1 .. SPDX-License-Identifier: GPL-2.0
7 /proc/sys/net/ipv4/* Variables
8 ==============================
11 - 0 - disabled (default)
14 Forward Packets between interfaces.
16 This variable is special, its change resets all configuration
17 parameters to their default state (RFC1122 for hosts, RFC1812
20 ip_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)
25 ip_no_pmtu_disc - INTEGER
26 Disable Path MTU Discovery. If enabled in mode 1 and a
27 fragmentation-required ICMP is received, the PMTU to this
28 destination will be set to the smallest of the old MTU to
29 this destination and min_pmtu (see below). You will need
30 to raise min_pmtu to the smallest interface MTU on your system
31 manually if you want to avoid locally generated fragments.
33 In mode 2 incoming Path MTU Discovery messages will be
34 discarded. Outgoing frames are handled the same as in mode 1,
35 implicitly setting IP_PMTUDISC_DONT on every created socket.
37 Mode 3 is a hardened pmtu discover mode. The kernel will only
38 accept fragmentation-needed errors if the underlying protocol
39 can verify them besides a plain socket lookup. Current
40 protocols for which pmtu events will be honored are TCP, SCTP
41 and DCCP as they verify e.g. the sequence number or the
42 association. This mode should not be enabled globally but is
43 only intended to secure e.g. name servers in namespaces where
44 TCP path mtu must still work but path MTU information of other
45 protocols should be discarded. If enabled globally this mode
46 could break other protocols.
53 default 552 - minimum Path MTU. Unless this is changed mannually,
54 each cached pmtu will never be lower than this setting.
56 ip_forward_use_pmtu - BOOLEAN
57 By default we don't trust protocol path MTUs while forwarding
58 because they could be easily forged and can lead to unwanted
59 fragmentation by the router.
60 You only need to enable this if you have user-space software
61 which tries to discover path mtus by itself and depends on the
62 kernel honoring this information. This is normally not the
72 fwmark_reflect - BOOLEAN
73 Controls the fwmark of kernel-generated IPv4 reply packets that are not
74 associated with a socket for example, TCP RSTs or ICMP echo replies).
75 If unset, these packets have a fwmark of zero. If set, they have the
76 fwmark of the packet they are replying to.
80 fib_multipath_use_neigh - BOOLEAN
81 Use status of existing neighbor entry when determining nexthop for
82 multipath routes. If disabled, neighbor information is not used and
83 packets could be directed to a failed nexthop. Only valid for kernels
84 built with CONFIG_IP_ROUTE_MULTIPATH enabled.
93 fib_multipath_hash_policy - INTEGER
94 Controls which hash policy to use for multipath routes. Only valid
95 for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
103 - 2 - Layer 3 or inner Layer 3 if present
104 - 3 - Custom multipath hash. Fields used for multipath hash calculation
105 are determined by fib_multipath_hash_fields sysctl
107 fib_multipath_hash_fields - UNSIGNED INTEGER
108 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
109 fields used for multipath hash calculation are determined by this
112 This value is a bitmask which enables various fields for multipath hash
117 ====== ============================
118 0x0001 Source IP address
119 0x0002 Destination IP address
121 0x0008 Unused (Flow Label)
123 0x0020 Destination port
124 0x0040 Inner source IP address
125 0x0080 Inner destination IP address
126 0x0100 Inner IP protocol
127 0x0200 Inner Flow Label
128 0x0400 Inner source port
129 0x0800 Inner destination port
130 ====== ============================
132 Default: 0x0007 (source IP, destination IP and IP protocol)
134 fib_sync_mem - UNSIGNED INTEGER
135 Amount of dirty memory from fib entries that can be backlogged before
136 synchronize_rcu is forced.
138 Default: 512kB Minimum: 64kB Maximum: 64MB
140 ip_forward_update_priority - INTEGER
141 Whether to update SKB priority from "TOS" field in IPv4 header after it
142 is forwarded. The new SKB priority is mapped from TOS field value
143 according to an rt_tos2priority table (see e.g. man tc-prio).
145 Default: 1 (Update priority.)
149 - 0 - Do not update priority.
150 - 1 - Update priority.
152 route/max_size - INTEGER
153 Maximum number of routes allowed in the kernel. Increase
154 this when using large numbers of interfaces and/or routes.
156 From linux kernel 3.6 onwards, this is deprecated for ipv4
157 as route cache is no longer used.
159 neigh/default/gc_thresh1 - INTEGER
160 Minimum number of entries to keep. Garbage collector will not
161 purge entries if there are fewer than this number.
