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 min_pmtu (see below). You will need 29 to raise min_pmtu to the smallest interface MTU on your system 30 manually if you want to avoid locally generated fragments. 31 32 In mode 2 incoming Path MTU Discovery messages will be 33 discarded. Outgoing frames are handled the same as in mode 1, 34 implicitly setting IP_PMTUDISC_DONT on every created socket. 35 36 Mode 3 is a hardened pmtu discover mode. The kernel will only 37 accept fragmentation-needed errors if the underlying protocol 38 can verify them besides a plain socket lookup. Current 39 protocols for which pmtu events will be honored are TCP, SCTP 40 and DCCP as they verify e.g. the sequence number or the 41 association. This mode should not be enabled globally but is 42 only intended to secure e.g. name servers in namespaces where 43 TCP path mtu must still work but path MTU information of other 44 protocols should be discarded. If enabled globally this mode 45 could break other protocols. 46 47 Possible values: 0-3 48 49 Default: FALSE 50 51min_pmtu - INTEGER 52 default 552 - minimum discovered Path MTU 53 54ip_forward_use_pmtu - BOOLEAN 55 By default we don't trust protocol path MTUs while forwarding 56 because they could be easily forged and can lead to unwanted 57 fragmentation by the router. 58 You only need to enable this if you have user-space software 59 which tries to discover path mtus by itself and depends on the 60 kernel honoring this information. This is normally not the 61 case. 62 63 Default: 0 (disabled) 64 65 Possible values: 66 67 - 0 - disabled 68 - 1 - enabled 69 70fwmark_reflect - BOOLEAN 71 Controls the fwmark of kernel-generated IPv4 reply packets that are not 72 associated with a socket for example, TCP RSTs or ICMP echo replies). 73 If unset, these packets have a fwmark of zero. If set, they have the 74 fwmark of the packet they are replying to. 75 76 Default: 0 77 78fib_multipath_use_neigh - BOOLEAN 79 Use status of existing neighbor entry when determining nexthop for 80 multipath routes. If disabled, neighbor information is not used and 81 packets could be directed to a failed nexthop. Only valid for kernels 82 built with CONFIG_IP_ROUTE_MULTIPATH enabled. 83 84 Default: 0 (disabled) 85 86 Possible values: 87 88 - 0 - disabled 89 - 1 - enabled 90 91fib_multipath_hash_policy - INTEGER 92 Controls which hash policy to use for multipath routes. Only valid 93 for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled. 94 95 Default: 0 (Layer 3) 96 97 Possible values: 98 99 - 0 - Layer 3 100 - 1 - Layer 4 101 - 2 - Layer 3 or inner Layer 3 if present 102 103fib_sync_mem - UNSIGNED INTEGER 104 Amount of dirty memory from fib entries that can be backlogged before 105 synchronize_rcu is forced. 106 107 Default: 512kB Minimum: 64kB Maximum: 64MB 108 109ip_forward_update_priority - INTEGER 110 Whether to update SKB priority from "TOS" field in IPv4 header after it 111 is forwarded. The new SKB priority is mapped from TOS field value 112 according to an rt_tos2priority table (see e.g. man tc-prio). 113 114 Default: 1 (Update priority.) 115 116 Possible values: 117 118 - 0 - Do not update priority. 119 - 1 - Update priority. 120 121route/max_size - INTEGER 122 Maximum number of routes allowed in the kernel. Increase 123 this when using large numbers of interfaces and/or routes. 124 125 From linux kernel 3.6 onwards, this is deprecated for ipv4 126 as route cache is no longer used. 127 128neigh/default/gc_thresh1 - INTEGER 129 Minimum number of entries to keep. Garbage collector will not 130 purge entries if there are fewer than this number. 131 132 Default: 128 133 134neigh/default/gc_thresh2 - INTEGER 135 Threshold when garbage collector becomes more aggressive about 136 purging entries. Entries older than 5 seconds will be cleared 137 when over this number. 138 139 Default: 512 140 141neigh/default/gc_thresh3 - INTEGER 142 Maximum number of non-PERMANENT neighbor entries allowed. Increase 143 this when using large numbers of interfaces and when communicating 144 with large numbers of directly-connected peers. 145 146 Default: 1024 147 148neigh/default/unres_qlen_bytes - INTEGER 149 The maximum number of bytes which may be used by packets 150 queued for each unresolved address by other network layers. 151 (added in linux 3.3) 152 153 Setting negative value is meaningless and will return error. 154 155 Default: SK_WMEM_MAX, (same as net.core.wmem_default). 156 157 Exact value depends on architecture and kernel options, 158 but should be enough to allow queuing 256 packets 159 of medium size. 160 161neigh/default/unres_qlen - INTEGER 162 The maximum number of packets which may be queued for each 163 unresolved address by other network layers. 164 165 (deprecated in linux 3.3) : use unres_qlen_bytes instead. 166 167 Prior to linux 3.3, the default value is 3 which may cause 168 unexpected packet loss. The current default value is calculated 169 according to default value of unres_qlen_bytes and true size of 170 packet. 171 172 Default: 101 173 174mtu_expires - INTEGER 175 Time, in seconds, that cached PMTU information is kept. 176 177min_adv_mss - INTEGER 178 The advertised MSS depends on the first hop route MTU, but will 179 never be lower than this setting. 180 181IP Fragmentation: 182 183ipfrag_high_thresh - LONG INTEGER 184 Maximum memory used to reassemble IP fragments. 185 186ipfrag_low_thresh - LONG INTEGER 187 (Obsolete since linux-4.17) 188 Maximum memory used to reassemble IP fragments before the kernel 189 begins to remove incomplete fragment queues to free up resources. 190 The kernel still accepts new fragments for defragmentation. 191 192ipfrag_time - INTEGER 193 Time in seconds to keep an IP fragment in memory. 194 195ipfrag_max_dist - INTEGER 196 ipfrag_max_dist is a non-negative integer value which defines the 197 maximum "disorder" which is allowed among fragments which share a 198 common IP source address. Note that reordering of packets is 199 not unusual, but if a large number of fragments arrive from a source 200 IP address while a particular fragment queue remains incomplete, it 201 probably indicates that one or more fragments belonging to that queue 202 have been lost. When ipfrag_max_dist is positive, an additional check 203 is done on fragments before they are added to a reassembly queue - if 204 ipfrag_max_dist (or more) fragments have arrived from a particular IP 205 address between additions to any IP fragment queue using that source 206 address, it's presumed that one or more fragments in the queue are 207 lost. The existing fragment queue will be dropped, and a new one 208 started. An ipfrag_max_dist value of zero disables this check. 209 210 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can 211 result in unnecessarily dropping fragment queues when normal 212 reordering of packets occurs, which could lead to poor application 213 performance. Using a very large value, e.g. 50000, increases the 214 likelihood of incorrectly reassembling IP fragments that originate 215 from different IP datagrams, which could result in data corruption. 216 Default: 64 217 218INET peer storage 219================= 220 221inet_peer_threshold - INTEGER 222 The approximate size of the storage. Starting from this threshold 223 entries will be thrown aggressively. This threshold also determines 224 entries' time-to-live and time intervals between garbage collection 225 passes. More entries, less time-to-live, less GC interval. 226 227inet_peer_minttl - INTEGER 228 Minimum time-to-live of entries. Should be enough to cover fragment 229 time-to-live on the reassembling side. This minimum time-to-live is 230 guaranteed if the pool size is less than inet_peer_threshold. 231 Measured in seconds. 232 233inet_peer_maxttl - INTEGER 234 Maximum time-to-live of entries. Unused entries will expire after 235 this period of time if there is no memory pressure on the pool (i.e. 236 when the number of entries in the pool is very small). 237 Measured in seconds. 238 239TCP variables 240============= 241 242somaxconn - INTEGER 243 Limit of socket listen() backlog, known in userspace as SOMAXCONN. 244 Defaults to 4096. (Was 128 before linux-5.4) 245 See also tcp_max_syn_backlog for additional tuning for TCP sockets. 246 247tcp_abort_on_overflow - BOOLEAN 248 If listening service is too slow to accept new connections, 249 reset them. Default state is FALSE. It means that if overflow 250 occurred due to a burst, connection will recover. Enable this 251 option _only_ if you are really sure that listening daemon 252 cannot be tuned to accept connections faster. Enabling this 253 option can harm clients of your server. 254 255tcp_adv_win_scale - INTEGER 256 Count buffering overhead as bytes/2^tcp_adv_win_scale 257 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale), 258 if it is <= 0. 259 260 Possible values are [-31, 31], inclusive. 261 262 Default: 1 263 264tcp_allowed_congestion_control - STRING 265 Show/set the congestion control choices available to non-privileged 266 processes. The list is a subset of those listed in 267 tcp_available_congestion_control. 268 269 Default is "reno" and the default setting (tcp_congestion_control). 270 271tcp_app_win - INTEGER 272 Reserve max(window/2^tcp_app_win, mss) of window for application 273 buffer. Value 0 is special, it means that nothing is reserved. 274 275 Default: 31 276 277tcp_autocorking - BOOLEAN 278 Enable TCP auto corking : 279 When applications do consecutive small write()/sendmsg() system calls, 280 we try to coalesce these small writes as much as possible, to lower 281 total amount of sent packets. This is done if at least one prior 282 packet for the flow is waiting in Qdisc queues or device transmit 283 queue. Applications can still use TCP_CORK for optimal behavior 284 when they know how/when to uncork their sockets. 285 286 Default : 1 287 288tcp_available_congestion_control - STRING 289 Shows the available congestion control choices that are registered. 290 More congestion control algorithms may be available as modules, 291 but not loaded. 292 293tcp_base_mss - INTEGER 294 The initial value of search_low to be used by the packetization layer 295 Path MTU discovery (MTU probing). If MTU probing is enabled, 296 this is the initial MSS used by the connection. 297 298tcp_mtu_probe_floor - INTEGER 299 If MTU probing is enabled this caps the minimum MSS used for search_low 300 for the connection. 301 302 Default : 48 303 304tcp_min_snd_mss - INTEGER 305 TCP SYN and SYNACK messages usually advertise an ADVMSS option, 306 as described in RFC 1122 and RFC 6691. 307 308 If this ADVMSS option is smaller than tcp_min_snd_mss, 309 it is silently capped to tcp_min_snd_mss. 310 311 Default : 48 (at least 8 bytes of payload per segment) 312 313tcp_congestion_control - STRING 314 Set the congestion control algorithm to be used for new 315 connections. The algorithm "reno" is always available, but 316 additional choices may be available based on kernel configuration. 317 Default is set as part of kernel configuration. 318 For passive connections, the listener congestion control choice 319 is inherited. 320 321 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ] 322 323tcp_dsack - BOOLEAN 324 Allows TCP to send "duplicate" SACKs. 325 326tcp_early_retrans - INTEGER 327 Tail loss probe (TLP) converts RTOs occurring due to tail 328 losses into fast recovery (draft-ietf-tcpm-rack). Note that 329 TLP requires RACK to function properly (see tcp_recovery below) 330 331 Possible values: 332 333 - 0 disables TLP 334 - 3 or 4 enables TLP 335 336 Default: 3 337 338tcp_ecn - INTEGER 339 Control use of Explicit Congestion Notification (ECN) by TCP. 340 ECN is used only when both ends of the TCP connection indicate 341 support for it. This feature is useful in avoiding losses due 342 to congestion by allowing supporting routers to signal 343 congestion before having to drop packets. 344 345 Possible values are: 346 347 = ===================================================== 348 0 Disable ECN. Neither initiate nor accept ECN. 349 1 Enable ECN when requested by incoming connections and 350 also request ECN on outgoing connection attempts. 351 2 Enable ECN when requested by incoming connections 352 but do not request ECN on outgoing connections. 