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 - 0 - disabled (default) 12 - not 0 - enabled 13 14 Forward Packets between interfaces. 15 16 This variable is special, its change resets all configuration 17 parameters to their default state (RFC1122 for hosts, RFC1812 18 for routers) 19 20ip_default_ttl - INTEGER 21 Default value of TTL field (Time To Live) for outgoing (but not 22 forwarded) IP packets. Should be between 1 and 255 inclusive. 23 Default: 64 (as recommended by RFC1700) 24 25ip_no_pmtu_disc - INTEGER 26 Disable Path MTU Discovery. If enabled in mode 1 and a 27 fragmentation-required ICMP is received, the PMTU to this 28 destination will be set to the smallest of the old MTU to 29 this destination and min_pmtu (see below). You will need 30 to raise min_pmtu to the smallest interface MTU on your system 31 manually if you want to avoid locally generated fragments. 32 33 In mode 2 incoming Path MTU Discovery messages will be 34 discarded. Outgoing frames are handled the same as in mode 1, 35 implicitly setting IP_PMTUDISC_DONT on every created socket. 36 37 Mode 3 is a hardened pmtu discover mode. The kernel will only 38 accept fragmentation-needed errors if the underlying protocol 39 can verify them besides a plain socket lookup. Current 40 protocols for which pmtu events will be honored are TCP, SCTP 41 and DCCP as they verify e.g. the sequence number or the 42 association. This mode should not be enabled globally but is 43 only intended to secure e.g. name servers in namespaces where 44 TCP path mtu must still work but path MTU information of other 45 protocols should be discarded. If enabled globally this mode 46 could break other protocols. 47 48 Possible values: 0-3 49 50 Default: FALSE 51 52min_pmtu - INTEGER 53 default 552 - minimum Path MTU. Unless this is changed mannually, 54 each cached pmtu will never be lower than this setting. 55 56ip_forward_use_pmtu - BOOLEAN 57 By default we don't trust protocol path MTUs while forwarding 58 because they could be easily forged and can lead to unwanted 59 fragmentation by the router. 60 You only need to enable this if you have user-space software 61 which tries to discover path mtus by itself and depends on the 62 kernel honoring this information. This is normally not the 63 case. 64 65 Default: 0 (disabled) 66 67 Possible values: 68 69 - 0 - disabled 70 - 1 - enabled 71 72fwmark_reflect - BOOLEAN 73 Controls the fwmark of kernel-generated IPv4 reply packets that are not 74 associated with a socket for example, TCP RSTs or ICMP echo replies). 75 If unset, these packets have a fwmark of zero. If set, they have the 76 fwmark of the packet they are replying to. 77 78 Default: 0 79 80fib_multipath_use_neigh - BOOLEAN 81 Use status of existing neighbor entry when determining nexthop for 82 multipath routes. If disabled, neighbor information is not used and 83 packets could be directed to a failed nexthop. Only valid for kernels 84 built with CONFIG_IP_ROUTE_MULTIPATH enabled. 85 86 Default: 0 (disabled) 87 88 Possible values: 89 90 - 0 - disabled 91 - 1 - enabled 92 93fib_multipath_hash_policy - INTEGER 94 Controls which hash policy to use for multipath routes. Only valid 95 for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled. 96 97 Default: 0 (Layer 3) 98 99 Possible values: 100 101 - 0 - Layer 3 102 - 1 - Layer 4 103 - 2 - Layer 3 or inner Layer 3 if present 104 - 3 - Custom multipath hash. Fields used for multipath hash calculation 105 are determined by fib_multipath_hash_fields sysctl 106 107fib_multipath_hash_fields - UNSIGNED INTEGER 108 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the 109 fields used for multipath hash calculation are determined by this 110 sysctl. 111 112 This value is a bitmask which enables various fields for multipath hash 113 calculation. 114 115 Possible fields are: 116 117 ====== ============================ 118 0x0001 Source IP address 119 0x0002 Destination IP address 120 0x0004 IP protocol 121 0x0008 Unused (Flow Label) 122 0x0010 Source port 123 0x0020 Destination port 124 0x0040 Inner source IP address 125 0x0080 Inner destination IP address 126 0x0100 Inner IP protocol 127 0x0200 Inner Flow Label 128 0x0400 Inner source port 129 0x0800 Inner destination port 130 ====== ============================ 131 132 Default: 0x0007 (source IP, destination IP and IP protocol) 133 134fib_sync_mem - UNSIGNED INTEGER 135 Amount of dirty memory from fib entries that can be backlogged before 136 synchronize_rcu is forced. 137 138 Default: 512kB Minimum: 64kB Maximum: 64MB 139 140ip_forward_update_priority - INTEGER 141 Whether to update SKB priority from "TOS" field in IPv4 header after it 142 is forwarded. The new SKB priority is mapped from TOS field value 143 according to an rt_tos2priority table (see e.g. man tc-prio). 144 145 Default: 1 (Update priority.) 146 147 Possible values: 148 149 - 0 - Do not update priority. 150 - 1 - Update priority. 151 152route/max_size - INTEGER 153 Maximum number of routes allowed in the kernel. Increase 154 this when using large numbers of interfaces and/or routes. 155 156 From linux kernel 3.6 onwards, this is deprecated for ipv4 157 as route cache is no longer used. 158 159neigh/default/gc_thresh1 - INTEGER 160 Minimum number of entries to keep. Garbage collector will not 161 purge entries if there are fewer than this number. 162 163 Default: 128 164 165neigh/default/gc_thresh2 - INTEGER 166 Threshold when garbage collector becomes more aggressive about 167 purging entries. Entries older than 5 seconds will be cleared 168 when over this number. 169 170 Default: 512 171 172neigh/default/gc_thresh3 - INTEGER 173 Maximum number of non-PERMANENT neighbor entries allowed. Increase 174 this when using large numbers of interfaces and when communicating 175 with large numbers of directly-connected peers. 176 177 Default: 1024 178 179neigh/default/unres_qlen_bytes - INTEGER 180 The maximum number of bytes which may be used by packets 181 queued for each unresolved address by other network layers. 182 (added in linux 3.3) 183 184 Setting negative value is meaningless and will return error. 185 186 Default: SK_WMEM_MAX, (same as net.core.wmem_default). 187 188 Exact value depends on architecture and kernel options, 189 but should be enough to allow queuing 256 packets 190 of medium size. 191 192neigh/default/unres_qlen - INTEGER 193 The maximum number of packets which may be queued for each 194 unresolved address by other network layers. 195 196 (deprecated in linux 3.3) : use unres_qlen_bytes instead. 197 198 Prior to linux 3.3, the default value is 3 which may cause 199 unexpected packet loss. The current default value is calculated 200 according to default value of unres_qlen_bytes and true size of 201 packet. 202 203 Default: 101 204 205neigh/default/interval_probe_time_ms - INTEGER 206 The probe interval for neighbor entries with NTF_MANAGED flag, 207 the min value is 1. 208 209 Default: 5000 210 211mtu_expires - INTEGER 212 Time, in seconds, that cached PMTU information is kept. 213 214min_adv_mss - INTEGER 215 The advertised MSS depends on the first hop route MTU, but will 216 never be lower than this setting. 217 218fib_notify_on_flag_change - INTEGER 219 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/ 220 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed. 221 222 After installing a route to the kernel, user space receives an 223 acknowledgment, which means the route was installed in the kernel, 224 but not necessarily in hardware. 225 It is also possible for a route already installed in hardware to change 226 its action and therefore its flags. For example, a host route that is 227 trapping packets can be "promoted" to perform decapsulation following 228 the installation of an IPinIP/VXLAN tunnel. 229 The notifications will indicate to user-space the state of the route. 230 231 Default: 0 (Do not emit notifications.) 232 233 Possible values: 234 235 - 0 - Do not emit notifications. 236 - 1 - Emit notifications. 237 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change. 238 239IP Fragmentation: 240 241ipfrag_high_thresh - LONG INTEGER 242 Maximum memory used to reassemble IP fragments. 243 244ipfrag_low_thresh - LONG INTEGER 245 (Obsolete since linux-4.17) 246 Maximum memory used to reassemble IP fragments before the kernel 247 begins to remove incomplete fragment queues to free up resources. 248 The kernel still accepts new fragments for defragmentation. 249 250ipfrag_time - INTEGER 251 Time in seconds to keep an IP fragment in memory. 252 253ipfrag_max_dist - INTEGER 254 ipfrag_max_dist is a non-negative integer value which defines the 255 maximum "disorder" which is allowed among fragments which share a 256 common IP source address. Note that reordering of packets is 257 not unusual, but if a large number of fragments arrive from a source 258 IP address while a particular fragment queue remains incomplete, it 259 probably indicates that one or more fragments belonging to that queue 260 have been lost. When ipfrag_max_dist is positive, an additional check 261 is done on fragments before they are added to a reassembly queue - if 262 ipfrag_max_dist (or more) fragments have arrived from a particular IP 263 address between additions to any IP fragment queue using that source 264 address, it's presumed that one or more fragments in the queue are 265 lost. The existing fragment queue will be dropped, and a new one 266 started. An ipfrag_max_dist value of zero disables this check. 267 268 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can 269 result in unnecessarily dropping fragment queues when normal 270 reordering of packets occurs, which could lead to poor application 271 performance. Using a very large value, e.g. 50000, increases the 272 likelihood of incorrectly reassembling IP fragments that originate 273 from different IP datagrams, which could result in data corruption. 274 Default: 64 275 276bc_forwarding - INTEGER 277 bc_forwarding enables the feature described in rfc1812#section-5.3.5.2 278 and rfc2644. It allows the router to forward directed broadcast. 279 To enable this feature, the 'all' entry and the input interface entry 280 should be set to 1. 281 Default: 0 282 283INET peer storage 284================= 285 286inet_peer_threshold - INTEGER 287 The approximate size of the storage. Starting from this threshold 288 entries will be thrown aggressively. This threshold also determines 289 entries' time-to-live and time intervals between garbage collection 290 passes. More entries, less time-to-live, less GC interval. 291 292inet_peer_minttl - INTEGER 293 Minimum time-to-live of entries. Should be enough to cover fragment 294 time-to-live on the reassembling side. This minimum time-to-live is 295 guaranteed if the pool size is less than inet_peer_threshold. 296 Measured in seconds. 297 298inet_peer_maxttl - INTEGER 299 Maximum time-to-live of entries. Unused entries will expire after 300 this period of time if there is no memory pressure on the pool (i.e. 301 when the number of entries in the pool is very small). 302 Measured in seconds. 303 304TCP variables 305============= 306 307somaxconn - INTEGER 308 Limit of socket listen() backlog, known in userspace as SOMAXCONN. 309 Defaults to 4096. (Was 128 before linux-5.4) 310 See also tcp_max_syn_backlog for additional tuning for TCP sockets. 311 312tcp_abort_on_overflow - BOOLEAN 313 If listening service is too slow to accept new connections, 314 reset them. Default state is FALSE. It means that if overflow 315 occurred due to a burst, connection will recover. Enable this 316 option _only_ if you are really sure that listening daemon 317 cannot be tuned to accept connections faster. Enabling this 318 option can harm clients of your server. 319 320tcp_adv_win_scale - INTEGER 321 Count buffering overhead as bytes/2^tcp_adv_win_scale 322 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale), 323 if it is <= 0. 324 325 Possible values are [-31, 31], inclusive. 326 327 Default: 1 328 329tcp_allowed_congestion_control - STRING 330 Show/set the congestion control choices available to non-privileged 331 processes. The list is a subset of those listed in 332 tcp_available_congestion_control. 333 334 Default is "reno" and the default setting (tcp_congestion_control). 335 336tcp_app_win - INTEGER 337 Reserve max(window/2^tcp_app_win, mss) of window for application 338 buffer. Value 0 is special, it means that nothing is reserved. 339 340 Default: 31 341 342tcp_autocorking - BOOLEAN 343 Enable TCP auto corking : 344 When applications do consecutive small write()/sendmsg() system calls, 345 we try to coalesce these small writes as much as possible, to lower 346 total amount of sent packets. This is done if at least one prior 347 packet for the flow is waiting in Qdisc queues or device transmit 348 queue. Applications can still use TCP_CORK for optimal behavior 349 when they know how/when to uncork their sockets. 350 351 Default : 1 352 353tcp_available_congestion_control - STRING 354 Shows the available congestion control choices that are registered. 355 More congestion control algorithms may be available as modules, 356 but not loaded. 357 358tcp_base_mss - INTEGER 359 The initial value of search_low to be used by the packetization layer 360 Path MTU discovery (MTU probing). If MTU probing is enabled, 361 this is the initial MSS used by the connection. 362 363tcp_mtu_probe_floor - INTEGER 364 If MTU probing is enabled this caps the minimum MSS used for search_low 365 for the connection. 366 367 Default : 48 368 369tcp_min_snd_mss - INTEGER 370 TCP SYN and SYNACK messages usually advertise an ADVMSS option, 371 as described in RFC 1122 and RFC 6691. 372 373 If this ADVMSS option is smaller than tcp_min_snd_mss, 374 it is silently capped to tcp_min_snd_mss. 375 376 Default : 48 (at least 8 bytes of payload per segment) 377 378tcp_congestion_control - STRING 379 Set the congestion control algorithm to be used for new 380 connections. The algorithm "reno" is always available, but 381 additional choices may be available based on kernel configuration. 382 Default is set as part of kernel configuration. 