165 neigh/default/gc_thresh2 - INTEGER
166 Threshold when garbage collector becomes more aggressive about
167 purging entries. Entries older than 5 seconds will be cleared
168 when over this number.
172 neigh/default/gc_thresh3 - INTEGER
173 Maximum number of non-PERMANENT neighbor entries allowed. Increase
174 this when using large numbers of interfaces and when communicating
175 with large numbers of directly-connected peers.
179 neigh/default/unres_qlen_bytes - INTEGER
180 The maximum number of bytes which may be used by packets
181 queued for each unresolved address by other network layers.
184 Setting negative value is meaningless and will return error.
186 Default: SK_WMEM_MAX, (same as net.core.wmem_default).
188 Exact value depends on architecture and kernel options,
189 but should be enough to allow queuing 256 packets
192 neigh/default/unres_qlen - INTEGER
193 The maximum number of packets which may be queued for each
194 unresolved address by other network layers.
196 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
198 Prior to linux 3.3, the default value is 3 which may cause
199 unexpected packet loss. The current default value is calculated
200 according to default value of unres_qlen_bytes and true size of
205 neigh/default/interval_probe_time_ms - INTEGER
206 The probe interval for neighbor entries with NTF_MANAGED flag,
211 mtu_expires - INTEGER
212 Time, in seconds, that cached PMTU information is kept.
214 min_adv_mss - INTEGER
215 The advertised MSS depends on the first hop route MTU, but will
216 never be lower than this setting.
218 fib_notify_on_flag_change - INTEGER
219 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
220 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
222 After installing a route to the kernel, user space receives an
223 acknowledgment, which means the route was installed in the kernel,
224 but not necessarily in hardware.
225 It is also possible for a route already installed in hardware to change
226 its action and therefore its flags. For example, a host route that is
227 trapping packets can be "promoted" to perform decapsulation following
228 the installation of an IPinIP/VXLAN tunnel.
229 The notifications will indicate to user-space the state of the route.
231 Default: 0 (Do not emit notifications.)
235 - 0 - Do not emit notifications.
236 - 1 - Emit notifications.
237 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
241 ipfrag_high_thresh - LONG INTEGER
242 Maximum memory used to reassemble IP fragments.
244 ipfrag_low_thresh - LONG INTEGER
245 (Obsolete since linux-4.17)
246 Maximum memory used to reassemble IP fragments before the kernel
247 begins to remove incomplete fragment queues to free up resources.
248 The kernel still accepts new fragments for defragmentation.
250 ipfrag_time - INTEGER
251 Time in seconds to keep an IP fragment in memory.
253 ipfrag_max_dist - INTEGER
254 ipfrag_max_dist is a non-negative integer value which defines the
255 maximum "disorder" which is allowed among fragments which share a
256 common IP source address. Note that reordering of packets is
257 not unusual, but if a large number of fragments arrive from a source
258 IP address while a particular fragment queue remains incomplete, it
259 probably indicates that one or more fragments belonging to that queue
260 have been lost. When ipfrag_max_dist is positive, an additional check
261 is done on fragments before they are added to a reassembly queue - if
262 ipfrag_max_dist (or more) fragments have arrived from a particular IP
263 address between additions to any IP fragment queue using that source
264 address, it's presumed that one or more fragments in the queue are
265 lost. The existing fragment queue will be dropped, and a new one
266 started. An ipfrag_max_dist value of zero disables this check.
268 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
269 result in unnecessarily dropping fragment queues when normal
270 reordering of packets occurs, which could lead to poor application
271 performance. Using a very large value, e.g. 50000, increases the
272 likelihood of incorrectly reassembling IP fragments that originate
273 from different IP datagrams, which could result in data corruption.
276 bc_forwarding - INTEGER
277 bc_forwarding enables the feature described in rfc1812#section-5.3.5.2
278 and rfc2644. It allows the router to forward directed broadcast.
279 To enable this feature, the 'all' entry and the input interface entry
286 inet_peer_threshold - INTEGER
287 The approximate size of the storage. Starting from this threshold
288 entries will be thrown aggressively. This threshold also determines
289 entries' time-to-live and time intervals between garbage collection
290 passes. More entries, less time-to-live, less GC interval.
292 inet_peer_minttl - INTEGER
293 Minimum time-to-live of entries. Should be enough to cover fragment
294 time-to-live on the reassembling side. This minimum time-to-live is
295 guaranteed if the pool size is less than inet_peer_threshold.
298 inet_peer_maxttl - INTEGER
299 Maximum time-to-live of entries. Unused entries will expire after
300 this period of time if there is no memory pressure on the pool (i.e.
301 when the number of entries in the pool is very small).
