351        Enabling this option allows a "global" listen socket to work
 352        across L3 master domains (e.g., VRFs) with connected sockets
 353        derived from the listen socket to be bound to the L3 domain in
 354        which the packets originated. Only valid when the kernel was
 355        compiled with CONFIG_NET_L3_MASTER_DEV.
 356
 357tcp_low_latency - BOOLEAN
 358        This is a legacy option, it has no effect anymore.
 359
 360tcp_max_orphans - INTEGER
 361        Maximal number of TCP sockets not attached to any user file handle,
 362        held by system. If this number is exceeded orphaned connections are
 363        reset immediately and warning is printed. This limit exists
 364        only to prevent simple DoS attacks, you _must_ not rely on this
 365        or lower the limit artificially, but rather increase it
 366        (probably, after increasing installed memory),
 367        if network conditions require more than default value,
 368        and tune network services to linger and kill such states
 369        more aggressively. Let me to remind again: each orphan eats
 370        up to ~64K of unswappable memory.
 371
 372tcp_max_syn_backlog - INTEGER
 373        Maximal number of remembered connection requests, which have not
 374        received an acknowledgment from connecting client.
 375        The minimal value is 128 for low memory machines, and it will
 376        increase in proportion to the memory of machine.
 377        If server suffers from overload, try increasing this number.
 378
 379tcp_max_tw_buckets - INTEGER
 380        Maximal number of timewait sockets held by system simultaneously.
 381        If this number is exceeded time-wait socket is immediately destroyed
 382        and warning is printed. This limit exists only to prevent
 383        simple DoS attacks, you _must_ not lower the limit artificially,
 384        but rather increase it (probably, after increasing installed memory),
 385        if network conditions require more than default value.
 386
 387tcp_mem - vector of 3 INTEGERs: min, pressure, max
 388        min: below this number of pages TCP is not bothered about its
 389        memory appetite.
 390
 391        pressure: when amount of memory allocated by TCP exceeds this number
 392        of pages, TCP moderates its memory consumption and enters memory
 393        pressure mode, which is exited when memory consumption falls
 394        under "min".
 395
 396        max: number of pages allowed for queueing by all TCP sockets.
 397
 398        Defaults are calculated at boot time from amount of available
 399        memory.
 400
 401tcp_min_rtt_wlen - INTEGER
 402        The window length of the windowed min filter to track the minimum RTT.
 403        A shorter window lets a flow more quickly pick up new (higher)
 404        minimum RTT when it is moved to a longer path (e.g., due to traffic
 405        engineering). A longer window makes the filter more resistant to RTT
 406        inflations such as transient congestion. The unit is seconds.
 407        Default: 300
 408
 409tcp_moderate_rcvbuf - BOOLEAN
 410        If set, TCP performs receive buffer auto-tuning, attempting to
 411        automatically size the buffer (no greater than tcp_rmem[2]) to
 412        match the size required by the path for full throughput.  Enabled by
 413        default.
 414
 415tcp_mtu_probing - INTEGER
 416        Controls TCP Packetization-Layer Path MTU Discovery.  Takes three
 417        values:
 418          0 - Disabled
 419          1 - Disabled by default, enabled when an ICMP black hole detected
 420          2 - Always enabled, use initial MSS of tcp_base_mss.
 421
 422tcp_probe_interval - INTEGER
 423        Controls how often to start TCP Packetization-Layer Path MTU
 424        Discovery reprobe. The default is reprobing every 10 minutes as
 425        per RFC4821.
 426
 427tcp_probe_threshold - INTEGER
 428        Controls when TCP Packetization-Layer Path MTU Discovery probing
 429        will stop in respect to the width of search range in bytes. Default
 430        is 8 bytes.
 431
 432tcp_no_metrics_save - BOOLEAN
 433        By default, TCP saves various connection metrics in the route cache
 434        when the connection closes, so that connections established in the
 435        near future can use these to set initial conditions.  Usually, this
 436        increases overall performance, but may sometimes cause performance
 437        degradation.  If set, TCP will not cache metrics on closing
 438        connections.
 439
 440tcp_orphan_retries - INTEGER
 441        This value influences the timeout of a locally closed TCP connection,
 442        when RTO retransmissions remain unacknowledged.
 443        See tcp_retries2 for more details.
 444
 445        The default value is 8.
 446        If your machine is a loaded WEB server,
 447        you should think about lowering this value, such sockets
 448        may consume significant resources. Cf. tcp_max_orphans.
 449
 450tcp_recovery - INTEGER
 451        This value is a bitmap to enable various experimental loss recovery
 452        features.
 453
 454        RACK: 0x1 enables the RACK loss detection for fast detection of lost
 455              retransmissions and tail drops.
 456        RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
 457
 458        Default: 0x1
 459
 460tcp_reordering - INTEGER
 461        Initial reordering level of packets in a TCP stream.
 462        TCP stack can then dynamically adjust flow reordering level
 463        between this initial value and tcp_max_reordering
 464        Default: 3
 465
 466tcp_max_reordering - INTEGER
 467        Maximal reordering level of packets in a TCP stream.
 468        300 is a fairly conservative value, but you might increase it
 469        if paths are using per packet load balancing (like bonding rr mode)
 470        Default: 300
 471
 472tcp_retrans_collapse - BOOLEAN
 473        Bug-to-bug compatibility with some broken printers.
 474        On retransmit try to send bigger packets to work around bugs in
 475        certain TCP stacks.
 476
 477tcp_retries1 - INTEGER
 478        This value influences the time, after which TCP decides, that
 479        something is wrong due to unacknowledged RTO retransmissions,
 480        and reports this suspicion to the network layer.
 481        See tcp_retries2 for more details.
 482
 483        RFC 1122 recommends at least 3 retransmissions, which is the
 484        default.
 485
 486tcp_retries2 - INTEGER
 487        This value influences the timeout of an alive TCP connection,
 488        when RTO retransmissions remain unacknowledged.
 489        Given a value of N, a hypothetical TCP connection following
 490        exponential backoff with an initial RTO of TCP_RTO_MIN would
 491        retransmit N times before killing the connection at the (N+1)th RTO.
 492
 493        The default value of 15 yields a hypothetical timeout of 924.6
 494        seconds and is a lower bound for the effective timeout.
 495        TCP will effectively time out at the first RTO which exceeds the
 496        hypothetical timeout.
 497
 498        RFC 1122 recommends at least 100 seconds for the timeout,
 499        which corresponds to a value of at least 8.
 500
 501tcp_rfc1337 - BOOLEAN
 502        If set, the TCP stack behaves conforming to RFC1337. If unset,
 503        we are not conforming to RFC, but prevent TCP TIME_WAIT
 504        assassination.
 505        Default: 0
 506
 507tcp_rmem - vector of 3 INTEGERs: min, default, max
 508        min: Minimal size of receive buffer used by TCP sockets.
 509        It is guaranteed to each TCP socket, even under moderate memory
 510        pressure.
 511        Default: 4K
 512
 513        default: initial size of receive buffer used by TCP sockets.
 514        This value overrides net.core.rmem_default used by other protocols.
 515        Default: 87380 bytes. This value results in window of 65535 with
 516        default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
 517        less for default tcp_app_win. See below about these variables.
 518
 519        max: maximal size of receive buffer allowed for automatically
 520        selected receiver buffers for TCP socket. This value does not override
 521        net.core.rmem_max.  Calling setsockopt() with SO_RCVBUF disables
 522        automatic tuning of that socket's receive buffer size, in which
 523        case this value is ignored.
 524        Default: between 87380B and 6MB, depending on RAM size.
 525
 526tcp_sack - BOOLEAN
 527        Enable select acknowledgments (SACKS).
 528
 529tcp_slow_start_after_idle - BOOLEAN
 530        If set, provide RFC2861 behavior and time out the congestion
 531        window after an idle period.  An idle period is defined at
 532        the current RTO.  If unset, the congestion window will not
 533        be timed out after an idle period.
 534        Default: 1
 535
 536tcp_stdurg - BOOLEAN
 537        Use the Host requirements interpretation of the TCP urgent pointer field.
 538        Most hosts use the older BSD interpretation, so if you turn this on
 539        Linux might not communicate correctly with them.
 540        Default: FALSE
 541
 542tcp_synack_retries - INTEGER
 543        Number of times SYNACKs for a passive TCP connection attempt will
 544        be retransmitted. Should not be higher than 255. Default value
 545        is 5, which corresponds to 31seconds till the last retransmission
 546        with the current initial RTO of 1second. With this the final timeout
 547        for a passive TCP connection will happen after 63seconds.
