1.. SPDX-License-Identifier: GPL-2.0 2 3========= 4IP Sysctl 5========= 6 7/proc/sys/net/ipv4/* Variables 8============================== 9 10ip_forward - BOOLEAN 11 Forward Packets between interfaces. 12 13 This variable is special, its change resets all configuration 14 parameters to their default state (RFC1122 for hosts, RFC1812 15 for routers) 16 17 Possible values: 18 19 - 0 (disabled) 20 - 1 (enabled) 21 22 Default: 0 (disabled) 23 24ip_default_ttl - INTEGER 25 Default value of TTL field (Time To Live) for outgoing (but not 26 forwarded) IP packets. Should be between 1 and 255 inclusive. 27 Default: 64 (as recommended by RFC1700) 28 29ip_no_pmtu_disc - INTEGER 30 Disable Path MTU Discovery. If enabled in mode 1 and a 31 fragmentation-required ICMP is received, the PMTU to this 32 destination will be set to the smallest of the old MTU to 33 this destination and min_pmtu (see below). You will need 34 to raise min_pmtu to the smallest interface MTU on your system 35 manually if you want to avoid locally generated fragments. 36 37 In mode 2 incoming Path MTU Discovery messages will be 38 discarded. Outgoing frames are handled the same as in mode 1, 39 implicitly setting IP_PMTUDISC_DONT on every created socket. 40 41 Mode 3 is a hardened pmtu discover mode. The kernel will only 42 accept fragmentation-needed errors if the underlying protocol 43 can verify them besides a plain socket lookup. Current 44 protocols for which pmtu events will be honored are TCP and 45 SCTP as they verify e.g. the sequence number or the 46 association. This mode should not be enabled globally but is 47 only intended to secure e.g. name servers in namespaces where 48 TCP path mtu must still work but path MTU information of other 49 protocols should be discarded. If enabled globally this mode 50 could break other protocols. 51 52 Possible values: 0-3 53 54 Default: FALSE 55 56min_pmtu - INTEGER 57 default 552 - minimum Path MTU. Unless this is changed manually, 58 each cached pmtu will never be lower than this setting. 59 60ip_forward_use_pmtu - BOOLEAN 61 By default we don't trust protocol path MTUs while forwarding 62 because they could be easily forged and can lead to unwanted 63 fragmentation by the router. 64 You only need to enable this if you have user-space software 65 which tries to discover path mtus by itself and depends on the 66 kernel honoring this information. This is normally not the 67 case. 68 69 Possible values: 70 71 - 0 (disabled) 72 - 1 (enabled) 73 74 Default: 0 (disabled) 75 76fwmark_reflect - BOOLEAN 77 Controls the fwmark of kernel-generated IPv4 reply packets that are not 78 associated with a socket for example, TCP RSTs or ICMP echo replies). 79 If disabled, these packets have a fwmark of zero. If enabled, they have the 80 fwmark of the packet they are replying to. 81 82 Possible values: 83 84 - 0 (disabled) 85 - 1 (enabled) 86 87 Default: 0 (disabled) 88 89fib_multipath_use_neigh - BOOLEAN 90 Use status of existing neighbor entry when determining nexthop for 91 multipath routes. If disabled, neighbor information is not used and 92 packets could be directed to a failed nexthop. Only valid for kernels 93 built with CONFIG_IP_ROUTE_MULTIPATH enabled. 94 95 Possible values: 96 97 - 0 (disabled) 98 - 1 (enabled) 99 100 Default: 0 (disabled) 101 102fib_multipath_hash_policy - INTEGER 103 Controls which hash policy to use for multipath routes. Only valid 104 for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled. 105 106 Default: 0 (Layer 3) 107 108 Possible values: 109 110 - 0 - Layer 3 111 - 1 - Layer 4 112 - 2 - Layer 3 or inner Layer 3 if present 113 - 3 - Custom multipath hash. Fields used for multipath hash calculation 114 are determined by fib_multipath_hash_fields sysctl 115 116fib_multipath_hash_fields - UNSIGNED INTEGER 117 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the 118 fields used for multipath hash calculation are determined by this 119 sysctl. 120 121 This value is a bitmask which enables various fields for multipath hash 122 calculation. 123 124 Possible fields are: 125 126 ====== ============================ 127 0x0001 Source IP address 128 0x0002 Destination IP address 129 0x0004 IP protocol 130 0x0008 Unused (Flow Label) 131 0x0010 Source port 132 0x0020 Destination port 133 0x0040 Inner source IP address 134 0x0080 Inner destination IP address 135 0x0100 Inner IP protocol 136 0x0200 Inner Flow Label 137 0x0400 Inner source port 138 0x0800 Inner destination port 139 ====== ============================ 140 141 Default: 0x0007 (source IP, destination IP and IP protocol) 142 143fib_multipath_hash_seed - UNSIGNED INTEGER 144 The seed value used when calculating hash for multipath routes. Applies 145 to both IPv4 and IPv6 datapath. Only present for kernels built with 146 CONFIG_IP_ROUTE_MULTIPATH enabled. 147 148 When set to 0, the seed value used for multipath routing defaults to an 149 internal random-generated one. 150 151 The actual hashing algorithm is not specified -- there is no guarantee 152 that a next hop distribution effected by a given seed will keep stable 153 across kernel versions. 154 155 Default: 0 (random) 156 157fib_sync_mem - UNSIGNED INTEGER 158 Amount of dirty memory from fib entries that can be backlogged before 159 synchronize_rcu is forced. 160 161 Default: 512kB Minimum: 64kB Maximum: 64MB 162 163ip_forward_update_priority - INTEGER 164 Whether to update SKB priority from "TOS" field in IPv4 header after it 165 is forwarded. The new SKB priority is mapped from TOS field value 166 according to an rt_tos2priority table (see e.g. man tc-prio). 167 168 Default: 1 (Update priority.) 169 170 Possible values: 171 172 - 0 - Do not update priority. 173 - 1 - Update priority. 174 175route/max_size - INTEGER 176 Maximum number of routes allowed in the kernel. Increase 177 this when using large numbers of interfaces and/or routes. 178 179 From linux kernel 3.6 onwards, this is deprecated for ipv4 180 as route cache is no longer used. 181 182 From linux kernel 6.3 onwards, this is deprecated for ipv6 183 as garbage collection manages cached route entries. 184 185neigh/default/gc_thresh1 - INTEGER 186 Minimum number of entries to keep. Garbage collector will not 187 purge entries if there are fewer than this number. 188 189 Default: 128 190 191neigh/default/gc_thresh2 - INTEGER 192 Threshold when garbage collector becomes more aggressive about 193 purging entries. Entries older than 5 seconds will be cleared 194 when over this number. 195 196 Default: 512 197 198neigh/default/gc_thresh3 - INTEGER 199 Maximum number of non-PERMANENT neighbor entries allowed. Increase 200 this when using large numbers of interfaces and when communicating 201 with large numbers of directly-connected peers. 202 203 Default: 1024 204 205neigh/default/gc_interval - INTEGER 206 Specifies how often the garbage collector for neighbor entries 207 should run. This value applies to the entire table, not 208 individual entries. Unused since kernel v2.6.8. 209 210 Default: 30 seconds 211 212neigh/default/gc_stale_time - INTEGER 213 Determines how long a neighbor entry can remain unused before it is 214 considered stale and eligible for garbage collection. Entries that have 215 not been used for longer than this time will be removed by the garbage 216 collector, unless they have active references, are marked as PERMANENT, 217 or carry the NTF_EXT_LEARNED or NTF_EXT_VALIDATED flag. Stale entries 218 are only removed by the periodic GC when there are at least gc_thresh1 219 neighbors in the table. 220 221 Default: 60 seconds 222 223neigh/default/unres_qlen_bytes - INTEGER 224 The maximum number of bytes which may be used by packets 225 queued for each unresolved address by other network layers. 226 (added in linux 3.3) 227 228 Setting negative value is meaningless and will return error. 229 230 Default: SK_WMEM_DEFAULT, (same as net.core.wmem_default). 231 232 Exact value depends on architecture and kernel options, 233 but should be enough to allow queuing 256 packets 234 of medium size. 235 236neigh/default/unres_qlen - INTEGER 237 The maximum number of packets which may be queued for each 238 unresolved address by other network layers. 239 240 (deprecated in linux 3.3) : use unres_qlen_bytes instead. 241 242 Prior to linux 3.3, the default value is 3 which may cause 243 unexpected packet loss. The current default value is calculated 244 according to default value of unres_qlen_bytes and true size of 245 packet. 246 247 Default: 101 248 249neigh/default/interval_probe_time_ms - INTEGER 250 The probe interval for neighbor entries with NTF_MANAGED flag, 251 the min value is 1, and the max value is 86400000 (1 day). 252 253 Default: 5000 254 255mtu_expires - INTEGER 256 Time, in seconds, that cached PMTU information is kept. 257 258min_adv_mss - INTEGER 259 The advertised MSS depends on the first hop route MTU, but will 260 never be lower than this setting. 261 262fib_notify_on_flag_change - INTEGER 263 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/ 264 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed. 265 266 After installing a route to the kernel, user space receives an 267 acknowledgment, which means the route was installed in the kernel, 268 but not necessarily in hardware. 269 It is also possible for a route already installed in hardware to change 270 its action and therefore its flags. For example, a host route that is 271 trapping packets can be "promoted" to perform decapsulation following 272 the installation of an IPinIP/VXLAN tunnel. 273 The notifications will indicate to user-space the state of the route. 274 275 Default: 0 (Do not emit notifications.) 276 277 Possible values: 278 279 - 0 - Do not emit notifications. 280 - 1 - Emit notifications. 281 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change. 282 283IP Fragmentation: 284 285ipfrag_high_thresh - LONG INTEGER 286 Maximum memory used to reassemble IP fragments. 287 288ipfrag_low_thresh - LONG INTEGER 289 (Obsolete since linux-4.17) 290 Maximum memory used to reassemble IP fragments before the kernel 291 begins to remove incomplete fragment queues to free up resources. 292 The kernel still accepts new fragments for defragmentation. 293 294ipfrag_time - INTEGER 295 Time in seconds to keep an IP fragment in memory. 296 297ipfrag_max_dist - INTEGER 298 ipfrag_max_dist is a non-negative integer value which defines the 299 maximum "disorder" which is allowed among fragments which share a 300 common IP source address. Note that reordering of packets is 301 not unusual, but if a large number of fragments arrive from a source 302 IP address while a particular fragment queue remains incomplete, it 303 probably indicates that one or more fragments belonging to that queue 304 have been lost. When ipfrag_max_dist is positive, an additional check 305 is done on fragments before they are added to a reassembly queue - if 306 ipfrag_max_dist (or more) fragments have arrived from a particular IP 307 address between additions to any IP fragment queue using that source 308 address, it's presumed that one or more fragments in the queue are 309 lost. The existing fragment queue will be dropped, and a new one 310 started. An ipfrag_max_dist value of zero disables this check. 311 312 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can 313 result in unnecessarily dropping fragment queues when normal 314 reordering of packets occurs, which could lead to poor application 315 performance. Using a very large value, e.g. 50000, increases the 316 likelihood of incorrectly reassembling IP fragments that originate 317 from different IP datagrams, which could result in data corruption. 318 Default: 64 319 320bc_forwarding - INTEGER 321 bc_forwarding enables the feature described in rfc1812#section-5.3.5.2 322 and rfc2644. It allows the router to forward directed broadcast. 323 To enable this feature, the 'all' entry and the input interface entry 324 should be set to 1. 325 Default: 0 326 327INET peer storage 328================= 329 330inet_peer_threshold - INTEGER 331 The approximate size of the storage. Starting from this threshold 332 entries will be thrown aggressively. This threshold also determines 333 entries' time-to-live and time intervals between garbage collection 334 passes. More entries, less time-to-live, less GC interval. 335 336inet_peer_minttl - INTEGER 337 Minimum time-to-live of entries. Should be enough to cover fragment 338 time-to-live on the reassembling side. This minimum time-to-live is 339 guaranteed if the pool size is less than inet_peer_threshold. 340 Measured in seconds. 341 342inet_peer_maxttl - INTEGER 343 Maximum time-to-live of entries. Unused entries will expire after 344 this period of time if there is no memory pressure on the pool (i.e. 345 when the number of entries in the pool is very small). 346 Measured in seconds. 347 348TCP variables 349============= 350 351somaxconn - INTEGER 352 Limit of socket listen() backlog, known in userspace as SOMAXCONN. 353 Defaults to 4096. (Was 128 before linux-5.4) 354 See also tcp_max_syn_backlog for additional tuning for TCP sockets. 355 356tcp_abort_on_overflow - BOOLEAN 357 If listening service is too slow to accept new connections, 358 reset them. Default state is FALSE. It means that if overflow 359 occurred due to a burst, connection will recover. Enable this 360 option _only_ if you are really sure that listening daemon 361 cannot be tuned to accept connections faster. Enabling this 362 option can harm clients of your server. 363 364tcp_adv_win_scale - INTEGER 365 Obsolete since linux-6.6 366 Count buffering overhead as bytes/2^tcp_adv_win_scale 367 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale), 368 if it is <= 0. 369 370 Possible values are [-31, 31], inclusive. 371 372 Default: 1 373 374tcp_allowed_congestion_control - STRING 375 Show/set the congestion control choices available to non-privileged 376 processes. The list is a subset of those listed in 377 tcp_available_congestion_control. 378 379 Default is "reno" and the default setting (tcp_congestion_control). 380 381tcp_app_win - INTEGER 382 Reserve max(window/2^tcp_app_win, mss) of window for application 383 buffer. Value 0 is special, it means that nothing is reserved. 384 385 Possible values are [0, 31], inclusive. 386 387 Default: 31 388 389tcp_autocorking - BOOLEAN 390 Enable TCP auto corking : 391 When applications do consecutive small write()/sendmsg() system calls, 392 we try to coalesce these small writes as much as possible, to lower 393 total amount of sent packets. This is done if at least one prior 394 packet for the flow is waiting in Qdisc queues or device transmit 395 queue. Applications can still use TCP_CORK for optimal behavior 396 when they know how/when to uncork their sockets. 397 398 Possible values: 399 400 - 0 (disabled) 401 - 1 (enabled) 402 403 Default: 1 (enabled) 404 405tcp_available_congestion_control - STRING 406 Shows the available congestion control choices that are registered. 407 More congestion control algorithms may be available as modules, 408 but not loaded. 409 410tcp_base_mss - INTEGER 411 The initial value of search_low to be used by the packetization layer 412 Path MTU discovery (MTU probing). If MTU probing is enabled, 413 this is the initial MSS used by the connection. 414 415tcp_mtu_probe_floor - INTEGER 416 If MTU probing is enabled this caps the minimum MSS used for search_low 417 for the connection. 418 419 Default : 48 420 421tcp_min_snd_mss - INTEGER 422 TCP SYN and SYNACK messages usually advertise an ADVMSS option, 423 as described in RFC 1122 and RFC 6691. 424 425 If this ADVMSS option is smaller than tcp_min_snd_mss, 426 it is silently capped to tcp_min_snd_mss. 427 428 Default : 48 (at least 8 bytes of payload per segment) 429 430tcp_congestion_control - STRING 431 Set the congestion control algorithm to be used for new 432 connections. The algorithm "reno" is always available, but 433 additional choices may be available based on kernel configuration. 434 Default is set as part of kernel configuration. 435 For passive connections, the listener congestion control choice 436 is inherited. 437 438 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ] 439 440tcp_dsack - BOOLEAN 441 Allows TCP to send "duplicate" SACKs. 442 443 Possible values: 444 445 - 0 (disabled) 446 - 1 (enabled) 447 448 Default: 1 (enabled) 449 450tcp_early_retrans - INTEGER 451 Tail loss probe (TLP) converts RTOs occurring due to tail 452 losses into fast recovery (RFC8985). Note that 453 TLP requires RACK to function properly (see tcp_recovery below) 454 455 Possible values: 456 457 - 0 disables TLP 458 - 3 or 4 enables TLP 459 460 Default: 3 461 462tcp_ecn - INTEGER 463 Control use of Explicit Congestion Notification (ECN) by TCP. 464 ECN is used only when both ends of the TCP connection indicate support 465 for it. This feature is useful in avoiding losses due to congestion by 466 allowing supporting routers to signal congestion before having to drop 467 packets. A host that supports ECN both sends ECN at the IP layer and 468 feeds back ECN at the TCP layer. The highest variant of ECN feedback 469 that both peers support is chosen by the ECN negotiation (Accurate ECN, 470 ECN, or no ECN). 