308 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
309 Defaults to 4096. (Was 128 before linux-5.4)
310 See also tcp_max_syn_backlog for additional tuning for TCP sockets.
312 tcp_abort_on_overflow - BOOLEAN
313 If listening service is too slow to accept new connections,
314 reset them. Default state is FALSE. It means that if overflow
315 occurred due to a burst, connection will recover. Enable this
316 option _only_ if you are really sure that listening daemon
317 cannot be tuned to accept connections faster. Enabling this
318 option can harm clients of your server.
320 tcp_adv_win_scale - INTEGER
321 Count buffering overhead as bytes/2^tcp_adv_win_scale
322 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
325 Possible values are [-31, 31], inclusive.
329 tcp_allowed_congestion_control - STRING
330 Show/set the congestion control choices available to non-privileged
331 processes. The list is a subset of those listed in
332 tcp_available_congestion_control.
334 Default is "reno" and the default setting (tcp_congestion_control).
336 tcp_app_win - INTEGER
337 Reserve max(window/2^tcp_app_win, mss) of window for application
338 buffer. Value 0 is special, it means that nothing is reserved.
340 Possible values are [0, 31], inclusive.
344 tcp_autocorking - BOOLEAN
345 Enable TCP auto corking :
346 When applications do consecutive small write()/sendmsg() system calls,
347 we try to coalesce these small writes as much as possible, to lower
348 total amount of sent packets. This is done if at least one prior
349 packet for the flow is waiting in Qdisc queues or device transmit
350 queue. Applications can still use TCP_CORK for optimal behavior
351 when they know how/when to uncork their sockets.
355 tcp_available_congestion_control - STRING
356 Shows the available congestion control choices that are registered.
357 More congestion control algorithms may be available as modules,
360 tcp_base_mss - INTEGER
361 The initial value of search_low to be used by the packetization layer
362 Path MTU discovery (MTU probing). If MTU probing is enabled,
363 this is the initial MSS used by the connection.
365 tcp_mtu_probe_floor - INTEGER
366 If MTU probing is enabled this caps the minimum MSS used for search_low
371 tcp_min_snd_mss - INTEGER
372 TCP SYN and SYNACK messages usually advertise an ADVMSS option,
373 as described in RFC 1122 and RFC 6691.
375 If this ADVMSS option is smaller than tcp_min_snd_mss,
376 it is silently capped to tcp_min_snd_mss.
378 Default : 48 (at least 8 bytes of payload per segment)
380 tcp_congestion_control - STRING
381 Set the congestion control algorithm to be used for new
382 connections. The algorithm "reno" is always available, but
383 additional choices may be available based on kernel configuration.
384 Default is set as part of kernel configuration.
385 For passive connections, the listener congestion control choice
388 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
391 Allows TCP to send "duplicate" SACKs.
393 tcp_early_retrans - INTEGER
394 Tail loss probe (TLP) converts RTOs occurring due to tail
395 losses into fast recovery (draft-ietf-tcpm-rack). Note that
396 TLP requires RACK to function properly (see tcp_recovery below)
406 Control use of Explicit Congestion Notification (ECN) by TCP.
407 ECN is used only when both ends of the TCP connection indicate
408 support for it. This feature is useful in avoiding losses due
409 to congestion by allowing supporting routers to signal
410 congestion before having to drop packets.
414 = =====================================================
415 0 Disable ECN. Neither initiate nor accept ECN.
416 1 Enable ECN when requested by incoming connections and
417 also request ECN on outgoing connection attempts.
418 2 Enable ECN when requested by incoming connections
419 but do not request ECN on outgoing connections.
420 = =====================================================
424 tcp_ecn_fallback - BOOLEAN
425 If the kernel detects that ECN connection misbehaves, enable fall
426 back to non-ECN. Currently, this knob implements the fallback
427 from RFC3168, section 6.1.1.1., but we reserve that in future,
428 additional detection mechanisms could be implemented under this
429 knob. The value is not used, if tcp_ecn or per route (or congestion
430 control) ECN settings are disabled.
432 Default: 1 (fallback enabled)
435 This is a legacy option, it has no effect anymore.
437 tcp_fin_timeout - INTEGER
438 The length of time an orphaned (no longer referenced by any
439 application) connection will remain in the FIN_WAIT_2 state
440 before it is aborted at the local end. While a perfectly
441 valid "receive only" state for an un-orphaned connection, an
442 orphaned connection in FIN_WAIT_2 state could otherwise wait
443 forever for the remote to close its end of the connection.
450 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
451 F-RTO is an enhanced recovery algorithm for TCP retransmission
452 timeouts. It is particularly beneficial in networks where the
453 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
454 modification. It does not require any support from the peer.