 548
 549tcp_syncookies - BOOLEAN
 550        Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
 551        Send out syncookies when the syn backlog queue of a socket
 552        overflows. This is to prevent against the common 'SYN flood attack'
 553        Default: 1
 554
 555        Note, that syncookies is fallback facility.
 556        It MUST NOT be used to help highly loaded servers to stand
 557        against legal connection rate. If you see SYN flood warnings
 558        in your logs, but investigation shows that they occur
 559        because of overload with legal connections, you should tune
 560        another parameters until this warning disappear.
 561        See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
 562
 563        syncookies seriously violate TCP protocol, do not allow
 564        to use TCP extensions, can result in serious degradation
 565        of some services (f.e. SMTP relaying), visible not by you,
 566        but your clients and relays, contacting you. While you see
 567        SYN flood warnings in logs not being really flooded, your server
 568        is seriously misconfigured.
 569
 570        If you want to test which effects syncookies have to your
 571        network connections you can set this knob to 2 to enable
 572        unconditionally generation of syncookies.
 573
 574tcp_fastopen - INTEGER
 575        Enable TCP Fast Open (RFC7413) to send and accept data in the opening
 576        SYN packet.
 577
 578        The client support is enabled by flag 0x1 (on by default). The client
 579        then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
 580        rather than connect() to send data in SYN.
 581
 582        The server support is enabled by flag 0x2 (off by default). Then
 583        either enable for all listeners with another flag (0x400) or
 584        enable individual listeners via TCP_FASTOPEN socket option with
 585        the option value being the length of the syn-data backlog.
 586
 587        The values (bitmap) are
 588          0x1: (client) enables sending data in the opening SYN on the client.
 589          0x2: (server) enables the server support, i.e., allowing data in
 590                        a SYN packet to be accepted and passed to the
 591                        application before 3-way handshake finishes.
 592          0x4: (client) send data in the opening SYN regardless of cookie
 593                        availability and without a cookie option.
 594        0x200: (server) accept data-in-SYN w/o any cookie option present.
 595        0x400: (server) enable all listeners to support Fast Open by
 596                        default without explicit TCP_FASTOPEN socket option.
 597
 598        Default: 0x1
 599
 600        Note that that additional client or server features are only
 601        effective if the basic support (0x1 and 0x2) are enabled respectively.
 602
 603tcp_fastopen_blackhole_timeout_sec - INTEGER
 604        Initial time period in second to disable Fastopen on active TCP sockets
 605        when a TFO firewall blackhole issue happens.
 606        This time period will grow exponentially when more blackhole issues
 607        get detected right after Fastopen is re-enabled and will reset to
 608        initial value when the blackhole issue goes away.
 609        By default, it is set to 1hr.
 610
 611tcp_syn_retries - INTEGER
 612        Number of times initial SYNs for an active TCP connection attempt
 613        will be retransmitted. Should not be higher than 127. Default value
 614        is 6, which corresponds to 63seconds till the last retransmission
 615        with the current initial RTO of 1second. With this the final timeout
 616        for an active TCP connection attempt will happen after 127seconds.
 617
 618tcp_timestamps - INTEGER
 619Enable timestamps as defined in RFC1323.
 620        0: Disabled.
 621        1: Enable timestamps as defined in RFC1323 and use random offset for
 622        each connection rather than only using the current time.
 623        2: Like 1, but without random offsets.
 624        Default: 1
 625
 626tcp_min_tso_segs - INTEGER
 627        Minimal number of segments per TSO frame.
 628        Since linux-3.12, TCP does an automatic sizing of TSO frames,
 629        depending on flow rate, instead of filling 64Kbytes packets.
 630        For specific usages, it's possible to force TCP to build big
 631        TSO frames. Note that TCP stack might split too big TSO packets
 632        if available window is too small.
 633        Default: 2
 634
 635tcp_pacing_ss_ratio - INTEGER
 636        sk->sk_pacing_rate is set by TCP stack using a ratio applied
 637        to current rate. (current_rate = cwnd * mss / srtt)
 638        If TCP is in slow start, tcp_pacing_ss_ratio is applied
 639        to let TCP probe for bigger speeds, assuming cwnd can be
 640        doubled every other RTT.
 641        Default: 200
 642
 643tcp_pacing_ca_ratio - INTEGER
 644        sk->sk_pacing_rate is set by TCP stack using a ratio applied
 645        to current rate. (current_rate = cwnd * mss / srtt)
 646        If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
 647        is applied to conservatively probe for bigger throughput.
 648        Default: 120
 649
 650tcp_tso_win_divisor - INTEGER
 651        This allows control over what percentage of the congestion window
 652        can be consumed by a single TSO frame.
 653        The setting of this parameter is a choice between burstiness and
 654        building larger TSO frames.
 655        Default: 3
 656
 657tcp_tw_reuse - BOOLEAN
 658        Allow to reuse TIME-WAIT sockets for new connections when it is
 659        safe from protocol viewpoint. Default value is 0.
 660        It should not be changed without advice/request of technical
 661        experts.
 662
 663tcp_window_scaling - BOOLEAN
 664        Enable window scaling as defined in RFC1323.
 665
 666tcp_wmem - vector of 3 INTEGERs: min, default, max
 667        min: Amount of memory reserved for send buffers for TCP sockets.
 668        Each TCP socket has rights to use it due to fact of its birth.
 669        Default: 4K
 670
 671        default: initial size of send buffer used by TCP sockets.  This
 672        value overrides net.core.wmem_default used by other protocols.
 673        It is usually lower than net.core.wmem_default.
 674        Default: 16K
 675
 676        max: Maximal amount of memory allowed for automatically tuned
 677        send buffers for TCP sockets. This value does not override
 678        net.core.wmem_max.  Calling setsockopt() with SO_SNDBUF disables
 679        automatic tuning of that socket's send buffer size, in which case
 680        this value is ignored.
 681        Default: between 64K and 4MB, depending on RAM size.
 682
 683tcp_notsent_lowat - UNSIGNED INTEGER
 684        A TCP socket can control the amount of unsent bytes in its write queue,
 685        thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
 686        reports POLLOUT events if the amount of unsent bytes is below a per
 687        socket value, and if the write queue is not full. sendmsg() will
 688        also not add new buffers if the limit is hit.
 689
 690        This global variable controls the amount of unsent data for
 691        sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
 692        to the global variable has immediate effect.
 693
 694        Default: UINT_MAX (0xFFFFFFFF)
 695
 696tcp_workaround_signed_windows - BOOLEAN
 697        If set, assume no receipt of a window scaling option means the
 698        remote TCP is broken and treats the window as a signed quantity.
 699        If unset, assume the remote TCP is not broken even if we do
 700        not receive a window scaling option from them.
 701        Default: 0
 702
 703tcp_thin_linear_timeouts - BOOLEAN
 704        Enable dynamic triggering of linear timeouts for thin streams.
 705        If set, a check is performed upon retransmission by timeout to
 706        determine if the stream is thin (less than 4 packets in flight).
 707        As long as the stream is found to be thin, up to 6 linear
 708        timeouts may be performed before exponential backoff mode is
 709        initiated. This improves retransmission latency for
 710        non-aggressive thin streams, often found to be time-dependent.
 711        For more information on thin streams, see
 712        Documentation/networking/tcp-thin.txt
 713        Default: 0
 714
 715tcp_limit_output_bytes - INTEGER
 716        Controls TCP Small Queue limit per tcp socket.
 717        TCP bulk sender tends to increase packets in flight until it
 718        gets losses notifications. With SNDBUF autotuning, this can
 719        result in a large amount of packets queued in qdisc/device
 720        on the local machine, hurting latency of other flows, for
 721        typical pfifo_fast qdiscs.
 722        tcp_limit_output_bytes limits the number of bytes on qdisc
 723        or device to reduce artificial RTT/cwnd and reduce bufferbloat.
 724        Default: 262144
 725
 726tcp_challenge_ack_limit - INTEGER
 727        Limits number of Challenge ACK sent per second, as recommended
 728        in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
 729        Default: 100
 730
 731UDP variables:
 732
 733udp_l3mdev_accept - BOOLEAN
 734        Enabling this option allows a "global" bound socket to work
 735        across L3 master domains (e.g., VRFs) with packets capable of
 736        being received regardless of the L3 domain in which they
 737        originated. Only valid when the kernel was compiled with
 738        CONFIG_NET_L3_MASTER_DEV.
 739
 740udp_mem - vector of 3 INTEGERs: min, pressure, max
 741        Number of pages allowed for queueing by all UDP sockets.
 742
 743        min: Below this number of pages UDP is not bothered about its
 744        memory appetite. When amount of memory allocated by UDP exceeds
 745        this number, UDP starts to moderate memory usage.
 746
 747        pressure: This value was introduced to follow format of tcp_mem.
 748
 749        max: Number of pages allowed for queueing by all UDP sockets.