471 472 The highest negotiated variant for incoming connection requests 473 and the highest variant requested by outgoing connection 474 attempts: 475 476 ===== ==================== ==================== 477 Value Incoming connections Outgoing connections 478 ===== ==================== ==================== 479 0 No ECN No ECN 480 1 ECN ECN 481 2 ECN No ECN 482 3 AccECN AccECN 483 4 AccECN ECN 484 5 AccECN No ECN 485 ===== ==================== ==================== 486 487 Default: 2 488 489tcp_ecn_option - INTEGER 490 Control Accurate ECN (AccECN) option sending when AccECN has been 491 successfully negotiated during handshake. Send logic inhibits 492 sending AccECN options regardless of this setting when no AccECN 493 option has been seen for the reverse direction. 494 495 Possible values are: 496 497 = ============================================================ 498 0 Never send AccECN option. This also disables sending AccECN 499 option in SYN/ACK during handshake. 500 1 Send AccECN option sparingly according to the minimum option 501 rules outlined in draft-ietf-tcpm-accurate-ecn. 502 2 Send AccECN option on every packet whenever it fits into TCP 503 option space except when AccECN fallback is triggered. 504 3 Send AccECN option on every packet whenever it fits into TCP 505 option space even when AccECN fallback is triggered. 506 = ============================================================ 507 508 Default: 2 509 510tcp_ecn_option_beacon - INTEGER 511 Control Accurate ECN (AccECN) option sending frequency per RTT and it 512 takes effect only when tcp_ecn_option is set to 2. 513 514 Default: 3 (AccECN will be send at least 3 times per RTT) 515 516tcp_ecn_fallback - BOOLEAN 517 If the kernel detects that ECN connection misbehaves, enable fall 518 back to non-ECN. Currently, this knob implements the fallback 519 from RFC3168, section 6.1.1.1., but we reserve that in future, 520 additional detection mechanisms could be implemented under this 521 knob. The value is not used, if tcp_ecn or per route (or congestion 522 control) ECN settings are disabled. 523 524 Possible values: 525 526 - 0 (disabled) 527 - 1 (enabled) 528 529 Default: 1 (enabled) 530 531tcp_fack - BOOLEAN 532 This is a legacy option, it has no effect anymore. 533 534tcp_fin_timeout - INTEGER 535 The length of time an orphaned (no longer referenced by any 536 application) connection will remain in the FIN_WAIT_2 state 537 before it is aborted at the local end. While a perfectly 538 valid "receive only" state for an un-orphaned connection, an 539 orphaned connection in FIN_WAIT_2 state could otherwise wait 540 forever for the remote to close its end of the connection. 541 542 Cf. tcp_max_orphans 543 544 Default: 60 seconds 545 546tcp_frto - INTEGER 547 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682. 548 F-RTO is an enhanced recovery algorithm for TCP retransmission 549 timeouts. It is particularly beneficial in networks where the 550 RTT fluctuates (e.g., wireless). F-RTO is sender-side only 551 modification. It does not require any support from the peer. 552 553 By default it's enabled with a non-zero value. 0 disables F-RTO. 554 555tcp_fwmark_accept - BOOLEAN 556 If enabled, incoming connections to listening sockets that do not have a 557 socket mark will set the mark of the accepting socket to the fwmark of 558 the incoming SYN packet. This will cause all packets on that connection 559 (starting from the first SYNACK) to be sent with that fwmark. The 560 listening socket's mark is unchanged. Listening sockets that already 561 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are 562 unaffected. 563 564 Possible values: 565 566 - 0 (disabled) 567 - 1 (enabled) 568 569 Default: 0 (disabled) 570 571tcp_invalid_ratelimit - INTEGER 572 Limit the maximal rate for sending duplicate acknowledgments 573 in response to incoming TCP packets that are for an existing 574 connection but that are invalid due to any of these reasons: 575 576 (a) out-of-window sequence number, 577 (b) out-of-window acknowledgment number, or 578 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure 579 580 This can help mitigate simple "ack loop" DoS attacks, wherein 581 a buggy or malicious middlebox or man-in-the-middle can 582 rewrite TCP header fields in manner that causes each endpoint 583 to think that the other is sending invalid TCP segments, thus 584 causing each side to send an unterminating stream of duplicate 585 acknowledgments for invalid segments. 586 587 Using 0 disables rate-limiting of dupacks in response to 588 invalid segments; otherwise this value specifies the minimal 589 space between sending such dupacks, in milliseconds. 590 591 Default: 500 (milliseconds). 592 593tcp_keepalive_time - INTEGER 594 How often TCP sends out keepalive messages when keepalive is enabled. 595 Default: 2hours. 596 597tcp_keepalive_probes - INTEGER 598 How many keepalive probes TCP sends out, until it decides that the 599 connection is broken. Default value: 9. 600 601tcp_keepalive_intvl - INTEGER 602 How frequently the probes are send out. Multiplied by 603 tcp_keepalive_probes it is time to kill not responding connection, 604 after probes started. Default value: 75sec i.e. connection 605 will be aborted after ~11 minutes of retries. 606 607tcp_l3mdev_accept - BOOLEAN 608 Enables child sockets to inherit the L3 master device index. 609 Enabling this option allows a "global" listen socket to work 610 across L3 master domains (e.g., VRFs) with connected sockets 611 derived from the listen socket to be bound to the L3 domain in 612 which the packets originated. Only valid when the kernel was 613 compiled with CONFIG_NET_L3_MASTER_DEV. 614 615 Possible values: 616 617 - 0 (disabled) 618 - 1 (enabled) 619 620 Default: 0 (disabled) 621 622tcp_low_latency - BOOLEAN 623 This is a legacy option, it has no effect anymore. 624 625tcp_max_orphans - INTEGER 626 Maximal number of TCP sockets not attached to any user file handle, 627 held by system. If this number is exceeded orphaned connections are 628 reset immediately and warning is printed. This limit exists 629 only to prevent simple DoS attacks, you _must_ not rely on this 630 or lower the limit artificially, but rather increase it 631 (probably, after increasing installed memory), 632 if network conditions require more than default value, 633 and tune network services to linger and kill such states 634 more aggressively. Let me to remind again: each orphan eats 635 up to ~64K of unswappable memory. 636 637tcp_max_syn_backlog - INTEGER 638 Maximal number of remembered connection requests (SYN_RECV), 639 which have not received an acknowledgment from connecting client. 640 641 This is a per-listener limit. 642 643 The minimal value is 128 for low memory machines, and it will 644 increase in proportion to the memory of machine. 645 646 If server suffers from overload, try increasing this number. 647 648 Remember to also check /proc/sys/net/core/somaxconn 649 A SYN_RECV request socket consumes about 304 bytes of memory. 650 651tcp_max_tw_buckets - INTEGER 652 Maximal number of timewait sockets held by system simultaneously. 653 If this number is exceeded time-wait socket is immediately destroyed 654 and warning is printed. This limit exists only to prevent 655 simple DoS attacks, you _must_ not lower the limit artificially, 656 but rather increase it (probably, after increasing installed memory), 657 if network conditions require more than default value. 658 659tcp_mem - vector of 3 INTEGERs: min, pressure, max 660 min: below this number of pages TCP is not bothered about its 661 memory appetite. 662 663 pressure: when amount of memory allocated by TCP exceeds this number 664 of pages, TCP moderates its memory consumption and enters memory 665 pressure mode, which is exited when memory consumption falls 666 under "min". 667 668 max: number of pages allowed for queueing by all TCP sockets. 669 670 Defaults are calculated at boot time from amount of available 671 memory. 672 673tcp_min_rtt_wlen - INTEGER 674 The window length of the windowed min filter to track the minimum RTT. 675 A shorter window lets a flow more quickly pick up new (higher) 676 minimum RTT when it is moved to a longer path (e.g., due to traffic 677 engineering). A longer window makes the filter more resistant to RTT 678 inflations such as transient congestion. The unit is seconds. 679 680 Possible values: 0 - 86400 (1 day) 681 682 Default: 300 683 684tcp_moderate_rcvbuf - BOOLEAN 685 If enabled, TCP performs receive buffer auto-tuning, attempting to 686 automatically size the buffer (no greater than tcp_rmem[2]) to 687 match the size required by the path for full throughput. 688 689 Possible values: 690 691 - 0 (disabled) 692 - 1 (enabled) 693 694 Default: 1 (enabled) 695 696tcp_rcvbuf_low_rtt - INTEGER 697 rcvbuf autotuning can over estimate final socket rcvbuf, which 698 can lead to cache trashing for high throughput flows. 699 700 For small RTT flows (below tcp_rcvbuf_low_rtt usecs), we can relax 701 rcvbuf growth: Few additional ms to reach the final (and smaller) 702 rcvbuf is a good tradeoff. 703 704 Default : 1000 (1 ms) 705 706tcp_mtu_probing - INTEGER 707 Controls TCP Packetization-Layer Path MTU Discovery. Takes three 708 values: 709 710 - 0 - Disabled 711 - 1 - Disabled by default, enabled when an ICMP black hole detected 712 - 2 - Always enabled, use initial MSS of tcp_base_mss. 713 714tcp_probe_interval - UNSIGNED INTEGER 715 Controls how often to start TCP Packetization-Layer Path MTU 716 Discovery reprobe. The default is reprobing every 10 minutes as 717 per RFC4821. 718 719tcp_probe_threshold - INTEGER 720 Controls when TCP Packetization-Layer Path MTU Discovery probing 721 will stop in respect to the width of search range in bytes. Default 722 is 8 bytes. 723 724tcp_no_metrics_save - BOOLEAN 725 By default, TCP saves various connection metrics in the route cache 726 when the connection closes, so that connections established in the 727 near future can use these to set initial conditions. Usually, this 728 increases overall performance, but may sometimes cause performance 729 degradation. If enabled, TCP will not cache metrics on closing 730 connections. 731 732 Possible values: 733 734 - 0 (disabled) 735 - 1 (enabled) 736 737 Default: 0 (disabled) 738 739tcp_no_ssthresh_metrics_save - BOOLEAN 740 Controls whether TCP saves ssthresh metrics in the route cache. 741 If enabled, ssthresh metrics are disabled. 742 743 Possible values: 744 745 - 0 (disabled) 746 - 1 (enabled) 747 748 Default: 1 (enabled) 749 750tcp_orphan_retries - INTEGER 751 This value influences the timeout of a locally closed TCP connection, 752 when RTO retransmissions remain unacknowledged. 753 See tcp_retries2 for more details. 754 755 The default value is 8. 756 757 If your machine is a loaded WEB server, 758 you should think about lowering this value, such sockets 759 may consume significant resources. Cf. tcp_max_orphans. 760 761tcp_recovery - INTEGER 762 This value is a bitmap to enable various experimental loss recovery 763 features. 764 765 ========= ============================================================= 766 RACK: 0x1 enables RACK loss detection, for fast detection of lost 767 retransmissions and tail drops, and resilience to 768 reordering. currently, setting this bit to 0 has no 769 effect, since RACK is the only supported loss detection 770 algorithm. 771 772 RACK: 0x2 makes RACK's reordering window static (min_rtt/4). 773 774 RACK: 0x4 disables RACK's DUPACK threshold heuristic 775 ========= ============================================================= 776 777 Default: 0x1 778 779tcp_reflect_tos - BOOLEAN 780 For listening sockets, reuse the DSCP value of the initial SYN message 781 for outgoing packets. This allows to have both directions of a TCP 782 stream to use the same DSCP value, assuming DSCP remains unchanged for 783 the lifetime of the connection. 784 785 This options affects both IPv4 and IPv6. 786 787 Possible values: 788 789 - 0 (disabled) 790 - 1 (enabled) 791 792 Default: 0 (disabled) 793 794tcp_reordering - INTEGER 795 Initial reordering level of packets in a TCP stream. 796 TCP stack can then dynamically adjust flow reordering level 797 between this initial value and tcp_max_reordering 798 799 Default: 3 800 801tcp_max_reordering - INTEGER 802 Maximal reordering level of packets in a TCP stream. 803 300 is a fairly conservative value, but you might increase it 804 if paths are using per packet load balancing (like bonding rr mode) 805 806 Default: 300 807 808tcp_retrans_collapse - BOOLEAN 809 Bug-to-bug compatibility with some broken printers. 810 On retransmit try to send bigger packets to work around bugs in 811 certain TCP stacks. 812 813 Possible values: 814 815 - 0 (disabled) 816 - 1 (enabled) 817 818 Default: 1 (enabled) 819 820tcp_retries1 - INTEGER 821 This value influences the time, after which TCP decides, that 822 something is wrong due to unacknowledged RTO retransmissions, 823 and reports this suspicion to the network layer. 824 See tcp_retries2 for more details. 825 826 RFC 1122 recommends at least 3 retransmissions, which is the 827 default. 828 829tcp_retries2 - INTEGER 830 This value influences the timeout of an alive TCP connection, 831 when RTO retransmissions remain unacknowledged. 832 Given a value of N, a hypothetical TCP connection following 833 exponential backoff with an initial RTO of TCP_RTO_MIN would 834 retransmit N times before killing the connection at the (N+1)th RTO. 835 836 The default value of 15 yields a hypothetical timeout of 924.6 837 seconds and is a lower bound for the effective timeout. 838 TCP will effectively time out at the first RTO which exceeds the 839 hypothetical timeout. 840 If tcp_rto_max_ms is decreased, it is recommended to also 841 change tcp_retries2. 842 843 RFC 1122 recommends at least 100 seconds for the timeout, 844 which corresponds to a value of at least 8. 845 846tcp_rfc1337 - BOOLEAN 847 If enabled, the TCP stack behaves conforming to RFC1337. If unset, 848 we are not conforming to RFC, but prevent TCP TIME_WAIT 849 assassination. 850 851 Possible values: 852 853 - 0 (disabled) 854 - 1 (enabled) 855 856 Default: 0 (disabled) 857 858tcp_rmem - vector of 3 INTEGERs: min, default, max 859 min: Minimal size of receive buffer used by TCP sockets. 860 It is guaranteed to each TCP socket, even under moderate memory 861 pressure. 862 863 Default: 4K 864 865 default: initial size of receive buffer used by TCP sockets. 866 This value overrides net.core.rmem_default used by other protocols. 867 Default: 131072 bytes. 868 This value results in initial window of 65535. 869 870 max: maximal size of receive buffer allowed for automatically 871 selected receiver buffers for TCP socket. 872 Calling setsockopt() with SO_RCVBUF disables 873 automatic tuning of that socket's receive buffer size, in which 874 case this value is ignored. 875 Default: between 131072 and 32MB, depending on RAM size. 876 877 Each of the three values cannot be set below 4096. 878 879tcp_sack - BOOLEAN 880 Enable select acknowledgments (SACKS). 881 882 Possible values: 883 884 - 0 (disabled) 885 - 1 (enabled) 886 887 Default: 1 (enabled) 888 889tcp_comp_sack_rtt_percent - INTEGER 890 Percentage of SRTT used for the compressed SACK feature. 891 See tcp_comp_sack_nr, tcp_comp_sack_delay_ns, tcp_comp_sack_slack_ns. 892 893 Possible values : 1 - 1000 894 895 Default : 33 % 896 897tcp_comp_sack_delay_ns - LONG INTEGER 898 TCP tries to reduce number of SACK sent, using a timer based 899 on tcp_comp_sack_rtt_percent of SRTT, capped by this sysctl 900 in nano seconds. 901 The default is 1ms, based on TSO autosizing period. 902 903 Default : 1,000,000 ns (1 ms) 904 905tcp_comp_sack_slack_ns - LONG INTEGER 906 This sysctl control the slack used when arming the 907 timer used by SACK compression. This gives extra time 908 for small RTT flows, and reduces system overhead by allowing 909 opportunistic reduction of timer interrupts. 910 Too big values might reduce goodput. 911 912 Default : 10,000 ns (10 us) 913 914tcp_comp_sack_nr - INTEGER 915 Max number of SACK that can be compressed. 916 Using 0 disables SACK compression. 917 918 Default : 44 919 920tcp_backlog_ack_defer - BOOLEAN 921 If enabled, user thread processing socket backlog tries sending 922 one ACK for the whole queue. This helps to avoid potential 923 long latencies at end of a TCP socket syscall. 924 925 Possible values: 926 927 - 0 (disabled) 928 - 1 (enabled) 929 930 Default: 1 (enabled) 931 932tcp_slow_start_after_idle - BOOLEAN 933 If enabled, provide RFC2861 behavior and time out the congestion 934 window after an idle period. An idle period is defined at 935 the current RTO. If unset, the congestion window will not 936 be timed out after an idle period. 937 938 Possible values: 939 940 - 0 (disabled) 941 - 1 (enabled) 942 943 Default: 1 (enabled) 944 945tcp_stdurg - BOOLEAN 946 Use the Host requirements interpretation of the TCP urgent pointer field. 947 Most hosts use the older BSD interpretation, so if enabled, 948 Linux might not communicate correctly with them. 949 950 Possible values: 951 952 - 0 (disabled) 953 - 1 (enabled) 954 955 Default: 0 (disabled) 956 957tcp_synack_retries - INTEGER 958 Number of times SYNACKs for a passive TCP connection attempt will 959 be retransmitted. Should not be higher than 255. Default value 960 is 5, which corresponds to 31seconds till the last retransmission 961 with the current initial RTO of 1second. With this the final timeout 962 for a passive TCP connection will happen after 63seconds. 