383 For passive connections, the listener congestion control choice 384 is inherited. 385 386 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ] 387 388tcp_dsack - BOOLEAN 389 Allows TCP to send "duplicate" SACKs. 390 391tcp_early_retrans - INTEGER 392 Tail loss probe (TLP) converts RTOs occurring due to tail 393 losses into fast recovery (draft-ietf-tcpm-rack). Note that 394 TLP requires RACK to function properly (see tcp_recovery below) 395 396 Possible values: 397 398 - 0 disables TLP 399 - 3 or 4 enables TLP 400 401 Default: 3 402 403tcp_ecn - INTEGER 404 Control use of Explicit Congestion Notification (ECN) by TCP. 405 ECN is used only when both ends of the TCP connection indicate 406 support for it. This feature is useful in avoiding losses due 407 to congestion by allowing supporting routers to signal 408 congestion before having to drop packets. 409 410 Possible values are: 411 412 = ===================================================== 413 0 Disable ECN. Neither initiate nor accept ECN. 414 1 Enable ECN when requested by incoming connections and 415 also request ECN on outgoing connection attempts. 416 2 Enable ECN when requested by incoming connections 417 but do not request ECN on outgoing connections. 418 = ===================================================== 419 420 Default: 2 421 422tcp_ecn_fallback - BOOLEAN 423 If the kernel detects that ECN connection misbehaves, enable fall 424 back to non-ECN. Currently, this knob implements the fallback 425 from RFC3168, section 6.1.1.1., but we reserve that in future, 426 additional detection mechanisms could be implemented under this 427 knob. The value is not used, if tcp_ecn or per route (or congestion 428 control) ECN settings are disabled. 429 430 Default: 1 (fallback enabled) 431 432tcp_fack - BOOLEAN 433 This is a legacy option, it has no effect anymore. 434 435tcp_fin_timeout - INTEGER 436 The length of time an orphaned (no longer referenced by any 437 application) connection will remain in the FIN_WAIT_2 state 438 before it is aborted at the local end. While a perfectly 439 valid "receive only" state for an un-orphaned connection, an 440 orphaned connection in FIN_WAIT_2 state could otherwise wait 441 forever for the remote to close its end of the connection. 442 443 Cf. tcp_max_orphans 444 445 Default: 60 seconds 446 447tcp_frto - INTEGER 448 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682. 449 F-RTO is an enhanced recovery algorithm for TCP retransmission 450 timeouts. It is particularly beneficial in networks where the 451 RTT fluctuates (e.g., wireless). F-RTO is sender-side only 452 modification. It does not require any support from the peer. 453 454 By default it's enabled with a non-zero value. 0 disables F-RTO. 455 456tcp_fwmark_accept - BOOLEAN 457 If set, incoming connections to listening sockets that do not have a 458 socket mark will set the mark of the accepting socket to the fwmark of 459 the incoming SYN packet. This will cause all packets on that connection 460 (starting from the first SYNACK) to be sent with that fwmark. The 461 listening socket's mark is unchanged. Listening sockets that already 462 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are 463 unaffected. 464 465 Default: 0 466 467tcp_invalid_ratelimit - INTEGER 468 Limit the maximal rate for sending duplicate acknowledgments 469 in response to incoming TCP packets that are for an existing 470 connection but that are invalid due to any of these reasons: 471 472 (a) out-of-window sequence number, 473 (b) out-of-window acknowledgment number, or 474 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure 475 476 This can help mitigate simple "ack loop" DoS attacks, wherein 477 a buggy or malicious middlebox or man-in-the-middle can 478 rewrite TCP header fields in manner that causes each endpoint 479 to think that the other is sending invalid TCP segments, thus 480 causing each side to send an unterminating stream of duplicate 481 acknowledgments for invalid segments. 482 483 Using 0 disables rate-limiting of dupacks in response to 484 invalid segments; otherwise this value specifies the minimal 485 space between sending such dupacks, in milliseconds. 486 487 Default: 500 (milliseconds). 488 489tcp_keepalive_time - INTEGER 490 How often TCP sends out keepalive messages when keepalive is enabled. 491 Default: 2hours. 492 493tcp_keepalive_probes - INTEGER 494 How many keepalive probes TCP sends out, until it decides that the 495 connection is broken. Default value: 9. 496 497tcp_keepalive_intvl - INTEGER 498 How frequently the probes are send out. Multiplied by 499 tcp_keepalive_probes it is time to kill not responding connection, 500 after probes started. Default value: 75sec i.e. connection 501 will be aborted after ~11 minutes of retries. 502 503tcp_l3mdev_accept - BOOLEAN 504 Enables child sockets to inherit the L3 master device index. 505 Enabling this option allows a "global" listen socket to work 506 across L3 master domains (e.g., VRFs) with connected sockets 507 derived from the listen socket to be bound to the L3 domain in 508 which the packets originated. Only valid when the kernel was 509 compiled with CONFIG_NET_L3_MASTER_DEV. 510 511 Default: 0 (disabled) 512 513tcp_low_latency - BOOLEAN 514 This is a legacy option, it has no effect anymore. 515 516tcp_max_orphans - INTEGER 517 Maximal number of TCP sockets not attached to any user file handle, 518 held by system. If this number is exceeded orphaned connections are 519 reset immediately and warning is printed. This limit exists 520 only to prevent simple DoS attacks, you _must_ not rely on this 521 or lower the limit artificially, but rather increase it 522 (probably, after increasing installed memory), 523 if network conditions require more than default value, 524 and tune network services to linger and kill such states 525 more aggressively. Let me to remind again: each orphan eats 526 up to ~64K of unswappable memory. 527 528tcp_max_syn_backlog - INTEGER 529 Maximal number of remembered connection requests (SYN_RECV), 530 which have not received an acknowledgment from connecting client. 531 532 This is a per-listener limit. 533 534 The minimal value is 128 for low memory machines, and it will 535 increase in proportion to the memory of machine. 536 537 If server suffers from overload, try increasing this number. 538 539 Remember to also check /proc/sys/net/core/somaxconn 540 A SYN_RECV request socket consumes about 304 bytes of memory. 541 542tcp_max_tw_buckets - INTEGER 543 Maximal number of timewait sockets held by system simultaneously. 544 If this number is exceeded time-wait socket is immediately destroyed 545 and warning is printed. This limit exists only to prevent 546 simple DoS attacks, you _must_ not lower the limit artificially, 547 but rather increase it (probably, after increasing installed memory), 548 if network conditions require more than default value. 549 550tcp_mem - vector of 3 INTEGERs: min, pressure, max 551 min: below this number of pages TCP is not bothered about its 552 memory appetite. 553 554 pressure: when amount of memory allocated by TCP exceeds this number 555 of pages, TCP moderates its memory consumption and enters memory 556 pressure mode, which is exited when memory consumption falls 557 under "min". 558 559 max: number of pages allowed for queueing by all TCP sockets. 560 561 Defaults are calculated at boot time from amount of available 562 memory. 563 564tcp_min_rtt_wlen - INTEGER 565 The window length of the windowed min filter to track the minimum RTT. 566 A shorter window lets a flow more quickly pick up new (higher) 567 minimum RTT when it is moved to a longer path (e.g., due to traffic 568 engineering). A longer window makes the filter more resistant to RTT 569 inflations such as transient congestion. The unit is seconds. 570 571 Possible values: 0 - 86400 (1 day) 572 573 Default: 300 574 575tcp_moderate_rcvbuf - BOOLEAN 576 If set, TCP performs receive buffer auto-tuning, attempting to 577 automatically size the buffer (no greater than tcp_rmem[2]) to 578 match the size required by the path for full throughput. Enabled by 579 default. 580 581tcp_mtu_probing - INTEGER 582 Controls TCP Packetization-Layer Path MTU Discovery. Takes three 583 values: 584 585 - 0 - Disabled 586 - 1 - Disabled by default, enabled when an ICMP black hole detected 587 - 2 - Always enabled, use initial MSS of tcp_base_mss. 588 589tcp_probe_interval - UNSIGNED INTEGER 590 Controls how often to start TCP Packetization-Layer Path MTU 591 Discovery reprobe. The default is reprobing every 10 minutes as 592 per RFC4821. 593 594tcp_probe_threshold - INTEGER 595 Controls when TCP Packetization-Layer Path MTU Discovery probing 596 will stop in respect to the width of search range in bytes. Default 597 is 8 bytes. 598 599tcp_no_metrics_save - BOOLEAN 600 By default, TCP saves various connection metrics in the route cache 601 when the connection closes, so that connections established in the 602 near future can use these to set initial conditions. Usually, this 603 increases overall performance, but may sometimes cause performance 604 degradation. If set, TCP will not cache metrics on closing 605 connections. 606 607tcp_no_ssthresh_metrics_save - BOOLEAN 608 Controls whether TCP saves ssthresh metrics in the route cache. 609 610 Default is 1, which disables ssthresh metrics. 611 612tcp_orphan_retries - INTEGER 613 This value influences the timeout of a locally closed TCP connection, 614 when RTO retransmissions remain unacknowledged. 615 See tcp_retries2 for more details. 616 617 The default value is 8. 618 619 If your machine is a loaded WEB server, 620 you should think about lowering this value, such sockets 621 may consume significant resources. Cf. tcp_max_orphans. 622 623tcp_recovery - INTEGER 624 This value is a bitmap to enable various experimental loss recovery 625 features. 626 627 ========= ============================================================= 628 RACK: 0x1 enables the RACK loss detection for fast detection of lost 629 retransmissions and tail drops. It also subsumes and disables 630 RFC6675 recovery for SACK connections. 631 632 RACK: 0x2 makes RACK's reordering window static (min_rtt/4). 633 634 RACK: 0x4 disables RACK's DUPACK threshold heuristic 635 ========= ============================================================= 636 637 Default: 0x1 638 639tcp_reflect_tos - BOOLEAN 640 For listening sockets, reuse the DSCP value of the initial SYN message 641 for outgoing packets. This allows to have both directions of a TCP 642 stream to use the same DSCP value, assuming DSCP remains unchanged for 643 the lifetime of the connection. 644 645 This options affects both IPv4 and IPv6. 646 647 Default: 0 (disabled) 648 649tcp_reordering - INTEGER 650 Initial reordering level of packets in a TCP stream. 651 TCP stack can then dynamically adjust flow reordering level 652 between this initial value and tcp_max_reordering 653 654 Default: 3 655 656tcp_max_reordering - INTEGER 657 Maximal reordering level of packets in a TCP stream. 658 300 is a fairly conservative value, but you might increase it 659 if paths are using per packet load balancing (like bonding rr mode) 660 661 Default: 300 662 663tcp_retrans_collapse - BOOLEAN 664 Bug-to-bug compatibility with some broken printers. 665 On retransmit try to send bigger packets to work around bugs in 666 certain TCP stacks. 667 668tcp_retries1 - INTEGER 669 This value influences the time, after which TCP decides, that 670 something is wrong due to unacknowledged RTO retransmissions, 671 and reports this suspicion to the network layer. 672 See tcp_retries2 for more details. 673 674 RFC 1122 recommends at least 3 retransmissions, which is the 675 default. 676 677tcp_retries2 - INTEGER 678 This value influences the timeout of an alive TCP connection, 679 when RTO retransmissions remain unacknowledged. 680 Given a value of N, a hypothetical TCP connection following 681 exponential backoff with an initial RTO of TCP_RTO_MIN would 682 retransmit N times before killing the connection at the (N+1)th RTO. 683 684 The default value of 15 yields a hypothetical timeout of 924.6 685 seconds and is a lower bound for the effective timeout. 686 TCP will effectively time out at the first RTO which exceeds the 687 hypothetical timeout. 688 689 RFC 1122 recommends at least 100 seconds for the timeout, 690 which corresponds to a value of at least 8. 691 692tcp_rfc1337 - BOOLEAN 693 If set, the TCP stack behaves conforming to RFC1337. If unset, 694 we are not conforming to RFC, but prevent TCP TIME_WAIT 695 assassination. 696 697 Default: 0 698 699tcp_rmem - vector of 3 INTEGERs: min, default, max 700 min: Minimal size of receive buffer used by TCP sockets. 701 It is guaranteed to each TCP socket, even under moderate memory 702 pressure. 703 704 Default: 4K 705 706 default: initial size of receive buffer used by TCP sockets. 707 This value overrides net.core.rmem_default used by other protocols. 708 Default: 131072 bytes. 709 This value results in initial window of 65535. 710 711 max: maximal size of receive buffer allowed for automatically 712 selected receiver buffers for TCP socket. This value does not override 713 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables 714 automatic tuning of that socket's receive buffer size, in which 715 case this value is ignored. 716 Default: between 131072 and 6MB, depending on RAM size. 717 718tcp_sack - BOOLEAN 719 Enable select acknowledgments (SACKS). 720 721tcp_comp_sack_delay_ns - LONG INTEGER 722 TCP tries to reduce number of SACK sent, using a timer 723 based on 5% of SRTT, capped by this sysctl, in nano seconds. 724 The default is 1ms, based on TSO autosizing period. 725 726 Default : 1,000,000 ns (1 ms) 727 728tcp_comp_sack_slack_ns - LONG INTEGER 729 This sysctl control the slack used when arming the 730 timer used by SACK compression. This gives extra time 731 for small RTT flows, and reduces system overhead by allowing 732 opportunistic reduction of timer interrupts. 733 734 Default : 100,000 ns (100 us) 735 736tcp_comp_sack_nr - INTEGER 737 Max number of SACK that can be compressed. 738 Using 0 disables SACK compression. 