353 = ===================================================== 354 355 Default: 2 356 357tcp_ecn_fallback - BOOLEAN 358 If the kernel detects that ECN connection misbehaves, enable fall 359 back to non-ECN. Currently, this knob implements the fallback 360 from RFC3168, section 6.1.1.1., but we reserve that in future, 361 additional detection mechanisms could be implemented under this 362 knob. The value is not used, if tcp_ecn or per route (or congestion 363 control) ECN settings are disabled. 364 365 Default: 1 (fallback enabled) 366 367tcp_fack - BOOLEAN 368 This is a legacy option, it has no effect anymore. 369 370tcp_fin_timeout - INTEGER 371 The length of time an orphaned (no longer referenced by any 372 application) connection will remain in the FIN_WAIT_2 state 373 before it is aborted at the local end. While a perfectly 374 valid "receive only" state for an un-orphaned connection, an 375 orphaned connection in FIN_WAIT_2 state could otherwise wait 376 forever for the remote to close its end of the connection. 377 378 Cf. tcp_max_orphans 379 380 Default: 60 seconds 381 382tcp_frto - INTEGER 383 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682. 384 F-RTO is an enhanced recovery algorithm for TCP retransmission 385 timeouts. It is particularly beneficial in networks where the 386 RTT fluctuates (e.g., wireless). F-RTO is sender-side only 387 modification. It does not require any support from the peer. 388 389 By default it's enabled with a non-zero value. 0 disables F-RTO. 390 391tcp_fwmark_accept - BOOLEAN 392 If set, incoming connections to listening sockets that do not have a 393 socket mark will set the mark of the accepting socket to the fwmark of 394 the incoming SYN packet. This will cause all packets on that connection 395 (starting from the first SYNACK) to be sent with that fwmark. The 396 listening socket's mark is unchanged. Listening sockets that already 397 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are 398 unaffected. 399 400 Default: 0 401 402tcp_invalid_ratelimit - INTEGER 403 Limit the maximal rate for sending duplicate acknowledgments 404 in response to incoming TCP packets that are for an existing 405 connection but that are invalid due to any of these reasons: 406 407 (a) out-of-window sequence number, 408 (b) out-of-window acknowledgment number, or 409 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure 410 411 This can help mitigate simple "ack loop" DoS attacks, wherein 412 a buggy or malicious middlebox or man-in-the-middle can 413 rewrite TCP header fields in manner that causes each endpoint 414 to think that the other is sending invalid TCP segments, thus 415 causing each side to send an unterminating stream of duplicate 416 acknowledgments for invalid segments. 417 418 Using 0 disables rate-limiting of dupacks in response to 419 invalid segments; otherwise this value specifies the minimal 420 space between sending such dupacks, in milliseconds. 421 422 Default: 500 (milliseconds). 423 424tcp_keepalive_time - INTEGER 425 How often TCP sends out keepalive messages when keepalive is enabled. 426 Default: 2hours. 427 428tcp_keepalive_probes - INTEGER 429 How many keepalive probes TCP sends out, until it decides that the 430 connection is broken. Default value: 9. 431 432tcp_keepalive_intvl - INTEGER 433 How frequently the probes are send out. Multiplied by 434 tcp_keepalive_probes it is time to kill not responding connection, 435 after probes started. Default value: 75sec i.e. connection 436 will be aborted after ~11 minutes of retries. 437 438tcp_l3mdev_accept - BOOLEAN 439 Enables child sockets to inherit the L3 master device index. 440 Enabling this option allows a "global" listen socket to work 441 across L3 master domains (e.g., VRFs) with connected sockets 442 derived from the listen socket to be bound to the L3 domain in 443 which the packets originated. Only valid when the kernel was 444 compiled with CONFIG_NET_L3_MASTER_DEV. 445 446 Default: 0 (disabled) 447 448tcp_low_latency - BOOLEAN 449 This is a legacy option, it has no effect anymore. 450 451tcp_max_orphans - INTEGER 452 Maximal number of TCP sockets not attached to any user file handle, 453 held by system. If this number is exceeded orphaned connections are 454 reset immediately and warning is printed. This limit exists 455 only to prevent simple DoS attacks, you _must_ not rely on this 456 or lower the limit artificially, but rather increase it 457 (probably, after increasing installed memory), 458 if network conditions require more than default value, 459 and tune network services to linger and kill such states 460 more aggressively. Let me to remind again: each orphan eats 461 up to ~64K of unswappable memory. 462 463tcp_max_syn_backlog - INTEGER 464 Maximal number of remembered connection requests (SYN_RECV), 465 which have not received an acknowledgment from connecting client. 466 467 This is a per-listener limit. 468 469 The minimal value is 128 for low memory machines, and it will 470 increase in proportion to the memory of machine. 471 472 If server suffers from overload, try increasing this number. 473 474 Remember to also check /proc/sys/net/core/somaxconn 475 A SYN_RECV request socket consumes about 304 bytes of memory. 476 477tcp_max_tw_buckets - INTEGER 478 Maximal number of timewait sockets held by system simultaneously. 479 If this number is exceeded time-wait socket is immediately destroyed 480 and warning is printed. This limit exists only to prevent 481 simple DoS attacks, you _must_ not lower the limit artificially, 482 but rather increase it (probably, after increasing installed memory), 483 if network conditions require more than default value. 484 485tcp_mem - vector of 3 INTEGERs: min, pressure, max 486 min: below this number of pages TCP is not bothered about its 487 memory appetite. 488 489 pressure: when amount of memory allocated by TCP exceeds this number 490 of pages, TCP moderates its memory consumption and enters memory 491 pressure mode, which is exited when memory consumption falls 492 under "min". 493 494 max: number of pages allowed for queueing by all TCP sockets. 495 496 Defaults are calculated at boot time from amount of available 497 memory. 498 499tcp_min_rtt_wlen - INTEGER 500 The window length of the windowed min filter to track the minimum RTT. 501 A shorter window lets a flow more quickly pick up new (higher) 502 minimum RTT when it is moved to a longer path (e.g., due to traffic 503 engineering). A longer window makes the filter more resistant to RTT 504 inflations such as transient congestion. The unit is seconds. 505 506 Possible values: 0 - 86400 (1 day) 507 508 Default: 300 509 510tcp_moderate_rcvbuf - BOOLEAN 511 If set, TCP performs receive buffer auto-tuning, attempting to 512 automatically size the buffer (no greater than tcp_rmem[2]) to 513 match the size required by the path for full throughput. Enabled by 514 default. 515 516tcp_mtu_probing - INTEGER 517 Controls TCP Packetization-Layer Path MTU Discovery. Takes three 518 values: 519 520 - 0 - Disabled 521 - 1 - Disabled by default, enabled when an ICMP black hole detected 522 - 2 - Always enabled, use initial MSS of tcp_base_mss. 523 524tcp_probe_interval - UNSIGNED INTEGER 525 Controls how often to start TCP Packetization-Layer Path MTU 526 Discovery reprobe. The default is reprobing every 10 minutes as 527 per RFC4821. 528 529tcp_probe_threshold - INTEGER 530 Controls when TCP Packetization-Layer Path MTU Discovery probing 531 will stop in respect to the width of search range in bytes. Default 532 is 8 bytes. 533 534tcp_no_metrics_save - BOOLEAN 535 By default, TCP saves various connection metrics in the route cache 536 when the connection closes, so that connections established in the 537 near future can use these to set initial conditions. Usually, this 538 increases overall performance, but may sometimes cause performance 539 degradation. If set, TCP will not cache metrics on closing 540 connections. 541 542tcp_no_ssthresh_metrics_save - BOOLEAN 543 Controls whether TCP saves ssthresh metrics in the route cache. 544 545 Default is 1, which disables ssthresh metrics. 546 547tcp_orphan_retries - INTEGER 548 This value influences the timeout of a locally closed TCP connection, 549 when RTO retransmissions remain unacknowledged. 550 See tcp_retries2 for more details. 551 552 The default value is 8. 553 554 If your machine is a loaded WEB server, 555 you should think about lowering this value, such sockets 556 may consume significant resources. Cf. tcp_max_orphans. 557 558tcp_recovery - INTEGER 559 This value is a bitmap to enable various experimental loss recovery 560 features. 561 562 ========= ============================================================= 563 RACK: 0x1 enables the RACK loss detection for fast detection of lost 564 retransmissions and tail drops. It also subsumes and disables 565 RFC6675 recovery for SACK connections. 566 567 RACK: 0x2 makes RACK's reordering window static (min_rtt/4). 568 569 RACK: 0x4 disables RACK's DUPACK threshold heuristic 570 ========= ============================================================= 571 572 Default: 0x1 573 574tcp_reordering - INTEGER 575 Initial reordering level of packets in a TCP stream. 576 TCP stack can then dynamically adjust flow reordering level 577 between this initial value and tcp_max_reordering 578 579 Default: 3 580 581tcp_max_reordering - INTEGER 582 Maximal reordering level of packets in a TCP stream. 583 300 is a fairly conservative value, but you might increase it 584 if paths are using per packet load balancing (like bonding rr mode) 585 586 Default: 300 587 588tcp_retrans_collapse - BOOLEAN 589 Bug-to-bug compatibility with some broken printers. 590 On retransmit try to send bigger packets to work around bugs in 591 certain TCP stacks. 592 593tcp_retries1 - INTEGER 594 This value influences the time, after which TCP decides, that 595 something is wrong due to unacknowledged RTO retransmissions, 596 and reports this suspicion to the network layer. 597 See tcp_retries2 for more details. 598 599 RFC 1122 recommends at least 3 retransmissions, which is the 600 default. 601 602tcp_retries2 - INTEGER 603 This value influences the timeout of an alive TCP connection, 604 when RTO retransmissions remain unacknowledged. 605 Given a value of N, a hypothetical TCP connection following 606 exponential backoff with an initial RTO of TCP_RTO_MIN would 607 retransmit N times before killing the connection at the (N+1)th RTO. 608 609 The default value of 15 yields a hypothetical timeout of 924.6 610 seconds and is a lower bound for the effective timeout. 611 TCP will effectively time out at the first RTO which exceeds the 612 hypothetical timeout. 613 614 RFC 1122 recommends at least 100 seconds for the timeout, 615 which corresponds to a value of at least 8. 616 617tcp_rfc1337 - BOOLEAN 618 If set, the TCP stack behaves conforming to RFC1337. If unset, 619 we are not conforming to RFC, but prevent TCP TIME_WAIT 620 assassination. 621 622 Default: 0 623 624tcp_rmem - vector of 3 INTEGERs: min, default, max 625 min: Minimal size of receive buffer used by TCP sockets. 626 It is guaranteed to each TCP socket, even under moderate memory 627 pressure. 628 629 Default: 4K 630 631 default: initial size of receive buffer used by TCP sockets. 632 This value overrides net.core.rmem_default used by other protocols. 633 Default: 87380 bytes. This value results in window of 65535 with 634 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit 635 less for default tcp_app_win. See below about these variables. 636 637 max: maximal size of receive buffer allowed for automatically 638 selected receiver buffers for TCP socket. This value does not override 639 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables 640 automatic tuning of that socket's receive buffer size, in which 641 case this value is ignored. 642 Default: between 87380B and 6MB, depending on RAM size. 643 644tcp_sack - BOOLEAN 645 Enable select acknowledgments (SACKS). 646 647tcp_comp_sack_delay_ns - LONG INTEGER 648 TCP tries to reduce number of SACK sent, using a timer 649 based on 5% of SRTT, capped by this sysctl, in nano seconds. 650 The default is 1ms, based on TSO autosizing period. 