963 964tcp_syncookies - INTEGER 965 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES 966 Send out syncookies when the syn backlog queue of a socket 967 overflows. This is to prevent against the common 'SYN flood attack' 968 Default: 1 969 970 Note, that syncookies is fallback facility. 971 It MUST NOT be used to help highly loaded servers to stand 972 against legal connection rate. If you see SYN flood warnings 973 in your logs, but investigation shows that they occur 974 because of overload with legal connections, you should tune 975 another parameters until this warning disappear. 976 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow. 977 978 syncookies seriously violate TCP protocol, do not allow 979 to use TCP extensions, can result in serious degradation 980 of some services (f.e. SMTP relaying), visible not by you, 981 but your clients and relays, contacting you. While you see 982 SYN flood warnings in logs not being really flooded, your server 983 is seriously misconfigured. 984 985 If you want to test which effects syncookies have to your 986 network connections you can set this knob to 2 to enable 987 unconditionally generation of syncookies. 988 989tcp_migrate_req - BOOLEAN 990 The incoming connection is tied to a specific listening socket when 991 the initial SYN packet is received during the three-way handshake. 992 When a listener is closed, in-flight request sockets during the 993 handshake and established sockets in the accept queue are aborted. 994 995 If the listener has SO_REUSEPORT enabled, other listeners on the 996 same port should have been able to accept such connections. This 997 option makes it possible to migrate such child sockets to another 998 listener after close() or shutdown(). 999 1000 The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should 1001 usually be used to define the policy to pick an alive listener. 1002 Otherwise, the kernel will randomly pick an alive listener only if 1003 this option is enabled. 1004 1005 Note that migration between listeners with different settings may 1006 crash applications. Let's say migration happens from listener A to 1007 B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from 1008 the requests migrated from A. To avoid such a situation, cancel 1009 migration by returning SK_DROP in the type of eBPF program, or 1010 disable this option. 1011 1012 Possible values: 1013 1014 - 0 (disabled) 1015 - 1 (enabled) 1016 1017 Default: 0 (disabled) 1018 1019tcp_fastopen - INTEGER 1020 Enable TCP Fast Open (RFC7413) to send and accept data in the opening 1021 SYN packet. 1022 1023 The client support is enabled by flag 0x1 (on by default). The client 1024 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag, 1025 rather than connect() to send data in SYN. 1026 1027 The server support is enabled by flag 0x2 (off by default). Then 1028 either enable for all listeners with another flag (0x400) or 1029 enable individual listeners via TCP_FASTOPEN socket option with 1030 the option value being the length of the syn-data backlog. 1031 1032 The values (bitmap) are 1033 1034 ===== ======== ====================================================== 1035 0x1 (client) enables sending data in the opening SYN on the client. 1036 0x2 (server) enables the server support, i.e., allowing data in 1037 a SYN packet to be accepted and passed to the 1038 application before 3-way handshake finishes. 1039 0x4 (client) send data in the opening SYN regardless of cookie 1040 availability and without a cookie option. 1041 0x200 (server) accept data-in-SYN w/o any cookie option present. 1042 0x400 (server) enable all listeners to support Fast Open by 1043 default without explicit TCP_FASTOPEN socket option. 1044 ===== ======== ====================================================== 1045 1046 Default: 0x1 1047 1048 Note that additional client or server features are only 1049 effective if the basic support (0x1 and 0x2) are enabled respectively. 1050 1051tcp_fastopen_blackhole_timeout_sec - INTEGER 1052 Initial time period in second to disable Fastopen on active TCP sockets 1053 when a TFO firewall blackhole issue happens. 1054 This time period will grow exponentially when more blackhole issues 1055 get detected right after Fastopen is re-enabled and will reset to 1056 initial value when the blackhole issue goes away. 1057 0 to disable the blackhole detection. 1058 1059 By default, it is set to 0 (feature is disabled). 1060 1061tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs 1062 The list consists of a primary key and an optional backup key. The 1063 primary key is used for both creating and validating cookies, while the 1064 optional backup key is only used for validating cookies. The purpose of 1065 the backup key is to maximize TFO validation when keys are rotated. 1066 1067 A randomly chosen primary key may be configured by the kernel if 1068 the tcp_fastopen sysctl is set to 0x400 (see above), or if the 1069 TCP_FASTOPEN setsockopt() optname is set and a key has not been 1070 previously configured via sysctl. If keys are configured via 1071 setsockopt() by using the TCP_FASTOPEN_KEY optname, then those 1072 per-socket keys will be used instead of any keys that are specified via 1073 sysctl. 1074 1075 A key is specified as 4 8-digit hexadecimal integers which are separated 1076 by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be 1077 omitted. A primary and a backup key may be specified by separating them 1078 by a comma. If only one key is specified, it becomes the primary key and 1079 any previously configured backup keys are removed. 1080 1081tcp_syn_retries - INTEGER 1082 Number of times initial SYNs for an active TCP connection attempt 1083 will be retransmitted. Should not be higher than 127. Default value 1084 is 6, which corresponds to 67seconds (with tcp_syn_linear_timeouts = 4) 1085 till the last retransmission with the current initial RTO of 1second. 1086 With this the final timeout for an active TCP connection attempt 1087 will happen after 131seconds. 1088 1089tcp_timestamps - INTEGER 1090 Enable timestamps as defined in RFC1323. 1091 1092 - 0: Disabled. 1093 - 1: Enable timestamps as defined in RFC1323 and use random offset for 1094 each connection rather than only using the current time. 1095 - 2: Like 1, but without random offsets. 1096 1097 Default: 1 1098 1099tcp_min_tso_segs - INTEGER 1100 Minimal number of segments per TSO frame. 1101 1102 Since linux-3.12, TCP does an automatic sizing of TSO frames, 1103 depending on flow rate, instead of filling 64Kbytes packets. 1104 For specific usages, it's possible to force TCP to build big 1105 TSO frames. Note that TCP stack might split too big TSO packets 1106 if available window is too small. 1107 1108 Default: 2 1109 1110tcp_tso_rtt_log - INTEGER 1111 Adjustment of TSO packet sizes based on min_rtt 1112 1113 Starting from linux-5.18, TCP autosizing can be tweaked 1114 for flows having small RTT. 1115 1116 Old autosizing was splitting the pacing budget to send 1024 TSO 1117 per second. 1118 1119 tso_packet_size = sk->sk_pacing_rate / 1024; 1120 1121 With the new mechanism, we increase this TSO sizing using: 1122 1123 distance = min_rtt_usec / (2^tcp_tso_rtt_log) 1124 tso_packet_size += gso_max_size >> distance; 1125 1126 This means that flows between very close hosts can use bigger 1127 TSO packets, reducing their cpu costs. 1128 1129 If you want to use the old autosizing, set this sysctl to 0. 1130 1131 Default: 9 (2^9 = 512 usec) 1132 1133tcp_pacing_ss_ratio - INTEGER 1134 sk->sk_pacing_rate is set by TCP stack using a ratio applied 1135 to current rate. (current_rate = cwnd * mss / srtt) 1136 If TCP is in slow start, tcp_pacing_ss_ratio is applied 1137 to let TCP probe for bigger speeds, assuming cwnd can be 1138 doubled every other RTT. 1139 1140 Default: 200 1141 1142tcp_pacing_ca_ratio - INTEGER 1143 sk->sk_pacing_rate is set by TCP stack using a ratio applied 1144 to current rate. (current_rate = cwnd * mss / srtt) 1145 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio 1146 is applied to conservatively probe for bigger throughput. 1147 1148 Default: 120 1149 1150tcp_syn_linear_timeouts - INTEGER 1151 The number of times for an active TCP connection to retransmit SYNs with 1152 a linear backoff timeout before defaulting to an exponential backoff 1153 timeout. This has no effect on SYNACK at the passive TCP side. 1154 1155 With an initial RTO of 1 and tcp_syn_linear_timeouts = 4 we would 1156 expect SYN RTOs to be: 1, 1, 1, 1, 1, 2, 4, ... (4 linear timeouts, 1157 and the first exponential backoff using 2^0 * initial_RTO). 1158 Default: 4 1159 1160tcp_tso_win_divisor - INTEGER 1161 This allows control over what percentage of the congestion window 1162 can be consumed by a single TSO frame. 1163 The setting of this parameter is a choice between burstiness and 1164 building larger TSO frames. 1165 1166 Default: 3 1167 1168tcp_tw_reuse - INTEGER 1169 Enable reuse of TIME-WAIT sockets for new connections when it is 1170 safe from protocol viewpoint. 1171 1172 - 0 - disable 1173 - 1 - global enable 1174 - 2 - enable for loopback traffic only 1175 1176 It should not be changed without advice/request of technical 1177 experts. 1178 1179 Default: 2 1180 1181tcp_tw_reuse_delay - UNSIGNED INTEGER 1182 The delay in milliseconds before a TIME-WAIT socket can be reused by a 1183 new connection, if TIME-WAIT socket reuse is enabled. The actual reuse 1184 threshold is within [N, N+1] range, where N is the requested delay in 1185 milliseconds, to ensure the delay interval is never shorter than the 1186 configured value. 1187 1188 This setting contains an assumption about the other TCP timestamp clock 1189 tick interval. It should not be set to a value lower than the peer's 1190 clock tick for PAWS (Protection Against Wrapped Sequence numbers) 1191 mechanism work correctly for the reused connection. 1192 1193 Default: 1000 (milliseconds) 1194 1195tcp_window_scaling - BOOLEAN 1196 Enable window scaling as defined in RFC1323. 1197 1198 Possible values: 1199 1200 - 0 (disabled) 1201 - 1 (enabled) 1202 1203 Default: 1 (enabled) 1204 1205tcp_shrink_window - BOOLEAN 1206 This changes how the TCP receive window is calculated. 1207 1208 RFC 7323, section 2.4, says there are instances when a retracted 1209 window can be offered, and that TCP implementations MUST ensure 1210 that they handle a shrinking window, as specified in RFC 1122. 1211 1212 Possible values: 1213 1214 - 0 (disabled) - The window is never shrunk. 1215 - 1 (enabled) - The window is shrunk when necessary to remain within 1216 the memory limit set by autotuning (sk_rcvbuf). 1217 This only occurs if a non-zero receive window 1218 scaling factor is also in effect. 1219 1220 Default: 0 (disabled) 1221 1222tcp_wmem - vector of 3 INTEGERs: min, default, max 1223 min: Amount of memory reserved for send buffers for TCP sockets. 1224 Each TCP socket has rights to use it due to fact of its birth. 1225 1226 Default: 4K 1227 1228 default: initial size of send buffer used by TCP sockets. This 1229 value overrides net.core.wmem_default used by other protocols. 1230 1231 It is usually lower than net.core.wmem_default. 1232 1233 Default: 16K 1234 1235 max: Maximal amount of memory allowed for automatically tuned 1236 send buffers for TCP sockets. This value does not override 1237 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables 1238 automatic tuning of that socket's send buffer size, in which case 1239 this value is ignored. 1240 1241 Default: between 64K and 4MB, depending on RAM size. 1242 1243tcp_notsent_lowat - UNSIGNED INTEGER 1244 A TCP socket can control the amount of unsent bytes in its write queue, 1245 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll() 1246 reports POLLOUT events if the amount of unsent bytes is below a per 1247 socket value, and if the write queue is not full. sendmsg() will 1248 also not add new buffers if the limit is hit. 1249 1250 This global variable controls the amount of unsent data for 1251 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change 1252 to the global variable has immediate effect. 1253 1254 Default: UINT_MAX (0xFFFFFFFF) 1255 1256tcp_workaround_signed_windows - BOOLEAN 1257 If enabled, assume no receipt of a window scaling option means the 1258 remote TCP is broken and treats the window as a signed quantity. 1259 If disabled, assume the remote TCP is not broken even if we do 1260 not receive a window scaling option from them. 1261 1262 Possible values: 1263 1264 - 0 (disabled) 1265 - 1 (enabled) 1266 1267 Default: 0 (disabled) 1268 1269tcp_thin_linear_timeouts - BOOLEAN 1270 Enable dynamic triggering of linear timeouts for thin streams. 1271 If enabled, a check is performed upon retransmission by timeout to 1272 determine if the stream is thin (less than 4 packets in flight). 1273 As long as the stream is found to be thin, up to 6 linear 1274 timeouts may be performed before exponential backoff mode is 1275 initiated. This improves retransmission latency for 1276 non-aggressive thin streams, often found to be time-dependent. 1277 For more information on thin streams, see 1278 Documentation/networking/tcp-thin.rst 1279 1280 Possible values: 1281 1282 - 0 (disabled) 1283 - 1 (enabled) 1284 1285 Default: 0 (disabled) 1286 1287tcp_limit_output_bytes - INTEGER 1288 Controls TCP Small Queue limit per tcp socket. 1289 TCP bulk sender tends to increase packets in flight until it 1290 gets losses notifications. With SNDBUF autotuning, this can 1291 result in a large amount of packets queued on the local machine 1292 (e.g.: qdiscs, CPU backlog, or device) hurting latency of other 1293 flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes 1294 limits the number of bytes on qdisc or device to reduce artificial 1295 RTT/cwnd and reduce bufferbloat. 1296 1297 Default: 4194304 (4 MB) 1298 1299tcp_challenge_ack_limit - INTEGER 1300 Limits number of Challenge ACK sent per second, as recommended 1301 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks) 1302 Note that this per netns rate limit can allow some side channel 1303 attacks and probably should not be enabled. 1304 TCP stack implements per TCP socket limits anyway. 1305 Default: INT_MAX (unlimited) 1306 1307tcp_ehash_entries - INTEGER 1308 Show the number of hash buckets for TCP sockets in the current 1309 networking namespace. 1310 1311 A negative value means the networking namespace does not own its 1312 hash buckets and shares the initial networking namespace's one. 1313 1314tcp_child_ehash_entries - INTEGER 1315 Control the number of hash buckets for TCP sockets in the child 1316 networking namespace, which must be set before clone() or unshare(). 1317 1318 If the value is not 0, the kernel uses a value rounded up to 2^n 1319 as the actual hash bucket size. 0 is a special value, meaning 1320 the child networking namespace will share the initial networking 1321 namespace's hash buckets. 1322 1323 Note that the child will use the global one in case the kernel 1324 fails to allocate enough memory. In addition, the global hash 1325 buckets are spread over available NUMA nodes, but the allocation 1326 of the child hash table depends on the current process's NUMA 1327 policy, which could result in performance differences. 1328 1329 Note also that the default value of tcp_max_tw_buckets and 1330 tcp_max_syn_backlog depend on the hash bucket size. 1331 1332 Possible values: 0, 2^n (n: 0 - 24 (16Mi)) 1333 1334 Default: 0 1335 1336tcp_plb_enabled - BOOLEAN 1337 If enabled and the underlying congestion control (e.g. DCTCP) supports 1338 and enables PLB feature, TCP PLB (Protective Load Balancing) is 1339 enabled. PLB is described in the following paper: 1340 https://doi.org/10.1145/3544216.3544226. Based on PLB parameters, 1341 upon sensing sustained congestion, TCP triggers a change in 1342 flow label field for outgoing IPv6 packets. A change in flow label 1343 field potentially changes the path of outgoing packets for switches 1344 that use ECMP/WCMP for routing. 1345 1346 PLB changes socket txhash which results in a change in IPv6 Flow Label 1347 field, and currently no-op for IPv4 headers. It is possible 1348 to apply PLB for IPv4 with other network header fields (e.g. TCP 1349 or IPv4 options) or using encapsulation where outer header is used 1350 by switches to determine next hop. In either case, further host 1351 and switch side changes will be needed. 1352 1353 If enabled, PLB assumes that congestion signal (e.g. ECN) is made 1354 available and used by congestion control module to estimate a 1355 congestion measure (e.g. ce_ratio). PLB needs a congestion measure to 1356 make repathing decisions. 1357 1358 Possible values: 1359 1360 - 0 (disabled) 1361 - 1 (enabled) 1362 1363 Default: 0 (disabled) 1364 1365tcp_plb_idle_rehash_rounds - INTEGER 1366 Number of consecutive congested rounds (RTT) seen after which 1367 a rehash can be performed, given there are no packets in flight. 1368 This is referred to as M in PLB paper: 1369 https://doi.org/10.1145/3544216.3544226. 