739 740 Default : 44 741 742tcp_slow_start_after_idle - BOOLEAN 743 If set, provide RFC2861 behavior and time out the congestion 744 window after an idle period. An idle period is defined at 745 the current RTO. If unset, the congestion window will not 746 be timed out after an idle period. 747 748 Default: 1 749 750tcp_stdurg - BOOLEAN 751 Use the Host requirements interpretation of the TCP urgent pointer field. 752 Most hosts use the older BSD interpretation, so if you turn this on 753 Linux might not communicate correctly with them. 754 755 Default: FALSE 756 757tcp_synack_retries - INTEGER 758 Number of times SYNACKs for a passive TCP connection attempt will 759 be retransmitted. Should not be higher than 255. Default value 760 is 5, which corresponds to 31seconds till the last retransmission 761 with the current initial RTO of 1second. With this the final timeout 762 for a passive TCP connection will happen after 63seconds. 763 764tcp_syncookies - INTEGER 765 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES 766 Send out syncookies when the syn backlog queue of a socket 767 overflows. This is to prevent against the common 'SYN flood attack' 768 Default: 1 769 770 Note, that syncookies is fallback facility. 771 It MUST NOT be used to help highly loaded servers to stand 772 against legal connection rate. If you see SYN flood warnings 773 in your logs, but investigation shows that they occur 774 because of overload with legal connections, you should tune 775 another parameters until this warning disappear. 776 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow. 777 778 syncookies seriously violate TCP protocol, do not allow 779 to use TCP extensions, can result in serious degradation 780 of some services (f.e. SMTP relaying), visible not by you, 781 but your clients and relays, contacting you. While you see 782 SYN flood warnings in logs not being really flooded, your server 783 is seriously misconfigured. 784 785 If you want to test which effects syncookies have to your 786 network connections you can set this knob to 2 to enable 787 unconditionally generation of syncookies. 788 789tcp_migrate_req - BOOLEAN 790 The incoming connection is tied to a specific listening socket when 791 the initial SYN packet is received during the three-way handshake. 792 When a listener is closed, in-flight request sockets during the 793 handshake and established sockets in the accept queue are aborted. 794 795 If the listener has SO_REUSEPORT enabled, other listeners on the 796 same port should have been able to accept such connections. This 797 option makes it possible to migrate such child sockets to another 798 listener after close() or shutdown(). 799 800 The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should 801 usually be used to define the policy to pick an alive listener. 802 Otherwise, the kernel will randomly pick an alive listener only if 803 this option is enabled. 804 805 Note that migration between listeners with different settings may 806 crash applications. Let's say migration happens from listener A to 807 B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from 808 the requests migrated from A. To avoid such a situation, cancel 809 migration by returning SK_DROP in the type of eBPF program, or 810 disable this option. 811 812 Default: 0 813 814tcp_fastopen - INTEGER 815 Enable TCP Fast Open (RFC7413) to send and accept data in the opening 816 SYN packet. 817 818 The client support is enabled by flag 0x1 (on by default). The client 819 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag, 820 rather than connect() to send data in SYN. 821 822 The server support is enabled by flag 0x2 (off by default). Then 823 either enable for all listeners with another flag (0x400) or 824 enable individual listeners via TCP_FASTOPEN socket option with 825 the option value being the length of the syn-data backlog. 826 827 The values (bitmap) are 828 829 ===== ======== ====================================================== 830 0x1 (client) enables sending data in the opening SYN on the client. 831 0x2 (server) enables the server support, i.e., allowing data in 832 a SYN packet to be accepted and passed to the 833 application before 3-way handshake finishes. 834 0x4 (client) send data in the opening SYN regardless of cookie 835 availability and without a cookie option. 836 0x200 (server) accept data-in-SYN w/o any cookie option present. 837 0x400 (server) enable all listeners to support Fast Open by 838 default without explicit TCP_FASTOPEN socket option. 839 ===== ======== ====================================================== 840 841 Default: 0x1 842 843 Note that additional client or server features are only 844 effective if the basic support (0x1 and 0x2) are enabled respectively. 845 846tcp_fastopen_blackhole_timeout_sec - INTEGER 847 Initial time period in second to disable Fastopen on active TCP sockets 848 when a TFO firewall blackhole issue happens. 849 This time period will grow exponentially when more blackhole issues 850 get detected right after Fastopen is re-enabled and will reset to 851 initial value when the blackhole issue goes away. 852 0 to disable the blackhole detection. 853 854 By default, it is set to 0 (feature is disabled). 855 856tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs 857 The list consists of a primary key and an optional backup key. The 858 primary key is used for both creating and validating cookies, while the 859 optional backup key is only used for validating cookies. The purpose of 860 the backup key is to maximize TFO validation when keys are rotated. 861 862 A randomly chosen primary key may be configured by the kernel if 863 the tcp_fastopen sysctl is set to 0x400 (see above), or if the 864 TCP_FASTOPEN setsockopt() optname is set and a key has not been 865 previously configured via sysctl. If keys are configured via 866 setsockopt() by using the TCP_FASTOPEN_KEY optname, then those 867 per-socket keys will be used instead of any keys that are specified via 868 sysctl. 869 870 A key is specified as 4 8-digit hexadecimal integers which are separated 871 by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be 872 omitted. A primary and a backup key may be specified by separating them 873 by a comma. If only one key is specified, it becomes the primary key and 874 any previously configured backup keys are removed. 875 876tcp_syn_retries - INTEGER 877 Number of times initial SYNs for an active TCP connection attempt 878 will be retransmitted. Should not be higher than 127. Default value 879 is 6, which corresponds to 63seconds till the last retransmission 880 with the current initial RTO of 1second. With this the final timeout 881 for an active TCP connection attempt will happen after 127seconds. 882 883tcp_timestamps - INTEGER 884 Enable timestamps as defined in RFC1323. 885 886 - 0: Disabled. 887 - 1: Enable timestamps as defined in RFC1323 and use random offset for 888 each connection rather than only using the current time. 889 - 2: Like 1, but without random offsets. 890 891 Default: 1 892 893tcp_min_tso_segs - INTEGER 894 Minimal number of segments per TSO frame. 895 896 Since linux-3.12, TCP does an automatic sizing of TSO frames, 897 depending on flow rate, instead of filling 64Kbytes packets. 898 For specific usages, it's possible to force TCP to build big 899 TSO frames. Note that TCP stack might split too big TSO packets 900 if available window is too small. 901 902 Default: 2 903 904tcp_tso_rtt_log - INTEGER 905 Adjustment of TSO packet sizes based on min_rtt 906 907 Starting from linux-5.18, TCP autosizing can be tweaked 908 for flows having small RTT. 909 910 Old autosizing was splitting the pacing budget to send 1024 TSO 911 per second. 912 913 tso_packet_size = sk->sk_pacing_rate / 1024; 914 915 With the new mechanism, we increase this TSO sizing using: 916 917 distance = min_rtt_usec / (2^tcp_tso_rtt_log) 918 tso_packet_size += gso_max_size >> distance; 919 920 This means that flows between very close hosts can use bigger 921 TSO packets, reducing their cpu costs. 922 923 If you want to use the old autosizing, set this sysctl to 0. 924 925 Default: 9 (2^9 = 512 usec) 926 927tcp_pacing_ss_ratio - INTEGER 928 sk->sk_pacing_rate is set by TCP stack using a ratio applied 929 to current rate. (current_rate = cwnd * mss / srtt) 930 If TCP is in slow start, tcp_pacing_ss_ratio is applied 931 to let TCP probe for bigger speeds, assuming cwnd can be 932 doubled every other RTT. 933 934 Default: 200 935 936tcp_pacing_ca_ratio - INTEGER 937 sk->sk_pacing_rate is set by TCP stack using a ratio applied 938 to current rate. (current_rate = cwnd * mss / srtt) 939 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio 940 is applied to conservatively probe for bigger throughput. 941 942 Default: 120 943 944tcp_tso_win_divisor - INTEGER 945 This allows control over what percentage of the congestion window 946 can be consumed by a single TSO frame. 947 The setting of this parameter is a choice between burstiness and 948 building larger TSO frames. 949 950 Default: 3 951 952tcp_tw_reuse - INTEGER 953 Enable reuse of TIME-WAIT sockets for new connections when it is 954 safe from protocol viewpoint. 955 956 - 0 - disable 957 - 1 - global enable 958 - 2 - enable for loopback traffic only 959 960 It should not be changed without advice/request of technical 961 experts. 962 963 Default: 2 964 965tcp_window_scaling - BOOLEAN 966 Enable window scaling as defined in RFC1323. 967 968tcp_wmem - vector of 3 INTEGERs: min, default, max 969 min: Amount of memory reserved for send buffers for TCP sockets. 970 Each TCP socket has rights to use it due to fact of its birth. 971 972 Default: 4K 973 974 default: initial size of send buffer used by TCP sockets. This 975 value overrides net.core.wmem_default used by other protocols. 976 977 It is usually lower than net.core.wmem_default. 978 979 Default: 16K 980 981 max: Maximal amount of memory allowed for automatically tuned 982 send buffers for TCP sockets. This value does not override 983 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables 984 automatic tuning of that socket's send buffer size, in which case 985 this value is ignored. 986 987 Default: between 64K and 4MB, depending on RAM size. 988 989tcp_notsent_lowat - UNSIGNED INTEGER 990 A TCP socket can control the amount of unsent bytes in its write queue, 991 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll() 992 reports POLLOUT events if the amount of unsent bytes is below a per 993 socket value, and if the write queue is not full. sendmsg() will 994 also not add new buffers if the limit is hit. 995 996 This global variable controls the amount of unsent data for 997 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change 998 to the global variable has immediate effect. 999 1000 Default: UINT_MAX (0xFFFFFFFF) 1001 1002tcp_workaround_signed_windows - BOOLEAN 1003 If set, assume no receipt of a window scaling option means the 1004 remote TCP is broken and treats the window as a signed quantity. 1005 If unset, assume the remote TCP is not broken even if we do 1006 not receive a window scaling option from them. 1007 1008 Default: 0 1009 1010tcp_thin_linear_timeouts - BOOLEAN 1011 Enable dynamic triggering of linear timeouts for thin streams. 1012 If set, a check is performed upon retransmission by timeout to 1013 determine if the stream is thin (less than 4 packets in flight). 1014 As long as the stream is found to be thin, up to 6 linear 1015 timeouts may be performed before exponential backoff mode is 1016 initiated. This improves retransmission latency for 1017 non-aggressive thin streams, often found to be time-dependent. 1018 For more information on thin streams, see 1019 Documentation/networking/tcp-thin.rst 1020 1021 Default: 0 1022 1023tcp_limit_output_bytes - INTEGER 1024 Controls TCP Small Queue limit per tcp socket. 1025 TCP bulk sender tends to increase packets in flight until it 1026 gets losses notifications. With SNDBUF autotuning, this can 1027 result in a large amount of packets queued on the local machine 1028 (e.g.: qdiscs, CPU backlog, or device) hurting latency of other 1029 flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes 1030 limits the number of bytes on qdisc or device to reduce artificial 1031 RTT/cwnd and reduce bufferbloat. 1032 1033 Default: 1048576 (16 * 65536) 1034 1035tcp_challenge_ack_limit - INTEGER 1036 Limits number of Challenge ACK sent per second, as recommended 1037 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks) 1038 Default: 1000 1039 1040UDP variables 1041============= 1042 1043udp_l3mdev_accept - BOOLEAN 1044 Enabling this option allows a "global" bound socket to work 1045 across L3 master domains (e.g., VRFs) with packets capable of 1046 being received regardless of the L3 domain in which they 1047 originated. Only valid when the kernel was compiled with 1048 CONFIG_NET_L3_MASTER_DEV. 1049 1050 Default: 0 (disabled) 1051 1052udp_mem - vector of 3 INTEGERs: min, pressure, max 1053 Number of pages allowed for queueing by all UDP sockets. 1054 1055 min: Number of pages allowed for queueing by all UDP sockets. 1056 1057 pressure: This value was introduced to follow format of tcp_mem. 1058 1059 max: This value was introduced to follow format of tcp_mem. 1060 1061 Default is calculated at boot time from amount of available memory. 1062 1063udp_rmem_min - INTEGER 1064 Minimal size of receive buffer used by UDP sockets in moderation. 1065 Each UDP socket is able to use the size for receiving data, even if 1066 total pages of UDP sockets exceed udp_mem pressure. The unit is byte. 1067 1068 Default: 4K 1069 1070udp_wmem_min - INTEGER 1071 UDP does not have tx memory accounting and this tunable has no effect. 