 750
 751        Default is calculated at boot time from amount of available memory.
 752
 753udp_rmem_min - INTEGER
 754        Minimal size of receive buffer used by UDP sockets in moderation.
 755        Each UDP socket is able to use the size for receiving data, even if
 756        total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
 757        Default: 1 page
 758
 759udp_wmem_min - INTEGER
 760        Minimal size of send buffer used by UDP sockets in moderation.
 761        Each UDP socket is able to use the size for sending data, even if
 762        total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
 763        Default: 1 page
 764
 765CIPSOv4 Variables:
 766
 767cipso_cache_enable - BOOLEAN
 768        If set, enable additions to and lookups from the CIPSO label mapping
 769        cache.  If unset, additions are ignored and lookups always result in a
 770        miss.  However, regardless of the setting the cache is still
 771        invalidated when required when means you can safely toggle this on and
 772        off and the cache will always be "safe".
 773        Default: 1
 774
 775cipso_cache_bucket_size - INTEGER
 776        The CIPSO label cache consists of a fixed size hash table with each
 777        hash bucket containing a number of cache entries.  This variable limits
 778        the number of entries in each hash bucket; the larger the value the
 779        more CIPSO label mappings that can be cached.  When the number of
 780        entries in a given hash bucket reaches this limit adding new entries
 781        causes the oldest entry in the bucket to be removed to make room.
 782        Default: 10
 783
 784cipso_rbm_optfmt - BOOLEAN
 785        Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
 786        the CIPSO draft specification (see Documentation/netlabel for details).
 787        This means that when set the CIPSO tag will be padded with empty
 788        categories in order to make the packet data 32-bit aligned.
 789        Default: 0
 790
 791cipso_rbm_structvalid - BOOLEAN
 792        If set, do a very strict check of the CIPSO option when
 793        ip_options_compile() is called.  If unset, relax the checks done during
 794        ip_options_compile().  Either way is "safe" as errors are caught else
 795        where in the CIPSO processing code but setting this to 0 (False) should
 796        result in less work (i.e. it should be faster) but could cause problems
 797        with other implementations that require strict checking.
 798        Default: 0
 799
 800IP Variables:
 801
 802ip_local_port_range - 2 INTEGERS
 803        Defines the local port range that is used by TCP and UDP to
 804        choose the local port. The first number is the first, the
 805        second the last local port number.
 806        If possible, it is better these numbers have different parity.
 807        (one even and one odd values)
 808        The default values are 32768 and 60999 respectively.
 809
 810ip_local_reserved_ports - list of comma separated ranges
 811        Specify the ports which are reserved for known third-party
 812        applications. These ports will not be used by automatic port
 813        assignments (e.g. when calling connect() or bind() with port
 814        number 0). Explicit port allocation behavior is unchanged.
 815
 816        The format used for both input and output is a comma separated
 817        list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
 818        10). Writing to the file will clear all previously reserved
 819        ports and update the current list with the one given in the
 820        input.
 821
 822        Note that ip_local_port_range and ip_local_reserved_ports
 823        settings are independent and both are considered by the kernel
 824        when determining which ports are available for automatic port
 825        assignments.
 826
 827        You can reserve ports which are not in the current
 828        ip_local_port_range, e.g.:
 829
 830        $ cat /proc/sys/net/ipv4/ip_local_port_range
 831        32000   60999
 832        $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
 833        8080,9148
 834
 835        although this is redundant. However such a setting is useful
 836        if later the port range is changed to a value that will
 837        include the reserved ports.
 838
 839        Default: Empty
 840
 841ip_unprivileged_port_start - INTEGER
 842        This is a per-namespace sysctl.  It defines the first
 843        unprivileged port in the network namespace.  Privileged ports
 844        require root or CAP_NET_BIND_SERVICE in order to bind to them.
 845        To disable all privileged ports, set this to 0.  It may not
 846        overlap with the ip_local_reserved_ports range.
 847
 848        Default: 1024
 849
 850ip_nonlocal_bind - BOOLEAN
 851        If set, allows processes to bind() to non-local IP addresses,
 852        which can be quite useful - but may break some applications.
 853        Default: 0
 854
 855ip_dynaddr - BOOLEAN
 856        If set non-zero, enables support for dynamic addresses.
 857        If set to a non-zero value larger than 1, a kernel log
 858        message will be printed when dynamic address rewriting
 859        occurs.
 860        Default: 0
 861
 862ip_early_demux - BOOLEAN
 863        Optimize input packet processing down to one demux for
 864        certain kinds of local sockets.  Currently we only do this
 865        for established TCP and connected UDP sockets.
 866
 867        It may add an additional cost for pure routing workloads that
 868        reduces overall throughput, in such case you should disable it.
 869        Default: 1
 870
 871tcp_early_demux - BOOLEAN
 872        Enable early demux for established TCP sockets.
 873        Default: 1
 874
 875udp_early_demux - BOOLEAN
 876        Enable early demux for connected UDP sockets. Disable this if
 877        your system could experience more unconnected load.
 878        Default: 1
 879
 880icmp_echo_ignore_all - BOOLEAN
 881        If set non-zero, then the kernel will ignore all ICMP ECHO
 882        requests sent to it.
 883        Default: 0
 884
 885icmp_echo_ignore_broadcasts - BOOLEAN
 886        If set non-zero, then the kernel will ignore all ICMP ECHO and
 887        TIMESTAMP requests sent to it via broadcast/multicast.
 888        Default: 1
 889
 890icmp_ratelimit - INTEGER
 891        Limit the maximal rates for sending ICMP packets whose type matches
 892        icmp_ratemask (see below) to specific targets.
 893        0 to disable any limiting,
 894        otherwise the minimal space between responses in milliseconds.
 895        Note that another sysctl, icmp_msgs_per_sec limits the number
 896        of ICMP packets sent on all targets.
 897        Default: 1000
 898
 899icmp_msgs_per_sec - INTEGER
 900        Limit maximal number of ICMP packets sent per second from this host.
 901        Only messages whose type matches icmp_ratemask (see below) are
 902        controlled by this limit.
 903        Default: 1000
 904
 905icmp_msgs_burst - INTEGER
 906        icmp_msgs_per_sec controls number of ICMP packets sent per second,
 907        while icmp_msgs_burst controls the burst size of these packets.
 908        Default: 50
 909
 910icmp_ratemask - INTEGER
 911        Mask made of ICMP types for which rates are being limited.
 912        Significant bits: IHGFEDCBA9876543210
 913        Default mask:     0000001100000011000 (6168)
 914
 915        Bit definitions (see include/linux/icmp.h):
 916                0 Echo Reply
 917                3 Destination Unreachable *
 918                4 Source Quench *
 919                5 Redirect
 920                8 Echo Request
 921                B Time Exceeded *
 922                C Parameter Problem *
 923                D Timestamp Request
 924                E Timestamp Reply
 925                F Info Request
 926                G Info Reply
 927                H Address Mask Request
 928                I Address Mask Reply
 929
 930        * These are rate limited by default (see default mask above)
 931
 932icmp_ignore_bogus_error_responses - BOOLEAN
 933        Some routers violate RFC1122 by sending bogus responses to broadcast
 934        frames.  Such violations are normally logged via a kernel warning.
 935        If this is set to TRUE, the kernel will not give such warnings, which
 936        will avoid log file clutter.
 937        Default: 1
 938
 939icmp_errors_use_inbound_ifaddr - BOOLEAN
 940
 941        If zero, icmp error messages are sent with the primary address of
 942        the exiting interface.
 943
 944        If non-zero, the message will be sent with the primary address of
 945        the interface that received the packet that caused the icmp error.
 946        This is the behaviour network many administrators will expect from
 947        a router. And it can make debugging complicated network layouts
 948        much easier.
 949
 950        Note that if no primary address exists for the interface selected,
 951        then the primary address of the first non-loopback interface that
 952        has one will be used regardless of this setting.
 953
 954        Default: 0
 955
 956igmp_max_memberships - INTEGER
 957        Change the maximum number of multicast groups we can subscribe to.
 958        Default: 20
 959
 960        Theoretical maximum value is bounded by having to send a membership
 961        report in a single datagram (i.e. the report can't span multiple
 962        datagrams, or risk confusing the switch and leaving groups you don't
 963        intend to).
 964
 965        The number of supported groups 'M' is bounded by the number of group
 966        report entries you can fit into a single datagram of 65535 bytes.
 967
 968        M = 65536-sizeof (ip header)/(sizeof(Group record))
 969
 970        Group records are variable length, with a minimum of 12 bytes.
 971        So net.ipv4.igmp_max_memberships should not be set higher than:
 972
 973        (65536-24) / 12 = 5459
 974
 975        The value 5459 assumes no IP header options, so in practice
 976        this number may be lower.