1370 1371 Possible Values: 0 - 31 1372 1373 Default: 3 1374 1375tcp_plb_rehash_rounds - INTEGER 1376 Number of consecutive congested rounds (RTT) seen after which 1377 a forced rehash can be performed. Be careful when setting this 1378 parameter, as a small value increases the risk of retransmissions. 1379 This is referred to as N in PLB paper: 1380 https://doi.org/10.1145/3544216.3544226. 1381 1382 Possible Values: 0 - 31 1383 1384 Default: 12 1385 1386tcp_plb_suspend_rto_sec - INTEGER 1387 Time, in seconds, to suspend PLB in event of an RTO. In order to avoid 1388 having PLB repath onto a connectivity "black hole", after an RTO a TCP 1389 connection suspends PLB repathing for a random duration between 1x and 1390 2x of this parameter. Randomness is added to avoid concurrent rehashing 1391 of multiple TCP connections. This should be set corresponding to the 1392 amount of time it takes to repair a failed link. 1393 1394 Possible Values: 0 - 255 1395 1396 Default: 60 1397 1398tcp_plb_cong_thresh - INTEGER 1399 Fraction of packets marked with congestion over a round (RTT) to 1400 tag that round as congested. This is referred to as K in the PLB paper: 1401 https://doi.org/10.1145/3544216.3544226. 1402 1403 The 0-1 fraction range is mapped to 0-256 range to avoid floating 1404 point operations. For example, 128 means that if at least 50% of 1405 the packets in a round were marked as congested then the round 1406 will be tagged as congested. 1407 1408 Setting threshold to 0 means that PLB repaths every RTT regardless 1409 of congestion. This is not intended behavior for PLB and should be 1410 used only for experimentation purpose. 1411 1412 Possible Values: 0 - 256 1413 1414 Default: 128 1415 1416tcp_pingpong_thresh - INTEGER 1417 The number of estimated data replies sent for estimated incoming data 1418 requests that must happen before TCP considers that a connection is a 1419 "ping-pong" (request-response) connection for which delayed 1420 acknowledgments can provide benefits. 1421 1422 This threshold is 1 by default, but some applications may need a higher 1423 threshold for optimal performance. 1424 1425 Possible Values: 1 - 255 1426 1427 Default: 1 1428 1429tcp_rto_min_us - INTEGER 1430 Minimal TCP retransmission timeout (in microseconds). Note that the 1431 rto_min route option has the highest precedence for configuring this 1432 setting, followed by the TCP_BPF_RTO_MIN and TCP_RTO_MIN_US socket 1433 options, followed by this tcp_rto_min_us sysctl. 1434 1435 The recommended practice is to use a value less or equal to 200000 1436 microseconds. 1437 1438 Possible Values: 1 - INT_MAX 1439 1440 Default: 200000 1441 1442tcp_rto_max_ms - INTEGER 1443 Maximal TCP retransmission timeout (in ms). 1444 Note that TCP_RTO_MAX_MS socket option has higher precedence. 1445 1446 When changing tcp_rto_max_ms, it is important to understand 1447 that tcp_retries2 might need a change. 1448 1449 Possible Values: 1000 - 120,000 1450 1451 Default: 120,000 1452 1453UDP variables 1454============= 1455 1456udp_l3mdev_accept - BOOLEAN 1457 Enabling this option allows a "global" bound socket to work 1458 across L3 master domains (e.g., VRFs) with packets capable of 1459 being received regardless of the L3 domain in which they 1460 originated. Only valid when the kernel was compiled with 1461 CONFIG_NET_L3_MASTER_DEV. 1462 1463 Possible values: 1464 1465 - 0 (disabled) 1466 - 1 (enabled) 1467 1468 Default: 0 (disabled) 1469 1470udp_mem - vector of 3 INTEGERs: min, pressure, max 1471 Number of pages allowed for queueing by all UDP sockets. 1472 1473 min: Number of pages allowed for queueing by all UDP sockets. 1474 1475 pressure: This value was introduced to follow format of tcp_mem. 1476 1477 max: This value was introduced to follow format of tcp_mem. 1478 1479 Default is calculated at boot time from amount of available memory. 1480 1481udp_rmem_min - INTEGER 1482 Minimal size of receive buffer used by UDP sockets in moderation. 1483 Each UDP socket is able to use the size for receiving data, even if 1484 total pages of UDP sockets exceed udp_mem pressure. The unit is byte. 1485 1486 Default: 4K 1487 1488udp_wmem_min - INTEGER 1489 UDP does not have tx memory accounting and this tunable has no effect. 1490 1491udp_hash_entries - INTEGER 1492 Show the number of hash buckets for UDP sockets in the current 1493 networking namespace. 1494 1495 A negative value means the networking namespace does not own its 1496 hash buckets and shares the initial networking namespace's one. 1497 1498udp_child_hash_entries - INTEGER 1499 Control the number of hash buckets for UDP sockets in the child 1500 networking namespace, which must be set before clone() or unshare(). 1501 1502 If the value is not 0, the kernel uses a value rounded up to 2^n 1503 as the actual hash bucket size. 0 is a special value, meaning 1504 the child networking namespace will share the initial networking 1505 namespace's hash buckets. 1506 1507 Note that the child will use the global one in case the kernel 1508 fails to allocate enough memory. In addition, the global hash 1509 buckets are spread over available NUMA nodes, but the allocation 1510 of the child hash table depends on the current process's NUMA 1511 policy, which could result in performance differences. 1512 1513 Possible values: 0, 2^n (n: 7 (128) - 16 (64K)) 1514 1515 Default: 0 1516 1517 1518RAW variables 1519============= 1520 1521raw_l3mdev_accept - BOOLEAN 1522 Enabling this option allows a "global" bound socket to work 1523 across L3 master domains (e.g., VRFs) with packets capable of 1524 being received regardless of the L3 domain in which they 1525 originated. Only valid when the kernel was compiled with 1526 CONFIG_NET_L3_MASTER_DEV. 1527 1528 Possible values: 1529 1530 - 0 (disabled) 1531 - 1 (enabled) 1532 1533 Default: 1 (enabled) 1534 1535CIPSOv4 Variables 1536================= 1537 1538cipso_cache_enable - BOOLEAN 1539 If enabled, enable additions to and lookups from the CIPSO label mapping 1540 cache. If disabled, additions are ignored and lookups always result in a 1541 miss. However, regardless of the setting the cache is still 1542 invalidated when required when means you can safely toggle this on and 1543 off and the cache will always be "safe". 1544 1545 Possible values: 1546 1547 - 0 (disabled) 1548 - 1 (enabled) 1549 1550 Default: 1 (enabled) 1551 1552cipso_cache_bucket_size - INTEGER 1553 The CIPSO label cache consists of a fixed size hash table with each 1554 hash bucket containing a number of cache entries. This variable limits 1555 the number of entries in each hash bucket; the larger the value is, the 1556 more CIPSO label mappings that can be cached. When the number of 1557 entries in a given hash bucket reaches this limit adding new entries 1558 causes the oldest entry in the bucket to be removed to make room. 1559 1560 Default: 10 1561 1562cipso_rbm_optfmt - BOOLEAN 1563 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of 1564 the CIPSO draft specification (see Documentation/netlabel for details). 1565 This means that when set the CIPSO tag will be padded with empty 1566 categories in order to make the packet data 32-bit aligned. 1567 1568 Possible values: 1569 1570 - 0 (disabled) 1571 - 1 (enabled) 1572 1573 Default: 0 (disabled) 1574 1575cipso_rbm_strictvalid - BOOLEAN 1576 If enabled, do a very strict check of the CIPSO option when 1577 ip_options_compile() is called. If disabled, relax the checks done during 1578 ip_options_compile(). Either way is "safe" as errors are caught else 1579 where in the CIPSO processing code but setting this to 0 (False) should 1580 result in less work (i.e. it should be faster) but could cause problems 1581 with other implementations that require strict checking. 1582 1583 Possible values: 1584 1585 - 0 (disabled) 1586 - 1 (enabled) 1587 1588 Default: 0 (disabled) 1589 1590IP Variables 1591============ 1592 1593ip_local_port_range - 2 INTEGERS 1594 Defines the local port range that is used by TCP and UDP to 1595 choose the local port. The first number is the first, the 1596 second the last local port number. 1597 If possible, it is better these numbers have different parity 1598 (one even and one odd value). 1599 Must be greater than or equal to ip_unprivileged_port_start. 1600 The default values are 32768 and 60999 respectively. 1601 1602ip_local_reserved_ports - list of comma separated ranges 1603 Specify the ports which are reserved for known third-party 1604 applications. These ports will not be used by automatic port 1605 assignments (e.g. when calling connect() or bind() with port 1606 number 0). Explicit port allocation behavior is unchanged. 1607 1608 The format used for both input and output is a comma separated 1609 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and 1610 10). Writing to the file will clear all previously reserved 1611 ports and update the current list with the one given in the 1612 input. 1613 1614 Note that ip_local_port_range and ip_local_reserved_ports 1615 settings are independent and both are considered by the kernel 1616 when determining which ports are available for automatic port 1617 assignments. 1618 1619 You can reserve ports which are not in the current 1620 ip_local_port_range, e.g.:: 1621 1622 $ cat /proc/sys/net/ipv4/ip_local_port_range 1623 32000 60999 1624 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports 1625 8080,9148 1626 1627 although this is redundant. However such a setting is useful 1628 if later the port range is changed to a value that will 1629 include the reserved ports. Also keep in mind, that overlapping 1630 of these ranges may affect probability of selecting ephemeral 1631 ports which are right after block of reserved ports. 1632 1633 Default: Empty 1634 1635ip_local_port_step_width - INTEGER 1636 Defines the numerical maximum increment between successive port 1637 allocations within the ephemeral port range when an unavailable port is 1638 reached. This can be used to mitigate accumulated nodes in port 1639 distribution when reserved ports have been configured. Please note that 1640 port collisions may be more frequent in a system with a very high load. 1641 1642 It is recommended to set this value strictly larger than the largest 1643 contiguous block of ports configure in ip_local_reserved_ports. For 1644 large reserved port ranges, setting this to 3x or 4x the size of the 1645 largest block is advised. Using a value equal or greater than the local 1646 port range size completely solves the uneven port distribution problem, 1647 but it can degrade performance under port exhaustion situations. 1648 1649 Default: 0 (disabled) 1650 1651ip_unprivileged_port_start - INTEGER 1652 This is a per-namespace sysctl. It defines the first 1653 unprivileged port in the network namespace. Privileged ports 1654 require root or CAP_NET_BIND_SERVICE in order to bind to them. 1655 To disable all privileged ports, set this to 0. They must not 1656 overlap with the ip_local_port_range. 1657 1658 Default: 1024 1659 1660ip_nonlocal_bind - BOOLEAN 1661 If enabled, allows processes to bind() to non-local IP addresses, 1662 which can be quite useful - but may break some applications. 1663 1664 Possible values: 1665 1666 - 0 (disabled) 1667 - 1 (enabled) 1668 1669 Default: 0 (disabled) 1670 1671ip_autobind_reuse - BOOLEAN 1672 By default, bind() does not select the ports automatically even if 1673 the new socket and all sockets bound to the port have SO_REUSEADDR. 1674 ip_autobind_reuse allows bind() to reuse the port and this is useful 1675 when you use bind()+connect(), but may break some applications. 1676 The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this 1677 option should only be set by experts. 1678 1679 Possible values: 1680 1681 - 0 (disabled) 1682 - 1 (enabled) 1683 1684 Default: 0 (disabled) 1685 1686ip_dynaddr - INTEGER 1687 If set non-zero, enables support for dynamic addresses. 1688 If set to a non-zero value larger than 1, a kernel log 1689 message will be printed when dynamic address rewriting 1690 occurs. 1691 1692 Default: 0 1693 1694ip_early_demux - BOOLEAN 1695 Optimize input packet processing down to one demux for 1696 certain kinds of local sockets. Currently we only do this 1697 for established TCP and connected UDP sockets. 1698 1699 It may add an additional cost for pure routing workloads that 1700 reduces overall throughput, in such case you should disable it. 1701 1702 Possible values: 1703 1704 - 0 (disabled) 1705 - 1 (enabled) 1706 1707 Default: 1 (enabled) 1708 1709ping_group_range - 2 INTEGERS 1710 Restrict ICMP_PROTO datagram sockets to users in the group range. 1711 The default is "1 0", meaning, that nobody (not even root) may 1712 create ping sockets. Setting it to "100 100" would grant permissions 1713 to the single group. "0 4294967294" would enable it for the world, "100 1714 4294967294" would enable it for the users, but not daemons. 1715 1716tcp_early_demux - BOOLEAN 1717 Enable early demux for established TCP sockets. 1718 1719 Possible values: 1720 1721 - 0 (disabled) 1722 - 1 (enabled) 1723 1724 Default: 1 (enabled) 1725 1726udp_early_demux - BOOLEAN 1727 Enable early demux for connected UDP sockets. Disable this if 1728 your system could experience more unconnected load. 1729 1730 Possible values: 1731 1732 - 0 (disabled) 1733 - 1 (enabled) 1734 1735 Default: 1 (enabled) 1736 1737icmp_echo_ignore_all - BOOLEAN 1738 If enabled, then the kernel will ignore all ICMP ECHO 1739 requests sent to it. 1740 1741 Possible values: 1742 1743 - 0 (disabled) 1744 - 1 (enabled) 1745 1746 Default: 0 (disabled) 1747 1748icmp_echo_enable_probe - BOOLEAN 1749 If enabled, then the kernel will respond to RFC 8335 PROBE 1750 requests sent to it. 1751 1752 Possible values: 1753 1754 - 0 (disabled) 1755 - 1 (enabled) 1756 1757 Default: 0 (disabled) 1758 1759icmp_echo_ignore_broadcasts - BOOLEAN 1760 If enabled, then the kernel will ignore all ICMP ECHO and 1761 TIMESTAMP requests sent to it via broadcast/multicast. 1762 1763 Possible values: 1764 1765 - 0 (disabled) 1766 - 1 (enabled) 1767 1768 Default: 1 (enabled) 1769 1770icmp_ratelimit - INTEGER 1771 Limit the maximal rates for sending ICMP packets whose type matches 1772 icmp_ratemask (see below) to specific targets. 1773 0 to disable any limiting, 1774 otherwise the minimal space between responses in milliseconds. 1775 Note that another sysctl, icmp_msgs_per_sec limits the number 1776 of ICMP packets sent on all targets. 1777 1778 Default: 1000 1779 1780icmp_msgs_per_sec - INTEGER 1781 Limit maximal number of ICMP packets sent per second from this host. 1782 Only messages whose type matches icmp_ratemask (see below) are 1783 controlled by this limit. For security reasons, the precise count 1784 of messages per second is randomized. 1785 1786 Default: 10000 1787 1788icmp_msgs_burst - INTEGER 1789 icmp_msgs_per_sec controls number of ICMP packets sent per second, 1790 while icmp_msgs_burst controls the token bucket size. 1791 For security reasons, the precise burst size is randomized. 1792 1793 Default: 10000 1794 1795icmp_ratemask - INTEGER 1796 Mask made of ICMP types for which rates are being limited. 1797 1798 Significant bits: IHGFEDCBA9876543210 1799 1800 Default mask: 0000001100000011000 (6168) 1801 1802 Bit definitions (see include/linux/icmp.h): 1803 1804 = ========================= 1805 0 Echo Reply 1806 3 Destination Unreachable [1]_ 1807 4 Source Quench [1]_ 1808 5 Redirect 1809 8 Echo Request 1810 B Time Exceeded [1]_ 1811 C Parameter Problem [1]_ 1812 D Timestamp Request 1813 E Timestamp Reply 1814 F Info Request 1815 G Info Reply 1816 H Address Mask Request 1817 I Address Mask Reply 1818 = ========================= 1819 1820 .. [1] These are rate limited by default (see default mask above) 1821 1822icmp_ignore_bogus_error_responses - BOOLEAN 1823 Some routers violate RFC1122 by sending bogus responses to broadcast 1824 frames. Such violations are normally logged via a kernel warning. 1825 If enabled, the kernel will not give such warnings, which 1826 will avoid log file clutter. 1827 1828 Possible values: 1829 1830 - 0 (disabled) 1831 - 1 (enabled) 1832 1833 Default: 1 (enabled) 1834 1835icmp_errors_use_inbound_ifaddr - BOOLEAN 1836 1837 If disabled, icmp error messages are sent with the primary address of 1838 the exiting interface. 1839 1840 If enabled, the message will be sent with the primary address of 1841 the interface that received the packet that caused the icmp error. 1842 This is the behaviour many network administrators will expect from 1843 a router. And it can make debugging complicated network layouts 1844 much easier. 1845 1846 Note that if no primary address exists for the interface selected, 1847 then the primary address of the first non-loopback interface that 1848 has one will be used regardless of this setting. 1849 1850 Possible values: 1851 1852 - 0 (disabled) 1853 - 1 (enabled) 1854 1855 Default: 0 (disabled) 1856 1857icmp_errors_extension_mask - UNSIGNED INTEGER 1858 Bitmask of ICMP extensions to append to ICMPv4 error messages 1859 ("Destination Unreachable", "Time Exceeded" and "Parameter Problem"). 1860 The original datagram is trimmed / padded to 128 bytes in order to be 1861 compatible with applications that do not comply with RFC 4884. 1862 1863 Possible extensions are: 1864 1865 ==== ============================================================== 1866 0x01 Incoming IP interface information according to RFC 5837. 