1072 1073RAW variables 1074============= 1075 1076raw_l3mdev_accept - BOOLEAN 1077 Enabling this option allows a "global" bound socket to work 1078 across L3 master domains (e.g., VRFs) with packets capable of 1079 being received regardless of the L3 domain in which they 1080 originated. Only valid when the kernel was compiled with 1081 CONFIG_NET_L3_MASTER_DEV. 1082 1083 Default: 1 (enabled) 1084 1085CIPSOv4 Variables 1086================= 1087 1088cipso_cache_enable - BOOLEAN 1089 If set, enable additions to and lookups from the CIPSO label mapping 1090 cache. If unset, additions are ignored and lookups always result in a 1091 miss. However, regardless of the setting the cache is still 1092 invalidated when required when means you can safely toggle this on and 1093 off and the cache will always be "safe". 1094 1095 Default: 1 1096 1097cipso_cache_bucket_size - INTEGER 1098 The CIPSO label cache consists of a fixed size hash table with each 1099 hash bucket containing a number of cache entries. This variable limits 1100 the number of entries in each hash bucket; the larger the value is, the 1101 more CIPSO label mappings that can be cached. When the number of 1102 entries in a given hash bucket reaches this limit adding new entries 1103 causes the oldest entry in the bucket to be removed to make room. 1104 1105 Default: 10 1106 1107cipso_rbm_optfmt - BOOLEAN 1108 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of 1109 the CIPSO draft specification (see Documentation/netlabel for details). 1110 This means that when set the CIPSO tag will be padded with empty 1111 categories in order to make the packet data 32-bit aligned. 1112 1113 Default: 0 1114 1115cipso_rbm_structvalid - BOOLEAN 1116 If set, do a very strict check of the CIPSO option when 1117 ip_options_compile() is called. If unset, relax the checks done during 1118 ip_options_compile(). Either way is "safe" as errors are caught else 1119 where in the CIPSO processing code but setting this to 0 (False) should 1120 result in less work (i.e. it should be faster) but could cause problems 1121 with other implementations that require strict checking. 1122 1123 Default: 0 1124 1125IP Variables 1126============ 1127 1128ip_local_port_range - 2 INTEGERS 1129 Defines the local port range that is used by TCP and UDP to 1130 choose the local port. The first number is the first, the 1131 second the last local port number. 1132 If possible, it is better these numbers have different parity 1133 (one even and one odd value). 1134 Must be greater than or equal to ip_unprivileged_port_start. 1135 The default values are 32768 and 60999 respectively. 1136 1137ip_local_reserved_ports - list of comma separated ranges 1138 Specify the ports which are reserved for known third-party 1139 applications. These ports will not be used by automatic port 1140 assignments (e.g. when calling connect() or bind() with port 1141 number 0). Explicit port allocation behavior is unchanged. 1142 1143 The format used for both input and output is a comma separated 1144 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and 1145 10). Writing to the file will clear all previously reserved 1146 ports and update the current list with the one given in the 1147 input. 1148 1149 Note that ip_local_port_range and ip_local_reserved_ports 1150 settings are independent and both are considered by the kernel 1151 when determining which ports are available for automatic port 1152 assignments. 1153 1154 You can reserve ports which are not in the current 1155 ip_local_port_range, e.g.:: 1156 1157 $ cat /proc/sys/net/ipv4/ip_local_port_range 1158 32000 60999 1159 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports 1160 8080,9148 1161 1162 although this is redundant. However such a setting is useful 1163 if later the port range is changed to a value that will 1164 include the reserved ports. Also keep in mind, that overlapping 1165 of these ranges may affect probability of selecting ephemeral 1166 ports which are right after block of reserved ports. 1167 1168 Default: Empty 1169 1170ip_unprivileged_port_start - INTEGER 1171 This is a per-namespace sysctl. It defines the first 1172 unprivileged port in the network namespace. Privileged ports 1173 require root or CAP_NET_BIND_SERVICE in order to bind to them. 1174 To disable all privileged ports, set this to 0. They must not 1175 overlap with the ip_local_port_range. 1176 1177 Default: 1024 1178 1179ip_nonlocal_bind - BOOLEAN 1180 If set, allows processes to bind() to non-local IP addresses, 1181 which can be quite useful - but may break some applications. 1182 1183 Default: 0 1184 1185ip_autobind_reuse - BOOLEAN 1186 By default, bind() does not select the ports automatically even if 1187 the new socket and all sockets bound to the port have SO_REUSEADDR. 1188 ip_autobind_reuse allows bind() to reuse the port and this is useful 1189 when you use bind()+connect(), but may break some applications. 1190 The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this 1191 option should only be set by experts. 1192 Default: 0 1193 1194ip_dynaddr - INTEGER 1195 If set non-zero, enables support for dynamic addresses. 1196 If set to a non-zero value larger than 1, a kernel log 1197 message will be printed when dynamic address rewriting 1198 occurs. 1199 1200 Default: 0 1201 1202ip_early_demux - BOOLEAN 1203 Optimize input packet processing down to one demux for 1204 certain kinds of local sockets. Currently we only do this 1205 for established TCP and connected UDP sockets. 1206 1207 It may add an additional cost for pure routing workloads that 1208 reduces overall throughput, in such case you should disable it. 1209 1210 Default: 1 1211 1212ping_group_range - 2 INTEGERS 1213 Restrict ICMP_PROTO datagram sockets to users in the group range. 1214 The default is "1 0", meaning, that nobody (not even root) may 1215 create ping sockets. Setting it to "100 100" would grant permissions 1216 to the single group. "0 4294967295" would enable it for the world, "100 1217 4294967295" would enable it for the users, but not daemons. 1218 1219tcp_early_demux - BOOLEAN 1220 Enable early demux for established TCP sockets. 1221 1222 Default: 1 1223 1224udp_early_demux - BOOLEAN 1225 Enable early demux for connected UDP sockets. Disable this if 1226 your system could experience more unconnected load. 1227 1228 Default: 1 1229 1230icmp_echo_ignore_all - BOOLEAN 1231 If set non-zero, then the kernel will ignore all ICMP ECHO 1232 requests sent to it. 1233 1234 Default: 0 1235 1236icmp_echo_enable_probe - BOOLEAN 1237 If set to one, then the kernel will respond to RFC 8335 PROBE 1238 requests sent to it. 1239 1240 Default: 0 1241 1242icmp_echo_ignore_broadcasts - BOOLEAN 1243 If set non-zero, then the kernel will ignore all ICMP ECHO and 1244 TIMESTAMP requests sent to it via broadcast/multicast. 1245 1246 Default: 1 1247 1248icmp_ratelimit - INTEGER 1249 Limit the maximal rates for sending ICMP packets whose type matches 1250 icmp_ratemask (see below) to specific targets. 1251 0 to disable any limiting, 1252 otherwise the minimal space between responses in milliseconds. 1253 Note that another sysctl, icmp_msgs_per_sec limits the number 1254 of ICMP packets sent on all targets. 1255 1256 Default: 1000 1257 1258icmp_msgs_per_sec - INTEGER 1259 Limit maximal number of ICMP packets sent per second from this host. 1260 Only messages whose type matches icmp_ratemask (see below) are 1261 controlled by this limit. For security reasons, the precise count 1262 of messages per second is randomized. 1263 1264 Default: 1000 1265 1266icmp_msgs_burst - INTEGER 1267 icmp_msgs_per_sec controls number of ICMP packets sent per second, 1268 while icmp_msgs_burst controls the burst size of these packets. 1269 For security reasons, the precise burst size is randomized. 1270 1271 Default: 50 1272 1273icmp_ratemask - INTEGER 1274 Mask made of ICMP types for which rates are being limited. 1275 1276 Significant bits: IHGFEDCBA9876543210 1277 1278 Default mask: 0000001100000011000 (6168) 1279 1280 Bit definitions (see include/linux/icmp.h): 1281 1282 = ========================= 1283 0 Echo Reply 1284 3 Destination Unreachable [1]_ 1285 4 Source Quench [1]_ 1286 5 Redirect 1287 8 Echo Request 1288 B Time Exceeded [1]_ 1289 C Parameter Problem [1]_ 1290 D Timestamp Request 1291 E Timestamp Reply 1292 F Info Request 1293 G Info Reply 1294 H Address Mask Request 1295 I Address Mask Reply 1296 = ========================= 1297 1298 .. [1] These are rate limited by default (see default mask above) 1299 1300icmp_ignore_bogus_error_responses - BOOLEAN 1301 Some routers violate RFC1122 by sending bogus responses to broadcast 1302 frames. Such violations are normally logged via a kernel warning. 1303 If this is set to TRUE, the kernel will not give such warnings, which 1304 will avoid log file clutter. 1305 1306 Default: 1 1307 1308icmp_errors_use_inbound_ifaddr - BOOLEAN 1309 1310 If zero, icmp error messages are sent with the primary address of 1311 the exiting interface. 1312 1313 If non-zero, the message will be sent with the primary address of 1314 the interface that received the packet that caused the icmp error. 1315 This is the behaviour many network administrators will expect from 1316 a router. And it can make debugging complicated network layouts 1317 much easier. 1318 1319 Note that if no primary address exists for the interface selected, 1320 then the primary address of the first non-loopback interface that 1321 has one will be used regardless of this setting. 1322 1323 Default: 0 1324 1325igmp_max_memberships - INTEGER 1326 Change the maximum number of multicast groups we can subscribe to. 1327 Default: 20 1328 1329 Theoretical maximum value is bounded by having to send a membership 1330 report in a single datagram (i.e. the report can't span multiple 1331 datagrams, or risk confusing the switch and leaving groups you don't 1332 intend to). 1333 1334 The number of supported groups 'M' is bounded by the number of group 1335 report entries you can fit into a single datagram of 65535 bytes. 1336 1337 M = 65536-sizeof (ip header)/(sizeof(Group record)) 1338 1339 Group records are variable length, with a minimum of 12 bytes. 1340 So net.ipv4.igmp_max_memberships should not be set higher than: 1341 1342 (65536-24) / 12 = 5459 1343 1344 The value 5459 assumes no IP header options, so in practice 1345 this number may be lower. 1346 1347igmp_max_msf - INTEGER 1348 Maximum number of addresses allowed in the source filter list for a 1349 multicast group. 1350 1351 Default: 10 1352 1353igmp_qrv - INTEGER 1354 Controls the IGMP query robustness variable (see RFC2236 8.1). 1355 1356 Default: 2 (as specified by RFC2236 8.1) 1357 1358 Minimum: 1 (as specified by RFC6636 4.5) 1359 1360force_igmp_version - INTEGER 1361 - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback 1362 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier 1363 Present timer expires. 1364 - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if 1365 receive IGMPv2/v3 query. 1366 - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive 1367 IGMPv1 query message. Will reply report if receive IGMPv3 query. 1368 - 3 - Enforce to use IGMP version 3. The same react with default 0. 1369 1370 .. note:: 1371 1372 this is not the same with force_mld_version because IGMPv3 RFC3376 1373 Security Considerations does not have clear description that we could 1374 ignore other version messages completely as MLDv2 RFC3810. So make 1375 this value as default 0 is recommended. 1376 1377``conf/interface/*`` 1378 changes special settings per interface (where 1379 interface" is the name of your network interface) 1380 1381``conf/all/*`` 1382 is special, changes the settings for all interfaces 1383 1384log_martians - BOOLEAN 1385 Log packets with impossible addresses to kernel log. 1386 log_martians for the interface will be enabled if at least one of 1387 conf/{all,interface}/log_martians is set to TRUE, 1388 it will be disabled otherwise 1389 1390accept_redirects - BOOLEAN 1391 Accept ICMP redirect messages. 1392 accept_redirects for the interface will be enabled if: 1393 1394 - both conf/{all,interface}/accept_redirects are TRUE in the case 1395 forwarding for the interface is enabled 1396 1397 or 1398 1399 - at least one of conf/{all,interface}/accept_redirects is TRUE in the 1400 case forwarding for the interface is disabled 1401 1402 accept_redirects for the interface will be disabled otherwise 1403 1404 default: 1405 1406 - TRUE (host) 1407 - FALSE (router) 1408 1409forwarding - BOOLEAN 1410 Enable IP forwarding on this interface. This controls whether packets 1411 received _on_ this interface can be forwarded. 1412 1413mc_forwarding - BOOLEAN 1414 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE 1415 and a multicast routing daemon is required. 1416 conf/all/mc_forwarding must also be set to TRUE to enable multicast 1417 routing for the interface 1418 1419medium_id - INTEGER 1420 Integer value used to differentiate the devices by the medium they 1421 are attached to. Two devices can have different id values when 1422 the broadcast packets are received only on one of them. 1423 The default value 0 means that the device is the only interface 1424 to its medium, value of -1 means that medium is not known. 1425 1426 Currently, it is used to change the proxy_arp behavior: 1427 the proxy_arp feature is enabled for packets forwarded between 1428 two devices attached to different media. 1429 1430proxy_arp - BOOLEAN 1431 Do proxy arp. 1432 1433 proxy_arp for the interface will be enabled if at least one of 1434 conf/{all,interface}/proxy_arp is set to TRUE, 1435 it will be disabled otherwise 1436 1437proxy_arp_pvlan - BOOLEAN 1438 Private VLAN proxy arp. 1439 1440 Basically allow proxy arp replies back to the same interface 1441 (from which the ARP request/solicitation was received). 1442 1443 This is done to support (ethernet) switch features, like RFC 1444 3069, where the individual ports are NOT allowed to 1445 communicate with each other, but they are allowed to talk to 1446 the upstream router. As described in RFC 3069, it is possible 1447 to allow these hosts to communicate through the upstream 1448 router by proxy_arp'ing. Don't need to be used together with 1449 proxy_arp. 1450 1451 This technology is known by different names: 1452 1453 In RFC 3069 it is called VLAN Aggregation. 1454 Cisco and Allied Telesyn call it Private VLAN. 1455 Hewlett-Packard call it Source-Port filtering or port-isolation. 1456 Ericsson call it MAC-Forced Forwarding (RFC Draft). 