456 By default it's enabled with a non-zero value. 0 disables F-RTO.
458 tcp_fwmark_accept - BOOLEAN
459 If set, incoming connections to listening sockets that do not have a
460 socket mark will set the mark of the accepting socket to the fwmark of
461 the incoming SYN packet. This will cause all packets on that connection
462 (starting from the first SYNACK) to be sent with that fwmark. The
463 listening socket's mark is unchanged. Listening sockets that already
464 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
469 tcp_invalid_ratelimit - INTEGER
470 Limit the maximal rate for sending duplicate acknowledgments
471 in response to incoming TCP packets that are for an existing
472 connection but that are invalid due to any of these reasons:
474 (a) out-of-window sequence number,
475 (b) out-of-window acknowledgment number, or
476 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
478 This can help mitigate simple "ack loop" DoS attacks, wherein
479 a buggy or malicious middlebox or man-in-the-middle can
480 rewrite TCP header fields in manner that causes each endpoint
481 to think that the other is sending invalid TCP segments, thus
482 causing each side to send an unterminating stream of duplicate
483 acknowledgments for invalid segments.
485 Using 0 disables rate-limiting of dupacks in response to
486 invalid segments; otherwise this value specifies the minimal
487 space between sending such dupacks, in milliseconds.
489 Default: 500 (milliseconds).
491 tcp_keepalive_time - INTEGER
492 How often TCP sends out keepalive messages when keepalive is enabled.
495 tcp_keepalive_probes - INTEGER
496 How many keepalive probes TCP sends out, until it decides that the
497 connection is broken. Default value: 9.
499 tcp_keepalive_intvl - INTEGER
500 How frequently the probes are send out. Multiplied by
501 tcp_keepalive_probes it is time to kill not responding connection,
502 after probes started. Default value: 75sec i.e. connection
503 will be aborted after ~11 minutes of retries.
505 tcp_l3mdev_accept - BOOLEAN
506 Enables child sockets to inherit the L3 master device index.
507 Enabling this option allows a "global" listen socket to work
508 across L3 master domains (e.g., VRFs) with connected sockets
509 derived from the listen socket to be bound to the L3 domain in
510 which the packets originated. Only valid when the kernel was
511 compiled with CONFIG_NET_L3_MASTER_DEV.
513 Default: 0 (disabled)
515 tcp_low_latency - BOOLEAN
516 This is a legacy option, it has no effect anymore.
518 tcp_max_orphans - INTEGER
519 Maximal number of TCP sockets not attached to any user file handle,
520 held by system. If this number is exceeded orphaned connections are
521 reset immediately and warning is printed. This limit exists
522 only to prevent simple DoS attacks, you _must_ not rely on this
523 or lower the limit artificially, but rather increase it
524 (probably, after increasing installed memory),
525 if network conditions require more than default value,
526 and tune network services to linger and kill such states
527 more aggressively. Let me to remind again: each orphan eats
528 up to ~64K of unswappable memory.
530 tcp_max_syn_backlog - INTEGER
531 Maximal number of remembered connection requests (SYN_RECV),
532 which have not received an acknowledgment from connecting client.
534 This is a per-listener limit.
536 The minimal value is 128 for low memory machines, and it will
537 increase in proportion to the memory of machine.
539 If server suffers from overload, try increasing this number.
541 Remember to also check /proc/sys/net/core/somaxconn
542 A SYN_RECV request socket consumes about 304 bytes of memory.
544 tcp_max_tw_buckets - INTEGER
545 Maximal number of timewait sockets held by system simultaneously.
546 If this number is exceeded time-wait socket is immediately destroyed
547 and warning is printed. This limit exists only to prevent
548 simple DoS attacks, you _must_ not lower the limit artificially,
549 but rather increase it (probably, after increasing installed memory),
550 if network conditions require more than default value.
552 tcp_mem - vector of 3 INTEGERs: min, pressure, max
553 min: below this number of pages TCP is not bothered about its
556 pressure: when amount of memory allocated by TCP exceeds this number
557 of pages, TCP moderates its memory consumption and enters memory
558 pressure mode, which is exited when memory consumption falls
561 max: number of pages allowed for queueing by all TCP sockets.
563 Defaults are calculated at boot time from amount of available
566 tcp_min_rtt_wlen - INTEGER
567 The window length of the windowed min filter to track the minimum RTT.
568 A shorter window lets a flow more quickly pick up new (higher)
569 minimum RTT when it is moved to a longer path (e.g., due to traffic
570 engineering). A longer window makes the filter more resistant to RTT
571 inflations such as transient congestion. The unit is seconds.