 977
 978igmp_max_msf - INTEGER
 979        Maximum number of addresses allowed in the source filter list for a
 980        multicast group.
 981        Default: 10
 982
 983igmp_qrv - INTEGER
 984        Controls the IGMP query robustness variable (see RFC2236 8.1).
 985        Default: 2 (as specified by RFC2236 8.1)
 986        Minimum: 1 (as specified by RFC6636 4.5)
 987
 988force_igmp_version - INTEGER
 989        0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
 990            allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
 991            Present timer expires.
 992        1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
 993            receive IGMPv2/v3 query.
 994        2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
 995            IGMPv1 query message. Will reply report if receive IGMPv3 query.
 996        3 - Enforce to use IGMP version 3. The same react with default 0.
 997
 998        Note: this is not the same with force_mld_version because IGMPv3 RFC3376
 999        Security Considerations does not have clear description that we could
1000        ignore other version messages completely as MLDv2 RFC3810. So make
1001        this value as default 0 is recommended.
1002
1003conf/interface/*  changes special settings per interface (where
1004"interface" is the name of your network interface)
1005
1006conf/all/*        is special, changes the settings for all interfaces
1007
1008log_martians - BOOLEAN
1009        Log packets with impossible addresses to kernel log.
1010        log_martians for the interface will be enabled if at least one of
1011        conf/{all,interface}/log_martians is set to TRUE,
1012        it will be disabled otherwise
1013
1014accept_redirects - BOOLEAN
1015        Accept ICMP redirect messages.
1016        accept_redirects for the interface will be enabled if:
1017        - both conf/{all,interface}/accept_redirects are TRUE in the case
1018          forwarding for the interface is enabled
1019        or
1020        - at least one of conf/{all,interface}/accept_redirects is TRUE in the
1021          case forwarding for the interface is disabled
1022        accept_redirects for the interface will be disabled otherwise
1023        default TRUE (host)
1024                FALSE (router)
1025
1026forwarding - BOOLEAN
1027        Enable IP forwarding on this interface.  This controls whether packets
1028        received _on_ this interface can be forwarded.
1029
1030mc_forwarding - BOOLEAN
1031        Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1032        and a multicast routing daemon is required.
1033        conf/all/mc_forwarding must also be set to TRUE to enable multicast
1034        routing for the interface
1035
1036medium_id - INTEGER
1037        Integer value used to differentiate the devices by the medium they
1038        are attached to. Two devices can have different id values when
1039        the broadcast packets are received only on one of them.
1040        The default value 0 means that the device is the only interface
1041        to its medium, value of -1 means that medium is not known.
1042
1043        Currently, it is used to change the proxy_arp behavior:
1044        the proxy_arp feature is enabled for packets forwarded between
1045        two devices attached to different media.
1046
1047proxy_arp - BOOLEAN
1048        Do proxy arp.
1049        proxy_arp for the interface will be enabled if at least one of
1050        conf/{all,interface}/proxy_arp is set to TRUE,
1051        it will be disabled otherwise
1052
1053proxy_arp_pvlan - BOOLEAN
1054        Private VLAN proxy arp.
1055        Basically allow proxy arp replies back to the same interface
1056        (from which the ARP request/solicitation was received).
1057
1058        This is done to support (ethernet) switch features, like RFC
1059        3069, where the individual ports are NOT allowed to
1060        communicate with each other, but they are allowed to talk to
1061        the upstream router.  As described in RFC 3069, it is possible
1062        to allow these hosts to communicate through the upstream
1063        router by proxy_arp'ing. Don't need to be used together with
1064        proxy_arp.
1065
1066        This technology is known by different names:
1067          In RFC 3069 it is called VLAN Aggregation.
1068          Cisco and Allied Telesyn call it Private VLAN.
1069          Hewlett-Packard call it Source-Port filtering or port-isolation.
1070          Ericsson call it MAC-Forced Forwarding (RFC Draft).
1071
1072shared_media - BOOLEAN
1073        Send(router) or accept(host) RFC1620 shared media redirects.
1074        Overrides secure_redirects.
1075        shared_media for the interface will be enabled if at least one of
1076        conf/{all,interface}/shared_media is set to TRUE,
1077        it will be disabled otherwise
1078        default TRUE
1079
1080secure_redirects - BOOLEAN
1081        Accept ICMP redirect messages only to gateways listed in the
1082        interface's current gateway list. Even if disabled, RFC1122 redirect
1083        rules still apply.
1084        Overridden by shared_media.
1085        secure_redirects for the interface will be enabled if at least one of
1086        conf/{all,interface}/secure_redirects is set to TRUE,
1087        it will be disabled otherwise
1088        default TRUE
1089
1090send_redirects - BOOLEAN
1091        Send redirects, if router.
1092        send_redirects for the interface will be enabled if at least one of
1093        conf/{all,interface}/send_redirects is set to TRUE,
1094        it will be disabled otherwise
1095        Default: TRUE
1096
1097bootp_relay - BOOLEAN
1098        Accept packets with source address 0.b.c.d destined
1099        not to this host as local ones. It is supposed, that
1100        BOOTP relay daemon will catch and forward such packets.
1101        conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1102        for the interface
1103        default FALSE
1104        Not Implemented Yet.
1105
1106accept_source_route - BOOLEAN
1107        Accept packets with SRR option.
1108        conf/all/accept_source_route must also be set to TRUE to accept packets
1109        with SRR option on the interface
1110        default TRUE (router)
1111                FALSE (host)
1112
1113accept_local - BOOLEAN
1114        Accept packets with local source addresses. In combination with
1115        suitable routing, this can be used to direct packets between two
1116        local interfaces over the wire and have them accepted properly.
1117        default FALSE
1118
1119route_localnet - BOOLEAN
1120        Do not consider loopback addresses as martian source or destination
1121        while routing. This enables the use of 127/8 for local routing purposes.
1122        default FALSE
1123
1124rp_filter - INTEGER
1125        0 - No source validation.
1126        1 - Strict mode as defined in RFC3704 Strict Reverse Path
1127            Each incoming packet is tested against the FIB and if the interface
1128            is not the best reverse path the packet check will fail.
1129            By default failed packets are discarded.
1130        2 - Loose mode as defined in RFC3704 Loose Reverse Path
1131            Each incoming packet's source address is also tested against the FIB
1132            and if the source address is not reachable via any interface
1133            the packet check will fail.
1134
1135        Current recommended practice in RFC3704 is to enable strict mode
1136        to prevent IP spoofing from DDos attacks. If using asymmetric routing
1137        or other complicated routing, then loose mode is recommended.
1138
1139        The max value from conf/{all,interface}/rp_filter is used
1140        when doing source validation on the {interface}.
1141
1142        Default value is 0. Note that some distributions enable it
1143        in startup scripts.
1144
1145arp_filter - BOOLEAN
1146        1 - Allows you to have multiple network interfaces on the same
1147        subnet, and have the ARPs for each interface be answered
1148        based on whether or not the kernel would route a packet from
1149        the ARP'd IP out that interface (therefore you must use source
1150        based routing for this to work). In other words it allows control
1151        of which cards (usually 1) will respond to an arp request.
1152
1153        0 - (default) The kernel can respond to arp requests with addresses
1154        from other interfaces. This may seem wrong but it usually makes
1155        sense, because it increases the chance of successful communication.
1156        IP addresses are owned by the complete host on Linux, not by
1157        particular interfaces. Only for more complex setups like load-
1158        balancing, does this behaviour cause problems.
1159
1160        arp_filter for the interface will be enabled if at least one of
1161        conf/{all,interface}/arp_filter is set to TRUE,
1162        it will be disabled otherwise
1163
1164arp_announce - INTEGER
1165        Define different restriction levels for announcing the local
1166        source IP address from IP packets in ARP requests sent on
1167        interface:
1168        0 - (default) Use any local address, configured on any interface
1169        1 - Try to avoid local addresses that are not in the target's
1170        subnet for this interface. This mode is useful when target
1171        hosts reachable via this interface require the source IP
1172        address in ARP requests to be part of their logical network
1173        configured on the receiving interface. When we generate the
1174        request we will check all our subnets that include the
1175        target IP and will preserve the source address if it is from
1176        such subnet. If there is no such subnet we select source
1177        address according to the rules for level 2.
1178        2 - Always use the best local address for this target.
1179        In this mode we ignore the source address in the IP packet
1180        and try to select local address that we prefer for talks with
1181        the target host. Such local address is selected by looking
1182        for primary IP addresses on all our subnets on the outgoing
1183        interface that include the target IP address. If no suitable
1184        local address is found we select the first local address
1185        we have on the outgoing interface or on all other interfaces,
1186        with the hope we will receive reply for our request and
1187        even sometimes no matter the source IP address we announce.