1457 1458shared_media - BOOLEAN 1459 Send(router) or accept(host) RFC1620 shared media redirects. 1460 Overrides secure_redirects. 1461 1462 shared_media for the interface will be enabled if at least one of 1463 conf/{all,interface}/shared_media is set to TRUE, 1464 it will be disabled otherwise 1465 1466 default TRUE 1467 1468secure_redirects - BOOLEAN 1469 Accept ICMP redirect messages only to gateways listed in the 1470 interface's current gateway list. Even if disabled, RFC1122 redirect 1471 rules still apply. 1472 1473 Overridden by shared_media. 1474 1475 secure_redirects for the interface will be enabled if at least one of 1476 conf/{all,interface}/secure_redirects is set to TRUE, 1477 it will be disabled otherwise 1478 1479 default TRUE 1480 1481send_redirects - BOOLEAN 1482 Send redirects, if router. 1483 1484 send_redirects for the interface will be enabled if at least one of 1485 conf/{all,interface}/send_redirects is set to TRUE, 1486 it will be disabled otherwise 1487 1488 Default: TRUE 1489 1490bootp_relay - BOOLEAN 1491 Accept packets with source address 0.b.c.d destined 1492 not to this host as local ones. It is supposed, that 1493 BOOTP relay daemon will catch and forward such packets. 1494 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay 1495 for the interface 1496 1497 default FALSE 1498 1499 Not Implemented Yet. 1500 1501accept_source_route - BOOLEAN 1502 Accept packets with SRR option. 1503 conf/all/accept_source_route must also be set to TRUE to accept packets 1504 with SRR option on the interface 1505 1506 default 1507 1508 - TRUE (router) 1509 - FALSE (host) 1510 1511accept_local - BOOLEAN 1512 Accept packets with local source addresses. In combination with 1513 suitable routing, this can be used to direct packets between two 1514 local interfaces over the wire and have them accepted properly. 1515 default FALSE 1516 1517route_localnet - BOOLEAN 1518 Do not consider loopback addresses as martian source or destination 1519 while routing. This enables the use of 127/8 for local routing purposes. 1520 1521 default FALSE 1522 1523rp_filter - INTEGER 1524 - 0 - No source validation. 1525 - 1 - Strict mode as defined in RFC3704 Strict Reverse Path 1526 Each incoming packet is tested against the FIB and if the interface 1527 is not the best reverse path the packet check will fail. 1528 By default failed packets are discarded. 1529 - 2 - Loose mode as defined in RFC3704 Loose Reverse Path 1530 Each incoming packet's source address is also tested against the FIB 1531 and if the source address is not reachable via any interface 1532 the packet check will fail. 1533 1534 Current recommended practice in RFC3704 is to enable strict mode 1535 to prevent IP spoofing from DDos attacks. If using asymmetric routing 1536 or other complicated routing, then loose mode is recommended. 1537 1538 The max value from conf/{all,interface}/rp_filter is used 1539 when doing source validation on the {interface}. 1540 1541 Default value is 0. Note that some distributions enable it 1542 in startup scripts. 1543 1544src_valid_mark - BOOLEAN 1545 - 0 - The fwmark of the packet is not included in reverse path 1546 route lookup. This allows for asymmetric routing configurations 1547 utilizing the fwmark in only one direction, e.g., transparent 1548 proxying. 1549 1550 - 1 - The fwmark of the packet is included in reverse path route 1551 lookup. This permits rp_filter to function when the fwmark is 1552 used for routing traffic in both directions. 1553 1554 This setting also affects the utilization of fmwark when 1555 performing source address selection for ICMP replies, or 1556 determining addresses stored for the IPOPT_TS_TSANDADDR and 1557 IPOPT_RR IP options. 1558 1559 The max value from conf/{all,interface}/src_valid_mark is used. 1560 1561 Default value is 0. 1562 1563arp_filter - BOOLEAN 1564 - 1 - Allows you to have multiple network interfaces on the same 1565 subnet, and have the ARPs for each interface be answered 1566 based on whether or not the kernel would route a packet from 1567 the ARP'd IP out that interface (therefore you must use source 1568 based routing for this to work). In other words it allows control 1569 of which cards (usually 1) will respond to an arp request. 1570 1571 - 0 - (default) The kernel can respond to arp requests with addresses 1572 from other interfaces. This may seem wrong but it usually makes 1573 sense, because it increases the chance of successful communication. 1574 IP addresses are owned by the complete host on Linux, not by 1575 particular interfaces. Only for more complex setups like load- 1576 balancing, does this behaviour cause problems. 1577 1578 arp_filter for the interface will be enabled if at least one of 1579 conf/{all,interface}/arp_filter is set to TRUE, 1580 it will be disabled otherwise 1581 1582arp_announce - INTEGER 1583 Define different restriction levels for announcing the local 1584 source IP address from IP packets in ARP requests sent on 1585 interface: 1586 1587 - 0 - (default) Use any local address, configured on any interface 1588 - 1 - Try to avoid local addresses that are not in the target's 1589 subnet for this interface. This mode is useful when target 1590 hosts reachable via this interface require the source IP 1591 address in ARP requests to be part of their logical network 1592 configured on the receiving interface. When we generate the 1593 request we will check all our subnets that include the 1594 target IP and will preserve the source address if it is from 1595 such subnet. If there is no such subnet we select source 1596 address according to the rules for level 2. 1597 - 2 - Always use the best local address for this target. 1598 In this mode we ignore the source address in the IP packet 1599 and try to select local address that we prefer for talks with 1600 the target host. Such local address is selected by looking 1601 for primary IP addresses on all our subnets on the outgoing 1602 interface that include the target IP address. If no suitable 1603 local address is found we select the first local address 1604 we have on the outgoing interface or on all other interfaces, 1605 with the hope we will receive reply for our request and 1606 even sometimes no matter the source IP address we announce. 1607 1608 The max value from conf/{all,interface}/arp_announce is used. 1609 1610 Increasing the restriction level gives more chance for 1611 receiving answer from the resolved target while decreasing 1612 the level announces more valid sender's information. 1613 1614arp_ignore - INTEGER 1615 Define different modes for sending replies in response to 1616 received ARP requests that resolve local target IP addresses: 1617 1618 - 0 - (default): reply for any local target IP address, configured 1619 on any interface 1620 - 1 - reply only if the target IP address is local address 1621 configured on the incoming interface 1622 - 2 - reply only if the target IP address is local address 1623 configured on the incoming interface and both with the 1624 sender's IP address are part from same subnet on this interface 1625 - 3 - do not reply for local addresses configured with scope host, 1626 only resolutions for global and link addresses are replied 1627 - 4-7 - reserved 1628 - 8 - do not reply for all local addresses 1629 1630 The max value from conf/{all,interface}/arp_ignore is used 1631 when ARP request is received on the {interface} 1632 1633arp_notify - BOOLEAN 1634 Define mode for notification of address and device changes. 1635 1636 == ========================================================== 1637 0 (default): do nothing 1638 1 Generate gratuitous arp requests when device is brought up 1639 or hardware address changes. 1640 == ========================================================== 1641 1642arp_accept - INTEGER 1643 Define behavior for accepting gratuitous ARP (garp) frames from devices 1644 that are not already present in the ARP table: 1645 1646 - 0 - don't create new entries in the ARP table 1647 - 1 - create new entries in the ARP table 1648 - 2 - create new entries only if the source IP address is in the same 1649 subnet as an address configured on the interface that received the 1650 garp message. 1651 1652 Both replies and requests type gratuitous arp will trigger the 1653 ARP table to be updated, if this setting is on. 1654 1655 If the ARP table already contains the IP address of the 1656 gratuitous arp frame, the arp table will be updated regardless 1657 if this setting is on or off. 1658 1659arp_evict_nocarrier - BOOLEAN 1660 Clears the ARP cache on NOCARRIER events. This option is important for 1661 wireless devices where the ARP cache should not be cleared when roaming 1662 between access points on the same network. In most cases this should 1663 remain as the default (1). 1664 1665 - 1 - (default): Clear the ARP cache on NOCARRIER events 1666 - 0 - Do not clear ARP cache on NOCARRIER events 1667 1668mcast_solicit - INTEGER 1669 The maximum number of multicast probes in INCOMPLETE state, 1670 when the associated hardware address is unknown. Defaults 1671 to 3. 1672 1673ucast_solicit - INTEGER 1674 The maximum number of unicast probes in PROBE state, when 1675 the hardware address is being reconfirmed. Defaults to 3. 1676 1677app_solicit - INTEGER 1678 The maximum number of probes to send to the user space ARP daemon 1679 via netlink before dropping back to multicast probes (see 1680 mcast_resolicit). Defaults to 0. 1681 1682mcast_resolicit - INTEGER 1683 The maximum number of multicast probes after unicast and 1684 app probes in PROBE state. Defaults to 0. 1685 1686disable_policy - BOOLEAN 1687 Disable IPSEC policy (SPD) for this interface 1688 1689disable_xfrm - BOOLEAN 1690 Disable IPSEC encryption on this interface, whatever the policy 1691 1692igmpv2_unsolicited_report_interval - INTEGER 1693 The interval in milliseconds in which the next unsolicited 1694 IGMPv1 or IGMPv2 report retransmit will take place. 1695 1696 Default: 10000 (10 seconds) 1697 1698igmpv3_unsolicited_report_interval - INTEGER 1699 The interval in milliseconds in which the next unsolicited 1700 IGMPv3 report retransmit will take place. 1701 1702 Default: 1000 (1 seconds) 1703 1704ignore_routes_with_linkdown - BOOLEAN 1705 Ignore routes whose link is down when performing a FIB lookup. 1706 1707promote_secondaries - BOOLEAN 1708 When a primary IP address is removed from this interface 1709 promote a corresponding secondary IP address instead of 1710 removing all the corresponding secondary IP addresses. 1711 1712drop_unicast_in_l2_multicast - BOOLEAN 1713 Drop any unicast IP packets that are received in link-layer 1714 multicast (or broadcast) frames. 1715 1716 This behavior (for multicast) is actually a SHOULD in RFC 1717 1122, but is disabled by default for compatibility reasons. 1718 1719 Default: off (0) 1720 1721drop_gratuitous_arp - BOOLEAN 1722 Drop all gratuitous ARP frames, for example if there's a known 1723 good ARP proxy on the network and such frames need not be used 1724 (or in the case of 802.11, must not be used to prevent attacks.) 1725 1726 Default: off (0) 1727 1728 1729tag - INTEGER 1730 Allows you to write a number, which can be used as required. 1731 1732 Default value is 0. 1733 1734xfrm4_gc_thresh - INTEGER 1735 (Obsolete since linux-4.14) 1736 The threshold at which we will start garbage collecting for IPv4 1737 destination cache entries. At twice this value the system will 1738 refuse new allocations. 1739 1740igmp_link_local_mcast_reports - BOOLEAN 1741 Enable IGMP reports for link local multicast groups in the 1742 224.0.0.X range. 1743 1744 Default TRUE 1745 1746Alexey Kuznetsov. 1747kuznet@ms2.inr.ac.ru 1748 1749Updated by: 1750 1751- Andi Kleen 1752 ak@muc.de 1753- Nicolas Delon 1754 delon.nicolas@wanadoo.fr 1755 1756 1757 1758 1759/proc/sys/net/ipv6/* Variables 1760============================== 1761 1762IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also 1763apply to IPv6 [XXX?]. 1764 1765bindv6only - BOOLEAN 1766 Default value for IPV6_V6ONLY socket option, 1767 which restricts use of the IPv6 socket to IPv6 communication 1768 only. 1769 1770 - TRUE: disable IPv4-mapped address feature 1771 - FALSE: enable IPv4-mapped address feature 1772 1773 Default: FALSE (as specified in RFC3493) 1774 1775flowlabel_consistency - BOOLEAN 1776 Protect the consistency (and unicity) of flow label. 1777 You have to disable it to use IPV6_FL_F_REFLECT flag on the 1778 flow label manager. 1779 1780 - TRUE: enabled 1781 - FALSE: disabled 1782 1783 Default: TRUE 1784 1785auto_flowlabels - INTEGER 1786 Automatically generate flow labels based on a flow hash of the 1787 packet. This allows intermediate devices, such as routers, to 1788 identify packet flows for mechanisms like Equal Cost Multipath 1789 Routing (see RFC 6438). 1790 1791 = =========================================================== 1792 0 automatic flow labels are completely disabled 1793 1 automatic flow labels are enabled by default, they can be 1794 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL 1795 socket option 1796 2 automatic flow labels are allowed, they may be enabled on a 1797 per socket basis using the IPV6_AUTOFLOWLABEL socket option 1798 3 automatic flow labels are enabled and enforced, they cannot 1799 be disabled by the socket option 1800 = =========================================================== 1801 1802 Default: 1 1803 1804flowlabel_state_ranges - BOOLEAN 1805 Split the flow label number space into two ranges. 0-0x7FFFF is 1806 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF 1807 is reserved for stateless flow labels as described in RFC6437. 1808 1809 - TRUE: enabled 1810 - FALSE: disabled 1811 1812 Default: true 1813 1814flowlabel_reflect - INTEGER 1815 Control flow label reflection. Needed for Path MTU 1816 Discovery to work with Equal Cost Multipath Routing in anycast 1817 environments. See RFC 7690 and: 1818 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01 1819 1820 This is a bitmask. 