651 652 Default : 1,000,000 ns (1 ms) 653 654tcp_comp_sack_slack_ns - LONG INTEGER 655 This sysctl control the slack used when arming the 656 timer used by SACK compression. This gives extra time 657 for small RTT flows, and reduces system overhead by allowing 658 opportunistic reduction of timer interrupts. 659 660 Default : 100,000 ns (100 us) 661 662tcp_comp_sack_nr - INTEGER 663 Max number of SACK that can be compressed. 664 Using 0 disables SACK compression. 665 666 Default : 44 667 668tcp_slow_start_after_idle - BOOLEAN 669 If set, provide RFC2861 behavior and time out the congestion 670 window after an idle period. An idle period is defined at 671 the current RTO. If unset, the congestion window will not 672 be timed out after an idle period. 673 674 Default: 1 675 676tcp_stdurg - BOOLEAN 677 Use the Host requirements interpretation of the TCP urgent pointer field. 678 Most hosts use the older BSD interpretation, so if you turn this on 679 Linux might not communicate correctly with them. 680 681 Default: FALSE 682 683tcp_synack_retries - INTEGER 684 Number of times SYNACKs for a passive TCP connection attempt will 685 be retransmitted. Should not be higher than 255. Default value 686 is 5, which corresponds to 31seconds till the last retransmission 687 with the current initial RTO of 1second. With this the final timeout 688 for a passive TCP connection will happen after 63seconds. 689 690tcp_syncookies - INTEGER 691 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES 692 Send out syncookies when the syn backlog queue of a socket 693 overflows. This is to prevent against the common 'SYN flood attack' 694 Default: 1 695 696 Note, that syncookies is fallback facility. 697 It MUST NOT be used to help highly loaded servers to stand 698 against legal connection rate. If you see SYN flood warnings 699 in your logs, but investigation shows that they occur 700 because of overload with legal connections, you should tune 701 another parameters until this warning disappear. 702 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow. 703 704 syncookies seriously violate TCP protocol, do not allow 705 to use TCP extensions, can result in serious degradation 706 of some services (f.e. SMTP relaying), visible not by you, 707 but your clients and relays, contacting you. While you see 708 SYN flood warnings in logs not being really flooded, your server 709 is seriously misconfigured. 710 711 If you want to test which effects syncookies have to your 712 network connections you can set this knob to 2 to enable 713 unconditionally generation of syncookies. 714 715tcp_fastopen - INTEGER 716 Enable TCP Fast Open (RFC7413) to send and accept data in the opening 717 SYN packet. 718 719 The client support is enabled by flag 0x1 (on by default). The client 720 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag, 721 rather than connect() to send data in SYN. 722 723 The server support is enabled by flag 0x2 (off by default). Then 724 either enable for all listeners with another flag (0x400) or 725 enable individual listeners via TCP_FASTOPEN socket option with 726 the option value being the length of the syn-data backlog. 727 728 The values (bitmap) are 729 730 ===== ======== ====================================================== 731 0x1 (client) enables sending data in the opening SYN on the client. 732 0x2 (server) enables the server support, i.e., allowing data in 733 a SYN packet to be accepted and passed to the 734 application before 3-way handshake finishes. 735 0x4 (client) send data in the opening SYN regardless of cookie 736 availability and without a cookie option. 737 0x200 (server) accept data-in-SYN w/o any cookie option present. 738 0x400 (server) enable all listeners to support Fast Open by 739 default without explicit TCP_FASTOPEN socket option. 740 ===== ======== ====================================================== 741 742 Default: 0x1 743 744 Note that additional client or server features are only 745 effective if the basic support (0x1 and 0x2) are enabled respectively. 746 747tcp_fastopen_blackhole_timeout_sec - INTEGER 748 Initial time period in second to disable Fastopen on active TCP sockets 749 when a TFO firewall blackhole issue happens. 750 This time period will grow exponentially when more blackhole issues 751 get detected right after Fastopen is re-enabled and will reset to 752 initial value when the blackhole issue goes away. 753 0 to disable the blackhole detection. 754 755 By default, it is set to 1hr. 756 757tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs 758 The list consists of a primary key and an optional backup key. The 759 primary key is used for both creating and validating cookies, while the 760 optional backup key is only used for validating cookies. The purpose of 761 the backup key is to maximize TFO validation when keys are rotated. 762 763 A randomly chosen primary key may be configured by the kernel if 764 the tcp_fastopen sysctl is set to 0x400 (see above), or if the 765 TCP_FASTOPEN setsockopt() optname is set and a key has not been 766 previously configured via sysctl. If keys are configured via 767 setsockopt() by using the TCP_FASTOPEN_KEY optname, then those 768 per-socket keys will be used instead of any keys that are specified via 769 sysctl. 770 771 A key is specified as 4 8-digit hexadecimal integers which are separated 772 by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be 773 omitted. A primary and a backup key may be specified by separating them 774 by a comma. If only one key is specified, it becomes the primary key and 775 any previously configured backup keys are removed. 776 777tcp_syn_retries - INTEGER 778 Number of times initial SYNs for an active TCP connection attempt 779 will be retransmitted. Should not be higher than 127. Default value 780 is 6, which corresponds to 63seconds till the last retransmission 781 with the current initial RTO of 1second. With this the final timeout 782 for an active TCP connection attempt will happen after 127seconds. 783 784tcp_timestamps - INTEGER 785 Enable timestamps as defined in RFC1323. 786 787 - 0: Disabled. 788 - 1: Enable timestamps as defined in RFC1323 and use random offset for 789 each connection rather than only using the current time. 790 - 2: Like 1, but without random offsets. 791 792 Default: 1 793 794tcp_min_tso_segs - INTEGER 795 Minimal number of segments per TSO frame. 796 797 Since linux-3.12, TCP does an automatic sizing of TSO frames, 798 depending on flow rate, instead of filling 64Kbytes packets. 799 For specific usages, it's possible to force TCP to build big 800 TSO frames. Note that TCP stack might split too big TSO packets 801 if available window is too small. 802 803 Default: 2 804 805tcp_pacing_ss_ratio - INTEGER 806 sk->sk_pacing_rate is set by TCP stack using a ratio applied 807 to current rate. (current_rate = cwnd * mss / srtt) 808 If TCP is in slow start, tcp_pacing_ss_ratio is applied 809 to let TCP probe for bigger speeds, assuming cwnd can be 810 doubled every other RTT. 811 812 Default: 200 813 814tcp_pacing_ca_ratio - INTEGER 815 sk->sk_pacing_rate is set by TCP stack using a ratio applied 816 to current rate. (current_rate = cwnd * mss / srtt) 817 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio 818 is applied to conservatively probe for bigger throughput. 819 820 Default: 120 821 822tcp_tso_win_divisor - INTEGER 823 This allows control over what percentage of the congestion window 824 can be consumed by a single TSO frame. 825 The setting of this parameter is a choice between burstiness and 826 building larger TSO frames. 827 828 Default: 3 829 830tcp_tw_reuse - INTEGER 831 Enable reuse of TIME-WAIT sockets for new connections when it is 832 safe from protocol viewpoint. 833 834 - 0 - disable 835 - 1 - global enable 836 - 2 - enable for loopback traffic only 837 838 It should not be changed without advice/request of technical 839 experts. 840 841 Default: 2 842 843tcp_window_scaling - BOOLEAN 844 Enable window scaling as defined in RFC1323. 845 846tcp_wmem - vector of 3 INTEGERs: min, default, max 847 min: Amount of memory reserved for send buffers for TCP sockets. 848 Each TCP socket has rights to use it due to fact of its birth. 849 850 Default: 4K 851 852 default: initial size of send buffer used by TCP sockets. This 853 value overrides net.core.wmem_default used by other protocols. 854 855 It is usually lower than net.core.wmem_default. 856 857 Default: 16K 858 859 max: Maximal amount of memory allowed for automatically tuned 860 send buffers for TCP sockets. This value does not override 861 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables 862 automatic tuning of that socket's send buffer size, in which case 863 this value is ignored. 864 865 Default: between 64K and 4MB, depending on RAM size. 866 867tcp_notsent_lowat - UNSIGNED INTEGER 868 A TCP socket can control the amount of unsent bytes in its write queue, 869 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll() 870 reports POLLOUT events if the amount of unsent bytes is below a per 871 socket value, and if the write queue is not full. sendmsg() will 872 also not add new buffers if the limit is hit. 873 874 This global variable controls the amount of unsent data for 875 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change 876 to the global variable has immediate effect. 877 878 Default: UINT_MAX (0xFFFFFFFF) 879 880tcp_workaround_signed_windows - BOOLEAN 881 If set, assume no receipt of a window scaling option means the 882 remote TCP is broken and treats the window as a signed quantity. 883 If unset, assume the remote TCP is not broken even if we do 884 not receive a window scaling option from them. 885 886 Default: 0 887 888tcp_thin_linear_timeouts - BOOLEAN 889 Enable dynamic triggering of linear timeouts for thin streams. 890 If set, a check is performed upon retransmission by timeout to 891 determine if the stream is thin (less than 4 packets in flight). 892 As long as the stream is found to be thin, up to 6 linear 893 timeouts may be performed before exponential backoff mode is 894 initiated. This improves retransmission latency for 895 non-aggressive thin streams, often found to be time-dependent. 896 For more information on thin streams, see 897 Documentation/networking/tcp-thin.rst 898 899 Default: 0 900 901tcp_limit_output_bytes - INTEGER 902 Controls TCP Small Queue limit per tcp socket. 903 TCP bulk sender tends to increase packets in flight until it 904 gets losses notifications. With SNDBUF autotuning, this can 905 result in a large amount of packets queued on the local machine 906 (e.g.: qdiscs, CPU backlog, or device) hurting latency of other 907 flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes 908 limits the number of bytes on qdisc or device to reduce artificial 909 RTT/cwnd and reduce bufferbloat. 910 911 Default: 1048576 (16 * 65536) 912 913tcp_challenge_ack_limit - INTEGER 914 Limits number of Challenge ACK sent per second, as recommended 915 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks) 916 Default: 1000 917 918tcp_rx_skb_cache - BOOLEAN 919 Controls a per TCP socket cache of one skb, that might help 920 performance of some workloads. This might be dangerous 921 on systems with a lot of TCP sockets, since it increases 922 memory usage. 923 924 Default: 0 (disabled) 925 926UDP variables 927============= 928 929udp_l3mdev_accept - BOOLEAN 930 Enabling this option allows a "global" bound socket to work 931 across L3 master domains (e.g., VRFs) with packets capable of 932 being received regardless of the L3 domain in which they 933 originated. Only valid when the kernel was compiled with 934 CONFIG_NET_L3_MASTER_DEV. 935 936 Default: 0 (disabled) 937 938udp_mem - vector of 3 INTEGERs: min, pressure, max 939 Number of pages allowed for queueing by all UDP sockets. 940 941 min: Below this number of pages UDP is not bothered about its 942 memory appetite. When amount of memory allocated by UDP exceeds 943 this number, UDP starts to moderate memory usage. 944 945 pressure: This value was introduced to follow format of tcp_mem. 946 947 max: Number of pages allowed for queueing by all UDP sockets. 948 949 Default is calculated at boot time from amount of available memory. 950 951udp_rmem_min - INTEGER 952 Minimal size of receive buffer used by UDP sockets in moderation. 