1188
1189        The max value from conf/{all,interface}/arp_announce is used.
1190
1191        Increasing the restriction level gives more chance for
1192        receiving answer from the resolved target while decreasing
1193        the level announces more valid sender's information.
1194
1195arp_ignore - INTEGER
1196        Define different modes for sending replies in response to
1197        received ARP requests that resolve local target IP addresses:
1198        0 - (default): reply for any local target IP address, configured
1199        on any interface
1200        1 - reply only if the target IP address is local address
1201        configured on the incoming interface
1202        2 - reply only if the target IP address is local address
1203        configured on the incoming interface and both with the
1204        sender's IP address are part from same subnet on this interface
1205        3 - do not reply for local addresses configured with scope host,
1206        only resolutions for global and link addresses are replied
1207        4-7 - reserved
1208        8 - do not reply for all local addresses
1209
1210        The max value from conf/{all,interface}/arp_ignore is used
1211        when ARP request is received on the {interface}
1212
1213arp_notify - BOOLEAN
1214        Define mode for notification of address and device changes.
1215        0 - (default): do nothing
1216        1 - Generate gratuitous arp requests when device is brought up
1217            or hardware address changes.
1218
1219arp_accept - BOOLEAN
1220        Define behavior for gratuitous ARP frames who's IP is not
1221        already present in the ARP table:
1222        0 - don't create new entries in the ARP table
1223        1 - create new entries in the ARP table
1224
1225        Both replies and requests type gratuitous arp will trigger the
1226        ARP table to be updated, if this setting is on.
1227
1228        If the ARP table already contains the IP address of the
1229        gratuitous arp frame, the arp table will be updated regardless
1230        if this setting is on or off.
1231
1232mcast_solicit - INTEGER
1233        The maximum number of multicast probes in INCOMPLETE state,
1234        when the associated hardware address is unknown.  Defaults
1235        to 3.
1236
1237ucast_solicit - INTEGER
1238        The maximum number of unicast probes in PROBE state, when
1239        the hardware address is being reconfirmed.  Defaults to 3.
1240
1241app_solicit - INTEGER
1242        The maximum number of probes to send to the user space ARP daemon
1243        via netlink before dropping back to multicast probes (see
1244        mcast_resolicit).  Defaults to 0.
1245
1246mcast_resolicit - INTEGER
1247        The maximum number of multicast probes after unicast and
1248        app probes in PROBE state.  Defaults to 0.
1249
1250disable_policy - BOOLEAN
1251        Disable IPSEC policy (SPD) for this interface
1252
1253disable_xfrm - BOOLEAN
1254        Disable IPSEC encryption on this interface, whatever the policy
1255
1256igmpv2_unsolicited_report_interval - INTEGER
1257        The interval in milliseconds in which the next unsolicited
1258        IGMPv1 or IGMPv2 report retransmit will take place.
1259        Default: 10000 (10 seconds)
1260
1261igmpv3_unsolicited_report_interval - INTEGER
1262        The interval in milliseconds in which the next unsolicited
1263        IGMPv3 report retransmit will take place.
1264        Default: 1000 (1 seconds)
1265
1266promote_secondaries - BOOLEAN
1267        When a primary IP address is removed from this interface
1268        promote a corresponding secondary IP address instead of
1269        removing all the corresponding secondary IP addresses.
1270
1271drop_unicast_in_l2_multicast - BOOLEAN
1272        Drop any unicast IP packets that are received in link-layer
1273        multicast (or broadcast) frames.
1274        This behavior (for multicast) is actually a SHOULD in RFC
1275        1122, but is disabled by default for compatibility reasons.
1276        Default: off (0)
1277
1278drop_gratuitous_arp - BOOLEAN
1279        Drop all gratuitous ARP frames, for example if there's a known
1280        good ARP proxy on the network and such frames need not be used
1281        (or in the case of 802.11, must not be used to prevent attacks.)
1282        Default: off (0)
1283
1284
1285tag - INTEGER
1286        Allows you to write a number, which can be used as required.
1287        Default value is 0.
1288
1289xfrm4_gc_thresh - INTEGER
1290        The threshold at which we will start garbage collecting for IPv4
1291        destination cache entries.  At twice this value the system will
1292        refuse new allocations.
1293
1294igmp_link_local_mcast_reports - BOOLEAN
1295        Enable IGMP reports for link local multicast groups in the
1296        224.0.0.X range.
1297        Default TRUE
1298
1299Alexey Kuznetsov.
1300kuznet@ms2.inr.ac.ru
1301
1302Updated by:
1303Andi Kleen
1304ak@muc.de
1305Nicolas Delon
1306delon.nicolas@wanadoo.fr
1307
1308
1309
1310
1311/proc/sys/net/ipv6/* Variables:
1312
1313IPv6 has no global variables such as tcp_*.  tcp_* settings under ipv4/ also
1314apply to IPv6 [XXX?].
1315
1316bindv6only - BOOLEAN
1317        Default value for IPV6_V6ONLY socket option,
1318        which restricts use of the IPv6 socket to IPv6 communication
1319        only.
1320                TRUE: disable IPv4-mapped address feature
1321                FALSE: enable IPv4-mapped address feature
1322
1323        Default: FALSE (as specified in RFC3493)
1324
1325flowlabel_consistency - BOOLEAN
1326        Protect the consistency (and unicity) of flow label.
1327        You have to disable it to use IPV6_FL_F_REFLECT flag on the
1328        flow label manager.
1329        TRUE: enabled
1330        FALSE: disabled
1331        Default: TRUE
1332
1333auto_flowlabels - INTEGER
1334        Automatically generate flow labels based on a flow hash of the
1335        packet. This allows intermediate devices, such as routers, to
1336        identify packet flows for mechanisms like Equal Cost Multipath
1337        Routing (see RFC 6438).
1338        0: automatic flow labels are completely disabled
1339        1: automatic flow labels are enabled by default, they can be
1340           disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1341           socket option
1342        2: automatic flow labels are allowed, they may be enabled on a
1343           per socket basis using the IPV6_AUTOFLOWLABEL socket option
1344        3: automatic flow labels are enabled and enforced, they cannot
1345           be disabled by the socket option
1346        Default: 1
1347
1348flowlabel_state_ranges - BOOLEAN
1349        Split the flow label number space into two ranges. 0-0x7FFFF is
1350        reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1351        is reserved for stateless flow labels as described in RFC6437.
1352        TRUE: enabled
1353        FALSE: disabled
1354        Default: true
1355
1356flowlabel_reflect - BOOLEAN
1357        Automatically reflect the flow label. Needed for Path MTU
1358        Discovery to work with Equal Cost Multipath Routing in anycast
1359        environments. See RFC 7690 and:
1360        https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
1361        TRUE: enabled
1362        FALSE: disabled
1363        Default: FALSE
1364
1365anycast_src_echo_reply - BOOLEAN
1366        Controls the use of anycast addresses as source addresses for ICMPv6
1367        echo reply
1368        TRUE:  enabled
1369        FALSE: disabled
1370        Default: FALSE
1371
1372idgen_delay - INTEGER
1373        Controls the delay in seconds after which time to retry
1374        privacy stable address generation if a DAD conflict is
1375        detected.
1376        Default: 1 (as specified in RFC7217)
1377
1378idgen_retries - INTEGER
1379        Controls the number of retries to generate a stable privacy
1380        address if a DAD conflict is detected.
1381        Default: 3 (as specified in RFC7217)
1382
1383mld_qrv - INTEGER
1384        Controls the MLD query robustness variable (see RFC3810 9.1).
1385        Default: 2 (as specified by RFC3810 9.1)
1386        Minimum: 1 (as specified by RFC6636 4.5)
1387
1388max_dst_opts_cnt - INTEGER
1389        Maximum number of non-padding TLVs allowed in a Destination
1390        options extension header. If this value is less than zero
1391        then unknown options are disallowed and the number of known
1392        TLVs allowed is the absolute value of this number.
1393        Default: 8
1394
1395max_hbh_opts_cnt - INTEGER
1396        Maximum number of non-padding TLVs allowed in a Hop-by-Hop
1397        options extension header. If this value is less than zero
1398        then unknown options are disallowed and the number of known
1399        TLVs allowed is the absolute value of this number.
1400        Default: 8
1401
1402max dst_opts_len - INTEGER
1403        Maximum length allowed for a Destination options extension
1404        header.
1405        Default: INT_MAX (unlimited)
1406
1407max hbh_opts_len - INTEGER
1408        Maximum length allowed for a Hop-by-Hop options extension
1409        header.
1410        Default: INT_MAX (unlimited)
1411
1412IPv6 Fragmentation:
1413
1414ip6frag_high_thresh - INTEGER
1415        Maximum memory used to reassemble IPv6 fragments. When
1416        ip6frag_high_thresh bytes of memory is allocated for this purpose,
1417        the fragment handler will toss packets until ip6frag_low_thresh
1418        is reached.