1821 1822 - 1: enabled for established flows 1823 1824 Note that this prevents automatic flowlabel changes, as done 1825 in "tcp: change IPv6 flow-label upon receiving spurious retransmission" 1826 and "tcp: Change txhash on every SYN and RTO retransmit" 1827 1828 - 2: enabled for TCP RESET packets (no active listener) 1829 If set, a RST packet sent in response to a SYN packet on a closed 1830 port will reflect the incoming flow label. 1831 1832 - 4: enabled for ICMPv6 echo reply messages. 1833 1834 Default: 0 1835 1836fib_multipath_hash_policy - INTEGER 1837 Controls which hash policy to use for multipath routes. 1838 1839 Default: 0 (Layer 3) 1840 1841 Possible values: 1842 1843 - 0 - Layer 3 (source and destination addresses plus flow label) 1844 - 1 - Layer 4 (standard 5-tuple) 1845 - 2 - Layer 3 or inner Layer 3 if present 1846 - 3 - Custom multipath hash. Fields used for multipath hash calculation 1847 are determined by fib_multipath_hash_fields sysctl 1848 1849fib_multipath_hash_fields - UNSIGNED INTEGER 1850 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the 1851 fields used for multipath hash calculation are determined by this 1852 sysctl. 1853 1854 This value is a bitmask which enables various fields for multipath hash 1855 calculation. 1856 1857 Possible fields are: 1858 1859 ====== ============================ 1860 0x0001 Source IP address 1861 0x0002 Destination IP address 1862 0x0004 IP protocol 1863 0x0008 Flow Label 1864 0x0010 Source port 1865 0x0020 Destination port 1866 0x0040 Inner source IP address 1867 0x0080 Inner destination IP address 1868 0x0100 Inner IP protocol 1869 0x0200 Inner Flow Label 1870 0x0400 Inner source port 1871 0x0800 Inner destination port 1872 ====== ============================ 1873 1874 Default: 0x0007 (source IP, destination IP and IP protocol) 1875 1876anycast_src_echo_reply - BOOLEAN 1877 Controls the use of anycast addresses as source addresses for ICMPv6 1878 echo reply 1879 1880 - TRUE: enabled 1881 - FALSE: disabled 1882 1883 Default: FALSE 1884 1885idgen_delay - INTEGER 1886 Controls the delay in seconds after which time to retry 1887 privacy stable address generation if a DAD conflict is 1888 detected. 1889 1890 Default: 1 (as specified in RFC7217) 1891 1892idgen_retries - INTEGER 1893 Controls the number of retries to generate a stable privacy 1894 address if a DAD conflict is detected. 1895 1896 Default: 3 (as specified in RFC7217) 1897 1898mld_qrv - INTEGER 1899 Controls the MLD query robustness variable (see RFC3810 9.1). 1900 1901 Default: 2 (as specified by RFC3810 9.1) 1902 1903 Minimum: 1 (as specified by RFC6636 4.5) 1904 1905max_dst_opts_number - INTEGER 1906 Maximum number of non-padding TLVs allowed in a Destination 1907 options extension header. If this value is less than zero 1908 then unknown options are disallowed and the number of known 1909 TLVs allowed is the absolute value of this number. 1910 1911 Default: 8 1912 1913max_hbh_opts_number - INTEGER 1914 Maximum number of non-padding TLVs allowed in a Hop-by-Hop 1915 options extension header. If this value is less than zero 1916 then unknown options are disallowed and the number of known 1917 TLVs allowed is the absolute value of this number. 1918 1919 Default: 8 1920 1921max_dst_opts_length - INTEGER 1922 Maximum length allowed for a Destination options extension 1923 header. 1924 1925 Default: INT_MAX (unlimited) 1926 1927max_hbh_length - INTEGER 1928 Maximum length allowed for a Hop-by-Hop options extension 1929 header. 1930 1931 Default: INT_MAX (unlimited) 1932 1933skip_notify_on_dev_down - BOOLEAN 1934 Controls whether an RTM_DELROUTE message is generated for routes 1935 removed when a device is taken down or deleted. IPv4 does not 1936 generate this message; IPv6 does by default. Setting this sysctl 1937 to true skips the message, making IPv4 and IPv6 on par in relying 1938 on userspace caches to track link events and evict routes. 1939 1940 Default: false (generate message) 1941 1942nexthop_compat_mode - BOOLEAN 1943 New nexthop API provides a means for managing nexthops independent of 1944 prefixes. Backwards compatibilty with old route format is enabled by 1945 default which means route dumps and notifications contain the new 1946 nexthop attribute but also the full, expanded nexthop definition. 1947 Further, updates or deletes of a nexthop configuration generate route 1948 notifications for each fib entry using the nexthop. Once a system 1949 understands the new API, this sysctl can be disabled to achieve full 1950 performance benefits of the new API by disabling the nexthop expansion 1951 and extraneous notifications. 1952 Default: true (backward compat mode) 1953 1954fib_notify_on_flag_change - INTEGER 1955 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/ 1956 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed. 1957 1958 After installing a route to the kernel, user space receives an 1959 acknowledgment, which means the route was installed in the kernel, 1960 but not necessarily in hardware. 1961 It is also possible for a route already installed in hardware to change 1962 its action and therefore its flags. For example, a host route that is 1963 trapping packets can be "promoted" to perform decapsulation following 1964 the installation of an IPinIP/VXLAN tunnel. 1965 The notifications will indicate to user-space the state of the route. 1966 1967 Default: 0 (Do not emit notifications.) 1968 1969 Possible values: 1970 1971 - 0 - Do not emit notifications. 1972 - 1 - Emit notifications. 1973 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change. 1974 1975ioam6_id - INTEGER 1976 Define the IOAM id of this node. Uses only 24 bits out of 32 in total. 1977 1978 Min: 0 1979 Max: 0xFFFFFF 1980 1981 Default: 0xFFFFFF 1982 1983ioam6_id_wide - LONG INTEGER 1984 Define the wide IOAM id of this node. Uses only 56 bits out of 64 in 1985 total. Can be different from ioam6_id. 1986 1987 Min: 0 1988 Max: 0xFFFFFFFFFFFFFF 1989 1990 Default: 0xFFFFFFFFFFFFFF 1991 1992IPv6 Fragmentation: 1993 1994ip6frag_high_thresh - INTEGER 1995 Maximum memory used to reassemble IPv6 fragments. When 1996 ip6frag_high_thresh bytes of memory is allocated for this purpose, 1997 the fragment handler will toss packets until ip6frag_low_thresh 1998 is reached. 1999 2000ip6frag_low_thresh - INTEGER 2001 See ip6frag_high_thresh 2002 2003ip6frag_time - INTEGER 2004 Time in seconds to keep an IPv6 fragment in memory. 2005 2006``conf/default/*``: 2007 Change the interface-specific default settings. 2008 2009 These settings would be used during creating new interfaces. 2010 2011 2012``conf/all/*``: 2013 Change all the interface-specific settings. 2014 2015 [XXX: Other special features than forwarding?] 2016 2017conf/all/disable_ipv6 - BOOLEAN 2018 Changing this value is same as changing ``conf/default/disable_ipv6`` 2019 setting and also all per-interface ``disable_ipv6`` settings to the same 2020 value. 2021 2022 Reading this value does not have any particular meaning. It does not say 2023 whether IPv6 support is enabled or disabled. Returned value can be 1 2024 also in the case when some interface has ``disable_ipv6`` set to 0 and 2025 has configured IPv6 addresses. 2026 2027conf/all/forwarding - BOOLEAN 2028 Enable global IPv6 forwarding between all interfaces. 2029 2030 IPv4 and IPv6 work differently here; e.g. netfilter must be used 2031 to control which interfaces may forward packets and which not. 2032 2033 This also sets all interfaces' Host/Router setting 2034 'forwarding' to the specified value. See below for details. 2035 2036 This referred to as global forwarding. 2037 2038proxy_ndp - BOOLEAN 2039 Do proxy ndp. 2040 2041fwmark_reflect - BOOLEAN 2042 Controls the fwmark of kernel-generated IPv6 reply packets that are not 2043 associated with a socket for example, TCP RSTs or ICMPv6 echo replies). 2044 If unset, these packets have a fwmark of zero. If set, they have the 2045 fwmark of the packet they are replying to. 2046 2047 Default: 0 2048 2049``conf/interface/*``: 2050 Change special settings per interface. 2051 2052 The functional behaviour for certain settings is different 2053 depending on whether local forwarding is enabled or not. 2054 2055accept_ra - INTEGER 2056 Accept Router Advertisements; autoconfigure using them. 2057 2058 It also determines whether or not to transmit Router 2059 Solicitations. If and only if the functional setting is to 2060 accept Router Advertisements, Router Solicitations will be 2061 transmitted. 2062 2063 Possible values are: 2064 2065 == =========================================================== 2066 0 Do not accept Router Advertisements. 2067 1 Accept Router Advertisements if forwarding is disabled. 2068 2 Overrule forwarding behaviour. Accept Router Advertisements 2069 even if forwarding is enabled. 2070 == =========================================================== 2071 2072 Functional default: 2073 2074 - enabled if local forwarding is disabled. 2075 - disabled if local forwarding is enabled. 2076 2077accept_ra_defrtr - BOOLEAN 2078 Learn default router in Router Advertisement. 2079 2080 Functional default: 2081 2082 - enabled if accept_ra is enabled. 2083 - disabled if accept_ra is disabled. 2084 2085ra_defrtr_metric - UNSIGNED INTEGER 2086 Route metric for default route learned in Router Advertisement. This value 2087 will be assigned as metric for the default route learned via IPv6 Router 2088 Advertisement. Takes affect only if accept_ra_defrtr is enabled. 2089 2090 Possible values: 2091 1 to 0xFFFFFFFF 2092 2093 Default: IP6_RT_PRIO_USER i.e. 1024. 2094 2095accept_ra_from_local - BOOLEAN 2096 Accept RA with source-address that is found on local machine 2097 if the RA is otherwise proper and able to be accepted. 2098 2099 Default is to NOT accept these as it may be an un-intended 2100 network loop. 2101 2102 Functional default: 2103 2104 - enabled if accept_ra_from_local is enabled 2105 on a specific interface. 2106 - disabled if accept_ra_from_local is disabled 2107 on a specific interface. 2108 2109accept_ra_min_hop_limit - INTEGER 2110 Minimum hop limit Information in Router Advertisement. 2111 2112 Hop limit Information in Router Advertisement less than this 2113 variable shall be ignored. 2114 2115 Default: 1 2116 2117accept_ra_pinfo - BOOLEAN 2118 Learn Prefix Information in Router Advertisement. 2119 2120 Functional default: 2121 2122 - enabled if accept_ra is enabled. 2123 - disabled if accept_ra is disabled. 2124 2125accept_ra_rt_info_min_plen - INTEGER 2126 Minimum prefix length of Route Information in RA. 2127 2128 Route Information w/ prefix smaller than this variable shall 2129 be ignored. 2130 2131 Functional default: 2132 2133 * 0 if accept_ra_rtr_pref is enabled. 2134 * -1 if accept_ra_rtr_pref is disabled. 2135 2136accept_ra_rt_info_max_plen - INTEGER 2137 Maximum prefix length of Route Information in RA. 2138 2139 Route Information w/ prefix larger than this variable shall 2140 be ignored. 2141 2142 Functional default: 2143 2144 * 0 if accept_ra_rtr_pref is enabled. 2145 * -1 if accept_ra_rtr_pref is disabled. 2146 2147accept_ra_rtr_pref - BOOLEAN 2148 Accept Router Preference in RA. 2149 2150 Functional default: 2151 2152 - enabled if accept_ra is enabled. 2153 - disabled if accept_ra is disabled. 2154 2155accept_ra_mtu - BOOLEAN 2156 Apply the MTU value specified in RA option 5 (RFC4861). If 2157 disabled, the MTU specified in the RA will be ignored. 2158 2159 Functional default: 2160 2161 - enabled if accept_ra is enabled. 2162 - disabled if accept_ra is disabled. 2163 2164accept_redirects - BOOLEAN 2165 Accept Redirects. 2166 2167 Functional default: 2168 2169 - enabled if local forwarding is disabled. 2170 - disabled if local forwarding is enabled. 2171 2172accept_source_route - INTEGER 2173 Accept source routing (routing extension header). 2174 2175 - >= 0: Accept only routing header type 2. 2176 - < 0: Do not accept routing header. 2177 2178 Default: 0 2179 2180autoconf - BOOLEAN 2181 Autoconfigure addresses using Prefix Information in Router 2182 Advertisements. 2183 2184 Functional default: 2185 2186 - enabled if accept_ra_pinfo is enabled. 2187 - disabled if accept_ra_pinfo is disabled. 2188 2189dad_transmits - INTEGER 2190 The amount of Duplicate Address Detection probes to send. 2191 2192 Default: 1 2193 2194forwarding - INTEGER 2195 Configure interface-specific Host/Router behaviour. 2196 2197 .. note:: 2198 2199 It is recommended to have the same setting on all 2200 interfaces; mixed router/host scenarios are rather uncommon. 2201 2202 Possible values are: 2203 2204 - 0 Forwarding disabled 2205 - 1 Forwarding enabled 2206 2207 **FALSE (0)**: 2208 2209 By default, Host behaviour is assumed. This means: 2210 2211 1. IsRouter flag is not set in Neighbour Advertisements. 2212 2. If accept_ra is TRUE (default), transmit Router 2213 Solicitations. 2214 3. If accept_ra is TRUE (default), accept Router 2215 Advertisements (and do autoconfiguration). 2216 4. If accept_redirects is TRUE (default), accept Redirects. 2217 2218 **TRUE (1)**: 2219 2220 If local forwarding is enabled, Router behaviour is assumed. 2221 This means exactly the reverse from the above: 2222 2223 1. IsRouter flag is set in Neighbour Advertisements. 2224 2. Router Solicitations are not sent unless accept_ra is 2. 2225 3. Router Advertisements are ignored unless accept_ra is 2. 2226 4. Redirects are ignored. 2227 2228 Default: 0 (disabled) if global forwarding is disabled (default), 2229 otherwise 1 (enabled). 2230 2231hop_limit - INTEGER 2232 Default Hop Limit to set. 2233 2234 Default: 64 2235 2236mtu - INTEGER 2237 Default Maximum Transfer Unit 2238 2239 Default: 1280 (IPv6 required minimum) 2240 2241ip_nonlocal_bind - BOOLEAN 2242 If set, allows processes to bind() to non-local IPv6 addresses, 2243 which can be quite useful - but may break some applications. 2244 2245 Default: 0 2246 2247router_probe_interval - INTEGER 2248 Minimum interval (in seconds) between Router Probing described 2249 in RFC4191. 2250 2251 Default: 60 2252 2253router_solicitation_delay - INTEGER 2254 Number of seconds to wait after interface is brought up 2255 before sending Router Solicitations. 2256 2257 Default: 1 2258 2259router_solicitation_interval - INTEGER 2260 Number of seconds to wait between Router Solicitations. 