953 Each UDP socket is able to use the size for receiving data, even if 954 total pages of UDP sockets exceed udp_mem pressure. The unit is byte. 955 956 Default: 4K 957 958udp_wmem_min - INTEGER 959 Minimal size of send buffer used by UDP sockets in moderation. 960 Each UDP socket is able to use the size for sending data, even if 961 total pages of UDP sockets exceed udp_mem pressure. The unit is byte. 962 963 Default: 4K 964 965RAW variables 966============= 967 968raw_l3mdev_accept - BOOLEAN 969 Enabling this option allows a "global" bound socket to work 970 across L3 master domains (e.g., VRFs) with packets capable of 971 being received regardless of the L3 domain in which they 972 originated. Only valid when the kernel was compiled with 973 CONFIG_NET_L3_MASTER_DEV. 974 975 Default: 1 (enabled) 976 977CIPSOv4 Variables 978================= 979 980cipso_cache_enable - BOOLEAN 981 If set, enable additions to and lookups from the CIPSO label mapping 982 cache. If unset, additions are ignored and lookups always result in a 983 miss. However, regardless of the setting the cache is still 984 invalidated when required when means you can safely toggle this on and 985 off and the cache will always be "safe". 986 987 Default: 1 988 989cipso_cache_bucket_size - INTEGER 990 The CIPSO label cache consists of a fixed size hash table with each 991 hash bucket containing a number of cache entries. This variable limits 992 the number of entries in each hash bucket; the larger the value the 993 more CIPSO label mappings that can be cached. When the number of 994 entries in a given hash bucket reaches this limit adding new entries 995 causes the oldest entry in the bucket to be removed to make room. 996 997 Default: 10 998 999cipso_rbm_optfmt - BOOLEAN 1000 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of 1001 the CIPSO draft specification (see Documentation/netlabel for details). 1002 This means that when set the CIPSO tag will be padded with empty 1003 categories in order to make the packet data 32-bit aligned. 1004 1005 Default: 0 1006 1007cipso_rbm_structvalid - BOOLEAN 1008 If set, do a very strict check of the CIPSO option when 1009 ip_options_compile() is called. If unset, relax the checks done during 1010 ip_options_compile(). Either way is "safe" as errors are caught else 1011 where in the CIPSO processing code but setting this to 0 (False) should 1012 result in less work (i.e. it should be faster) but could cause problems 1013 with other implementations that require strict checking. 1014 1015 Default: 0 1016 1017IP Variables 1018============ 1019 1020ip_local_port_range - 2 INTEGERS 1021 Defines the local port range that is used by TCP and UDP to 1022 choose the local port. The first number is the first, the 1023 second the last local port number. 1024 If possible, it is better these numbers have different parity 1025 (one even and one odd value). 1026 Must be greater than or equal to ip_unprivileged_port_start. 1027 The default values are 32768 and 60999 respectively. 1028 1029ip_local_reserved_ports - list of comma separated ranges 1030 Specify the ports which are reserved for known third-party 1031 applications. These ports will not be used by automatic port 1032 assignments (e.g. when calling connect() or bind() with port 1033 number 0). Explicit port allocation behavior is unchanged. 1034 1035 The format used for both input and output is a comma separated 1036 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and 1037 10). Writing to the file will clear all previously reserved 1038 ports and update the current list with the one given in the 1039 input. 1040 1041 Note that ip_local_port_range and ip_local_reserved_ports 1042 settings are independent and both are considered by the kernel 1043 when determining which ports are available for automatic port 1044 assignments. 1045 1046 You can reserve ports which are not in the current 1047 ip_local_port_range, e.g.:: 1048 1049 $ cat /proc/sys/net/ipv4/ip_local_port_range 1050 32000 60999 1051 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports 1052 8080,9148 1053 1054 although this is redundant. However such a setting is useful 1055 if later the port range is changed to a value that will 1056 include the reserved ports. 1057 1058 Default: Empty 1059 1060ip_unprivileged_port_start - INTEGER 1061 This is a per-namespace sysctl. It defines the first 1062 unprivileged port in the network namespace. Privileged ports 1063 require root or CAP_NET_BIND_SERVICE in order to bind to them. 1064 To disable all privileged ports, set this to 0. They must not 1065 overlap with the ip_local_port_range. 1066 1067 Default: 1024 1068 1069ip_nonlocal_bind - BOOLEAN 1070 If set, allows processes to bind() to non-local IP addresses, 1071 which can be quite useful - but may break some applications. 1072 1073 Default: 0 1074 1075ip_autobind_reuse - BOOLEAN 1076 By default, bind() does not select the ports automatically even if 1077 the new socket and all sockets bound to the port have SO_REUSEADDR. 1078 ip_autobind_reuse allows bind() to reuse the port and this is useful 1079 when you use bind()+connect(), but may break some applications. 1080 The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this 1081 option should only be set by experts. 1082 Default: 0 1083 1084ip_dynaddr - BOOLEAN 1085 If set non-zero, enables support for dynamic addresses. 1086 If set to a non-zero value larger than 1, a kernel log 1087 message will be printed when dynamic address rewriting 1088 occurs. 1089 1090 Default: 0 1091 1092ip_early_demux - BOOLEAN 1093 Optimize input packet processing down to one demux for 1094 certain kinds of local sockets. Currently we only do this 1095 for established TCP and connected UDP sockets. 1096 1097 It may add an additional cost for pure routing workloads that 1098 reduces overall throughput, in such case you should disable it. 1099 1100 Default: 1 1101 1102ping_group_range - 2 INTEGERS 1103 Restrict ICMP_PROTO datagram sockets to users in the group range. 1104 The default is "1 0", meaning, that nobody (not even root) may 1105 create ping sockets. Setting it to "100 100" would grant permissions 1106 to the single group. "0 4294967295" would enable it for the world, "100 1107 4294967295" would enable it for the users, but not daemons. 1108 1109tcp_early_demux - BOOLEAN 1110 Enable early demux for established TCP sockets. 1111 1112 Default: 1 1113 1114udp_early_demux - BOOLEAN 1115 Enable early demux for connected UDP sockets. Disable this if 1116 your system could experience more unconnected load. 1117 1118 Default: 1 1119 1120icmp_echo_ignore_all - BOOLEAN 1121 If set non-zero, then the kernel will ignore all ICMP ECHO 1122 requests sent to it. 1123 1124 Default: 0 1125 1126icmp_echo_ignore_broadcasts - BOOLEAN 1127 If set non-zero, then the kernel will ignore all ICMP ECHO and 1128 TIMESTAMP requests sent to it via broadcast/multicast. 1129 1130 Default: 1 1131 1132icmp_ratelimit - INTEGER 1133 Limit the maximal rates for sending ICMP packets whose type matches 1134 icmp_ratemask (see below) to specific targets. 1135 0 to disable any limiting, 1136 otherwise the minimal space between responses in milliseconds. 1137 Note that another sysctl, icmp_msgs_per_sec limits the number 1138 of ICMP packets sent on all targets. 1139 1140 Default: 1000 1141 1142icmp_msgs_per_sec - INTEGER 1143 Limit maximal number of ICMP packets sent per second from this host. 1144 Only messages whose type matches icmp_ratemask (see below) are 1145 controlled by this limit. 1146 1147 Default: 1000 1148 1149icmp_msgs_burst - INTEGER 1150 icmp_msgs_per_sec controls number of ICMP packets sent per second, 1151 while icmp_msgs_burst controls the burst size of these packets. 1152 1153 Default: 50 1154 1155icmp_ratemask - INTEGER 1156 Mask made of ICMP types for which rates are being limited. 1157 1158 Significant bits: IHGFEDCBA9876543210 1159 1160 Default mask: 0000001100000011000 (6168) 1161 1162 Bit definitions (see include/linux/icmp.h): 1163 1164 = ========================= 1165 0 Echo Reply 1166 3 Destination Unreachable [1]_ 1167 4 Source Quench [1]_ 1168 5 Redirect 1169 8 Echo Request 1170 B Time Exceeded [1]_ 1171 C Parameter Problem [1]_ 1172 D Timestamp Request 1173 E Timestamp Reply 1174 F Info Request 1175 G Info Reply 1176 H Address Mask Request 1177 I Address Mask Reply 1178 = ========================= 1179 1180 .. [1] These are rate limited by default (see default mask above) 1181 1182icmp_ignore_bogus_error_responses - BOOLEAN 1183 Some routers violate RFC1122 by sending bogus responses to broadcast 1184 frames. Such violations are normally logged via a kernel warning. 1185 If this is set to TRUE, the kernel will not give such warnings, which 1186 will avoid log file clutter. 1187 1188 Default: 1 1189 1190icmp_errors_use_inbound_ifaddr - BOOLEAN 1191 1192 If zero, icmp error messages are sent with the primary address of 1193 the exiting interface. 1194 1195 If non-zero, the message will be sent with the primary address of 1196 the interface that received the packet that caused the icmp error. 1197 This is the behaviour network many administrators will expect from 1198 a router. And it can make debugging complicated network layouts 1199 much easier. 1200 1201 Note that if no primary address exists for the interface selected, 1202 then the primary address of the first non-loopback interface that 1203 has one will be used regardless of this setting. 1204 1205 Default: 0 1206 1207igmp_max_memberships - INTEGER 1208 Change the maximum number of multicast groups we can subscribe to. 1209 Default: 20 1210 1211 Theoretical maximum value is bounded by having to send a membership 1212 report in a single datagram (i.e. the report can't span multiple 1213 datagrams, or risk confusing the switch and leaving groups you don't 1214 intend to). 1215 1216 The number of supported groups 'M' is bounded by the number of group 1217 report entries you can fit into a single datagram of 65535 bytes. 1218 1219 M = 65536-sizeof (ip header)/(sizeof(Group record)) 1220 1221 Group records are variable length, with a minimum of 12 bytes. 1222 So net.ipv4.igmp_max_memberships should not be set higher than: 1223 1224 (65536-24) / 12 = 5459 1225 1226 The value 5459 assumes no IP header options, so in practice 1227 this number may be lower. 1228 1229igmp_max_msf - INTEGER 1230 Maximum number of addresses allowed in the source filter list for a 1231 multicast group. 1232 1233 Default: 10 1234 1235igmp_qrv - INTEGER 1236 Controls the IGMP query robustness variable (see RFC2236 8.1). 1237 1238 Default: 2 (as specified by RFC2236 8.1) 1239 1240 Minimum: 1 (as specified by RFC6636 4.5) 1241 1242force_igmp_version - INTEGER 1243 - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback 1244 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier 1245 Present timer expires. 1246 - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if 1247 receive IGMPv2/v3 query. 1248 - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive 1249 IGMPv1 query message. Will reply report if receive IGMPv3 query. 1250 - 3 - Enforce to use IGMP version 3. The same react with default 0. 1251 1252 .. note:: 1253 1254 this is not the same with force_mld_version because IGMPv3 RFC3376 1255 Security Considerations does not have clear description that we could 1256 ignore other version messages completely as MLDv2 RFC3810. So make 1257 this value as default 0 is recommended. 1258 1259``conf/interface/*`` 1260 changes special settings per interface (where 1261 interface" is the name of your network interface) 1262 1263``conf/all/*`` 1264 is special, changes the settings for all interfaces 1265 1266log_martians - BOOLEAN 1267 Log packets with impossible addresses to kernel log. 1268 log_martians for the interface will be enabled if at least one of 1269 conf/{all,interface}/log_martians is set to TRUE, 1270 it will be disabled otherwise 1271 1272accept_redirects - BOOLEAN 1273 Accept ICMP redirect messages. 1274 accept_redirects for the interface will be enabled if: 1275 1276 - both conf/{all,interface}/accept_redirects are TRUE in the case 1277 forwarding for the interface is enabled 1278 1279 or 1280 1281 - at least one of conf/{all,interface}/accept_redirects is TRUE in the 1282 case forwarding for the interface is disabled 1283 1284 accept_redirects for the interface will be disabled otherwise 1285 1286 default: 1287 1288 - TRUE (host) 1289 - FALSE (router) 1290 1291forwarding - BOOLEAN 1292 Enable IP forwarding on this interface. This controls whether packets 1293 received _on_ this interface can be forwarded. 