1419
1420ip6frag_low_thresh - INTEGER
1421        See ip6frag_high_thresh
1422
1423ip6frag_time - INTEGER
1424        Time in seconds to keep an IPv6 fragment in memory.
1425
1426conf/default/*:
1427        Change the interface-specific default settings.
1428
1429
1430conf/all/*:
1431        Change all the interface-specific settings.
1432
1433        [XXX:  Other special features than forwarding?]
1434
1435conf/all/forwarding - BOOLEAN
1436        Enable global IPv6 forwarding between all interfaces.
1437
1438        IPv4 and IPv6 work differently here; e.g. netfilter must be used
1439        to control which interfaces may forward packets and which not.
1440
1441        This also sets all interfaces' Host/Router setting
1442        'forwarding' to the specified value.  See below for details.
1443
1444        This referred to as global forwarding.
1445
1446proxy_ndp - BOOLEAN
1447        Do proxy ndp.
1448
1449fwmark_reflect - BOOLEAN
1450        Controls the fwmark of kernel-generated IPv6 reply packets that are not
1451        associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
1452        If unset, these packets have a fwmark of zero. If set, they have the
1453        fwmark of the packet they are replying to.
1454        Default: 0
1455
1456conf/interface/*:
1457        Change special settings per interface.
1458
1459        The functional behaviour for certain settings is different
1460        depending on whether local forwarding is enabled or not.
1461
1462accept_ra - INTEGER
1463        Accept Router Advertisements; autoconfigure using them.
1464
1465        It also determines whether or not to transmit Router
1466        Solicitations. If and only if the functional setting is to
1467        accept Router Advertisements, Router Solicitations will be
1468        transmitted.
1469
1470        Possible values are:
1471                0 Do not accept Router Advertisements.
1472                1 Accept Router Advertisements if forwarding is disabled.
1473                2 Overrule forwarding behaviour. Accept Router Advertisements
1474                  even if forwarding is enabled.
1475
1476        Functional default: enabled if local forwarding is disabled.
1477                            disabled if local forwarding is enabled.
1478
1479accept_ra_defrtr - BOOLEAN
1480        Learn default router in Router Advertisement.
1481
1482        Functional default: enabled if accept_ra is enabled.
1483                            disabled if accept_ra is disabled.
1484
1485accept_ra_from_local - BOOLEAN
1486        Accept RA with source-address that is found on local machine
1487        if the RA is otherwise proper and able to be accepted.
1488        Default is to NOT accept these as it may be an un-intended
1489        network loop.
1490
1491        Functional default:
1492           enabled if accept_ra_from_local is enabled
1493               on a specific interface.
1494           disabled if accept_ra_from_local is disabled
1495               on a specific interface.
1496
1497accept_ra_min_hop_limit - INTEGER
1498        Minimum hop limit Information in Router Advertisement.
1499
1500        Hop limit Information in Router Advertisement less than this
1501        variable shall be ignored.
1502
1503        Default: 1
1504
1505accept_ra_pinfo - BOOLEAN
1506        Learn Prefix Information in Router Advertisement.
1507
1508        Functional default: enabled if accept_ra is enabled.
1509                            disabled if accept_ra is disabled.
1510
1511accept_ra_rt_info_min_plen - INTEGER
1512        Minimum prefix length of Route Information in RA.
1513
1514        Route Information w/ prefix smaller than this variable shall
1515        be ignored.
1516
1517        Functional default: 0 if accept_ra_rtr_pref is enabled.
1518                            -1 if accept_ra_rtr_pref is disabled.
1519
1520accept_ra_rt_info_max_plen - INTEGER
1521        Maximum prefix length of Route Information in RA.
1522
1523        Route Information w/ prefix larger than this variable shall
1524        be ignored.
1525
1526        Functional default: 0 if accept_ra_rtr_pref is enabled.
1527                            -1 if accept_ra_rtr_pref is disabled.
1528
1529accept_ra_rtr_pref - BOOLEAN
1530        Accept Router Preference in RA.
1531
1532        Functional default: enabled if accept_ra is enabled.
1533                            disabled if accept_ra is disabled.
1534
1535accept_ra_mtu - BOOLEAN
1536        Apply the MTU value specified in RA option 5 (RFC4861). If
1537        disabled, the MTU specified in the RA will be ignored.
1538
1539        Functional default: enabled if accept_ra is enabled.
1540                            disabled if accept_ra is disabled.
1541
1542accept_redirects - BOOLEAN
1543        Accept Redirects.
1544
1545        Functional default: enabled if local forwarding is disabled.
1546                            disabled if local forwarding is enabled.
1547
1548accept_source_route - INTEGER
1549        Accept source routing (routing extension header).
1550
1551        >= 0: Accept only routing header type 2.
1552        < 0: Do not accept routing header.
1553
1554        Default: 0
1555
1556autoconf - BOOLEAN
1557        Autoconfigure addresses using Prefix Information in Router
1558        Advertisements.
1559
1560        Functional default: enabled if accept_ra_pinfo is enabled.
1561                            disabled if accept_ra_pinfo is disabled.
1562
1563dad_transmits - INTEGER
1564        The amount of Duplicate Address Detection probes to send.
1565        Default: 1
1566
1567forwarding - INTEGER
1568        Configure interface-specific Host/Router behaviour.
1569
1570        Note: It is recommended to have the same setting on all
1571        interfaces; mixed router/host scenarios are rather uncommon.
1572
1573        Possible values are:
1574                0 Forwarding disabled
1575                1 Forwarding enabled
1576
1577        FALSE (0):
1578
1579        By default, Host behaviour is assumed.  This means:
1580
1581        1. IsRouter flag is not set in Neighbour Advertisements.
1582        2. If accept_ra is TRUE (default), transmit Router
1583           Solicitations.
1584        3. If accept_ra is TRUE (default), accept Router
1585           Advertisements (and do autoconfiguration).
1586        4. If accept_redirects is TRUE (default), accept Redirects.
1587
1588        TRUE (1):
1589
1590        If local forwarding is enabled, Router behaviour is assumed.
1591        This means exactly the reverse from the above:
1592
1593        1. IsRouter flag is set in Neighbour Advertisements.
1594        2. Router Solicitations are not sent unless accept_ra is 2.
1595        3. Router Advertisements are ignored unless accept_ra is 2.
1596        4. Redirects are ignored.
1597
1598        Default: 0 (disabled) if global forwarding is disabled (default),
1599                 otherwise 1 (enabled).
1600
1601hop_limit - INTEGER
1602        Default Hop Limit to set.
1603        Default: 64
1604
1605mtu - INTEGER
1606        Default Maximum Transfer Unit
1607        Default: 1280 (IPv6 required minimum)
1608
1609ip_nonlocal_bind - BOOLEAN
1610        If set, allows processes to bind() to non-local IPv6 addresses,
1611        which can be quite useful - but may break some applications.
1612        Default: 0
1613
1614router_probe_interval - INTEGER
1615        Minimum interval (in seconds) between Router Probing described
1616        in RFC4191.
1617
1618        Default: 60
1619
1620router_solicitation_delay - INTEGER
1621        Number of seconds to wait after interface is brought up
1622        before sending Router Solicitations.
1623        Default: 1
1624
1625router_solicitation_interval - INTEGER
1626        Number of seconds to wait between Router Solicitations.
1627        Default: 4
1628
1629router_solicitations - INTEGER
1630        Number of Router Solicitations to send until assuming no
1631        routers are present.
1632        Default: 3
1633
1634use_oif_addrs_only - BOOLEAN
1635        When enabled, the candidate source addresses for destinations
1636        routed via this interface are restricted to the set of addresses
1637        configured on this interface (vis. RFC 6724, section 4).
1638
1639        Default: false
1640
1641use_tempaddr - INTEGER
1642        Preference for Privacy Extensions (RFC3041).
1643          <= 0 : disable Privacy Extensions
1644          == 1 : enable Privacy Extensions, but prefer public
1645                 addresses over temporary addresses.
1646          >  1 : enable Privacy Extensions and prefer temporary
1647                 addresses over public addresses.
1648        Default:  0 (for most devices)
1649                 -1 (for point-to-point devices and loopback devices)
1650
1651temp_valid_lft - INTEGER
1652        valid lifetime (in seconds) for temporary addresses.
1653        Default: 604800 (7 days)
1654
1655temp_prefered_lft - INTEGER
1656        Preferred lifetime (in seconds) for temporary addresses.
1657        Default: 86400 (1 day)
1658
1659keep_addr_on_down - INTEGER
1660        Keep all IPv6 addresses on an interface down event. If set static
1661        global addresses with no expiration time are not flushed.