2261 2262 Default: 4 2263 2264router_solicitations - INTEGER 2265 Number of Router Solicitations to send until assuming no 2266 routers are present. 2267 2268 Default: 3 2269 2270use_oif_addrs_only - BOOLEAN 2271 When enabled, the candidate source addresses for destinations 2272 routed via this interface are restricted to the set of addresses 2273 configured on this interface (vis. RFC 6724, section 4). 2274 2275 Default: false 2276 2277use_tempaddr - INTEGER 2278 Preference for Privacy Extensions (RFC3041). 2279 2280 * <= 0 : disable Privacy Extensions 2281 * == 1 : enable Privacy Extensions, but prefer public 2282 addresses over temporary addresses. 2283 * > 1 : enable Privacy Extensions and prefer temporary 2284 addresses over public addresses. 2285 2286 Default: 2287 2288 * 0 (for most devices) 2289 * -1 (for point-to-point devices and loopback devices) 2290 2291temp_valid_lft - INTEGER 2292 valid lifetime (in seconds) for temporary addresses. 2293 2294 Default: 172800 (2 days) 2295 2296temp_prefered_lft - INTEGER 2297 Preferred lifetime (in seconds) for temporary addresses. 2298 2299 Default: 86400 (1 day) 2300 2301keep_addr_on_down - INTEGER 2302 Keep all IPv6 addresses on an interface down event. If set static 2303 global addresses with no expiration time are not flushed. 2304 2305 * >0 : enabled 2306 * 0 : system default 2307 * <0 : disabled 2308 2309 Default: 0 (addresses are removed) 2310 2311max_desync_factor - INTEGER 2312 Maximum value for DESYNC_FACTOR, which is a random value 2313 that ensures that clients don't synchronize with each 2314 other and generate new addresses at exactly the same time. 2315 value is in seconds. 2316 2317 Default: 600 2318 2319regen_max_retry - INTEGER 2320 Number of attempts before give up attempting to generate 2321 valid temporary addresses. 2322 2323 Default: 5 2324 2325max_addresses - INTEGER 2326 Maximum number of autoconfigured addresses per interface. Setting 2327 to zero disables the limitation. It is not recommended to set this 2328 value too large (or to zero) because it would be an easy way to 2329 crash the kernel by allowing too many addresses to be created. 2330 2331 Default: 16 2332 2333disable_ipv6 - BOOLEAN 2334 Disable IPv6 operation. If accept_dad is set to 2, this value 2335 will be dynamically set to TRUE if DAD fails for the link-local 2336 address. 2337 2338 Default: FALSE (enable IPv6 operation) 2339 2340 When this value is changed from 1 to 0 (IPv6 is being enabled), 2341 it will dynamically create a link-local address on the given 2342 interface and start Duplicate Address Detection, if necessary. 2343 2344 When this value is changed from 0 to 1 (IPv6 is being disabled), 2345 it will dynamically delete all addresses and routes on the given 2346 interface. From now on it will not possible to add addresses/routes 2347 to the selected interface. 2348 2349accept_dad - INTEGER 2350 Whether to accept DAD (Duplicate Address Detection). 2351 2352 == ============================================================== 2353 0 Disable DAD 2354 1 Enable DAD (default) 2355 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate 2356 link-local address has been found. 2357 == ============================================================== 2358 2359 DAD operation and mode on a given interface will be selected according 2360 to the maximum value of conf/{all,interface}/accept_dad. 2361 2362force_tllao - BOOLEAN 2363 Enable sending the target link-layer address option even when 2364 responding to a unicast neighbor solicitation. 2365 2366 Default: FALSE 2367 2368 Quoting from RFC 2461, section 4.4, Target link-layer address: 2369 2370 "The option MUST be included for multicast solicitations in order to 2371 avoid infinite Neighbor Solicitation "recursion" when the peer node 2372 does not have a cache entry to return a Neighbor Advertisements 2373 message. When responding to unicast solicitations, the option can be 2374 omitted since the sender of the solicitation has the correct link- 2375 layer address; otherwise it would not have be able to send the unicast 2376 solicitation in the first place. However, including the link-layer 2377 address in this case adds little overhead and eliminates a potential 2378 race condition where the sender deletes the cached link-layer address 2379 prior to receiving a response to a previous solicitation." 2380 2381ndisc_notify - BOOLEAN 2382 Define mode for notification of address and device changes. 2383 2384 * 0 - (default): do nothing 2385 * 1 - Generate unsolicited neighbour advertisements when device is brought 2386 up or hardware address changes. 2387 2388ndisc_tclass - INTEGER 2389 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor 2390 Discovery (Router Solicitation, Router Advertisement, Neighbor 2391 Solicitation, Neighbor Advertisement, Redirect) messages. 2392 These 8 bits can be interpreted as 6 high order bits holding the DSCP 2393 value and 2 low order bits representing ECN (which you probably want 2394 to leave cleared). 2395 2396 * 0 - (default) 2397 2398ndisc_evict_nocarrier - BOOLEAN 2399 Clears the neighbor discovery table on NOCARRIER events. This option is 2400 important for wireless devices where the neighbor discovery cache should 2401 not be cleared when roaming between access points on the same network. 2402 In most cases this should remain as the default (1). 2403 2404 - 1 - (default): Clear neighbor discover cache on NOCARRIER events. 2405 - 0 - Do not clear neighbor discovery cache on NOCARRIER events. 2406 2407mldv1_unsolicited_report_interval - INTEGER 2408 The interval in milliseconds in which the next unsolicited 2409 MLDv1 report retransmit will take place. 2410 2411 Default: 10000 (10 seconds) 2412 2413mldv2_unsolicited_report_interval - INTEGER 2414 The interval in milliseconds in which the next unsolicited 2415 MLDv2 report retransmit will take place. 2416 2417 Default: 1000 (1 second) 2418 2419force_mld_version - INTEGER 2420 * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed 2421 * 1 - Enforce to use MLD version 1 2422 * 2 - Enforce to use MLD version 2 2423 2424suppress_frag_ndisc - INTEGER 2425 Control RFC 6980 (Security Implications of IPv6 Fragmentation 2426 with IPv6 Neighbor Discovery) behavior: 2427 2428 * 1 - (default) discard fragmented neighbor discovery packets 2429 * 0 - allow fragmented neighbor discovery packets 2430 2431optimistic_dad - BOOLEAN 2432 Whether to perform Optimistic Duplicate Address Detection (RFC 4429). 2433 2434 * 0: disabled (default) 2435 * 1: enabled 2436 2437 Optimistic Duplicate Address Detection for the interface will be enabled 2438 if at least one of conf/{all,interface}/optimistic_dad is set to 1, 2439 it will be disabled otherwise. 2440 2441use_optimistic - BOOLEAN 2442 If enabled, do not classify optimistic addresses as deprecated during 2443 source address selection. Preferred addresses will still be chosen 2444 before optimistic addresses, subject to other ranking in the source 2445 address selection algorithm. 2446 2447 * 0: disabled (default) 2448 * 1: enabled 2449 2450 This will be enabled if at least one of 2451 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise. 2452 2453stable_secret - IPv6 address 2454 This IPv6 address will be used as a secret to generate IPv6 2455 addresses for link-local addresses and autoconfigured 2456 ones. All addresses generated after setting this secret will 2457 be stable privacy ones by default. This can be changed via the 2458 addrgenmode ip-link. conf/default/stable_secret is used as the 2459 secret for the namespace, the interface specific ones can 2460 overwrite that. Writes to conf/all/stable_secret are refused. 2461 2462 It is recommended to generate this secret during installation 2463 of a system and keep it stable after that. 2464 2465 By default the stable secret is unset. 2466 2467addr_gen_mode - INTEGER 2468 Defines how link-local and autoconf addresses are generated. 2469 2470 = ================================================================= 2471 0 generate address based on EUI64 (default) 2472 1 do no generate a link-local address, use EUI64 for addresses 2473 generated from autoconf 2474 2 generate stable privacy addresses, using the secret from 2475 stable_secret (RFC7217) 2476 3 generate stable privacy addresses, using a random secret if unset 2477 = ================================================================= 2478 2479drop_unicast_in_l2_multicast - BOOLEAN 2480 Drop any unicast IPv6 packets that are received in link-layer 2481 multicast (or broadcast) frames. 2482 2483 By default this is turned off. 2484 2485drop_unsolicited_na - BOOLEAN 2486 Drop all unsolicited neighbor advertisements, for example if there's 2487 a known good NA proxy on the network and such frames need not be used 2488 (or in the case of 802.11, must not be used to prevent attacks.) 2489 2490 By default this is turned off. 2491 2492accept_untracked_na - INTEGER 2493 Define behavior for accepting neighbor advertisements from devices that 2494 are absent in the neighbor cache: 2495 2496 - 0 - (default) Do not accept unsolicited and untracked neighbor 2497 advertisements. 2498 2499 - 1 - Add a new neighbor cache entry in STALE state for routers on 2500 receiving a neighbor advertisement (either solicited or unsolicited) 2501 with target link-layer address option specified if no neighbor entry 2502 is already present for the advertised IPv6 address. Without this knob, 2503 NAs received for untracked addresses (absent in neighbor cache) are 2504 silently ignored. 2505 2506 This is as per router-side behavior documented in RFC9131. 2507 2508 This has lower precedence than drop_unsolicited_na. 2509 2510 This will optimize the return path for the initial off-link 2511 communication that is initiated by a directly connected host, by 2512 ensuring that the first-hop router which turns on this setting doesn't 2513 have to buffer the initial return packets to do neighbor-solicitation. 2514 The prerequisite is that the host is configured to send unsolicited 2515 neighbor advertisements on interface bringup. This setting should be 2516 used in conjunction with the ndisc_notify setting on the host to 2517 satisfy this prerequisite. 2518 2519 - 2 - Extend option (1) to add a new neighbor cache entry only if the 2520 source IP address is in the same subnet as an address configured on 2521 the interface that received the neighbor advertisement. 2522 2523enhanced_dad - BOOLEAN 2524 Include a nonce option in the IPv6 neighbor solicitation messages used for 2525 duplicate address detection per RFC7527. A received DAD NS will only signal 2526 a duplicate address if the nonce is different. This avoids any false 2527 detection of duplicates due to loopback of the NS messages that we send. 2528 The nonce option will be sent on an interface unless both of 2529 conf/{all,interface}/enhanced_dad are set to FALSE. 2530 2531 Default: TRUE 2532 2533``icmp/*``: 2534=========== 2535 2536ratelimit - INTEGER 2537 Limit the maximal rates for sending ICMPv6 messages. 2538 2539 0 to disable any limiting, 2540 otherwise the minimal space between responses in milliseconds. 2541 2542 Default: 1000 2543 2544ratemask - list of comma separated ranges 2545 For ICMPv6 message types matching the ranges in the ratemask, limit 2546 the sending of the message according to ratelimit parameter. 2547 2548 The format used for both input and output is a comma separated 2549 list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and 2550 129). Writing to the file will clear all previous ranges of ICMPv6 2551 message types and update the current list with the input. 2552 2553 Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml 2554 for numerical values of ICMPv6 message types, e.g. echo request is 128 2555 and echo reply is 129. 2556 2557 Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big) 2558 2559echo_ignore_all - BOOLEAN 2560 If set non-zero, then the kernel will ignore all ICMP ECHO 2561 requests sent to it over the IPv6 protocol. 2562 2563 Default: 0 2564 2565echo_ignore_multicast - BOOLEAN 2566 If set non-zero, then the kernel will ignore all ICMP ECHO 2567 requests sent to it over the IPv6 protocol via multicast. 2568 2569 Default: 0 2570 2571echo_ignore_anycast - BOOLEAN 2572 If set non-zero, then the kernel will ignore all ICMP ECHO 2573 requests sent to it over the IPv6 protocol destined to anycast address. 2574 2575 Default: 0 2576 2577xfrm6_gc_thresh - INTEGER 2578 (Obsolete since linux-4.14) 2579 The threshold at which we will start garbage collecting for IPv6 2580 destination cache entries. At twice this value the system will 2581 refuse new allocations. 2582 2583 2584IPv6 Update by: 2585Pekka Savola <pekkas@netcore.fi> 2586YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org> 2587 2588 2589/proc/sys/net/bridge/* Variables: 2590================================= 2591 2592bridge-nf-call-arptables - BOOLEAN 2593 - 1 : pass bridged ARP traffic to arptables' FORWARD chain. 2594 - 0 : disable this. 2595 2596 Default: 1 2597 2598bridge-nf-call-iptables - BOOLEAN 2599 - 1 : pass bridged IPv4 traffic to iptables' chains. 2600 - 0 : disable this. 2601 2602 Default: 1 2603 2604bridge-nf-call-ip6tables - BOOLEAN 2605 - 1 : pass bridged IPv6 traffic to ip6tables' chains. 2606 - 0 : disable this. 2607 2608 Default: 1 2609 2610bridge-nf-filter-vlan-tagged - BOOLEAN 2611 - 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables. 2612 - 0 : disable this. 2613 2614 Default: 0 2615 2616bridge-nf-filter-pppoe-tagged - BOOLEAN 2617 - 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables. 2618 - 0 : disable this. 2619 2620 Default: 0 2621 2622bridge-nf-pass-vlan-input-dev - BOOLEAN 2623 - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan 2624 interface on the bridge and set the netfilter input device to the 2625 vlan. This allows use of e.g. "iptables -i br0.1" and makes the 2626 REDIRECT target work with vlan-on-top-of-bridge interfaces. When no 2627 matching vlan interface is found, or this switch is off, the input 2628 device is set to the bridge interface. 2629 2630 - 0: disable bridge netfilter vlan interface lookup. 