1294 1295mc_forwarding - BOOLEAN 1296 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE 1297 and a multicast routing daemon is required. 1298 conf/all/mc_forwarding must also be set to TRUE to enable multicast 1299 routing for the interface 1300 1301medium_id - INTEGER 1302 Integer value used to differentiate the devices by the medium they 1303 are attached to. Two devices can have different id values when 1304 the broadcast packets are received only on one of them. 1305 The default value 0 means that the device is the only interface 1306 to its medium, value of -1 means that medium is not known. 1307 1308 Currently, it is used to change the proxy_arp behavior: 1309 the proxy_arp feature is enabled for packets forwarded between 1310 two devices attached to different media. 1311 1312proxy_arp - BOOLEAN 1313 Do proxy arp. 1314 1315 proxy_arp for the interface will be enabled if at least one of 1316 conf/{all,interface}/proxy_arp is set to TRUE, 1317 it will be disabled otherwise 1318 1319proxy_arp_pvlan - BOOLEAN 1320 Private VLAN proxy arp. 1321 1322 Basically allow proxy arp replies back to the same interface 1323 (from which the ARP request/solicitation was received). 1324 1325 This is done to support (ethernet) switch features, like RFC 1326 3069, where the individual ports are NOT allowed to 1327 communicate with each other, but they are allowed to talk to 1328 the upstream router. As described in RFC 3069, it is possible 1329 to allow these hosts to communicate through the upstream 1330 router by proxy_arp'ing. Don't need to be used together with 1331 proxy_arp. 1332 1333 This technology is known by different names: 1334 1335 In RFC 3069 it is called VLAN Aggregation. 1336 Cisco and Allied Telesyn call it Private VLAN. 1337 Hewlett-Packard call it Source-Port filtering or port-isolation. 1338 Ericsson call it MAC-Forced Forwarding (RFC Draft). 1339 1340shared_media - BOOLEAN 1341 Send(router) or accept(host) RFC1620 shared media redirects. 1342 Overrides secure_redirects. 1343 1344 shared_media for the interface will be enabled if at least one of 1345 conf/{all,interface}/shared_media is set to TRUE, 1346 it will be disabled otherwise 1347 1348 default TRUE 1349 1350secure_redirects - BOOLEAN 1351 Accept ICMP redirect messages only to gateways listed in the 1352 interface's current gateway list. Even if disabled, RFC1122 redirect 1353 rules still apply. 1354 1355 Overridden by shared_media. 1356 1357 secure_redirects for the interface will be enabled if at least one of 1358 conf/{all,interface}/secure_redirects is set to TRUE, 1359 it will be disabled otherwise 1360 1361 default TRUE 1362 1363send_redirects - BOOLEAN 1364 Send redirects, if router. 1365 1366 send_redirects for the interface will be enabled if at least one of 1367 conf/{all,interface}/send_redirects is set to TRUE, 1368 it will be disabled otherwise 1369 1370 Default: TRUE 1371 1372bootp_relay - BOOLEAN 1373 Accept packets with source address 0.b.c.d destined 1374 not to this host as local ones. It is supposed, that 1375 BOOTP relay daemon will catch and forward such packets. 1376 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay 1377 for the interface 1378 1379 default FALSE 1380 1381 Not Implemented Yet. 1382 1383accept_source_route - BOOLEAN 1384 Accept packets with SRR option. 1385 conf/all/accept_source_route must also be set to TRUE to accept packets 1386 with SRR option on the interface 1387 1388 default 1389 1390 - TRUE (router) 1391 - FALSE (host) 1392 1393accept_local - BOOLEAN 1394 Accept packets with local source addresses. In combination with 1395 suitable routing, this can be used to direct packets between two 1396 local interfaces over the wire and have them accepted properly. 1397 default FALSE 1398 1399route_localnet - BOOLEAN 1400 Do not consider loopback addresses as martian source or destination 1401 while routing. This enables the use of 127/8 for local routing purposes. 1402 1403 default FALSE 1404 1405rp_filter - INTEGER 1406 - 0 - No source validation. 1407 - 1 - Strict mode as defined in RFC3704 Strict Reverse Path 1408 Each incoming packet is tested against the FIB and if the interface 1409 is not the best reverse path the packet check will fail. 1410 By default failed packets are discarded. 1411 - 2 - Loose mode as defined in RFC3704 Loose Reverse Path 1412 Each incoming packet's source address is also tested against the FIB 1413 and if the source address is not reachable via any interface 1414 the packet check will fail. 1415 1416 Current recommended practice in RFC3704 is to enable strict mode 1417 to prevent IP spoofing from DDos attacks. If using asymmetric routing 1418 or other complicated routing, then loose mode is recommended. 1419 1420 The max value from conf/{all,interface}/rp_filter is used 1421 when doing source validation on the {interface}. 1422 1423 Default value is 0. Note that some distributions enable it 1424 in startup scripts. 1425 1426arp_filter - BOOLEAN 1427 - 1 - Allows you to have multiple network interfaces on the same 1428 subnet, and have the ARPs for each interface be answered 1429 based on whether or not the kernel would route a packet from 1430 the ARP'd IP out that interface (therefore you must use source 1431 based routing for this to work). In other words it allows control 1432 of which cards (usually 1) will respond to an arp request. 1433 1434 - 0 - (default) The kernel can respond to arp requests with addresses 1435 from other interfaces. This may seem wrong but it usually makes 1436 sense, because it increases the chance of successful communication. 1437 IP addresses are owned by the complete host on Linux, not by 1438 particular interfaces. Only for more complex setups like load- 1439 balancing, does this behaviour cause problems. 1440 1441 arp_filter for the interface will be enabled if at least one of 1442 conf/{all,interface}/arp_filter is set to TRUE, 1443 it will be disabled otherwise 1444 1445arp_announce - INTEGER 1446 Define different restriction levels for announcing the local 1447 source IP address from IP packets in ARP requests sent on 1448 interface: 1449 1450 - 0 - (default) Use any local address, configured on any interface 1451 - 1 - Try to avoid local addresses that are not in the target's 1452 subnet for this interface. This mode is useful when target 1453 hosts reachable via this interface require the source IP 1454 address in ARP requests to be part of their logical network 1455 configured on the receiving interface. When we generate the 1456 request we will check all our subnets that include the 1457 target IP and will preserve the source address if it is from 1458 such subnet. If there is no such subnet we select source 1459 address according to the rules for level 2. 1460 - 2 - Always use the best local address for this target. 1461 In this mode we ignore the source address in the IP packet 1462 and try to select local address that we prefer for talks with 1463 the target host. Such local address is selected by looking 1464 for primary IP addresses on all our subnets on the outgoing 1465 interface that include the target IP address. If no suitable 1466 local address is found we select the first local address 1467 we have on the outgoing interface or on all other interfaces, 1468 with the hope we will receive reply for our request and 1469 even sometimes no matter the source IP address we announce. 1470 1471 The max value from conf/{all,interface}/arp_announce is used. 1472 1473 Increasing the restriction level gives more chance for 1474 receiving answer from the resolved target while decreasing 1475 the level announces more valid sender's information. 1476 1477arp_ignore - INTEGER 1478 Define different modes for sending replies in response to 1479 received ARP requests that resolve local target IP addresses: 1480 1481 - 0 - (default): reply for any local target IP address, configured 1482 on any interface 1483 - 1 - reply only if the target IP address is local address 1484 configured on the incoming interface 1485 - 2 - reply only if the target IP address is local address 1486 configured on the incoming interface and both with the 1487 sender's IP address are part from same subnet on this interface 1488 - 3 - do not reply for local addresses configured with scope host, 1489 only resolutions for global and link addresses are replied 1490 - 4-7 - reserved 1491 - 8 - do not reply for all local addresses 1492 1493 The max value from conf/{all,interface}/arp_ignore is used 1494 when ARP request is received on the {interface} 1495 1496arp_notify - BOOLEAN 1497 Define mode for notification of address and device changes. 1498 1499 == ========================================================== 1500 0 (default): do nothing 1501 1 Generate gratuitous arp requests when device is brought up 1502 or hardware address changes. 1503 == ========================================================== 1504 1505arp_accept - BOOLEAN 1506 Define behavior for gratuitous ARP frames who's IP is not 1507 already present in the ARP table: 1508 1509 - 0 - don't create new entries in the ARP table 1510 - 1 - create new entries in the ARP table 1511 1512 Both replies and requests type gratuitous arp will trigger the 1513 ARP table to be updated, if this setting is on. 1514 1515 If the ARP table already contains the IP address of the 1516 gratuitous arp frame, the arp table will be updated regardless 1517 if this setting is on or off. 1518 1519mcast_solicit - INTEGER 1520 The maximum number of multicast probes in INCOMPLETE state, 1521 when the associated hardware address is unknown. Defaults 1522 to 3. 1523 1524ucast_solicit - INTEGER 1525 The maximum number of unicast probes in PROBE state, when 1526 the hardware address is being reconfirmed. Defaults to 3. 1527 1528app_solicit - INTEGER 1529 The maximum number of probes to send to the user space ARP daemon 1530 via netlink before dropping back to multicast probes (see 1531 mcast_resolicit). Defaults to 0. 1532 1533mcast_resolicit - INTEGER 1534 The maximum number of multicast probes after unicast and 1535 app probes in PROBE state. Defaults to 0. 1536 1537disable_policy - BOOLEAN 1538 Disable IPSEC policy (SPD) for this interface 1539 1540disable_xfrm - BOOLEAN 1541 Disable IPSEC encryption on this interface, whatever the policy 1542 1543igmpv2_unsolicited_report_interval - INTEGER 1544 The interval in milliseconds in which the next unsolicited 1545 IGMPv1 or IGMPv2 report retransmit will take place. 1546 1547 Default: 10000 (10 seconds) 1548 1549igmpv3_unsolicited_report_interval - INTEGER 1550 The interval in milliseconds in which the next unsolicited 1551 IGMPv3 report retransmit will take place. 1552 1553 Default: 1000 (1 seconds) 1554 1555promote_secondaries - BOOLEAN 1556 When a primary IP address is removed from this interface 1557 promote a corresponding secondary IP address instead of 1558 removing all the corresponding secondary IP addresses. 1559 1560drop_unicast_in_l2_multicast - BOOLEAN 1561 Drop any unicast IP packets that are received in link-layer 1562 multicast (or broadcast) frames. 1563 1564 This behavior (for multicast) is actually a SHOULD in RFC 1565 1122, but is disabled by default for compatibility reasons. 1566 1567 Default: off (0) 1568 1569drop_gratuitous_arp - BOOLEAN 1570 Drop all gratuitous ARP frames, for example if there's a known 1571 good ARP proxy on the network and such frames need not be used 1572 (or in the case of 802.11, must not be used to prevent attacks.) 1573 1574 Default: off (0) 1575 1576 1577tag - INTEGER 1578 Allows you to write a number, which can be used as required. 1579 1580 Default value is 0. 1581 1582xfrm4_gc_thresh - INTEGER 1583 (Obsolete since linux-4.14) 1584 The threshold at which we will start garbage collecting for IPv4 1585 destination cache entries. At twice this value the system will 1586 refuse new allocations. 1587 1588igmp_link_local_mcast_reports - BOOLEAN 1589 Enable IGMP reports for link local multicast groups in the 1590 224.0.0.X range. 