573 Possible values: 0 - 86400 (1 day)
577 tcp_moderate_rcvbuf - BOOLEAN
578 If set, TCP performs receive buffer auto-tuning, attempting to
579 automatically size the buffer (no greater than tcp_rmem[2]) to
580 match the size required by the path for full throughput. Enabled by
583 tcp_mtu_probing - INTEGER
584 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
588 - 1 - Disabled by default, enabled when an ICMP black hole detected
589 - 2 - Always enabled, use initial MSS of tcp_base_mss.
591 tcp_probe_interval - UNSIGNED INTEGER
592 Controls how often to start TCP Packetization-Layer Path MTU
593 Discovery reprobe. The default is reprobing every 10 minutes as
596 tcp_probe_threshold - INTEGER
597 Controls when TCP Packetization-Layer Path MTU Discovery probing
598 will stop in respect to the width of search range in bytes. Default
601 tcp_no_metrics_save - BOOLEAN
602 By default, TCP saves various connection metrics in the route cache
603 when the connection closes, so that connections established in the
604 near future can use these to set initial conditions. Usually, this
605 increases overall performance, but may sometimes cause performance
606 degradation. If set, TCP will not cache metrics on closing
609 tcp_no_ssthresh_metrics_save - BOOLEAN
610 Controls whether TCP saves ssthresh metrics in the route cache.
612 Default is 1, which disables ssthresh metrics.
614 tcp_orphan_retries - INTEGER
615 This value influences the timeout of a locally closed TCP connection,
616 when RTO retransmissions remain unacknowledged.
617 See tcp_retries2 for more details.
619 The default value is 8.
621 If your machine is a loaded WEB server,
622 you should think about lowering this value, such sockets
623 may consume significant resources. Cf. tcp_max_orphans.
625 tcp_recovery - INTEGER
626 This value is a bitmap to enable various experimental loss recovery
629 ========= =============================================================
630 RACK: 0x1 enables the RACK loss detection for fast detection of lost
631 retransmissions and tail drops. It also subsumes and disables
632 RFC6675 recovery for SACK connections.
634 RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
636 RACK: 0x4 disables RACK's DUPACK threshold heuristic
637 ========= =============================================================
641 tcp_reflect_tos - BOOLEAN
642 For listening sockets, reuse the DSCP value of the initial SYN message
643 for outgoing packets. This allows to have both directions of a TCP
644 stream to use the same DSCP value, assuming DSCP remains unchanged for
645 the lifetime of the connection.
647 This options affects both IPv4 and IPv6.
649 Default: 0 (disabled)
651 tcp_reordering - INTEGER
652 Initial reordering level of packets in a TCP stream.
653 TCP stack can then dynamically adjust flow reordering level
654 between this initial value and tcp_max_reordering
658 tcp_max_reordering - INTEGER
659 Maximal reordering level of packets in a TCP stream.
660 300 is a fairly conservative value, but you might increase it
661 if paths are using per packet load balancing (like bonding rr mode)
665 tcp_retrans_collapse - BOOLEAN
666 Bug-to-bug compatibility with some broken printers.
667 On retransmit try to send bigger packets to work around bugs in
670 tcp_retries1 - INTEGER
671 This value influences the time, after which TCP decides, that
672 something is wrong due to unacknowledged RTO retransmissions,
673 and reports this suspicion to the network layer.
674 See tcp_retries2 for more details.
676 RFC 1122 recommends at least 3 retransmissions, which is the
679 tcp_retries2 - INTEGER
680 This value influences the timeout of an alive TCP connection,
681 when RTO retransmissions remain unacknowledged.
682 Given a value of N, a hypothetical TCP connection following
683 exponential backoff with an initial RTO of TCP_RTO_MIN would
684 retransmit N times before killing the connection at the (N+1)th RTO.
686 The default value of 15 yields a hypothetical timeout of 924.6
687 seconds and is a lower bound for the effective timeout.
688 TCP will effectively time out at the first RTO which exceeds the
689 hypothetical timeout.
691 RFC 1122 recommends at least 100 seconds for the timeout,
692 which corresponds to a value of at least 8.
694 tcp_rfc1337 - BOOLEAN
695 If set, the TCP stack behaves conforming to RFC1337. If unset,
696 we are not conforming to RFC, but prevent TCP TIME_WAIT
701 tcp_rmem - vector of 3 INTEGERs: min, default, max
702 min: Minimal size of receive buffer used by TCP sockets.
703 It is guaranteed to each TCP socket, even under moderate memory
708 default: initial size of receive buffer used by TCP sockets.