1867 Extension will include the index, IPv4 address (if present), 1868 name and MTU of the IP interface that received the datagram 1869 which elicited the ICMP error. 1870 ==== ============================================================== 1871 1872 Default: 0x00 (no extensions) 1873 1874igmp_max_memberships - INTEGER 1875 Change the maximum number of multicast groups we can subscribe to. 1876 Default: 20 1877 1878 Theoretical maximum value is bounded by having to send a membership 1879 report in a single datagram (i.e. the report can't span multiple 1880 datagrams, or risk confusing the switch and leaving groups you don't 1881 intend to). 1882 1883 The number of supported groups 'M' is bounded by the number of group 1884 report entries you can fit into a single datagram of 65535 bytes. 1885 1886 M = 65536-sizeof (ip header)/(sizeof(Group record)) 1887 1888 Group records are variable length, with a minimum of 12 bytes. 1889 So net.ipv4.igmp_max_memberships should not be set higher than: 1890 1891 (65536-24) / 12 = 5459 1892 1893 The value 5459 assumes no IP header options, so in practice 1894 this number may be lower. 1895 1896igmp_max_msf - INTEGER 1897 Maximum number of addresses allowed in the source filter list for a 1898 multicast group. 1899 1900 Default: 10 1901 1902igmp_qrv - INTEGER 1903 Controls the IGMP query robustness variable (see RFC2236 8.1). 1904 1905 Default: 2 (as specified by RFC2236 8.1) 1906 1907 Minimum: 1 (as specified by RFC6636 4.5) 1908 1909force_igmp_version - INTEGER 1910 - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback 1911 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier 1912 Present timer expires. 1913 - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if 1914 receive IGMPv2/v3 query. 1915 - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive 1916 IGMPv1 query message. Will reply report if receive IGMPv3 query. 1917 - 3 - Enforce to use IGMP version 3. The same react with default 0. 1918 1919 .. note:: 1920 1921 this is not the same with force_mld_version because IGMPv3 RFC3376 1922 Security Considerations does not have clear description that we could 1923 ignore other version messages completely as MLDv2 RFC3810. So make 1924 this value as default 0 is recommended. 1925 1926``conf/interface/*`` 1927 changes special settings per interface (where 1928 interface" is the name of your network interface) 1929 1930``conf/all/*`` 1931 is special, changes the settings for all interfaces 1932 1933log_martians - BOOLEAN 1934 Log packets with impossible addresses to kernel log. 1935 log_martians for the interface will be enabled if at least one of 1936 conf/{all,interface}/log_martians is set to TRUE, 1937 it will be disabled otherwise 1938 1939accept_redirects - BOOLEAN 1940 Accept ICMP redirect messages. 1941 accept_redirects for the interface will be enabled if: 1942 1943 - both conf/{all,interface}/accept_redirects are TRUE in the case 1944 forwarding for the interface is enabled 1945 1946 or 1947 1948 - at least one of conf/{all,interface}/accept_redirects is TRUE in the 1949 case forwarding for the interface is disabled 1950 1951 accept_redirects for the interface will be disabled otherwise 1952 1953 default: 1954 1955 - TRUE (host) 1956 - FALSE (router) 1957 1958forwarding - BOOLEAN 1959 Enable IP forwarding on this interface. This controls whether packets 1960 received _on_ this interface can be forwarded. 1961 1962mc_forwarding - BOOLEAN 1963 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE 1964 and a multicast routing daemon is required. 1965 conf/all/mc_forwarding must also be set to TRUE to enable multicast 1966 routing for the interface 1967 1968medium_id - INTEGER 1969 Integer value used to differentiate the devices by the medium they 1970 are attached to. Two devices can have different id values when 1971 the broadcast packets are received only on one of them. 1972 The default value 0 means that the device is the only interface 1973 to its medium, value of -1 means that medium is not known. 1974 1975 Currently, it is used to change the proxy_arp behavior: 1976 the proxy_arp feature is enabled for packets forwarded between 1977 two devices attached to different media. 1978 1979proxy_arp - BOOLEAN 1980 Do proxy arp. 1981 1982 proxy_arp for the interface will be enabled if at least one of 1983 conf/{all,interface}/proxy_arp is set to TRUE, 1984 it will be disabled otherwise 1985 1986proxy_arp_pvlan - BOOLEAN 1987 Private VLAN proxy arp. 1988 1989 Basically allow proxy arp replies back to the same interface 1990 (from which the ARP request/solicitation was received). 1991 1992 This is done to support (ethernet) switch features, like RFC 1993 3069, where the individual ports are NOT allowed to 1994 communicate with each other, but they are allowed to talk to 1995 the upstream router. As described in RFC 3069, it is possible 1996 to allow these hosts to communicate through the upstream 1997 router by proxy_arp'ing. Don't need to be used together with 1998 proxy_arp. 1999 2000 This technology is known by different names: 2001 2002 - In RFC 3069 it is called VLAN Aggregation. 2003 - Cisco and Allied Telesyn call it Private VLAN. 2004 - Hewlett-Packard call it Source-Port filtering or port-isolation. 2005 - Ericsson call it MAC-Forced Forwarding (RFC Draft). 2006 2007proxy_delay - INTEGER 2008 Delay proxy response. 2009 2010 Delay response to a neighbor solicitation when proxy_arp 2011 or proxy_ndp is enabled. A random value between [0, proxy_delay) 2012 will be chosen, setting to zero means reply with no delay. 2013 Value in jiffies. Defaults to 80. 2014 2015shared_media - BOOLEAN 2016 Send(router) or accept(host) RFC1620 shared media redirects. 2017 Overrides secure_redirects. 2018 2019 shared_media for the interface will be enabled if at least one of 2020 conf/{all,interface}/shared_media is set to TRUE, 2021 it will be disabled otherwise 2022 2023 default TRUE 2024 2025secure_redirects - BOOLEAN 2026 Accept ICMP redirect messages only to gateways listed in the 2027 interface's current gateway list. Even if disabled, RFC1122 redirect 2028 rules still apply. 2029 2030 Overridden by shared_media. 2031 2032 secure_redirects for the interface will be enabled if at least one of 2033 conf/{all,interface}/secure_redirects is set to TRUE, 2034 it will be disabled otherwise 2035 2036 default TRUE 2037 2038send_redirects - BOOLEAN 2039 Send redirects, if router. 2040 2041 send_redirects for the interface will be enabled if at least one of 2042 conf/{all,interface}/send_redirects is set to TRUE, 2043 it will be disabled otherwise 2044 2045 Default: TRUE 2046 2047bootp_relay - BOOLEAN 2048 Accept packets with source address 0.b.c.d destined 2049 not to this host as local ones. It is supposed, that 2050 BOOTP relay daemon will catch and forward such packets. 2051 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay 2052 for the interface 2053 2054 default FALSE 2055 2056 Not Implemented Yet. 2057 2058accept_source_route - BOOLEAN 2059 Accept packets with SRR option. 2060 conf/all/accept_source_route must also be set to TRUE to accept packets 2061 with SRR option on the interface 2062 2063 default 2064 2065 - TRUE (router) 2066 - FALSE (host) 2067 2068accept_local - BOOLEAN 2069 Accept packets with local source addresses. In combination with 2070 suitable routing, this can be used to direct packets between two 2071 local interfaces over the wire and have them accepted properly. 2072 default FALSE 2073 2074route_localnet - BOOLEAN 2075 Do not consider loopback addresses as martian source or destination 2076 while routing. This enables the use of 127/8 for local routing purposes. 2077 2078 default FALSE 2079 2080rp_filter - INTEGER 2081 - 0 - No source validation. 2082 - 1 - Strict mode as defined in RFC3704 Strict Reverse Path 2083 Each incoming packet is tested against the FIB and if the interface 2084 is not the best reverse path the packet check will fail. 2085 By default failed packets are discarded. 2086 - 2 - Loose mode as defined in RFC3704 Loose Reverse Path 2087 Each incoming packet's source address is also tested against the FIB 2088 and if the source address is not reachable via any interface 2089 the packet check will fail. 2090 2091 Current recommended practice in RFC3704 is to enable strict mode 2092 to prevent IP spoofing from DDos attacks. If using asymmetric routing 2093 or other complicated routing, then loose mode is recommended. 2094 2095 The max value from conf/{all,interface}/rp_filter is used 2096 when doing source validation on the {interface}. 2097 2098 Default value is 0. Note that some distributions enable it 2099 in startup scripts. 2100 2101src_valid_mark - BOOLEAN 2102 - 0 - The fwmark of the packet is not included in reverse path 2103 route lookup. This allows for asymmetric routing configurations 2104 utilizing the fwmark in only one direction, e.g., transparent 2105 proxying. 2106 2107 - 1 - The fwmark of the packet is included in reverse path route 2108 lookup. This permits rp_filter to function when the fwmark is 2109 used for routing traffic in both directions. 2110 2111 This setting also affects the utilization of fmwark when 2112 performing source address selection for ICMP replies, or 2113 determining addresses stored for the IPOPT_TS_TSANDADDR and 2114 IPOPT_RR IP options. 2115 2116 The max value from conf/{all,interface}/src_valid_mark is used. 2117 2118 Default value is 0. 2119 2120arp_filter - BOOLEAN 2121 - 1 - Allows you to have multiple network interfaces on the same 2122 subnet, and have the ARPs for each interface be answered 2123 based on whether or not the kernel would route a packet from 2124 the ARP'd IP out that interface (therefore you must use source 2125 based routing for this to work). In other words it allows control 2126 of which cards (usually 1) will respond to an arp request. 2127 2128 - 0 - (default) The kernel can respond to arp requests with addresses 2129 from other interfaces. This may seem wrong but it usually makes 2130 sense, because it increases the chance of successful communication. 2131 IP addresses are owned by the complete host on Linux, not by 2132 particular interfaces. Only for more complex setups like load- 2133 balancing, does this behaviour cause problems. 2134 2135 arp_filter for the interface will be enabled if at least one of 2136 conf/{all,interface}/arp_filter is set to TRUE, 2137 it will be disabled otherwise 2138 2139arp_announce - INTEGER 2140 Define different restriction levels for announcing the local 2141 source IP address from IP packets in ARP requests sent on 2142 interface: 2143 2144 - 0 - (default) Use any local address, configured on any interface 2145 - 1 - Try to avoid local addresses that are not in the target's 2146 subnet for this interface. This mode is useful when target 2147 hosts reachable via this interface require the source IP 2148 address in ARP requests to be part of their logical network 2149 configured on the receiving interface. When we generate the 2150 request we will check all our subnets that include the 2151 target IP and will preserve the source address if it is from 2152 such subnet. If there is no such subnet we select source 2153 address according to the rules for level 2. 2154 - 2 - Always use the best local address for this target. 2155 In this mode we ignore the source address in the IP packet 2156 and try to select local address that we prefer for talks with 2157 the target host. Such local address is selected by looking 2158 for primary IP addresses on all our subnets on the outgoing 2159 interface that include the target IP address. If no suitable 2160 local address is found we select the first local address 2161 we have on the outgoing interface or on all other interfaces, 2162 with the hope we will receive reply for our request and 2163 even sometimes no matter the source IP address we announce. 2164 2165 The max value from conf/{all,interface}/arp_announce is used. 2166 2167 Increasing the restriction level gives more chance for 2168 receiving answer from the resolved target while decreasing 2169 the level announces more valid sender's information. 2170 2171arp_ignore - INTEGER 2172 Define different modes for sending replies in response to 2173 received ARP requests that resolve local target IP addresses: 2174 2175 - 0 - (default): reply for any local target IP address, configured 2176 on any interface 2177 - 1 - reply only if the target IP address is local address 2178 configured on the incoming interface 2179 - 2 - reply only if the target IP address is local address 2180 configured on the incoming interface and both with the 2181 sender's IP address are part from same subnet on this interface 2182 - 3 - do not reply for local addresses configured with scope host, 2183 only resolutions for global and link addresses are replied 2184 - 4-7 - reserved 2185 - 8 - do not reply for all local addresses 2186 2187 The max value from conf/{all,interface}/arp_ignore is used 2188 when ARP request is received on the {interface} 2189 2190arp_notify - BOOLEAN 2191 Define mode for notification of address and device changes. 2192 2193 == ========================================================== 2194 0 (default): do nothing 2195 1 Generate gratuitous arp requests when device is brought up 2196 or hardware address changes. 2197 == ========================================================== 2198 2199arp_accept - INTEGER 2200 Define behavior for accepting gratuitous ARP (garp) frames from devices 2201 that are not already present in the ARP table: 2202 2203 - 0 - don't create new entries in the ARP table 2204 - 1 - create new entries in the ARP table 2205 - 2 - create new entries only if the source IP address is in the same 2206 subnet as an address configured on the interface that received the 2207 garp message. 2208 2209 Both replies and requests type gratuitous arp will trigger the 2210 ARP table to be updated, if this setting is on. 2211 2212 If the ARP table already contains the IP address of the 2213 gratuitous arp frame, the arp table will be updated regardless 2214 if this setting is on or off. 2215 2216arp_evict_nocarrier - BOOLEAN 2217 Clears the ARP cache on NOCARRIER events. This option is important for 2218 wireless devices where the ARP cache should not be cleared when roaming 2219 between access points on the same network. In most cases this should 2220 remain as the default (1). 2221 2222 Possible values: 2223 2224 - 0 (disabled) - Do not clear ARP cache on NOCARRIER events 2225 - 1 (enabled) - Clear the ARP cache on NOCARRIER events 2226 2227 Default: 1 (enabled) 2228 2229mcast_solicit - INTEGER 2230 The maximum number of multicast probes in INCOMPLETE state, 2231 when the associated hardware address is unknown. Defaults 2232 to 3. 2233 2234ucast_solicit - INTEGER 2235 The maximum number of unicast probes in PROBE state, when 2236 the hardware address is being reconfirmed. Defaults to 3. 2237 2238app_solicit - INTEGER 2239 The maximum number of probes to send to the user space ARP daemon 2240 via netlink before dropping back to multicast probes (see 2241 mcast_resolicit). Defaults to 0. 2242 2243mcast_resolicit - INTEGER 2244 The maximum number of multicast probes after unicast and 2245 app probes in PROBE state. Defaults to 0. 2246 2247disable_policy - BOOLEAN 2248 Disable IPSEC policy (SPD) for this interface 2249 2250 Possible values: 2251 2252 - 0 (disabled) 2253 - 1 (enabled) 2254 2255 Default: 0 (disabled) 2256 2257disable_xfrm - BOOLEAN 2258 Disable IPSEC encryption on this interface, whatever the policy 2259 2260 Possible values: 2261 2262 - 0 (disabled) 2263 - 1 (enabled) 2264 2265 Default: 0 (disabled) 2266 2267igmpv2_unsolicited_report_interval - INTEGER 2268 The interval in milliseconds in which the next unsolicited 2269 IGMPv1 or IGMPv2 report retransmit will take place. 2270 2271 Default: 10000 (10 seconds) 2272 2273igmpv3_unsolicited_report_interval - INTEGER 2274 The interval in milliseconds in which the next unsolicited 2275 IGMPv3 report retransmit will take place. 2276 2277 Default: 1000 (1 seconds) 2278 2279ignore_routes_with_linkdown - BOOLEAN 2280 Ignore routes whose link is down when performing a FIB lookup. 2281 2282 Possible values: 2283 2284 - 0 (disabled) 2285 - 1 (enabled) 2286 2287 Default: 0 (disabled) 2288 2289promote_secondaries - BOOLEAN 2290 When a primary IP address is removed from this interface 2291 promote a corresponding secondary IP address instead of 2292 removing all the corresponding secondary IP addresses. 2293 2294 Possible values: 2295 2296 - 0 (disabled) 2297 - 1 (enabled) 2298 2299 Default: 0 (disabled) 2300 2301drop_unicast_in_l2_multicast - BOOLEAN 2302 Drop any unicast IP packets that are received in link-layer 2303 multicast (or broadcast) frames. 2304 2305 This behavior (for multicast) is actually a SHOULD in RFC 2306 1122, but is disabled by default for compatibility reasons. 2307 2308 Possible values: 2309 2310 - 0 (disabled) 2311 - 1 (enabled) 2312 2313 Default: 0 (disabled) 2314 2315drop_gratuitous_arp - BOOLEAN 2316 Drop all gratuitous ARP frames, for example if there's a known 2317 good ARP proxy on the network and such frames need not be used 2318 (or in the case of 802.11, must not be used to prevent attacks.) 