1591 1592 Default TRUE 1593 1594Alexey Kuznetsov. 1595kuznet@ms2.inr.ac.ru 1596 1597Updated by: 1598 1599- Andi Kleen 1600 ak@muc.de 1601- Nicolas Delon 1602 delon.nicolas@wanadoo.fr 1603 1604 1605 1606 1607/proc/sys/net/ipv6/* Variables 1608============================== 1609 1610IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also 1611apply to IPv6 [XXX?]. 1612 1613bindv6only - BOOLEAN 1614 Default value for IPV6_V6ONLY socket option, 1615 which restricts use of the IPv6 socket to IPv6 communication 1616 only. 1617 1618 - TRUE: disable IPv4-mapped address feature 1619 - FALSE: enable IPv4-mapped address feature 1620 1621 Default: FALSE (as specified in RFC3493) 1622 1623flowlabel_consistency - BOOLEAN 1624 Protect the consistency (and unicity) of flow label. 1625 You have to disable it to use IPV6_FL_F_REFLECT flag on the 1626 flow label manager. 1627 1628 - TRUE: enabled 1629 - FALSE: disabled 1630 1631 Default: TRUE 1632 1633auto_flowlabels - INTEGER 1634 Automatically generate flow labels based on a flow hash of the 1635 packet. This allows intermediate devices, such as routers, to 1636 identify packet flows for mechanisms like Equal Cost Multipath 1637 Routing (see RFC 6438). 1638 1639 = =========================================================== 1640 0 automatic flow labels are completely disabled 1641 1 automatic flow labels are enabled by default, they can be 1642 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL 1643 socket option 1644 2 automatic flow labels are allowed, they may be enabled on a 1645 per socket basis using the IPV6_AUTOFLOWLABEL socket option 1646 3 automatic flow labels are enabled and enforced, they cannot 1647 be disabled by the socket option 1648 = =========================================================== 1649 1650 Default: 1 1651 1652flowlabel_state_ranges - BOOLEAN 1653 Split the flow label number space into two ranges. 0-0x7FFFF is 1654 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF 1655 is reserved for stateless flow labels as described in RFC6437. 1656 1657 - TRUE: enabled 1658 - FALSE: disabled 1659 1660 Default: true 1661 1662flowlabel_reflect - INTEGER 1663 Control flow label reflection. Needed for Path MTU 1664 Discovery to work with Equal Cost Multipath Routing in anycast 1665 environments. See RFC 7690 and: 1666 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01 1667 1668 This is a bitmask. 1669 1670 - 1: enabled for established flows 1671 1672 Note that this prevents automatic flowlabel changes, as done 1673 in "tcp: change IPv6 flow-label upon receiving spurious retransmission" 1674 and "tcp: Change txhash on every SYN and RTO retransmit" 1675 1676 - 2: enabled for TCP RESET packets (no active listener) 1677 If set, a RST packet sent in response to a SYN packet on a closed 1678 port will reflect the incoming flow label. 1679 1680 - 4: enabled for ICMPv6 echo reply messages. 1681 1682 Default: 0 1683 1684fib_multipath_hash_policy - INTEGER 1685 Controls which hash policy to use for multipath routes. 1686 1687 Default: 0 (Layer 3) 1688 1689 Possible values: 1690 1691 - 0 - Layer 3 (source and destination addresses plus flow label) 1692 - 1 - Layer 4 (standard 5-tuple) 1693 - 2 - Layer 3 or inner Layer 3 if present 1694 1695anycast_src_echo_reply - BOOLEAN 1696 Controls the use of anycast addresses as source addresses for ICMPv6 1697 echo reply 1698 1699 - TRUE: enabled 1700 - FALSE: disabled 1701 1702 Default: FALSE 1703 1704idgen_delay - INTEGER 1705 Controls the delay in seconds after which time to retry 1706 privacy stable address generation if a DAD conflict is 1707 detected. 1708 1709 Default: 1 (as specified in RFC7217) 1710 1711idgen_retries - INTEGER 1712 Controls the number of retries to generate a stable privacy 1713 address if a DAD conflict is detected. 1714 1715 Default: 3 (as specified in RFC7217) 1716 1717mld_qrv - INTEGER 1718 Controls the MLD query robustness variable (see RFC3810 9.1). 1719 1720 Default: 2 (as specified by RFC3810 9.1) 1721 1722 Minimum: 1 (as specified by RFC6636 4.5) 1723 1724max_dst_opts_number - INTEGER 1725 Maximum number of non-padding TLVs allowed in a Destination 1726 options extension header. If this value is less than zero 1727 then unknown options are disallowed and the number of known 1728 TLVs allowed is the absolute value of this number. 1729 1730 Default: 8 1731 1732max_hbh_opts_number - INTEGER 1733 Maximum number of non-padding TLVs allowed in a Hop-by-Hop 1734 options extension header. If this value is less than zero 1735 then unknown options are disallowed and the number of known 1736 TLVs allowed is the absolute value of this number. 1737 1738 Default: 8 1739 1740max_dst_opts_length - INTEGER 1741 Maximum length allowed for a Destination options extension 1742 header. 1743 1744 Default: INT_MAX (unlimited) 1745 1746max_hbh_length - INTEGER 1747 Maximum length allowed for a Hop-by-Hop options extension 1748 header. 1749 1750 Default: INT_MAX (unlimited) 1751 1752skip_notify_on_dev_down - BOOLEAN 1753 Controls whether an RTM_DELROUTE message is generated for routes 1754 removed when a device is taken down or deleted. IPv4 does not 1755 generate this message; IPv6 does by default. Setting this sysctl 1756 to true skips the message, making IPv4 and IPv6 on par in relying 1757 on userspace caches to track link events and evict routes. 1758 1759 Default: false (generate message) 1760 1761nexthop_compat_mode - BOOLEAN 1762 New nexthop API provides a means for managing nexthops independent of 1763 prefixes. Backwards compatibilty with old route format is enabled by 1764 default which means route dumps and notifications contain the new 1765 nexthop attribute but also the full, expanded nexthop definition. 1766 Further, updates or deletes of a nexthop configuration generate route 1767 notifications for each fib entry using the nexthop. Once a system 1768 understands the new API, this sysctl can be disabled to achieve full 1769 performance benefits of the new API by disabling the nexthop expansion 1770 and extraneous notifications. 1771 Default: true (backward compat mode) 1772 1773IPv6 Fragmentation: 1774 1775ip6frag_high_thresh - INTEGER 1776 Maximum memory used to reassemble IPv6 fragments. When 1777 ip6frag_high_thresh bytes of memory is allocated for this purpose, 1778 the fragment handler will toss packets until ip6frag_low_thresh 1779 is reached. 1780 1781ip6frag_low_thresh - INTEGER 1782 See ip6frag_high_thresh 1783 1784ip6frag_time - INTEGER 1785 Time in seconds to keep an IPv6 fragment in memory. 1786 1787IPv6 Segment Routing: 1788 1789seg6_flowlabel - INTEGER 1790 Controls the behaviour of computing the flowlabel of outer 1791 IPv6 header in case of SR T.encaps 1792 1793 == ======================================================= 1794 -1 set flowlabel to zero. 1795 0 copy flowlabel from Inner packet in case of Inner IPv6 1796 (Set flowlabel to 0 in case IPv4/L2) 1797 1 Compute the flowlabel using seg6_make_flowlabel() 1798 == ======================================================= 1799 1800 Default is 0. 1801 1802``conf/default/*``: 1803 Change the interface-specific default settings. 1804 1805 1806``conf/all/*``: 1807 Change all the interface-specific settings. 1808 1809 [XXX: Other special features than forwarding?] 1810 1811conf/all/forwarding - BOOLEAN 1812 Enable global IPv6 forwarding between all interfaces. 1813 1814 IPv4 and IPv6 work differently here; e.g. netfilter must be used 1815 to control which interfaces may forward packets and which not. 1816 1817 This also sets all interfaces' Host/Router setting 1818 'forwarding' to the specified value. See below for details. 1819 1820 This referred to as global forwarding. 1821 1822proxy_ndp - BOOLEAN 1823 Do proxy ndp. 1824 1825fwmark_reflect - BOOLEAN 1826 Controls the fwmark of kernel-generated IPv6 reply packets that are not 1827 associated with a socket for example, TCP RSTs or ICMPv6 echo replies). 1828 If unset, these packets have a fwmark of zero. If set, they have the 1829 fwmark of the packet they are replying to. 1830 1831 Default: 0 1832 1833``conf/interface/*``: 1834 Change special settings per interface. 1835 1836 The functional behaviour for certain settings is different 1837 depending on whether local forwarding is enabled or not. 1838 1839accept_ra - INTEGER 1840 Accept Router Advertisements; autoconfigure using them. 1841 1842 It also determines whether or not to transmit Router 1843 Solicitations. If and only if the functional setting is to 1844 accept Router Advertisements, Router Solicitations will be 1845 transmitted. 1846 1847 Possible values are: 1848 1849 == =========================================================== 1850 0 Do not accept Router Advertisements. 1851 1 Accept Router Advertisements if forwarding is disabled. 1852 2 Overrule forwarding behaviour. Accept Router Advertisements 1853 even if forwarding is enabled. 1854 == =========================================================== 1855 1856 Functional default: 1857 1858 - enabled if local forwarding is disabled. 1859 - disabled if local forwarding is enabled. 1860 1861accept_ra_defrtr - BOOLEAN 1862 Learn default router in Router Advertisement. 1863 1864 Functional default: 1865 1866 - enabled if accept_ra is enabled. 1867 - disabled if accept_ra is disabled. 1868 1869accept_ra_from_local - BOOLEAN 1870 Accept RA with source-address that is found on local machine 1871 if the RA is otherwise proper and able to be accepted. 1872 1873 Default is to NOT accept these as it may be an un-intended 1874 network loop. 1875 1876 Functional default: 1877 1878 - enabled if accept_ra_from_local is enabled 1879 on a specific interface. 1880 - disabled if accept_ra_from_local is disabled 1881 on a specific interface. 1882 1883accept_ra_min_hop_limit - INTEGER 1884 Minimum hop limit Information in Router Advertisement. 1885 1886 Hop limit Information in Router Advertisement less than this 1887 variable shall be ignored. 1888 1889 Default: 1 1890 1891accept_ra_pinfo - BOOLEAN 1892 Learn Prefix Information in Router Advertisement. 1893 1894 Functional default: 1895 1896 - enabled if accept_ra is enabled. 1897 - disabled if accept_ra is disabled. 1898 1899accept_ra_rt_info_min_plen - INTEGER 1900 Minimum prefix length of Route Information in RA. 1901 1902 Route Information w/ prefix smaller than this variable shall 1903 be ignored. 1904 1905 Functional default: 1906 1907 * 0 if accept_ra_rtr_pref is enabled. 1908 * -1 if accept_ra_rtr_pref is disabled. 1909 1910accept_ra_rt_info_max_plen - INTEGER 1911 Maximum prefix length of Route Information in RA. 1912 1913 Route Information w/ prefix larger than this variable shall 1914 be ignored. 1915 1916 Functional default: 1917 1918 * 0 if accept_ra_rtr_pref is enabled. 1919 * -1 if accept_ra_rtr_pref is disabled. 1920 1921accept_ra_rtr_pref - BOOLEAN 1922 Accept Router Preference in RA. 1923 1924 Functional default: 1925 1926 - enabled if accept_ra is enabled. 1927 - disabled if accept_ra is disabled. 1928 1929accept_ra_mtu - BOOLEAN 1930 Apply the MTU value specified in RA option 5 (RFC4861). If 1931 disabled, the MTU specified in the RA will be ignored. 1932 1933 Functional default: 1934 1935 - enabled if accept_ra is enabled. 1936 - disabled if accept_ra is disabled. 1937 1938accept_redirects - BOOLEAN 1939 Accept Redirects. 1940 1941 Functional default: 1942 1943 - enabled if local forwarding is disabled. 1944 - disabled if local forwarding is enabled. 1945 1946accept_source_route - INTEGER 1947 Accept source routing (routing extension header). 1948 1949 - >= 0: Accept only routing header type 2. 1950 - < 0: Do not accept routing header. 1951 1952 Default: 0 1953 1954autoconf - BOOLEAN 1955 Autoconfigure addresses using Prefix Information in Router 1956 Advertisements. 1957 1958 Functional default: 1959 1960 - enabled if accept_ra_pinfo is enabled. 1961 - disabled if accept_ra_pinfo is disabled. 1962 1963dad_transmits - INTEGER 1964 The amount of Duplicate Address Detection probes to send. 1965 1966 Default: 1 1967 1968forwarding - INTEGER 1969 Configure interface-specific Host/Router behaviour. 1970 1971 .. note:: 1972 1973 It is recommended to have the same setting on all 1974 interfaces; mixed router/host scenarios are rather uncommon. 