1662          >0 : enabled
1663           0 : system default
1664          <0 : disabled
1665
1666        Default: 0 (addresses are removed)
1667
1668max_desync_factor - INTEGER
1669        Maximum value for DESYNC_FACTOR, which is a random value
1670        that ensures that clients don't synchronize with each
1671        other and generate new addresses at exactly the same time.
1672        value is in seconds.
1673        Default: 600
1674
1675regen_max_retry - INTEGER
1676        Number of attempts before give up attempting to generate
1677        valid temporary addresses.
1678        Default: 5
1679
1680max_addresses - INTEGER
1681        Maximum number of autoconfigured addresses per interface.  Setting
1682        to zero disables the limitation.  It is not recommended to set this
1683        value too large (or to zero) because it would be an easy way to
1684        crash the kernel by allowing too many addresses to be created.
1685        Default: 16
1686
1687disable_ipv6 - BOOLEAN
1688        Disable IPv6 operation.  If accept_dad is set to 2, this value
1689        will be dynamically set to TRUE if DAD fails for the link-local
1690        address.
1691        Default: FALSE (enable IPv6 operation)
1692
1693        When this value is changed from 1 to 0 (IPv6 is being enabled),
1694        it will dynamically create a link-local address on the given
1695        interface and start Duplicate Address Detection, if necessary.
1696
1697        When this value is changed from 0 to 1 (IPv6 is being disabled),
1698        it will dynamically delete all address on the given interface.
1699
1700accept_dad - INTEGER
1701        Whether to accept DAD (Duplicate Address Detection).
1702        0: Disable DAD
1703        1: Enable DAD (default)
1704        2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1705           link-local address has been found.
1706
1707        DAD operation and mode on a given interface will be selected according
1708        to the maximum value of conf/{all,interface}/accept_dad.
1709
1710force_tllao - BOOLEAN
1711        Enable sending the target link-layer address option even when
1712        responding to a unicast neighbor solicitation.
1713        Default: FALSE
1714
1715        Quoting from RFC 2461, section 4.4, Target link-layer address:
1716
1717        "The option MUST be included for multicast solicitations in order to
1718        avoid infinite Neighbor Solicitation "recursion" when the peer node
1719        does not have a cache entry to return a Neighbor Advertisements
1720        message.  When responding to unicast solicitations, the option can be
1721        omitted since the sender of the solicitation has the correct link-
1722        layer address; otherwise it would not have be able to send the unicast
1723        solicitation in the first place. However, including the link-layer
1724        address in this case adds little overhead and eliminates a potential
1725        race condition where the sender deletes the cached link-layer address
1726        prior to receiving a response to a previous solicitation."
1727
1728ndisc_notify - BOOLEAN
1729        Define mode for notification of address and device changes.
1730        0 - (default): do nothing
1731        1 - Generate unsolicited neighbour advertisements when device is brought
1732            up or hardware address changes.
1733
1734ndisc_tclass - INTEGER
1735        The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
1736        Discovery (Router Solicitation, Router Advertisement, Neighbor
1737        Solicitation, Neighbor Advertisement, Redirect) messages.
1738        These 8 bits can be interpreted as 6 high order bits holding the DSCP
1739        value and 2 low order bits representing ECN (which you probably want
1740        to leave cleared).
1741        0 - (default)
1742
1743mldv1_unsolicited_report_interval - INTEGER
1744        The interval in milliseconds in which the next unsolicited
1745        MLDv1 report retransmit will take place.
1746        Default: 10000 (10 seconds)
1747
1748mldv2_unsolicited_report_interval - INTEGER
1749        The interval in milliseconds in which the next unsolicited
1750        MLDv2 report retransmit will take place.
1751        Default: 1000 (1 second)
1752
1753force_mld_version - INTEGER
1754        0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1755        1 - Enforce to use MLD version 1
1756        2 - Enforce to use MLD version 2
1757
1758suppress_frag_ndisc - INTEGER
1759        Control RFC 6980 (Security Implications of IPv6 Fragmentation
1760        with IPv6 Neighbor Discovery) behavior:
1761        1 - (default) discard fragmented neighbor discovery packets
1762        0 - allow fragmented neighbor discovery packets
1763
1764optimistic_dad - BOOLEAN
1765        Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
1766        0: disabled (default)
1767        1: enabled
1768
1769        Optimistic Duplicate Address Detection for the interface will be enabled
1770        if at least one of conf/{all,interface}/optimistic_dad is set to 1,
1771        it will be disabled otherwise.
1772
1773use_optimistic - BOOLEAN
1774        If enabled, do not classify optimistic addresses as deprecated during
1775        source address selection.  Preferred addresses will still be chosen
1776        before optimistic addresses, subject to other ranking in the source
1777        address selection algorithm.
1778        0: disabled (default)
1779        1: enabled
1780
1781        This will be enabled if at least one of
1782        conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
1783
1784stable_secret - IPv6 address
1785        This IPv6 address will be used as a secret to generate IPv6
1786        addresses for link-local addresses and autoconfigured
1787        ones. All addresses generated after setting this secret will
1788        be stable privacy ones by default. This can be changed via the
1789        addrgenmode ip-link. conf/default/stable_secret is used as the
1790        secret for the namespace, the interface specific ones can
1791        overwrite that. Writes to conf/all/stable_secret are refused.
1792
1793        It is recommended to generate this secret during installation
1794        of a system and keep it stable after that.
1795
1796        By default the stable secret is unset.
1797
1798drop_unicast_in_l2_multicast - BOOLEAN
1799        Drop any unicast IPv6 packets that are received in link-layer
1800        multicast (or broadcast) frames.
1801
1802        By default this is turned off.
1803
1804drop_unsolicited_na - BOOLEAN
1805        Drop all unsolicited neighbor advertisements, for example if there's
1806        a known good NA proxy on the network and such frames need not be used
1807        (or in the case of 802.11, must not be used to prevent attacks.)
1808
1809        By default this is turned off.
1810
1811enhanced_dad - BOOLEAN
1812        Include a nonce option in the IPv6 neighbor solicitation messages used for
1813        duplicate address detection per RFC7527. A received DAD NS will only signal
1814        a duplicate address if the nonce is different. This avoids any false
1815        detection of duplicates due to loopback of the NS messages that we send.
1816        The nonce option will be sent on an interface unless both of
1817        conf/{all,interface}/enhanced_dad are set to FALSE.
1818        Default: TRUE
1819
1820icmp/*:
1821ratelimit - INTEGER
1822        Limit the maximal rates for sending ICMPv6 packets.
1823        0 to disable any limiting,
1824        otherwise the minimal space between responses in milliseconds.
1825        Default: 1000
1826
1827xfrm6_gc_thresh - INTEGER
1828        The threshold at which we will start garbage collecting for IPv6
1829        destination cache entries.  At twice this value the system will
1830        refuse new allocations.
1831
1832
1833IPv6 Update by:
1834Pekka Savola <pekkas@netcore.fi>
1835YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1836
1837
1838/proc/sys/net/bridge/* Variables:
1839
1840bridge-nf-call-arptables - BOOLEAN
1841        1 : pass bridged ARP traffic to arptables' FORWARD chain.
1842        0 : disable this.
1843        Default: 1
1844
1845bridge-nf-call-iptables - BOOLEAN
1846        1 : pass bridged IPv4 traffic to iptables' chains.
1847        0 : disable this.
1848        Default: 1
1849
1850bridge-nf-call-ip6tables - BOOLEAN
1851        1 : pass bridged IPv6 traffic to ip6tables' chains.
1852        0 : disable this.
1853        Default: 1
1854
1855bridge-nf-filter-vlan-tagged - BOOLEAN
1856        1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1857        0 : disable this.
1858        Default: 0
1859
1860bridge-nf-filter-pppoe-tagged - BOOLEAN
1861        1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1862        0 : disable this.
1863        Default: 0
1864
1865bridge-nf-pass-vlan-input-dev - BOOLEAN
1866        1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1867        interface on the bridge and set the netfilter input device to the vlan.
1868        This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1869        target work with vlan-on-top-of-bridge interfaces.  When no matching
1870        vlan interface is found, or this switch is off, the input device is
1871        set to the bridge interface.
1872        0: disable bridge netfilter vlan interface lookup.
1873        Default: 0
1874
1875proc/sys/net/sctp/* Variables:
1876
1877addip_enable - BOOLEAN
1878        Enable or disable extension of  Dynamic Address Reconfiguration
1879        (ADD-IP) functionality specified in RFC5061.  This extension provides
1880        the ability to dynamically add and remove new addresses for the SCTP
1881        associations.
1882
1883        1: Enable extension.
1884
1885        0: Disable extension.