2631 2632 Default: 0 2633 2634``proc/sys/net/sctp/*`` Variables: 2635================================== 2636 2637addip_enable - BOOLEAN 2638 Enable or disable extension of Dynamic Address Reconfiguration 2639 (ADD-IP) functionality specified in RFC5061. This extension provides 2640 the ability to dynamically add and remove new addresses for the SCTP 2641 associations. 2642 2643 1: Enable extension. 2644 2645 0: Disable extension. 2646 2647 Default: 0 2648 2649pf_enable - INTEGER 2650 Enable or disable pf (pf is short for potentially failed) state. A value 2651 of pf_retrans > path_max_retrans also disables pf state. That is, one of 2652 both pf_enable and pf_retrans > path_max_retrans can disable pf state. 2653 Since pf_retrans and path_max_retrans can be changed by userspace 2654 application, sometimes user expects to disable pf state by the value of 2655 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans 2656 or path_max_retrans is changed by the user application, this pf state is 2657 enabled. As such, it is necessary to add this to dynamically enable 2658 and disable pf state. See: 2659 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for 2660 details. 2661 2662 1: Enable pf. 2663 2664 0: Disable pf. 2665 2666 Default: 1 2667 2668pf_expose - INTEGER 2669 Unset or enable/disable pf (pf is short for potentially failed) state 2670 exposure. Applications can control the exposure of the PF path state 2671 in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO 2672 sockopt. When it's unset, no SCTP_PEER_ADDR_CHANGE event with 2673 SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info 2674 can be got via SCTP_GET_PEER_ADDR_INFO sockopt; When it's enabled, 2675 a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming 2676 SCTP_PF state and a SCTP_PF-state transport info can be got via 2677 SCTP_GET_PEER_ADDR_INFO sockopt; When it's diabled, no 2678 SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when 2679 trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO 2680 sockopt. 2681 2682 0: Unset pf state exposure, Compatible with old applications. 2683 2684 1: Disable pf state exposure. 2685 2686 2: Enable pf state exposure. 2687 2688 Default: 0 2689 2690addip_noauth_enable - BOOLEAN 2691 Dynamic Address Reconfiguration (ADD-IP) requires the use of 2692 authentication to protect the operations of adding or removing new 2693 addresses. This requirement is mandated so that unauthorized hosts 2694 would not be able to hijack associations. However, older 2695 implementations may not have implemented this requirement while 2696 allowing the ADD-IP extension. For reasons of interoperability, 2697 we provide this variable to control the enforcement of the 2698 authentication requirement. 2699 2700 == =============================================================== 2701 1 Allow ADD-IP extension to be used without authentication. This 2702 should only be set in a closed environment for interoperability 2703 with older implementations. 2704 2705 0 Enforce the authentication requirement 2706 == =============================================================== 2707 2708 Default: 0 2709 2710auth_enable - BOOLEAN 2711 Enable or disable Authenticated Chunks extension. This extension 2712 provides the ability to send and receive authenticated chunks and is 2713 required for secure operation of Dynamic Address Reconfiguration 2714 (ADD-IP) extension. 2715 2716 - 1: Enable this extension. 2717 - 0: Disable this extension. 2718 2719 Default: 0 2720 2721prsctp_enable - BOOLEAN 2722 Enable or disable the Partial Reliability extension (RFC3758) which 2723 is used to notify peers that a given DATA should no longer be expected. 2724 2725 - 1: Enable extension 2726 - 0: Disable 2727 2728 Default: 1 2729 2730max_burst - INTEGER 2731 The limit of the number of new packets that can be initially sent. It 2732 controls how bursty the generated traffic can be. 2733 2734 Default: 4 2735 2736association_max_retrans - INTEGER 2737 Set the maximum number for retransmissions that an association can 2738 attempt deciding that the remote end is unreachable. If this value 2739 is exceeded, the association is terminated. 2740 2741 Default: 10 2742 2743max_init_retransmits - INTEGER 2744 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks 2745 that an association will attempt before declaring the destination 2746 unreachable and terminating. 2747 2748 Default: 8 2749 2750path_max_retrans - INTEGER 2751 The maximum number of retransmissions that will be attempted on a given 2752 path. Once this threshold is exceeded, the path is considered 2753 unreachable, and new traffic will use a different path when the 2754 association is multihomed. 2755 2756 Default: 5 2757 2758pf_retrans - INTEGER 2759 The number of retransmissions that will be attempted on a given path 2760 before traffic is redirected to an alternate transport (should one 2761 exist). Note this is distinct from path_max_retrans, as a path that 2762 passes the pf_retrans threshold can still be used. Its only 2763 deprioritized when a transmission path is selected by the stack. This 2764 setting is primarily used to enable fast failover mechanisms without 2765 having to reduce path_max_retrans to a very low value. See: 2766 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt 2767 for details. Note also that a value of pf_retrans > path_max_retrans 2768 disables this feature. Since both pf_retrans and path_max_retrans can 2769 be changed by userspace application, a variable pf_enable is used to 2770 disable pf state. 2771 2772 Default: 0 2773 2774ps_retrans - INTEGER 2775 Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming 2776 from section-5 "Primary Path Switchover" in rfc7829. The primary path 2777 will be changed to another active path when the path error counter on 2778 the old primary path exceeds PSMR, so that "the SCTP sender is allowed 2779 to continue data transmission on a new working path even when the old 2780 primary destination address becomes active again". Note this feature 2781 is disabled by initializing 'ps_retrans' per netns as 0xffff by default, 2782 and its value can't be less than 'pf_retrans' when changing by sysctl. 2783 2784 Default: 0xffff 2785 2786rto_initial - INTEGER 2787 The initial round trip timeout value in milliseconds that will be used 2788 in calculating round trip times. This is the initial time interval 2789 for retransmissions. 2790 2791 Default: 3000 2792 2793rto_max - INTEGER 2794 The maximum value (in milliseconds) of the round trip timeout. This 2795 is the largest time interval that can elapse between retransmissions. 2796 2797 Default: 60000 2798 2799rto_min - INTEGER 2800 The minimum value (in milliseconds) of the round trip timeout. This 2801 is the smallest time interval the can elapse between retransmissions. 2802 2803 Default: 1000 2804 2805hb_interval - INTEGER 2806 The interval (in milliseconds) between HEARTBEAT chunks. These chunks 2807 are sent at the specified interval on idle paths to probe the state of 2808 a given path between 2 associations. 2809 2810 Default: 30000 2811 2812sack_timeout - INTEGER 2813 The amount of time (in milliseconds) that the implementation will wait 2814 to send a SACK. 2815 2816 Default: 200 2817 2818valid_cookie_life - INTEGER 2819 The default lifetime of the SCTP cookie (in milliseconds). The cookie 2820 is used during association establishment. 2821 2822 Default: 60000 2823 2824cookie_preserve_enable - BOOLEAN 2825 Enable or disable the ability to extend the lifetime of the SCTP cookie 2826 that is used during the establishment phase of SCTP association 2827 2828 - 1: Enable cookie lifetime extension. 2829 - 0: Disable 2830 2831 Default: 1 2832 2833cookie_hmac_alg - STRING 2834 Select the hmac algorithm used when generating the cookie value sent by 2835 a listening sctp socket to a connecting client in the INIT-ACK chunk. 2836 Valid values are: 2837 2838 * md5 2839 * sha1 2840 * none 2841 2842 Ability to assign md5 or sha1 as the selected alg is predicated on the 2843 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and 2844 CONFIG_CRYPTO_SHA1). 2845 2846 Default: Dependent on configuration. MD5 if available, else SHA1 if 2847 available, else none. 2848 2849rcvbuf_policy - INTEGER 2850 Determines if the receive buffer is attributed to the socket or to 2851 association. SCTP supports the capability to create multiple 2852 associations on a single socket. When using this capability, it is 2853 possible that a single stalled association that's buffering a lot 2854 of data may block other associations from delivering their data by 2855 consuming all of the receive buffer space. To work around this, 2856 the rcvbuf_policy could be set to attribute the receiver buffer space 2857 to each association instead of the socket. This prevents the described 2858 blocking. 2859 2860 - 1: rcvbuf space is per association 2861 - 0: rcvbuf space is per socket 2862 2863 Default: 0 2864 2865sndbuf_policy - INTEGER 2866 Similar to rcvbuf_policy above, this applies to send buffer space. 2867 2868 - 1: Send buffer is tracked per association 2869 - 0: Send buffer is tracked per socket. 2870 2871 Default: 0 2872 2873sctp_mem - vector of 3 INTEGERs: min, pressure, max 2874 Number of pages allowed for queueing by all SCTP sockets. 2875 2876 min: Below this number of pages SCTP is not bothered about its 2877 memory appetite. When amount of memory allocated by SCTP exceeds 2878 this number, SCTP starts to moderate memory usage. 2879 2880 pressure: This value was introduced to follow format of tcp_mem. 2881 2882 max: Number of pages allowed for queueing by all SCTP sockets. 2883 2884 Default is calculated at boot time from amount of available memory. 2885 2886sctp_rmem - vector of 3 INTEGERs: min, default, max 2887 Only the first value ("min") is used, "default" and "max" are 2888 ignored. 2889 2890 min: Minimal size of receive buffer used by SCTP socket. 2891 It is guaranteed to each SCTP socket (but not association) even 2892 under moderate memory pressure. 2893 2894 Default: 4K 2895 2896sctp_wmem - vector of 3 INTEGERs: min, default, max 2897 Only the first value ("min") is used, "default" and "max" are 2898 ignored. 2899 2900 min: Minimum size of send buffer that can be used by SCTP sockets. 2901 It is guaranteed to each SCTP socket (but not association) even 2902 under moderate memory pressure. 2903 2904 Default: 4K 2905 2906addr_scope_policy - INTEGER 2907 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00 2908 2909 - 0 - Disable IPv4 address scoping 2910 - 1 - Enable IPv4 address scoping 2911 - 2 - Follow draft but allow IPv4 private addresses 2912 - 3 - Follow draft but allow IPv4 link local addresses 2913 2914 Default: 1 2915 2916udp_port - INTEGER 2917 The listening port for the local UDP tunneling sock. Normally it's 2918 using the IANA-assigned UDP port number 9899 (sctp-tunneling). 2919 2920 This UDP sock is used for processing the incoming UDP-encapsulated 2921 SCTP packets (from RFC6951), and shared by all applications in the 2922 same net namespace. This UDP sock will be closed when the value is 2923 set to 0. 2924 2925 The value will also be used to set the src port of the UDP header 2926 for the outgoing UDP-encapsulated SCTP packets. For the dest port, 2927 please refer to 'encap_port' below. 2928 2929 Default: 0 2930 2931encap_port - INTEGER 2932 The default remote UDP encapsulation port. 2933 2934 This value is used to set the dest port of the UDP header for the 2935 outgoing UDP-encapsulated SCTP packets by default. Users can also 2936 change the value for each sock/asoc/transport by using setsockopt. 2937 For further information, please refer to RFC6951. 2938 2939 Note that when connecting to a remote server, the client should set 2940 this to the port that the UDP tunneling sock on the peer server is 2941 listening to and the local UDP tunneling sock on the client also 2942 must be started. On the server, it would get the encap_port from 2943 the incoming packet's source port. 2944 2945 Default: 0 2946 2947plpmtud_probe_interval - INTEGER 2948 The time interval (in milliseconds) for the PLPMTUD probe timer, 2949 which is configured to expire after this period to receive an 2950 acknowledgment to a probe packet. This is also the time interval 2951 between the probes for the current pmtu when the probe search 2952 is done. 2953 2954 PLPMTUD will be disabled when 0 is set, and other values for it 2955 must be >= 5000. 2956 2957 Default: 0 2958 2959reconf_enable - BOOLEAN 2960 Enable or disable extension of Stream Reconfiguration functionality 2961 specified in RFC6525. This extension provides the ability to "reset" 2962 a stream, and it includes the Parameters of "Outgoing/Incoming SSN 2963 Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams". 2964 2965 - 1: Enable extension. 2966 - 0: Disable extension. 2967 2968 Default: 0 2969 2970intl_enable - BOOLEAN 2971 Enable or disable extension of User Message Interleaving functionality 2972 specified in RFC8260. This extension allows the interleaving of user 2973 messages sent on different streams. With this feature enabled, I-DATA 2974 chunk will replace DATA chunk to carry user messages if also supported 2975 by the peer. Note that to use this feature, one needs to set this option 2976 to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2 2977 and SCTP_INTERLEAVING_SUPPORTED to 1. 2978 2979 - 1: Enable extension. 2980 - 0: Disable extension. 2981 2982 Default: 0 2983 2984ecn_enable - BOOLEAN 2985 Control use of Explicit Congestion Notification (ECN) by SCTP. 2986 Like in TCP, ECN is used only when both ends of the SCTP connection 2987 indicate support for it. This feature is useful in avoiding losses 2988 due to congestion by allowing supporting routers to signal congestion 2989 before having to drop packets. 2990 2991 1: Enable ecn. 2992 0: Disable ecn. 2993 2994 Default: 1 2995 2996 2997``/proc/sys/net/core/*`` 2998======================== 2999 3000 Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries. 3001 3002 3003``/proc/sys/net/unix/*`` 3004======================== 3005 3006max_dgram_qlen - INTEGER 3007 The maximum length of dgram socket receive queue 3008 3009 Default: 10 3010 3011