2319 2320 Possible values: 2321 2322 - 0 (disabled) 2323 - 1 (enabled) 2324 2325 Default: 0 (disabled) 2326 2327 2328tag - INTEGER 2329 Allows you to write a number, which can be used as required. 2330 2331 Default value is 0. 2332 2333xfrm4_gc_thresh - INTEGER 2334 (Obsolete since linux-4.14) 2335 The threshold at which we will start garbage collecting for IPv4 2336 destination cache entries. At twice this value the system will 2337 refuse new allocations. 2338 2339igmp_link_local_mcast_reports - BOOLEAN 2340 Enable IGMP reports for link local multicast groups in the 2341 224.0.0.X range. 2342 2343 Default TRUE 2344 2345Alexey Kuznetsov. 2346kuznet@ms2.inr.ac.ru 2347 2348Updated by: 2349 2350- Andi Kleen 2351 ak@muc.de 2352- Nicolas Delon 2353 delon.nicolas@wanadoo.fr 2354 2355 2356 2357 2358/proc/sys/net/ipv6/* Variables 2359============================== 2360 2361IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also 2362apply to IPv6 [XXX?]. 2363 2364bindv6only - BOOLEAN 2365 Default value for IPV6_V6ONLY socket option, 2366 which restricts use of the IPv6 socket to IPv6 communication 2367 only. 2368 2369 Possible values: 2370 2371 - 0 (disabled) - enable IPv4-mapped address feature 2372 - 1 (enabled) - disable IPv4-mapped address feature 2373 2374 Default: 0 (disabled) 2375 2376flowlabel_consistency - BOOLEAN 2377 Protect the consistency (and unicity) of flow label. 2378 You have to disable it to use IPV6_FL_F_REFLECT flag on the 2379 flow label manager. 2380 2381 Possible values: 2382 2383 - 0 (disabled) 2384 - 1 (enabled) 2385 2386 Default: 1 (enabled) 2387 2388auto_flowlabels - INTEGER 2389 Automatically generate flow labels based on a flow hash of the 2390 packet. This allows intermediate devices, such as routers, to 2391 identify packet flows for mechanisms like Equal Cost Multipath 2392 Routing (see RFC 6438). 2393 2394 = =========================================================== 2395 0 automatic flow labels are completely disabled 2396 1 automatic flow labels are enabled by default, they can be 2397 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL 2398 socket option 2399 2 automatic flow labels are allowed, they may be enabled on a 2400 per socket basis using the IPV6_AUTOFLOWLABEL socket option 2401 3 automatic flow labels are enabled and enforced, they cannot 2402 be disabled by the socket option 2403 = =========================================================== 2404 2405 Default: 1 2406 2407flowlabel_state_ranges - BOOLEAN 2408 Split the flow label number space into two ranges. 0-0x7FFFF is 2409 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF 2410 is reserved for stateless flow labels as described in RFC6437. 2411 2412 Possible values: 2413 2414 - 0 (disabled) 2415 - 1 (enabled) 2416 2417 Default: 1 (enabled) 2418 2419 2420flowlabel_reflect - INTEGER 2421 Control flow label reflection. Needed for Path MTU 2422 Discovery to work with Equal Cost Multipath Routing in anycast 2423 environments. See RFC 7690 and: 2424 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01 2425 2426 This is a bitmask. 2427 2428 - 1: enabled for established flows 2429 2430 Note that this prevents automatic flowlabel changes, as done 2431 in "tcp: change IPv6 flow-label upon receiving spurious retransmission" 2432 and "tcp: Change txhash on every SYN and RTO retransmit" 2433 2434 - 2: enabled for TCP RESET packets (no active listener) 2435 If set, a RST packet sent in response to a SYN packet on a closed 2436 port will reflect the incoming flow label. 2437 2438 - 4: enabled for ICMPv6 echo reply messages. 2439 2440 Default: 0 2441 2442fib_multipath_hash_policy - INTEGER 2443 Controls which hash policy to use for multipath routes. 2444 2445 Default: 0 (Layer 3) 2446 2447 Possible values: 2448 2449 - 0 - Layer 3 (source and destination addresses plus flow label). 2450 For IPv6 TCP, the local ECMP path is selected from the socket 2451 txhash rather than the flow label, and may change after a TCP 2452 rehash event (such as a retransmission timeout) to recover from 2453 path failure. The on-wire flow label is unaffected. 2454 - 1 - Layer 4 (standard 5-tuple) 2455 - 2 - Layer 3 or inner Layer 3 if present 2456 - 3 - Custom multipath hash. Fields used for multipath hash calculation 2457 are determined by fib_multipath_hash_fields sysctl 2458 2459fib_multipath_hash_fields - UNSIGNED INTEGER 2460 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the 2461 fields used for multipath hash calculation are determined by this 2462 sysctl. 2463 2464 This value is a bitmask which enables various fields for multipath hash 2465 calculation. 2466 2467 Possible fields are: 2468 2469 ====== ============================ 2470 0x0001 Source IP address 2471 0x0002 Destination IP address 2472 0x0004 IP protocol 2473 0x0008 Flow Label 2474 0x0010 Source port 2475 0x0020 Destination port 2476 0x0040 Inner source IP address 2477 0x0080 Inner destination IP address 2478 0x0100 Inner IP protocol 2479 0x0200 Inner Flow Label 2480 0x0400 Inner source port 2481 0x0800 Inner destination port 2482 ====== ============================ 2483 2484 Default: 0x0007 (source IP, destination IP and IP protocol) 2485 2486anycast_src_echo_reply - BOOLEAN 2487 Controls the use of anycast addresses as source addresses for ICMPv6 2488 echo reply 2489 2490 Possible values: 2491 2492 - 0 (disabled) 2493 - 1 (enabled) 2494 2495 Default: 0 (disabled) 2496 2497 2498idgen_delay - INTEGER 2499 Controls the delay in seconds after which time to retry 2500 privacy stable address generation if a DAD conflict is 2501 detected. 2502 2503 Default: 1 (as specified in RFC7217) 2504 2505idgen_retries - INTEGER 2506 Controls the number of retries to generate a stable privacy 2507 address if a DAD conflict is detected. 2508 2509 Default: 3 (as specified in RFC7217) 2510 2511mld_qrv - INTEGER 2512 Controls the MLD query robustness variable (see RFC3810 9.1). 2513 2514 Default: 2 (as specified by RFC3810 9.1) 2515 2516 Minimum: 1 (as specified by RFC6636 4.5) 2517 2518max_dst_opts_number - INTEGER 2519 Maximum number of non-padding TLVs allowed in a Destination 2520 options extension header. If this value is less than zero 2521 then unknown options are disallowed and the number of known 2522 TLVs allowed is the absolute value of this number. 2523 2524 Default: 8 2525 2526max_hbh_opts_number - INTEGER 2527 Maximum number of non-padding TLVs allowed in a Hop-by-Hop 2528 options extension header. If this value is less than zero 2529 then unknown options are disallowed and the number of known 2530 TLVs allowed is the absolute value of this number. 2531 2532 Default: 8 2533 2534max_dst_opts_length - INTEGER 2535 Maximum length allowed for a Destination options extension 2536 header. 2537 2538 Default: INT_MAX (unlimited) 2539 2540max_hbh_length - INTEGER 2541 Maximum length allowed for a Hop-by-Hop options extension 2542 header. 2543 2544 Default: INT_MAX (unlimited) 2545 2546skip_notify_on_dev_down - BOOLEAN 2547 Controls whether an RTM_DELROUTE message is generated for routes 2548 removed when a device is taken down or deleted. IPv4 does not 2549 generate this message; IPv6 does by default. Setting this sysctl 2550 to true skips the message, making IPv4 and IPv6 on par in relying 2551 on userspace caches to track link events and evict routes. 2552 2553 Possible values: 2554 2555 - 0 (disabled) - generate the message 2556 - 1 (enabled) - skip generating the message 2557 2558 Default: 0 (disabled) 2559 2560nexthop_compat_mode - BOOLEAN 2561 New nexthop API provides a means for managing nexthops independent of 2562 prefixes. Backwards compatibility with old route format is enabled by 2563 default which means route dumps and notifications contain the new 2564 nexthop attribute but also the full, expanded nexthop definition. 2565 Further, updates or deletes of a nexthop configuration generate route 2566 notifications for each fib entry using the nexthop. Once a system 2567 understands the new API, this sysctl can be disabled to achieve full 2568 performance benefits of the new API by disabling the nexthop expansion 2569 and extraneous notifications. 2570 2571 Note that as a backward-compatible mode, dumping of modern features 2572 might be incomplete or wrong. For example, resilient groups will not be 2573 shown as such, but rather as just a list of next hops. Also weights that 2574 do not fit into 8 bits will show incorrectly. 2575 2576 Default: true (backward compat mode) 2577 2578fib_notify_on_flag_change - INTEGER 2579 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/ 2580 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed. 2581 2582 After installing a route to the kernel, user space receives an 2583 acknowledgment, which means the route was installed in the kernel, 2584 but not necessarily in hardware. 2585 It is also possible for a route already installed in hardware to change 2586 its action and therefore its flags. For example, a host route that is 2587 trapping packets can be "promoted" to perform decapsulation following 2588 the installation of an IPinIP/VXLAN tunnel. 2589 The notifications will indicate to user-space the state of the route. 2590 2591 Default: 0 (Do not emit notifications.) 2592 2593 Possible values: 2594 2595 - 0 - Do not emit notifications. 2596 - 1 - Emit notifications. 2597 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change. 2598 2599ioam6_id - INTEGER 2600 Define the IOAM id of this node. Uses only 24 bits out of 32 in total. 2601 2602 Possible value range: 2603 2604 - Min: 0 2605 - Max: 0xFFFFFF 2606 2607 Default: 0xFFFFFF 2608 2609ioam6_id_wide - LONG INTEGER 2610 Define the wide IOAM id of this node. Uses only 56 bits out of 64 in 2611 total. Can be different from ioam6_id. 2612 2613 Possible value range: 2614 2615 - Min: 0 2616 - Max: 0xFFFFFFFFFFFFFF 2617 2618 Default: 0xFFFFFFFFFFFFFF 2619 2620IPv6 Fragmentation: 2621 2622ip6frag_high_thresh - INTEGER 2623 Maximum memory used to reassemble IPv6 fragments. When 2624 ip6frag_high_thresh bytes of memory is allocated for this purpose, 2625 the fragment handler will toss packets until ip6frag_low_thresh 2626 is reached. 2627 2628ip6frag_low_thresh - INTEGER 2629 See ip6frag_high_thresh 2630 2631ip6frag_time - INTEGER 2632 Time in seconds to keep an IPv6 fragment in memory. 2633 2634``conf/default/*``: 2635 Change the interface-specific default settings. 2636 2637 These settings would be used during creating new interfaces. 2638 2639 2640``conf/all/*``: 2641 Change all the interface-specific settings. 2642 2643 [XXX: Other special features than forwarding?] 2644 2645conf/all/disable_ipv6 - BOOLEAN 2646 Changing this value is same as changing ``conf/default/disable_ipv6`` 2647 setting and also all per-interface ``disable_ipv6`` settings to the same 2648 value. 2649 2650 Reading this value does not have any particular meaning. It does not say 2651 whether IPv6 support is enabled or disabled. Returned value can be 1 2652 also in the case when some interface has ``disable_ipv6`` set to 0 and 2653 has configured IPv6 addresses. 2654 2655conf/all/forwarding - BOOLEAN 2656 Enable global IPv6 forwarding between all interfaces. 2657 2658 IPv4 and IPv6 work differently here; the ``force_forwarding`` flag must 2659 be used to control which interfaces may forward packets. 2660 2661 This also sets all interfaces' Host/Router setting 2662 'forwarding' to the specified value. See below for details. 2663 2664 This referred to as global forwarding. 2665 2666proxy_ndp - BOOLEAN 2667 Do proxy ndp. 2668 2669 Possible values: 2670 2671 - 0 (disabled) 2672 - 1 (enabled) 2673 2674 Default: 0 (disabled) 2675 2676force_forwarding - BOOLEAN 2677 Enable forwarding on this interface only -- regardless of the setting on 2678 ``conf/all/forwarding``. When setting ``conf.all.forwarding`` to 0, 2679 the ``force_forwarding`` flag will be reset on all interfaces. 2680 2681fwmark_reflect - BOOLEAN 2682 Controls the fwmark of kernel-generated IPv6 reply packets that are not 2683 associated with a socket for example, TCP RSTs or ICMPv6 echo replies). 2684 If disabled, these packets have a fwmark of zero. If enabled, they have the 2685 fwmark of the packet they are replying to. 2686 2687 Possible values: 2688 2689 - 0 (disabled) 2690 - 1 (enabled) 2691 2692 Default: 0 (disabled) 2693 2694``conf/interface/*``: 2695 Change special settings per interface. 2696 2697 The functional behaviour for certain settings is different 2698 depending on whether local forwarding is enabled or not. 2699 2700accept_ra - INTEGER 2701 Accept Router Advertisements; autoconfigure using them. 2702 2703 It also determines whether or not to transmit Router 2704 Solicitations. If and only if the functional setting is to 2705 accept Router Advertisements, Router Solicitations will be 2706 transmitted. 2707 2708 Possible values are: 2709 2710 == =========================================================== 2711 0 Do not accept Router Advertisements. 2712 1 Accept Router Advertisements if forwarding is disabled. 2713 2 Overrule forwarding behaviour. Accept Router Advertisements 2714 even if forwarding is enabled. 2715 == =========================================================== 2716 2717 Functional default: 2718 2719 - enabled if local forwarding is disabled. 2720 - disabled if local forwarding is enabled. 2721 2722accept_ra_defrtr - BOOLEAN 2723 Learn default router in Router Advertisement. 2724 2725 Functional default: 2726 2727 - enabled if accept_ra is enabled. 2728 - disabled if accept_ra is disabled. 2729 2730ra_defrtr_metric - UNSIGNED INTEGER 2731 Route metric for default route learned in Router Advertisement. This value 2732 will be assigned as metric for the default route learned via IPv6 Router 2733 Advertisement. Takes affect only if accept_ra_defrtr is enabled. 2734 2735 Possible values: 2736 1 to 0xFFFFFFFF 2737 2738 Default: IP6_RT_PRIO_USER i.e. 1024. 2739 2740accept_ra_from_local - BOOLEAN 2741 Accept RA with source-address that is found on local machine 2742 if the RA is otherwise proper and able to be accepted. 2743 2744 Default is to NOT accept these as it may be an un-intended 2745 network loop. 2746 2747 Functional default: 2748 2749 - enabled if accept_ra_from_local is enabled 2750 on a specific interface. 2751 - disabled if accept_ra_from_local is disabled 2752 on a specific interface. 2753 2754accept_ra_min_hop_limit - INTEGER 2755 Minimum hop limit Information in Router Advertisement. 2756 2757 Hop limit Information in Router Advertisement less than this 2758 variable shall be ignored. 2759 2760 Default: 1 2761 2762accept_ra_min_lft - INTEGER 2763 Minimum acceptable lifetime value in Router Advertisement. 2764 2765 RA sections with a lifetime less than this value shall be 2766 ignored. Zero lifetimes stay unaffected. 2767 2768 Default: 0 2769 2770accept_ra_pinfo - BOOLEAN 2771 Learn Prefix Information in Router Advertisement. 2772 2773 Functional default: 2774 2775 - enabled if accept_ra is enabled. 2776 - disabled if accept_ra is disabled. 2777 2778ra_honor_pio_life - BOOLEAN 2779 Whether to use RFC4862 Section 5.5.3e to determine the valid 2780 lifetime of an address matching a prefix sent in a Router 2781 Advertisement Prefix Information Option. 2782 2783 Possible values: 2784 2785 - 0 (disabled) - RFC4862 section 5.5.3e is used to determine 2786 the valid lifetime of the address. 2787 - 1 (enabled) - the PIO valid lifetime will always be honored. 2788 2789 Default: 0 (disabled) 2790 2791ra_honor_pio_pflag - BOOLEAN 2792 The Prefix Information Option P-flag indicates the network can 2793 allocate a unique IPv6 prefix per client using DHCPv6-PD. 2794 This sysctl can be enabled when a userspace DHCPv6-PD client 2795 is running to cause the P-flag to take effect: i.e. the 2796 P-flag suppresses any effects of the A-flag within the same 2797 PIO. For a given PIO, P=1 and A=1 is treated as A=0. 2798 2799 Possible values: 2800 2801 - 0 (disabled) - the P-flag is ignored. 2802 - 1 (enabled) - the P-flag will disable SLAAC autoconfiguration 2803 for the given Prefix Information Option. 2804 2805 Default: 0 (disabled) 2806 2807accept_ra_rt_info_min_plen - INTEGER 2808 Minimum prefix length of Route Information in RA. 2809 2810 Route Information w/ prefix smaller than this variable shall 2811 be ignored. 2812 2813 Functional default: 2814 2815 * 0 if accept_ra_rtr_pref is enabled. 2816 * -1 if accept_ra_rtr_pref is disabled. 2817 2818accept_ra_rt_info_max_plen - INTEGER 2819 Maximum prefix length of Route Information in RA. 2820 2821 Route Information w/ prefix larger than this variable shall 2822 be ignored. 2823 2824 Functional default: 2825 2826 * 0 if accept_ra_rtr_pref is enabled. 2827 * -1 if accept_ra_rtr_pref is disabled. 2828 2829accept_ra_rtr_pref - BOOLEAN 2830 Accept Router Preference in RA. 2831 2832 Functional default: 2833 2834 - enabled if accept_ra is enabled. 2835 - disabled if accept_ra is disabled. 2836 2837accept_ra_mtu - BOOLEAN 2838 Apply the MTU value specified in RA option 5 (RFC4861). If 2839 disabled, the MTU specified in the RA will be ignored. 2840 2841 Functional default: 2842 2843 - enabled if accept_ra is enabled. 2844 - disabled if accept_ra is disabled. 2845 2846accept_redirects - BOOLEAN 2847 Accept Redirects. 2848 2849 Functional default: 2850 2851 - enabled if local forwarding is disabled. 