1886
1887        Default: 0
1888
1889pf_enable - INTEGER
1890        Enable or disable pf (pf is short for potentially failed) state. A value
1891        of pf_retrans > path_max_retrans also disables pf state. That is, one of
1892        both pf_enable and pf_retrans > path_max_retrans can disable pf state.
1893        Since pf_retrans and path_max_retrans can be changed by userspace
1894        application, sometimes user expects to disable pf state by the value of
1895        pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
1896        or path_max_retrans is changed by the user application, this pf state is
1897        enabled. As such, it is necessary to add this to dynamically enable
1898        and disable pf state. See:
1899        https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
1900        details.
1901
1902        1: Enable pf.
1903
1904        0: Disable pf.
1905
1906        Default: 1
1907
1908addip_noauth_enable - BOOLEAN
1909        Dynamic Address Reconfiguration (ADD-IP) requires the use of
1910        authentication to protect the operations of adding or removing new
1911        addresses.  This requirement is mandated so that unauthorized hosts
1912        would not be able to hijack associations.  However, older
1913        implementations may not have implemented this requirement while
1914        allowing the ADD-IP extension.  For reasons of interoperability,
1915        we provide this variable to control the enforcement of the
1916        authentication requirement.
1917
1918        1: Allow ADD-IP extension to be used without authentication.  This
1919           should only be set in a closed environment for interoperability
1920           with older implementations.
1921
1922        0: Enforce the authentication requirement
1923
1924        Default: 0
1925
1926auth_enable - BOOLEAN
1927        Enable or disable Authenticated Chunks extension.  This extension
1928        provides the ability to send and receive authenticated chunks and is
1929        required for secure operation of Dynamic Address Reconfiguration
1930        (ADD-IP) extension.
1931
1932        1: Enable this extension.
1933        0: Disable this extension.
1934
1935        Default: 0
1936
1937prsctp_enable - BOOLEAN
1938        Enable or disable the Partial Reliability extension (RFC3758) which
1939        is used to notify peers that a given DATA should no longer be expected.
1940
1941        1: Enable extension
1942        0: Disable
1943
1944        Default: 1
1945
1946max_burst - INTEGER
1947        The limit of the number of new packets that can be initially sent.  It
1948        controls how bursty the generated traffic can be.
1949
1950        Default: 4
1951
1952association_max_retrans - INTEGER
1953        Set the maximum number for retransmissions that an association can
1954        attempt deciding that the remote end is unreachable.  If this value
1955        is exceeded, the association is terminated.
1956
1957        Default: 10
1958
1959max_init_retransmits - INTEGER
1960        The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1961        that an association will attempt before declaring the destination
1962        unreachable and terminating.
1963
1964        Default: 8
1965
1966path_max_retrans - INTEGER
1967        The maximum number of retransmissions that will be attempted on a given
1968        path.  Once this threshold is exceeded, the path is considered
1969        unreachable, and new traffic will use a different path when the
1970        association is multihomed.
1971
1972        Default: 5
1973
1974pf_retrans - INTEGER
1975        The number of retransmissions that will be attempted on a given path
1976        before traffic is redirected to an alternate transport (should one
1977        exist).  Note this is distinct from path_max_retrans, as a path that
1978        passes the pf_retrans threshold can still be used.  Its only
1979        deprioritized when a transmission path is selected by the stack.  This
1980        setting is primarily used to enable fast failover mechanisms without
1981        having to reduce path_max_retrans to a very low value.  See:
1982        http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1983        for details.  Note also that a value of pf_retrans > path_max_retrans
1984        disables this feature. Since both pf_retrans and path_max_retrans can
1985        be changed by userspace application, a variable pf_enable is used to
1986        disable pf state.
1987
1988        Default: 0
1989
1990rto_initial - INTEGER
1991        The initial round trip timeout value in milliseconds that will be used
1992        in calculating round trip times.  This is the initial time interval
1993        for retransmissions.
1994
1995        Default: 3000
1996
1997rto_max - INTEGER
1998        The maximum value (in milliseconds) of the round trip timeout.  This
1999        is the largest time interval that can elapse between retransmissions.
2000
2001        Default: 60000
2002
2003rto_min - INTEGER
2004        The minimum value (in milliseconds) of the round trip timeout.  This
2005        is the smallest time interval the can elapse between retransmissions.
2006
2007        Default: 1000
2008
2009hb_interval - INTEGER
2010        The interval (in milliseconds) between HEARTBEAT chunks.  These chunks
2011        are sent at the specified interval on idle paths to probe the state of
2012        a given path between 2 associations.
2013
2014        Default: 30000
2015
2016sack_timeout - INTEGER
2017        The amount of time (in milliseconds) that the implementation will wait
2018        to send a SACK.
2019
2020        Default: 200
2021
2022valid_cookie_life - INTEGER
2023        The default lifetime of the SCTP cookie (in milliseconds).  The cookie
2024        is used during association establishment.
2025
2026        Default: 60000
2027
2028cookie_preserve_enable - BOOLEAN
2029        Enable or disable the ability to extend the lifetime of the SCTP cookie
2030        that is used during the establishment phase of SCTP association
2031
2032        1: Enable cookie lifetime extension.
2033        0: Disable
2034
2035        Default: 1
2036
2037cookie_hmac_alg - STRING
2038        Select the hmac algorithm used when generating the cookie value sent by
2039        a listening sctp socket to a connecting client in the INIT-ACK chunk.
2040        Valid values are:
2041        * md5
2042        * sha1
2043        * none
2044        Ability to assign md5 or sha1 as the selected alg is predicated on the
2045        configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
2046        CONFIG_CRYPTO_SHA1).
2047
2048        Default: Dependent on configuration.  MD5 if available, else SHA1 if
2049        available, else none.
2050
2051rcvbuf_policy - INTEGER
2052        Determines if the receive buffer is attributed to the socket or to
2053        association.   SCTP supports the capability to create multiple
2054        associations on a single socket.  When using this capability, it is
2055        possible that a single stalled association that's buffering a lot
2056        of data may block other associations from delivering their data by
2057        consuming all of the receive buffer space.  To work around this,
2058        the rcvbuf_policy could be set to attribute the receiver buffer space
2059        to each association instead of the socket.  This prevents the described
2060        blocking.
2061
2062        1: rcvbuf space is per association
2063        0: rcvbuf space is per socket
2064
2065        Default: 0
2066
2067sndbuf_policy - INTEGER
2068        Similar to rcvbuf_policy above, this applies to send buffer space.
2069
2070        1: Send buffer is tracked per association
2071        0: Send buffer is tracked per socket.
2072
2073        Default: 0
2074
2075sctp_mem - vector of 3 INTEGERs: min, pressure, max
2076        Number of pages allowed for queueing by all SCTP sockets.
2077
2078        min: Below this number of pages SCTP is not bothered about its
2079        memory appetite. When amount of memory allocated by SCTP exceeds
2080        this number, SCTP starts to moderate memory usage.
2081
2082        pressure: This value was introduced to follow format of tcp_mem.
2083
2084        max: Number of pages allowed for queueing by all SCTP sockets.
2085
2086        Default is calculated at boot time from amount of available memory.
2087
2088sctp_rmem - vector of 3 INTEGERs: min, default, max
2089        Only the first value ("min") is used, "default" and "max" are
2090        ignored.
2091
2092        min: Minimal size of receive buffer used by SCTP socket.
2093        It is guaranteed to each SCTP socket (but not association) even
2094        under moderate memory pressure.
2095
2096        Default: 1 page
2097
2098sctp_wmem  - vector of 3 INTEGERs: min, default, max
2099        Currently this tunable has no effect.
2100
2101addr_scope_policy - INTEGER
2102        Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
2103
2104        0   - Disable IPv4 address scoping
2105        1   - Enable IPv4 address scoping
2106        2   - Follow draft but allow IPv4 private addresses
2107        3   - Follow draft but allow IPv4 link local addresses
2108
2109        Default: 1
2110
2111
2112/proc/sys/net/core/*
2113        Please see: Documentation/sysctl/net.txt for descriptions of these entries.
2114
2115
2116/proc/sys/net/unix/*
2117max_dgram_qlen - INTEGER
2118        The maximum length of dgram socket receive queue
2119
2120        Default: 10
2121
2122
2123UNDOCUMENTED:
2124
2125/proc/sys/net/irda/*
2126        fast_poll_increase FIXME
2127        warn_noreply_time FIXME
2128        discovery_slots FIXME
2129        slot_timeout FIXME
2130        max_baud_rate FIXME
2131        discovery_timeout FIXME
2132        lap_keepalive_time FIXME
2133        max_noreply_time FIXME
2134        max_tx_data_size FIXME
2135        max_tx_window FIXME
2136        min_tx_turn_time FIXME
2137
lxr.linux.no kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.