709 This value overrides net.core.rmem_default used by other protocols.
710 Default: 131072 bytes.
711 This value results in initial window of 65535.
713 max: maximal size of receive buffer allowed for automatically
714 selected receiver buffers for TCP socket. This value does not override
715 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
716 automatic tuning of that socket's receive buffer size, in which
717 case this value is ignored.
718 Default: between 131072 and 6MB, depending on RAM size.
721 Enable select acknowledgments (SACKS).
723 tcp_comp_sack_delay_ns - LONG INTEGER
724 TCP tries to reduce number of SACK sent, using a timer
725 based on 5% of SRTT, capped by this sysctl, in nano seconds.
726 The default is 1ms, based on TSO autosizing period.
728 Default : 1,000,000 ns (1 ms)
730 tcp_comp_sack_slack_ns - LONG INTEGER
731 This sysctl control the slack used when arming the
732 timer used by SACK compression. This gives extra time
733 for small RTT flows, and reduces system overhead by allowing
734 opportunistic reduction of timer interrupts.
736 Default : 100,000 ns (100 us)
738 tcp_comp_sack_nr - INTEGER
739 Max number of SACK that can be compressed.
740 Using 0 disables SACK compression.
744 tcp_slow_start_after_idle - BOOLEAN
745 If set, provide RFC2861 behavior and time out the congestion
746 window after an idle period. An idle period is defined at
747 the current RTO. If unset, the congestion window will not
748 be timed out after an idle period.
753 Use the Host requirements interpretation of the TCP urgent pointer field.
754 Most hosts use the older BSD interpretation, so if you turn this on
755 Linux might not communicate correctly with them.
759 tcp_synack_retries - INTEGER
760 Number of times SYNACKs for a passive TCP connection attempt will
761 be retransmitted. Should not be higher than 255. Default value
762 is 5, which corresponds to 31seconds till the last retransmission
763 with the current initial RTO of 1second. With this the final timeout
764 for a passive TCP connection will happen after 63seconds.
766 tcp_syncookies - INTEGER
767 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
768 Send out syncookies when the syn backlog queue of a socket
769 overflows. This is to prevent against the common 'SYN flood attack'
772 Note, that syncookies is fallback facility.
773 It MUST NOT be used to help highly loaded servers to stand
774 against legal connection rate. If you see SYN flood warnings
775 in your logs, but investigation shows that they occur
776 because of overload with legal connections, you should tune
777 another parameters until this warning disappear.
778 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
780 syncookies seriously violate TCP protocol, do not allow
781 to use TCP extensions, can result in serious degradation
782 of some services (f.e. SMTP relaying), visible not by you,
783 but your clients and relays, contacting you. While you see
784 SYN flood warnings in logs not being really flooded, your server
785 is seriously misconfigured.
787 If you want to test which effects syncookies have to your
788 network connections you can set this knob to 2 to enable
789 unconditionally generation of syncookies.
791 tcp_migrate_req - BOOLEAN
792 The incoming connection is tied to a specific listening socket when
793 the initial SYN packet is received during the three-way handshake.
794 When a listener is closed, in-flight request sockets during the
795 handshake and established sockets in the accept queue are aborted.
797 If the listener has SO_REUSEPORT enabled, other listeners on the
798 same port should have been able to accept such connections. This
799 option makes it possible to migrate such child sockets to another
800 listener after close() or shutdown().
802 The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should
803 usually be used to define the policy to pick an alive listener.
804 Otherwise, the kernel will randomly pick an alive listener only if
805 this option is enabled.
807 Note that migration between listeners with different settings may
808 crash applications. Let's say migration happens from listener A to
809 B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from
810 the requests migrated from A. To avoid such a situation, cancel
811 migration by returning SK_DROP in the type of eBPF program, or
816 tcp_fastopen - INTEGER
817 Enable TCP Fast Open (RFC7413) to send and accept data in the opening
820 The client support is enabled by flag 0x1 (on by default). The client
821 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
822 rather than connect() to send data in SYN.
824 The server support is enabled by flag 0x2 (off by default). Then
825 either enable for all listeners with another flag (0x400) or
826 enable individual listeners via TCP_FASTOPEN socket option with
827 the option value being the length of the syn-data backlog.
829 The values (bitmap) are
831 ===== ======== ======================================================
832 0x1 (client) enables sending data in the opening SYN on the client.
833 0x2 (server) enables the server support, i.e., allowing data in
834 a SYN packet to be accepted and passed to the
835 application before 3-way handshake finishes.
836 0x4 (client) send data in the opening SYN regardless of cookie
837 availability and without a cookie option.