2852 - disabled if local forwarding is enabled. 2853 2854accept_source_route - INTEGER 2855 Accept source routing (routing extension header). 2856 2857 - >= 0: Accept only routing header type 2. 2858 - < 0: Do not accept routing header. 2859 2860 Default: 0 2861 2862autoconf - BOOLEAN 2863 Autoconfigure addresses using Prefix Information in Router 2864 Advertisements. 2865 2866 Functional default: 2867 2868 - enabled if accept_ra_pinfo is enabled. 2869 - disabled if accept_ra_pinfo is disabled. 2870 2871dad_transmits - INTEGER 2872 The amount of Duplicate Address Detection probes to send. 2873 2874 Default: 1 2875 2876forwarding - INTEGER 2877 Configure interface-specific Host/Router behaviour. 2878 2879 .. note:: 2880 2881 It is recommended to have the same setting on all 2882 interfaces; mixed router/host scenarios are rather uncommon. 2883 2884 Possible values are: 2885 2886 - 0 Forwarding disabled 2887 - 1 Forwarding enabled 2888 2889 **FALSE (0)**: 2890 2891 By default, Host behaviour is assumed. This means: 2892 2893 1. IsRouter flag is not set in Neighbour Advertisements. 2894 2. If accept_ra is TRUE (default), transmit Router 2895 Solicitations. 2896 3. If accept_ra is TRUE (default), accept Router 2897 Advertisements (and do autoconfiguration). 2898 4. If accept_redirects is TRUE (default), accept Redirects. 2899 2900 **TRUE (1)**: 2901 2902 If local forwarding is enabled, Router behaviour is assumed. 2903 This means exactly the reverse from the above: 2904 2905 1. IsRouter flag is set in Neighbour Advertisements. 2906 2. Router Solicitations are not sent unless accept_ra is 2. 2907 3. Router Advertisements are ignored unless accept_ra is 2. 2908 4. Redirects are ignored. 2909 2910 Default: 0 (disabled) if global forwarding is disabled (default), 2911 otherwise 1 (enabled). 2912 2913hop_limit - INTEGER 2914 Default Hop Limit to set. 2915 2916 Default: 64 2917 2918mtu - INTEGER 2919 Default Maximum Transfer Unit 2920 2921 Default: 1280 (IPv6 required minimum) 2922 2923ip_nonlocal_bind - BOOLEAN 2924 If enabled, allows processes to bind() to non-local IPv6 addresses, 2925 which can be quite useful - but may break some applications. 2926 2927 Possible values: 2928 2929 - 0 (disabled) 2930 - 1 (enabled) 2931 2932 Default: 0 (disabled) 2933 2934router_probe_interval - INTEGER 2935 Minimum interval (in seconds) between Router Probing described 2936 in RFC4191. 2937 2938 Default: 60 2939 2940router_solicitation_delay - INTEGER 2941 Number of seconds to wait after interface is brought up 2942 before sending Router Solicitations. 2943 2944 Default: 1 2945 2946router_solicitation_interval - INTEGER 2947 Number of seconds to wait between Router Solicitations. 2948 2949 Default: 4 2950 2951router_solicitations - INTEGER 2952 Number of Router Solicitations to send until assuming no 2953 routers are present. 2954 2955 Default: 3 2956 2957use_oif_addrs_only - BOOLEAN 2958 When enabled, the candidate source addresses for destinations 2959 routed via this interface are restricted to the set of addresses 2960 configured on this interface (vis. RFC 6724, section 4). 2961 2962 Possible values: 2963 2964 - 0 (disabled) 2965 - 1 (enabled) 2966 2967 Default: 0 (disabled) 2968 2969use_tempaddr - INTEGER 2970 Preference for Privacy Extensions (RFC3041). 2971 2972 * <= 0 : disable Privacy Extensions 2973 * == 1 : enable Privacy Extensions, but prefer public 2974 addresses over temporary addresses. 2975 * > 1 : enable Privacy Extensions and prefer temporary 2976 addresses over public addresses. 2977 2978 Default: 2979 2980 * 0 (for most devices) 2981 * -1 (for point-to-point devices and loopback devices) 2982 2983temp_valid_lft - INTEGER 2984 valid lifetime (in seconds) for temporary addresses. If less than the 2985 minimum required lifetime (typically 5-7 seconds), temporary addresses 2986 will not be created. 2987 2988 Default: 172800 (2 days) 2989 2990temp_prefered_lft - INTEGER 2991 Preferred lifetime (in seconds) for temporary addresses. If 2992 temp_prefered_lft is less than the minimum required lifetime (typically 2993 5-7 seconds), the preferred lifetime is the minimum required. If 2994 temp_prefered_lft is greater than temp_valid_lft, the preferred lifetime 2995 is temp_valid_lft. 2996 2997 Default: 86400 (1 day) 2998 2999keep_addr_on_down - INTEGER 3000 Keep all IPv6 addresses on an interface down event. If set static 3001 global addresses with no expiration time are not flushed. 3002 3003 * >0 : enabled 3004 * 0 : system default 3005 * <0 : disabled 3006 3007 Default: 0 (addresses are removed) 3008 3009max_desync_factor - INTEGER 3010 Maximum value for DESYNC_FACTOR, which is a random value 3011 that ensures that clients don't synchronize with each 3012 other and generate new addresses at exactly the same time. 3013 value is in seconds. 3014 3015 Default: 600 3016 3017regen_min_advance - INTEGER 3018 How far in advance (in seconds), at minimum, to create a new temporary 3019 address before the current one is deprecated. This value is added to 3020 the amount of time that may be required for duplicate address detection 3021 to determine when to create a new address. Linux permits setting this 3022 value to less than the default of 2 seconds, but a value less than 2 3023 does not conform to RFC 8981. 3024 3025 Default: 2 3026 3027regen_max_retry - INTEGER 3028 Number of attempts before give up attempting to generate 3029 valid temporary addresses. 3030 3031 Default: 5 3032 3033max_addresses - INTEGER 3034 Maximum number of autoconfigured addresses per interface. Setting 3035 to zero disables the limitation. It is not recommended to set this 3036 value too large (or to zero) because it would be an easy way to 3037 crash the kernel by allowing too many addresses to be created. 3038 3039 Default: 16 3040 3041disable_ipv6 - BOOLEAN 3042 Disable IPv6 operation. If accept_dad is set to 2, this value 3043 will be dynamically set to TRUE if DAD fails for the link-local 3044 address. 3045 3046 Default: FALSE (enable IPv6 operation) 3047 3048 When this value is changed from 1 to 0 (IPv6 is being enabled), 3049 it will dynamically create a link-local address on the given 3050 interface and start Duplicate Address Detection, if necessary. 3051 3052 When this value is changed from 0 to 1 (IPv6 is being disabled), 3053 it will dynamically delete all addresses and routes on the given 3054 interface. From now on it will not possible to add addresses/routes 3055 to the selected interface. 3056 3057accept_dad - INTEGER 3058 Whether to accept DAD (Duplicate Address Detection). 3059 3060 == ============================================================== 3061 0 Disable DAD 3062 1 Enable DAD (default) 3063 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate 3064 link-local address has been found. 3065 == ============================================================== 3066 3067 DAD operation and mode on a given interface will be selected according 3068 to the maximum value of conf/{all,interface}/accept_dad. 3069 3070force_tllao - BOOLEAN 3071 Enable sending the target link-layer address option even when 3072 responding to a unicast neighbor solicitation. 3073 3074 Default: FALSE 3075 3076 Quoting from RFC 2461, section 4.4, Target link-layer address: 3077 3078 "The option MUST be included for multicast solicitations in order to 3079 avoid infinite Neighbor Solicitation "recursion" when the peer node 3080 does not have a cache entry to return a Neighbor Advertisements 3081 message. When responding to unicast solicitations, the option can be 3082 omitted since the sender of the solicitation has the correct link- 3083 layer address; otherwise it would not have be able to send the unicast 3084 solicitation in the first place. However, including the link-layer 3085 address in this case adds little overhead and eliminates a potential 3086 race condition where the sender deletes the cached link-layer address 3087 prior to receiving a response to a previous solicitation." 3088 3089ndisc_notify - BOOLEAN 3090 Define mode for notification of address and device changes. 3091 3092 Possible values: 3093 3094 - 0 (disabled) - do nothing 3095 - 1 (enabled) - Generate unsolicited neighbour advertisements when device is brought 3096 up or hardware address changes. 3097 3098 Default: 0 (disabled) 3099 3100ndisc_tclass - INTEGER 3101 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor 3102 Discovery (Router Solicitation, Router Advertisement, Neighbor 3103 Solicitation, Neighbor Advertisement, Redirect) messages. 3104 These 8 bits can be interpreted as 6 high order bits holding the DSCP 3105 value and 2 low order bits representing ECN (which you probably want 3106 to leave cleared). 3107 3108 * 0 - (default) 3109 3110ndisc_evict_nocarrier - BOOLEAN 3111 Clears the neighbor discovery table on NOCARRIER events. This option is 3112 important for wireless devices where the neighbor discovery cache should 3113 not be cleared when roaming between access points on the same network. 3114 In most cases this should remain as the default (1). 3115 3116 Possible values: 3117 3118 - 0 (disabled) - Do not clear neighbor discovery cache on NOCARRIER events. 3119 - 1 (enabled) - Clear neighbor discover cache on NOCARRIER events. 3120 3121 Default: 1 (enabled) 3122 3123mldv1_unsolicited_report_interval - INTEGER 3124 The interval in milliseconds in which the next unsolicited 3125 MLDv1 report retransmit will take place. 3126 3127 Default: 10000 (10 seconds) 3128 3129mldv2_unsolicited_report_interval - INTEGER 3130 The interval in milliseconds in which the next unsolicited 3131 MLDv2 report retransmit will take place. 3132 3133 Default: 1000 (1 second) 3134 3135force_mld_version - INTEGER 3136 * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed 3137 * 1 - Enforce to use MLD version 1 3138 * 2 - Enforce to use MLD version 2 3139 3140suppress_frag_ndisc - INTEGER 3141 Control RFC 6980 (Security Implications of IPv6 Fragmentation 3142 with IPv6 Neighbor Discovery) behavior: 3143 3144 * 1 - (default) discard fragmented neighbor discovery packets 3145 * 0 - allow fragmented neighbor discovery packets 3146 3147optimistic_dad - BOOLEAN 3148 Whether to perform Optimistic Duplicate Address Detection (RFC 4429). 3149 3150 Optimistic Duplicate Address Detection for the interface will be enabled 3151 if at least one of conf/{all,interface}/optimistic_dad is set to 1, 3152 it will be disabled otherwise. 3153 3154 Possible values: 3155 3156 - 0 (disabled) 3157 - 1 (enabled) 3158 3159 Default: 0 (disabled) 3160 3161 3162use_optimistic - BOOLEAN 3163 If enabled, do not classify optimistic addresses as deprecated during 3164 source address selection. Preferred addresses will still be chosen 3165 before optimistic addresses, subject to other ranking in the source 3166 address selection algorithm. 3167 3168 This will be enabled if at least one of 3169 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise. 3170 3171 Possible values: 3172 3173 - 0 (disabled) 3174 - 1 (enabled) 3175 3176 Default: 0 (disabled) 3177 3178stable_secret - IPv6 address 3179 This IPv6 address will be used as a secret to generate IPv6 3180 addresses for link-local addresses and autoconfigured 3181 ones. All addresses generated after setting this secret will 3182 be stable privacy ones by default. This can be changed via the 3183 addrgenmode ip-link. conf/default/stable_secret is used as the 3184 secret for the namespace, the interface specific ones can 3185 overwrite that. Writes to conf/all/stable_secret are refused. 3186 3187 It is recommended to generate this secret during installation 3188 of a system and keep it stable after that. 3189 3190 By default the stable secret is unset. 3191 3192addr_gen_mode - INTEGER 3193 Defines how link-local and autoconf addresses are generated. 3194 3195 = ================================================================= 3196 0 generate address based on EUI64 (default) 3197 1 do no generate a link-local address, use EUI64 for addresses 3198 generated from autoconf 3199 2 generate stable privacy addresses, using the secret from 3200 stable_secret (RFC7217) 3201 3 generate stable privacy addresses, using a random secret if unset 3202 = ================================================================= 3203 3204drop_unicast_in_l2_multicast - BOOLEAN 3205 Drop any unicast IPv6 packets that are received in link-layer 3206 multicast (or broadcast) frames. 3207 3208 Possible values: 3209 3210 - 0 (disabled) 3211 - 1 (enabled) 3212 3213 Default: 0 (disabled) 3214 3215drop_unsolicited_na - BOOLEAN 3216 Drop all unsolicited neighbor advertisements, for example if there's 3217 a known good NA proxy on the network and such frames need not be used 3218 (or in the case of 802.11, must not be used to prevent attacks.) 3219 3220 Possible values: 3221 3222 - 0 (disabled) 3223 - 1 (enabled) 3224 3225 Default: 0 (disabled). 3226 3227accept_untracked_na - INTEGER 3228 Define behavior for accepting neighbor advertisements from devices that 3229 are absent in the neighbor cache: 3230 3231 - 0 - (default) Do not accept unsolicited and untracked neighbor 3232 advertisements. 3233 3234 - 1 - Add a new neighbor cache entry in STALE state for routers on 3235 receiving a neighbor advertisement (either solicited or unsolicited) 3236 with target link-layer address option specified if no neighbor entry 3237 is already present for the advertised IPv6 address. Without this knob, 3238 NAs received for untracked addresses (absent in neighbor cache) are 3239 silently ignored. 3240 3241 This is as per router-side behavior documented in RFC9131. 3242 3243 This has lower precedence than drop_unsolicited_na. 3244 3245 This will optimize the return path for the initial off-link 3246 communication that is initiated by a directly connected host, by 3247 ensuring that the first-hop router which turns on this setting doesn't 3248 have to buffer the initial return packets to do neighbor-solicitation. 3249 The prerequisite is that the host is configured to send unsolicited 3250 neighbor advertisements on interface bringup. This setting should be 3251 used in conjunction with the ndisc_notify setting on the host to 3252 satisfy this prerequisite. 3253 3254 - 2 - Extend option (1) to add a new neighbor cache entry only if the 3255 source IP address is in the same subnet as an address configured on 3256 the interface that received the neighbor advertisement. 3257 3258enhanced_dad - BOOLEAN 3259 Include a nonce option in the IPv6 neighbor solicitation messages used for 3260 duplicate address detection per RFC7527. A received DAD NS will only signal 3261 a duplicate address if the nonce is different. This avoids any false 3262 detection of duplicates due to loopback of the NS messages that we send. 3263 The nonce option will be sent on an interface unless both of 3264 conf/{all,interface}/enhanced_dad are set to FALSE. 3265 3266 Possible values: 3267 3268 - 0 (disabled) 3269 - 1 (enabled) 3270 3271 Default: 1 (enabled) 3272 3273``icmp/*``: 3274=========== 3275 3276ratelimit - INTEGER 3277 Limit the maximal rates for sending ICMPv6 messages to a particular 3278 peer. 3279 3280 0 to disable any limiting, 3281 otherwise the space between responses in milliseconds. 3282 3283 Default: 100 3284 3285ratemask - list of comma separated ranges 3286 For ICMPv6 message types matching the ranges in the ratemask, limit 3287 the sending of the message according to ratelimit parameter. 3288 3289 The format used for both input and output is a comma separated 3290 list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and 3291 129). Writing to the file will clear all previous ranges of ICMPv6 3292 message types and update the current list with the input. 3293 3294 Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml 3295 for numerical values of ICMPv6 message types, e.g. echo request is 128 3296 and echo reply is 129. 3297 3298 Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big) 3299 3300echo_ignore_all - BOOLEAN 3301 If enabled, then the kernel will ignore all ICMP ECHO 3302 requests sent to it over the IPv6 protocol. 3303 3304 Possible values: 3305 3306 - 0 (disabled) 3307 - 1 (enabled) 3308 3309 Default: 0 (disabled) 3310 3311echo_ignore_multicast - BOOLEAN 3312 If enabled, then the kernel will ignore all ICMP ECHO 3313 requests sent to it over the IPv6 protocol via multicast. 3314 3315 Possible values: 3316 3317 - 0 (disabled) 3318 - 1 (enabled) 3319 3320 Default: 0 (disabled) 3321 3322echo_ignore_anycast - BOOLEAN 3323 If enabled, then the kernel will ignore all ICMP ECHO 3324 requests sent to it over the IPv6 protocol destined to anycast address. 3325 3326 Possible values: 3327 3328 - 0 (disabled) 3329 - 1 (enabled) 3330 3331 Default: 0 (disabled) 3332 3333error_anycast_as_unicast - BOOLEAN 3334 If enabled, then the kernel will respond with ICMP Errors 3335 resulting from requests sent to it over the IPv6 protocol destined 3336 to anycast address essentially treating anycast as unicast. 3337 3338 Possible values: 3339 3340 - 0 (disabled) 3341 - 1 (enabled) 3342 3343 Default: 0 (disabled) 3344 3345errors_extension_mask - UNSIGNED INTEGER 3346 Bitmask of ICMP extensions to append to ICMPv6 error messages 3347 ("Destination Unreachable" and "Time Exceeded"). The original datagram 3348 is trimmed / padded to 128 bytes in order to be compatible with 3349 applications that do not comply with RFC 4884. 