1975 1976 Possible values are: 1977 1978 - 0 Forwarding disabled 1979 - 1 Forwarding enabled 1980 1981 **FALSE (0)**: 1982 1983 By default, Host behaviour is assumed. This means: 1984 1985 1. IsRouter flag is not set in Neighbour Advertisements. 1986 2. If accept_ra is TRUE (default), transmit Router 1987 Solicitations. 1988 3. If accept_ra is TRUE (default), accept Router 1989 Advertisements (and do autoconfiguration). 1990 4. If accept_redirects is TRUE (default), accept Redirects. 1991 1992 **TRUE (1)**: 1993 1994 If local forwarding is enabled, Router behaviour is assumed. 1995 This means exactly the reverse from the above: 1996 1997 1. IsRouter flag is set in Neighbour Advertisements. 1998 2. Router Solicitations are not sent unless accept_ra is 2. 1999 3. Router Advertisements are ignored unless accept_ra is 2. 2000 4. Redirects are ignored. 2001 2002 Default: 0 (disabled) if global forwarding is disabled (default), 2003 otherwise 1 (enabled). 2004 2005hop_limit - INTEGER 2006 Default Hop Limit to set. 2007 2008 Default: 64 2009 2010mtu - INTEGER 2011 Default Maximum Transfer Unit 2012 2013 Default: 1280 (IPv6 required minimum) 2014 2015ip_nonlocal_bind - BOOLEAN 2016 If set, allows processes to bind() to non-local IPv6 addresses, 2017 which can be quite useful - but may break some applications. 2018 2019 Default: 0 2020 2021router_probe_interval - INTEGER 2022 Minimum interval (in seconds) between Router Probing described 2023 in RFC4191. 2024 2025 Default: 60 2026 2027router_solicitation_delay - INTEGER 2028 Number of seconds to wait after interface is brought up 2029 before sending Router Solicitations. 2030 2031 Default: 1 2032 2033router_solicitation_interval - INTEGER 2034 Number of seconds to wait between Router Solicitations. 2035 2036 Default: 4 2037 2038router_solicitations - INTEGER 2039 Number of Router Solicitations to send until assuming no 2040 routers are present. 2041 2042 Default: 3 2043 2044use_oif_addrs_only - BOOLEAN 2045 When enabled, the candidate source addresses for destinations 2046 routed via this interface are restricted to the set of addresses 2047 configured on this interface (vis. RFC 6724, section 4). 2048 2049 Default: false 2050 2051use_tempaddr - INTEGER 2052 Preference for Privacy Extensions (RFC3041). 2053 2054 * <= 0 : disable Privacy Extensions 2055 * == 1 : enable Privacy Extensions, but prefer public 2056 addresses over temporary addresses. 2057 * > 1 : enable Privacy Extensions and prefer temporary 2058 addresses over public addresses. 2059 2060 Default: 2061 2062 * 0 (for most devices) 2063 * -1 (for point-to-point devices and loopback devices) 2064 2065temp_valid_lft - INTEGER 2066 valid lifetime (in seconds) for temporary addresses. 2067 2068 Default: 172800 (2 days) 2069 2070temp_prefered_lft - INTEGER 2071 Preferred lifetime (in seconds) for temporary addresses. 2072 2073 Default: 86400 (1 day) 2074 2075keep_addr_on_down - INTEGER 2076 Keep all IPv6 addresses on an interface down event. If set static 2077 global addresses with no expiration time are not flushed. 2078 2079 * >0 : enabled 2080 * 0 : system default 2081 * <0 : disabled 2082 2083 Default: 0 (addresses are removed) 2084 2085max_desync_factor - INTEGER 2086 Maximum value for DESYNC_FACTOR, which is a random value 2087 that ensures that clients don't synchronize with each 2088 other and generate new addresses at exactly the same time. 2089 value is in seconds. 2090 2091 Default: 600 2092 2093regen_max_retry - INTEGER 2094 Number of attempts before give up attempting to generate 2095 valid temporary addresses. 2096 2097 Default: 5 2098 2099max_addresses - INTEGER 2100 Maximum number of autoconfigured addresses per interface. Setting 2101 to zero disables the limitation. It is not recommended to set this 2102 value too large (or to zero) because it would be an easy way to 2103 crash the kernel by allowing too many addresses to be created. 2104 2105 Default: 16 2106 2107disable_ipv6 - BOOLEAN 2108 Disable IPv6 operation. If accept_dad is set to 2, this value 2109 will be dynamically set to TRUE if DAD fails for the link-local 2110 address. 2111 2112 Default: FALSE (enable IPv6 operation) 2113 2114 When this value is changed from 1 to 0 (IPv6 is being enabled), 2115 it will dynamically create a link-local address on the given 2116 interface and start Duplicate Address Detection, if necessary. 2117 2118 When this value is changed from 0 to 1 (IPv6 is being disabled), 2119 it will dynamically delete all addresses and routes on the given 2120 interface. From now on it will not possible to add addresses/routes 2121 to the selected interface. 2122 2123accept_dad - INTEGER 2124 Whether to accept DAD (Duplicate Address Detection). 2125 2126 == ============================================================== 2127 0 Disable DAD 2128 1 Enable DAD (default) 2129 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate 2130 link-local address has been found. 2131 == ============================================================== 2132 2133 DAD operation and mode on a given interface will be selected according 2134 to the maximum value of conf/{all,interface}/accept_dad. 2135 2136force_tllao - BOOLEAN 2137 Enable sending the target link-layer address option even when 2138 responding to a unicast neighbor solicitation. 2139 2140 Default: FALSE 2141 2142 Quoting from RFC 2461, section 4.4, Target link-layer address: 2143 2144 "The option MUST be included for multicast solicitations in order to 2145 avoid infinite Neighbor Solicitation "recursion" when the peer node 2146 does not have a cache entry to return a Neighbor Advertisements 2147 message. When responding to unicast solicitations, the option can be 2148 omitted since the sender of the solicitation has the correct link- 2149 layer address; otherwise it would not have be able to send the unicast 2150 solicitation in the first place. However, including the link-layer 2151 address in this case adds little overhead and eliminates a potential 2152 race condition where the sender deletes the cached link-layer address 2153 prior to receiving a response to a previous solicitation." 2154 2155ndisc_notify - BOOLEAN 2156 Define mode for notification of address and device changes. 2157 2158 * 0 - (default): do nothing 2159 * 1 - Generate unsolicited neighbour advertisements when device is brought 2160 up or hardware address changes. 2161 2162ndisc_tclass - INTEGER 2163 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor 2164 Discovery (Router Solicitation, Router Advertisement, Neighbor 2165 Solicitation, Neighbor Advertisement, Redirect) messages. 2166 These 8 bits can be interpreted as 6 high order bits holding the DSCP 2167 value and 2 low order bits representing ECN (which you probably want 2168 to leave cleared). 2169 2170 * 0 - (default) 2171 2172mldv1_unsolicited_report_interval - INTEGER 2173 The interval in milliseconds in which the next unsolicited 2174 MLDv1 report retransmit will take place. 2175 2176 Default: 10000 (10 seconds) 2177 2178mldv2_unsolicited_report_interval - INTEGER 2179 The interval in milliseconds in which the next unsolicited 2180 MLDv2 report retransmit will take place. 2181 2182 Default: 1000 (1 second) 2183 2184force_mld_version - INTEGER 2185 * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed 2186 * 1 - Enforce to use MLD version 1 2187 * 2 - Enforce to use MLD version 2 2188 2189suppress_frag_ndisc - INTEGER 2190 Control RFC 6980 (Security Implications of IPv6 Fragmentation 2191 with IPv6 Neighbor Discovery) behavior: 2192 2193 * 1 - (default) discard fragmented neighbor discovery packets 2194 * 0 - allow fragmented neighbor discovery packets 2195 2196optimistic_dad - BOOLEAN 2197 Whether to perform Optimistic Duplicate Address Detection (RFC 4429). 2198 2199 * 0: disabled (default) 2200 * 1: enabled 2201 2202 Optimistic Duplicate Address Detection for the interface will be enabled 2203 if at least one of conf/{all,interface}/optimistic_dad is set to 1, 2204 it will be disabled otherwise. 2205 2206use_optimistic - BOOLEAN 2207 If enabled, do not classify optimistic addresses as deprecated during 2208 source address selection. Preferred addresses will still be chosen 2209 before optimistic addresses, subject to other ranking in the source 2210 address selection algorithm. 2211 2212 * 0: disabled (default) 2213 * 1: enabled 2214 2215 This will be enabled if at least one of 2216 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise. 2217 2218stable_secret - IPv6 address 2219 This IPv6 address will be used as a secret to generate IPv6 2220 addresses for link-local addresses and autoconfigured 2221 ones. All addresses generated after setting this secret will 2222 be stable privacy ones by default. This can be changed via the 2223 addrgenmode ip-link. conf/default/stable_secret is used as the 2224 secret for the namespace, the interface specific ones can 2225 overwrite that. Writes to conf/all/stable_secret are refused. 2226 2227 It is recommended to generate this secret during installation 2228 of a system and keep it stable after that. 2229 2230 By default the stable secret is unset. 2231 2232addr_gen_mode - INTEGER 2233 Defines how link-local and autoconf addresses are generated. 2234 2235 = ================================================================= 2236 0 generate address based on EUI64 (default) 2237 1 do no generate a link-local address, use EUI64 for addresses 2238 generated from autoconf 2239 2 generate stable privacy addresses, using the secret from 2240 stable_secret (RFC7217) 2241 3 generate stable privacy addresses, using a random secret if unset 2242 = ================================================================= 2243 2244drop_unicast_in_l2_multicast - BOOLEAN 2245 Drop any unicast IPv6 packets that are received in link-layer 2246 multicast (or broadcast) frames. 2247 2248 By default this is turned off. 2249 2250drop_unsolicited_na - BOOLEAN 2251 Drop all unsolicited neighbor advertisements, for example if there's 2252 a known good NA proxy on the network and such frames need not be used 2253 (or in the case of 802.11, must not be used to prevent attacks.) 2254 2255 By default this is turned off. 2256 2257enhanced_dad - BOOLEAN 2258 Include a nonce option in the IPv6 neighbor solicitation messages used for 2259 duplicate address detection per RFC7527. A received DAD NS will only signal 2260 a duplicate address if the nonce is different. This avoids any false 2261 detection of duplicates due to loopback of the NS messages that we send. 2262 The nonce option will be sent on an interface unless both of 2263 conf/{all,interface}/enhanced_dad are set to FALSE. 2264 2265 Default: TRUE 2266 2267``icmp/*``: 2268=========== 2269 2270ratelimit - INTEGER 2271 Limit the maximal rates for sending ICMPv6 messages. 2272 2273 0 to disable any limiting, 2274 otherwise the minimal space between responses in milliseconds. 2275 2276 Default: 1000 2277 2278ratemask - list of comma separated ranges 2279 For ICMPv6 message types matching the ranges in the ratemask, limit 2280 the sending of the message according to ratelimit parameter. 2281 2282 The format used for both input and output is a comma separated 2283 list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and 2284 129). Writing to the file will clear all previous ranges of ICMPv6 2285 message types and update the current list with the input. 2286 2287 Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml 2288 for numerical values of ICMPv6 message types, e.g. echo request is 128 2289 and echo reply is 129. 2290 2291 Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big) 2292 2293echo_ignore_all - BOOLEAN 2294 If set non-zero, then the kernel will ignore all ICMP ECHO 2295 requests sent to it over the IPv6 protocol. 2296 2297 Default: 0 2298 2299echo_ignore_multicast - BOOLEAN 2300 If set non-zero, then the kernel will ignore all ICMP ECHO 2301 requests sent to it over the IPv6 protocol via multicast. 2302 2303 Default: 0 2304 2305echo_ignore_anycast - BOOLEAN 2306 If set non-zero, then the kernel will ignore all ICMP ECHO 2307 requests sent to it over the IPv6 protocol destined to anycast address. 2308 2309 Default: 0 2310 2311xfrm6_gc_thresh - INTEGER 2312 (Obsolete since linux-4.14) 2313 The threshold at which we will start garbage collecting for IPv6 2314 destination cache entries. At twice this value the system will 2315 refuse new allocations. 