838 0x200 (server) accept data-in-SYN w/o any cookie option present.
839 0x400 (server) enable all listeners to support Fast Open by
840 default without explicit TCP_FASTOPEN socket option.
841 ===== ======== ======================================================
845 Note that additional client or server features are only
846 effective if the basic support (0x1 and 0x2) are enabled respectively.
848 tcp_fastopen_blackhole_timeout_sec - INTEGER
849 Initial time period in second to disable Fastopen on active TCP sockets
850 when a TFO firewall blackhole issue happens.
851 This time period will grow exponentially when more blackhole issues
852 get detected right after Fastopen is re-enabled and will reset to
853 initial value when the blackhole issue goes away.
854 0 to disable the blackhole detection.
856 By default, it is set to 0 (feature is disabled).
858 tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
859 The list consists of a primary key and an optional backup key. The
860 primary key is used for both creating and validating cookies, while the
861 optional backup key is only used for validating cookies. The purpose of
862 the backup key is to maximize TFO validation when keys are rotated.
864 A randomly chosen primary key may be configured by the kernel if
865 the tcp_fastopen sysctl is set to 0x400 (see above), or if the
866 TCP_FASTOPEN setsockopt() optname is set and a key has not been
867 previously configured via sysctl. If keys are configured via
868 setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
869 per-socket keys will be used instead of any keys that are specified via
872 A key is specified as 4 8-digit hexadecimal integers which are separated
873 by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
874 omitted. A primary and a backup key may be specified by separating them
875 by a comma. If only one key is specified, it becomes the primary key and
876 any previously configured backup keys are removed.
878 tcp_syn_retries - INTEGER
879 Number of times initial SYNs for an active TCP connection attempt
880 will be retransmitted. Should not be higher than 127. Default value
881 is 6, which corresponds to 63seconds till the last retransmission
882 with the current initial RTO of 1second. With this the final timeout
883 for an active TCP connection attempt will happen after 127seconds.
885 tcp_timestamps - INTEGER
886 Enable timestamps as defined in RFC1323.
889 - 1: Enable timestamps as defined in RFC1323 and use random offset for
890 each connection rather than only using the current time.
891 - 2: Like 1, but without random offsets.
895 tcp_min_tso_segs - INTEGER
896 Minimal number of segments per TSO frame.
898 Since linux-3.12, TCP does an automatic sizing of TSO frames,
899 depending on flow rate, instead of filling 64Kbytes packets.
900 For specific usages, it's possible to force TCP to build big
901 TSO frames. Note that TCP stack might split too big TSO packets
902 if available window is too small.
906 tcp_tso_rtt_log - INTEGER
907 Adjustment of TSO packet sizes based on min_rtt
909 Starting from linux-5.18, TCP autosizing can be tweaked
910 for flows having small RTT.
912 Old autosizing was splitting the pacing budget to send 1024 TSO
915 tso_packet_size = sk->sk_pacing_rate / 1024;
917 With the new mechanism, we increase this TSO sizing using:
919 distance = min_rtt_usec / (2^tcp_tso_rtt_log)
920 tso_packet_size += gso_max_size >> distance;
922 This means that flows between very close hosts can use bigger
923 TSO packets, reducing their cpu costs.
925 If you want to use the old autosizing, set this sysctl to 0.
927 Default: 9 (2^9 = 512 usec)
929 tcp_pacing_ss_ratio - INTEGER
930 sk->sk_pacing_rate is set by TCP stack using a ratio applied
931 to current rate. (current_rate = cwnd * mss / srtt)
932 If TCP is in slow start, tcp_pacing_ss_ratio is applied
933 to let TCP probe for bigger speeds, assuming cwnd can be
934 doubled every other RTT.
938 tcp_pacing_ca_ratio - INTEGER
939 sk->sk_pacing_rate is set by TCP stack using a ratio applied
940 to current rate. (current_rate = cwnd * mss / srtt)
941 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
942 is applied to conservatively probe for bigger throughput.
946 tcp_tso_win_divisor - INTEGER
947 This allows control over what percentage of the congestion window
948 can be consumed by a single TSO frame.
949 The setting of this parameter is a choice between burstiness and
950 building larger TSO frames.
954 tcp_tw_reuse - INTEGER
955 Enable reuse of TIME-WAIT sockets for new connections when it is
956 safe from protocol viewpoint.
960 - 2 - enable for loopback traffic only
962 It should not be changed without advice/request of technical
967 tcp_window_scaling - BOOLEAN
968 Enable window scaling as defined in RFC1323.
970 tcp_wmem - vector of 3 INTEGERs: min, default, max
971 min: Amount of memory reserved for send buffers for TCP sockets.