3350 3351 Possible extensions are: 3352 3353 ==== ============================================================== 3354 0x01 Incoming IP interface information according to RFC 5837. 3355 Extension will include the index, IPv6 address (if present), 3356 name and MTU of the IP interface that received the datagram 3357 which elicited the ICMP error. 3358 ==== ============================================================== 3359 3360 Default: 0x00 (no extensions) 3361 3362xfrm6_gc_thresh - INTEGER 3363 (Obsolete since linux-4.14) 3364 The threshold at which we will start garbage collecting for IPv6 3365 destination cache entries. At twice this value the system will 3366 refuse new allocations. 3367 3368 3369IPv6 Update by: 3370Pekka Savola <pekkas@netcore.fi> 3371YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org> 3372 3373 3374/proc/sys/net/bridge/* Variables: 3375================================= 3376 3377bridge-nf-call-arptables - BOOLEAN 3378 3379 Possible values: 3380 3381 - 0 (disabled) - disable this. 3382 - 1 (enabled) - pass bridged ARP traffic to arptables' FORWARD chain. 3383 3384 Default: 1 (enabled) 3385 3386bridge-nf-call-iptables - BOOLEAN 3387 3388 Possible values: 3389 3390 - 0 (disabled) - disable this. 3391 - 1 (enabled) - pass bridged IPv4 traffic to iptables' chains. 3392 3393 Default: 1 (enabled) 3394 3395bridge-nf-call-ip6tables - BOOLEAN 3396 3397 Possible values: 3398 3399 - 0 (disabled) - disable this. 3400 - 1 (enabled) - pass bridged IPv6 traffic to ip6tables' chains. 3401 3402 Default: 1 (enabled) 3403 3404bridge-nf-filter-vlan-tagged - BOOLEAN 3405 3406 Possible values: 3407 3408 - 0 (disabled) - disable this. 3409 - 1 (enabled) - pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables 3410 3411 Default: 0 (disabled) 3412 3413bridge-nf-filter-pppoe-tagged - BOOLEAN 3414 3415 Possible values: 3416 3417 - 0 (disabled) - disable this. 3418 - 1 (enabled) - pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables. 3419 3420 Default: 0 (disabled) 3421 3422bridge-nf-pass-vlan-input-dev - BOOLEAN 3423 - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan 3424 interface on the bridge and set the netfilter input device to the 3425 vlan. This allows use of e.g. "iptables -i br0.1" and makes the 3426 REDIRECT target work with vlan-on-top-of-bridge interfaces. When no 3427 matching vlan interface is found, or this switch is off, the input 3428 device is set to the bridge interface. 3429 3430 - 0: disable bridge netfilter vlan interface lookup. 3431 3432 Default: 0 3433 3434``proc/sys/net/sctp/*`` Variables: 3435================================== 3436 3437addip_enable - BOOLEAN 3438 Enable or disable extension of Dynamic Address Reconfiguration 3439 (ADD-IP) functionality specified in RFC5061. This extension provides 3440 the ability to dynamically add and remove new addresses for the SCTP 3441 associations. 3442 3443 Possible values: 3444 3445 - 0 (disabled) - disable extension. 3446 - 1 (enabled) - enable extension 3447 3448 Default: 0 (disabled) 3449 3450pf_enable - INTEGER 3451 Enable or disable pf (pf is short for potentially failed) state. A value 3452 of pf_retrans > path_max_retrans also disables pf state. That is, one of 3453 both pf_enable and pf_retrans > path_max_retrans can disable pf state. 3454 Since pf_retrans and path_max_retrans can be changed by userspace 3455 application, sometimes user expects to disable pf state by the value of 3456 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans 3457 or path_max_retrans is changed by the user application, this pf state is 3458 enabled. As such, it is necessary to add this to dynamically enable 3459 and disable pf state. See: 3460 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for 3461 details. 3462 3463 Possible values: 3464 3465 - 1: Enable pf. 3466 - 0: Disable pf. 3467 3468 Default: 1 3469 3470pf_expose - INTEGER 3471 Unset or enable/disable pf (pf is short for potentially failed) state 3472 exposure. Applications can control the exposure of the PF path state 3473 in the SCTP_PEER_ADDR_CHANGE event and access of SCTP_PF-state 3474 transport info via SCTP_GET_PEER_ADDR_INFO sockopt. 3475 3476 Possible values: 3477 3478 - 0: Unset pf state exposure (compatible with old applications). No 3479 event will be sent but the transport info can be queried. 3480 - 1: Disable pf state exposure. No event will be sent and trying to 3481 obtain transport info will return -EACCESS. 3482 - 2: Enable pf state exposure. The event will be sent for a transport 3483 becoming SCTP_PF state and transport info can be obtained. 3484 3485 Default: 0 3486 3487addip_noauth_enable - BOOLEAN 3488 Dynamic Address Reconfiguration (ADD-IP) requires the use of 3489 authentication to protect the operations of adding or removing new 3490 addresses. This requirement is mandated so that unauthorized hosts 3491 would not be able to hijack associations. However, older 3492 implementations may not have implemented this requirement while 3493 allowing the ADD-IP extension. For reasons of interoperability, 3494 we provide this variable to control the enforcement of the 3495 authentication requirement. 3496 3497 == =============================================================== 3498 1 Allow ADD-IP extension to be used without authentication. This 3499 should only be set in a closed environment for interoperability 3500 with older implementations. 3501 3502 0 Enforce the authentication requirement 3503 == =============================================================== 3504 3505 Default: 0 3506 3507auth_enable - BOOLEAN 3508 Enable or disable Authenticated Chunks extension. This extension 3509 provides the ability to send and receive authenticated chunks and is 3510 required for secure operation of Dynamic Address Reconfiguration 3511 (ADD-IP) extension. 3512 3513 Possible values: 3514 3515 - 0 (disabled) - disable extension. 3516 - 1 (enabled) - enable extension 3517 3518 Default: 0 (disabled) 3519 3520prsctp_enable - BOOLEAN 3521 Enable or disable the Partial Reliability extension (RFC3758) which 3522 is used to notify peers that a given DATA should no longer be expected. 3523 3524 Possible values: 3525 3526 - 0 (disabled) - disable extension. 3527 - 1 (enabled) - enable extension 3528 3529 Default: 1 (enabled) 3530 3531max_burst - INTEGER 3532 The limit of the number of new packets that can be initially sent. It 3533 controls how bursty the generated traffic can be. 3534 3535 Default: 4 3536 3537association_max_retrans - INTEGER 3538 Set the maximum number for retransmissions that an association can 3539 attempt deciding that the remote end is unreachable. If this value 3540 is exceeded, the association is terminated. 3541 3542 Default: 10 3543 3544max_init_retransmits - INTEGER 3545 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks 3546 that an association will attempt before declaring the destination 3547 unreachable and terminating. 3548 3549 Default: 8 3550 3551path_max_retrans - INTEGER 3552 The maximum number of retransmissions that will be attempted on a given 3553 path. Once this threshold is exceeded, the path is considered 3554 unreachable, and new traffic will use a different path when the 3555 association is multihomed. 3556 3557 Default: 5 3558 3559pf_retrans - INTEGER 3560 The number of retransmissions that will be attempted on a given path 3561 before traffic is redirected to an alternate transport (should one 3562 exist). Note this is distinct from path_max_retrans, as a path that 3563 passes the pf_retrans threshold can still be used. Its only 3564 deprioritized when a transmission path is selected by the stack. This 3565 setting is primarily used to enable fast failover mechanisms without 3566 having to reduce path_max_retrans to a very low value. See: 3567 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt 3568 for details. Note also that a value of pf_retrans > path_max_retrans 3569 disables this feature. Since both pf_retrans and path_max_retrans can 3570 be changed by userspace application, a variable pf_enable is used to 3571 disable pf state. 3572 3573 Default: 0 3574 3575ps_retrans - INTEGER 3576 Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming 3577 from section-5 "Primary Path Switchover" in rfc7829. The primary path 3578 will be changed to another active path when the path error counter on 3579 the old primary path exceeds PSMR, so that "the SCTP sender is allowed 3580 to continue data transmission on a new working path even when the old 3581 primary destination address becomes active again". Note this feature 3582 is disabled by initializing 'ps_retrans' per netns as 0xffff by default, 3583 and its value can't be less than 'pf_retrans' when changing by sysctl. 3584 3585 Default: 0xffff 3586 3587rto_initial - INTEGER 3588 The initial round trip timeout value in milliseconds that will be used 3589 in calculating round trip times. This is the initial time interval 3590 for retransmissions. 3591 3592 Default: 3000 3593 3594rto_max - INTEGER 3595 The maximum value (in milliseconds) of the round trip timeout. This 3596 is the largest time interval that can elapse between retransmissions. 3597 3598 Default: 60000 3599 3600rto_min - INTEGER 3601 The minimum value (in milliseconds) of the round trip timeout. This 3602 is the smallest time interval the can elapse between retransmissions. 3603 3604 Default: 1000 3605 3606hb_interval - INTEGER 3607 The interval (in milliseconds) between HEARTBEAT chunks. These chunks 3608 are sent at the specified interval on idle paths to probe the state of 3609 a given path between 2 associations. 3610 3611 Default: 30000 3612 3613sack_timeout - INTEGER 3614 The amount of time (in milliseconds) that the implementation will wait 3615 to send a SACK. 3616 3617 Default: 200 3618 3619valid_cookie_life - INTEGER 3620 The default lifetime of the SCTP cookie (in milliseconds). The cookie 3621 is used during association establishment. 3622 3623 Default: 60000 3624 3625cookie_preserve_enable - BOOLEAN 3626 Enable or disable the ability to extend the lifetime of the SCTP cookie 3627 that is used during the establishment phase of SCTP association 3628 3629 Possible values: 3630 3631 - 0 (disabled) - disable. 3632 - 1 (enabled) - enable cookie lifetime extension. 3633 3634 Default: 1 (enabled) 3635 3636cookie_hmac_alg - STRING 3637 Select the hmac algorithm used when generating the cookie value sent by 3638 a listening sctp socket to a connecting client in the INIT-ACK chunk. 3639 Valid values are: 3640 3641 * sha256 3642 * none 3643 3644 Default: sha256 3645 3646rcvbuf_policy - INTEGER 3647 Determines if the receive buffer is attributed to the socket or to 3648 association. SCTP supports the capability to create multiple 3649 associations on a single socket. When using this capability, it is 3650 possible that a single stalled association that's buffering a lot 3651 of data may block other associations from delivering their data by 3652 consuming all of the receive buffer space. To work around this, 3653 the rcvbuf_policy could be set to attribute the receiver buffer space 3654 to each association instead of the socket. This prevents the described 3655 blocking. 3656 3657 - 1: rcvbuf space is per association 3658 - 0: rcvbuf space is per socket 3659 3660 Default: 0 3661 3662sndbuf_policy - INTEGER 3663 Similar to rcvbuf_policy above, this applies to send buffer space. 3664 3665 - 1: Send buffer is tracked per association 3666 - 0: Send buffer is tracked per socket. 3667 3668 Default: 0 3669 3670sctp_mem - vector of 3 INTEGERs: min, pressure, max 3671 Number of pages allowed for queueing by all SCTP sockets. 3672 3673 * min: Below this number of pages SCTP is not bothered about its 3674 memory usage. When amount of memory allocated by SCTP exceeds 3675 this number, SCTP starts to moderate memory usage. 3676 * pressure: This value was introduced to follow format of tcp_mem. 3677 * max: Maximum number of allowed pages. 3678 3679 Default is calculated at boot time from amount of available memory. 3680 3681sctp_rmem - vector of 3 INTEGERs: min, default, max 3682 Only the first value ("min") is used, "default" and "max" are 3683 ignored. 3684 3685 * min: Minimal size of receive buffer used by SCTP socket. 3686 It is guaranteed to each SCTP socket (but not association) even 3687 under moderate memory pressure. 3688 3689 Default: 4K 3690 3691sctp_wmem - vector of 3 INTEGERs: min, default, max 3692 Only the first value ("min") is used, "default" and "max" are 3693 ignored. 3694 3695 * min: Minimum size of send buffer that can be used by SCTP sockets. 3696 It is guaranteed to each SCTP socket (but not association) even 3697 under moderate memory pressure. 3698 3699 Default: 4K 3700 3701addr_scope_policy - INTEGER 3702 Control IPv4 address scoping (see 3703 https://datatracker.ietf.org/doc/draft-stewart-tsvwg-sctp-ipv4/00/ 3704 for details). 3705 3706 - 0 - Disable IPv4 address scoping 3707 - 1 - Enable IPv4 address scoping 3708 - 2 - Follow draft but allow IPv4 private addresses 3709 - 3 - Follow draft but allow IPv4 link local addresses 3710 3711 Default: 1 3712 3713udp_port - INTEGER 3714 The listening port for the local UDP tunneling sock. Normally it's 3715 using the IANA-assigned UDP port number 9899 (sctp-tunneling). 3716 3717 This UDP sock is used for processing the incoming UDP-encapsulated 3718 SCTP packets (from RFC6951), and shared by all applications in the 3719 same net namespace. This UDP sock will be closed when the value is 3720 set to 0. 3721 3722 The value will also be used to set the src port of the UDP header 3723 for the outgoing UDP-encapsulated SCTP packets. For the dest port, 3724 please refer to 'encap_port' below. 3725 3726 Default: 0 3727 3728encap_port - INTEGER 3729 The default remote UDP encapsulation port. 3730 3731 This value is used to set the dest port of the UDP header for the 3732 outgoing UDP-encapsulated SCTP packets by default. Users can also 3733 change the value for each sock/asoc/transport by using setsockopt. 3734 For further information, please refer to RFC6951. 3735 3736 Note that when connecting to a remote server, the client should set 3737 this to the port that the UDP tunneling sock on the peer server is 3738 listening to and the local UDP tunneling sock on the client also 3739 must be started. On the server, it would get the encap_port from 3740 the incoming packet's source port. 3741 3742 Default: 0 3743 3744plpmtud_probe_interval - INTEGER 3745 The time interval (in milliseconds) for the PLPMTUD probe timer, 3746 which is configured to expire after this period to receive an 3747 acknowledgment to a probe packet. This is also the time interval 3748 between the probes for the current pmtu when the probe search 3749 is done. 3750 3751 PLPMTUD will be disabled when 0 is set, and other values for it 3752 must be >= 5000. 3753 3754 Default: 0 3755 3756reconf_enable - BOOLEAN 3757 Enable or disable extension of Stream Reconfiguration functionality 3758 specified in RFC6525. This extension provides the ability to "reset" 3759 a stream, and it includes the Parameters of "Outgoing/Incoming SSN 3760 Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams". 3761 3762 Possible values: 3763 3764 - 0 (disabled) - Disable extension. 3765 - 1 (enabled) - Enable extension. 3766 3767 Default: 0 (disabled) 3768 3769intl_enable - BOOLEAN 3770 Enable or disable extension of User Message Interleaving functionality 3771 specified in RFC8260. This extension allows the interleaving of user 3772 messages sent on different streams. With this feature enabled, I-DATA 3773 chunk will replace DATA chunk to carry user messages if also supported 3774 by the peer. Note that to use this feature, one needs to set this option 3775 to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2 3776 and SCTP_INTERLEAVING_SUPPORTED to 1. 3777 3778 Possible values: 3779 3780 - 0 (disabled) - Disable extension. 3781 - 1 (enabled) - Enable extension. 3782 3783 Default: 0 (disabled) 3784 3785ecn_enable - BOOLEAN 3786 Control use of Explicit Congestion Notification (ECN) by SCTP. 3787 Like in TCP, ECN is used only when both ends of the SCTP connection 3788 indicate support for it. This feature is useful in avoiding losses 3789 due to congestion by allowing supporting routers to signal congestion 3790 before having to drop packets. 3791 3792 Possible values: 3793 3794 - 0 (disabled) - Disable ecn. 3795 - 1 (enabled) - Enable ecn. 3796 3797 Default: 1 (enabled) 3798 3799l3mdev_accept - BOOLEAN 3800 Enabling this option allows a "global" bound socket to work 3801 across L3 master domains (e.g., VRFs) with packets capable of 3802 being received regardless of the L3 domain in which they 3803 originated. Only valid when the kernel was compiled with 3804 CONFIG_NET_L3_MASTER_DEV. 3805 3806 Possible values: 3807 3808 - 0 (disabled) 3809 - 1 (enabled) 3810 3811 Default: 1 (enabled) 3812 3813 3814``/proc/sys/net/core/*`` 3815======================== 3816 3817 Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries. 3818 3819 3820``/proc/sys/net/unix/*`` 3821======================== 3822 3823max_dgram_qlen - INTEGER 3824 The maximum length of dgram socket receive queue 3825 3826 Default: 10 3827 3828