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