2316 2317 2318IPv6 Update by: 2319Pekka Savola <pekkas@netcore.fi> 2320YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org> 2321 2322 2323/proc/sys/net/bridge/* Variables: 2324================================= 2325 2326bridge-nf-call-arptables - BOOLEAN 2327 - 1 : pass bridged ARP traffic to arptables' FORWARD chain. 2328 - 0 : disable this. 2329 2330 Default: 1 2331 2332bridge-nf-call-iptables - BOOLEAN 2333 - 1 : pass bridged IPv4 traffic to iptables' chains. 2334 - 0 : disable this. 2335 2336 Default: 1 2337 2338bridge-nf-call-ip6tables - BOOLEAN 2339 - 1 : pass bridged IPv6 traffic to ip6tables' chains. 2340 - 0 : disable this. 2341 2342 Default: 1 2343 2344bridge-nf-filter-vlan-tagged - BOOLEAN 2345 - 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables. 2346 - 0 : disable this. 2347 2348 Default: 0 2349 2350bridge-nf-filter-pppoe-tagged - BOOLEAN 2351 - 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables. 2352 - 0 : disable this. 2353 2354 Default: 0 2355 2356bridge-nf-pass-vlan-input-dev - BOOLEAN 2357 - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan 2358 interface on the bridge and set the netfilter input device to the 2359 vlan. This allows use of e.g. "iptables -i br0.1" and makes the 2360 REDIRECT target work with vlan-on-top-of-bridge interfaces. When no 2361 matching vlan interface is found, or this switch is off, the input 2362 device is set to the bridge interface. 2363 2364 - 0: disable bridge netfilter vlan interface lookup. 2365 2366 Default: 0 2367 2368``proc/sys/net/sctp/*`` Variables: 2369================================== 2370 2371addip_enable - BOOLEAN 2372 Enable or disable extension of Dynamic Address Reconfiguration 2373 (ADD-IP) functionality specified in RFC5061. This extension provides 2374 the ability to dynamically add and remove new addresses for the SCTP 2375 associations. 2376 2377 1: Enable extension. 2378 2379 0: Disable extension. 2380 2381 Default: 0 2382 2383pf_enable - INTEGER 2384 Enable or disable pf (pf is short for potentially failed) state. A value 2385 of pf_retrans > path_max_retrans also disables pf state. That is, one of 2386 both pf_enable and pf_retrans > path_max_retrans can disable pf state. 2387 Since pf_retrans and path_max_retrans can be changed by userspace 2388 application, sometimes user expects to disable pf state by the value of 2389 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans 2390 or path_max_retrans is changed by the user application, this pf state is 2391 enabled. As such, it is necessary to add this to dynamically enable 2392 and disable pf state. See: 2393 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for 2394 details. 2395 2396 1: Enable pf. 2397 2398 0: Disable pf. 2399 2400 Default: 1 2401 2402pf_expose - INTEGER 2403 Unset or enable/disable pf (pf is short for potentially failed) state 2404 exposure. Applications can control the exposure of the PF path state 2405 in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO 2406 sockopt. When it's unset, no SCTP_PEER_ADDR_CHANGE event with 2407 SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info 2408 can be got via SCTP_GET_PEER_ADDR_INFO sockopt; When it's enabled, 2409 a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming 2410 SCTP_PF state and a SCTP_PF-state transport info can be got via 2411 SCTP_GET_PEER_ADDR_INFO sockopt; When it's diabled, no 2412 SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when 2413 trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO 2414 sockopt. 2415 2416 0: Unset pf state exposure, Compatible with old applications. 2417 2418 1: Disable pf state exposure. 2419 2420 2: Enable pf state exposure. 2421 2422 Default: 0 2423 2424addip_noauth_enable - BOOLEAN 2425 Dynamic Address Reconfiguration (ADD-IP) requires the use of 2426 authentication to protect the operations of adding or removing new 2427 addresses. This requirement is mandated so that unauthorized hosts 2428 would not be able to hijack associations. However, older 2429 implementations may not have implemented this requirement while 2430 allowing the ADD-IP extension. For reasons of interoperability, 2431 we provide this variable to control the enforcement of the 2432 authentication requirement. 2433 2434 == =============================================================== 2435 1 Allow ADD-IP extension to be used without authentication. This 2436 should only be set in a closed environment for interoperability 2437 with older implementations. 2438 2439 0 Enforce the authentication requirement 2440 == =============================================================== 2441 2442 Default: 0 2443 2444auth_enable - BOOLEAN 2445 Enable or disable Authenticated Chunks extension. This extension 2446 provides the ability to send and receive authenticated chunks and is 2447 required for secure operation of Dynamic Address Reconfiguration 2448 (ADD-IP) extension. 2449 2450 - 1: Enable this extension. 2451 - 0: Disable this extension. 2452 2453 Default: 0 2454 2455prsctp_enable - BOOLEAN 2456 Enable or disable the Partial Reliability extension (RFC3758) which 2457 is used to notify peers that a given DATA should no longer be expected. 2458 2459 - 1: Enable extension 2460 - 0: Disable 2461 2462 Default: 1 2463 2464max_burst - INTEGER 2465 The limit of the number of new packets that can be initially sent. It 2466 controls how bursty the generated traffic can be. 2467 2468 Default: 4 2469 2470association_max_retrans - INTEGER 2471 Set the maximum number for retransmissions that an association can 2472 attempt deciding that the remote end is unreachable. If this value 2473 is exceeded, the association is terminated. 2474 2475 Default: 10 2476 2477max_init_retransmits - INTEGER 2478 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks 2479 that an association will attempt before declaring the destination 2480 unreachable and terminating. 2481 2482 Default: 8 2483 2484path_max_retrans - INTEGER 2485 The maximum number of retransmissions that will be attempted on a given 2486 path. Once this threshold is exceeded, the path is considered 2487 unreachable, and new traffic will use a different path when the 2488 association is multihomed. 2489 2490 Default: 5 2491 2492pf_retrans - INTEGER 2493 The number of retransmissions that will be attempted on a given path 2494 before traffic is redirected to an alternate transport (should one 2495 exist). Note this is distinct from path_max_retrans, as a path that 2496 passes the pf_retrans threshold can still be used. Its only 2497 deprioritized when a transmission path is selected by the stack. This 2498 setting is primarily used to enable fast failover mechanisms without 2499 having to reduce path_max_retrans to a very low value. See: 2500 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt 2501 for details. Note also that a value of pf_retrans > path_max_retrans 2502 disables this feature. Since both pf_retrans and path_max_retrans can 2503 be changed by userspace application, a variable pf_enable is used to 2504 disable pf state. 2505 2506 Default: 0 2507 2508ps_retrans - INTEGER 2509 Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming 2510 from section-5 "Primary Path Switchover" in rfc7829. The primary path 2511 will be changed to another active path when the path error counter on 2512 the old primary path exceeds PSMR, so that "the SCTP sender is allowed 2513 to continue data transmission on a new working path even when the old 2514 primary destination address becomes active again". Note this feature 2515 is disabled by initializing 'ps_retrans' per netns as 0xffff by default, 2516 and its value can't be less than 'pf_retrans' when changing by sysctl. 2517 2518 Default: 0xffff 2519 2520rto_initial - INTEGER 2521 The initial round trip timeout value in milliseconds that will be used 2522 in calculating round trip times. This is the initial time interval 2523 for retransmissions. 2524 2525 Default: 3000 2526 2527rto_max - INTEGER 2528 The maximum value (in milliseconds) of the round trip timeout. This 2529 is the largest time interval that can elapse between retransmissions. 2530 2531 Default: 60000 2532 2533rto_min - INTEGER 2534 The minimum value (in milliseconds) of the round trip timeout. This 2535 is the smallest time interval the can elapse between retransmissions. 2536 2537 Default: 1000 2538 2539hb_interval - INTEGER 2540 The interval (in milliseconds) between HEARTBEAT chunks. These chunks 2541 are sent at the specified interval on idle paths to probe the state of 2542 a given path between 2 associations. 2543 2544 Default: 30000 2545 2546sack_timeout - INTEGER 2547 The amount of time (in milliseconds) that the implementation will wait 2548 to send a SACK. 2549 2550 Default: 200 2551 2552valid_cookie_life - INTEGER 2553 The default lifetime of the SCTP cookie (in milliseconds). The cookie 2554 is used during association establishment. 2555 2556 Default: 60000 2557 2558cookie_preserve_enable - BOOLEAN 2559 Enable or disable the ability to extend the lifetime of the SCTP cookie 2560 that is used during the establishment phase of SCTP association 2561 2562 - 1: Enable cookie lifetime extension. 2563 - 0: Disable 2564 2565 Default: 1 2566 2567cookie_hmac_alg - STRING 2568 Select the hmac algorithm used when generating the cookie value sent by 2569 a listening sctp socket to a connecting client in the INIT-ACK chunk. 2570 Valid values are: 2571 2572 * md5 2573 * sha1 2574 * none 2575 2576 Ability to assign md5 or sha1 as the selected alg is predicated on the 2577 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and 2578 CONFIG_CRYPTO_SHA1). 2579 2580 Default: Dependent on configuration. MD5 if available, else SHA1 if 2581 available, else none. 2582 2583rcvbuf_policy - INTEGER 2584 Determines if the receive buffer is attributed to the socket or to 2585 association. SCTP supports the capability to create multiple 2586 associations on a single socket. When using this capability, it is 2587 possible that a single stalled association that's buffering a lot 2588 of data may block other associations from delivering their data by 2589 consuming all of the receive buffer space. To work around this, 2590 the rcvbuf_policy could be set to attribute the receiver buffer space 2591 to each association instead of the socket. This prevents the described 2592 blocking. 2593 2594 - 1: rcvbuf space is per association 2595 - 0: rcvbuf space is per socket 2596 2597 Default: 0 2598 2599sndbuf_policy - INTEGER 2600 Similar to rcvbuf_policy above, this applies to send buffer space. 2601 2602 - 1: Send buffer is tracked per association 2603 - 0: Send buffer is tracked per socket. 2604 2605 Default: 0 2606 2607sctp_mem - vector of 3 INTEGERs: min, pressure, max 2608 Number of pages allowed for queueing by all SCTP sockets. 2609 2610 min: Below this number of pages SCTP is not bothered about its 2611 memory appetite. When amount of memory allocated by SCTP exceeds 2612 this number, SCTP starts to moderate memory usage. 2613 2614 pressure: This value was introduced to follow format of tcp_mem. 2615 2616 max: Number of pages allowed for queueing by all SCTP sockets. 2617 2618 Default is calculated at boot time from amount of available memory. 2619 2620sctp_rmem - vector of 3 INTEGERs: min, default, max 2621 Only the first value ("min") is used, "default" and "max" are 2622 ignored. 2623 2624 min: Minimal size of receive buffer used by SCTP socket. 2625 It is guaranteed to each SCTP socket (but not association) even 2626 under moderate memory pressure. 2627 2628 Default: 4K 2629 2630sctp_wmem - vector of 3 INTEGERs: min, default, max 2631 Currently this tunable has no effect. 2632 2633addr_scope_policy - INTEGER 2634 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00 2635 2636 - 0 - Disable IPv4 address scoping 2637 - 1 - Enable IPv4 address scoping 2638 - 2 - Follow draft but allow IPv4 private addresses 2639 - 3 - Follow draft but allow IPv4 link local addresses 2640 2641 Default: 1 2642 2643 2644``/proc/sys/net/core/*`` 2645======================== 2646 2647 Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries. 2648 2649 2650``/proc/sys/net/unix/*`` 2651======================== 2652 2653max_dgram_qlen - INTEGER 2654 The maximum length of dgram socket receive queue 2655 2656 Default: 10 2657 2658