972 Each TCP socket has rights to use it due to fact of its birth.
976 default: initial size of send buffer used by TCP sockets. This
977 value overrides net.core.wmem_default used by other protocols.
979 It is usually lower than net.core.wmem_default.
983 max: Maximal amount of memory allowed for automatically tuned
984 send buffers for TCP sockets. This value does not override
985 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
986 automatic tuning of that socket's send buffer size, in which case
987 this value is ignored.
989 Default: between 64K and 4MB, depending on RAM size.
991 tcp_notsent_lowat - UNSIGNED INTEGER
992 A TCP socket can control the amount of unsent bytes in its write queue,
993 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
994 reports POLLOUT events if the amount of unsent bytes is below a per
995 socket value, and if the write queue is not full. sendmsg() will
996 also not add new buffers if the limit is hit.
998 This global variable controls the amount of unsent data for
999 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
1000 to the global variable has immediate effect.
1002 Default: UINT_MAX (0xFFFFFFFF)
1004 tcp_workaround_signed_windows - BOOLEAN
1005 If set, assume no receipt of a window scaling option means the
1006 remote TCP is broken and treats the window as a signed quantity.
1007 If unset, assume the remote TCP is not broken even if we do
1008 not receive a window scaling option from them.
1012 tcp_thin_linear_timeouts - BOOLEAN
1013 Enable dynamic triggering of linear timeouts for thin streams.
1014 If set, a check is performed upon retransmission by timeout to
1015 determine if the stream is thin (less than 4 packets in flight).
1016 As long as the stream is found to be thin, up to 6 linear
1017 timeouts may be performed before exponential backoff mode is
1018 initiated. This improves retransmission latency for
1019 non-aggressive thin streams, often found to be time-dependent.
1020 For more information on thin streams, see
1021 Documentation/networking/tcp-thin.rst
1025 tcp_limit_output_bytes - INTEGER
1026 Controls TCP Small Queue limit per tcp socket.
1027 TCP bulk sender tends to increase packets in flight until it
1028 gets losses notifications. With SNDBUF autotuning, this can
1029 result in a large amount of packets queued on the local machine
1030 (e.g.: qdiscs, CPU backlog, or device) hurting latency of other
1031 flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes
1032 limits the number of bytes on qdisc or device to reduce artificial
1033 RTT/cwnd and reduce bufferbloat.
1035 Default: 1048576 (16 * 65536)
1037 tcp_challenge_ack_limit - INTEGER
1038 Limits number of Challenge ACK sent per second, as recommended
1039 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
1040 Note that this per netns rate limit can allow some side channel
1041 attacks and probably should not be enabled.
1042 TCP stack implements per TCP socket limits anyway.
1043 Default: INT_MAX (unlimited)
1045 tcp_ehash_entries - INTEGER
1046 Show the number of hash buckets for TCP sockets in the current
1047 networking namespace.
1049 A negative value means the networking namespace does not own its
1050 hash buckets and shares the initial networking namespace's one.
1052 tcp_child_ehash_entries - INTEGER
1053 Control the number of hash buckets for TCP sockets in the child
1054 networking namespace, which must be set before clone() or unshare().
1056 If the value is not 0, the kernel uses a value rounded up to 2^n
1057 as the actual hash bucket size. 0 is a special value, meaning
1058 the child networking namespace will share the initial networking
1059 namespace's hash buckets.
1061 Note that the child will use the global one in case the kernel
1062 fails to allocate enough memory. In addition, the global hash
1063 buckets are spread over available NUMA nodes, but the allocation
1064 of the child hash table depends on the current process's NUMA
1065 policy, which could result in performance differences.
1067 Note also that the default value of tcp_max_tw_buckets and
1068 tcp_max_syn_backlog depend on the hash bucket size.
1070 Possible values: 0, 2^n (n: 0 - 24 (16Mi))
1077 udp_l3mdev_accept - BOOLEAN
1078 Enabling this option allows a "global" bound socket to work
1079 across L3 master domains (e.g., VRFs) with packets capable of
1080 being received regardless of the L3 domain in which they
1081 originated. Only valid when the kernel was compiled with
1082 CONFIG_NET_L3_MASTER_DEV.
1084 Default: 0 (disabled)
1086 udp_mem - vector of 3 INTEGERs: min, pressure, max
1087 Number of pages allowed for queueing by all UDP sockets.
1089 min: Number of pages allowed for queueing by all UDP sockets.
1091 pressure: This value was introduced to follow format of tcp_mem.
1093 max: This value was introduced to follow format of tcp_mem.
1095 Default is calculated at boot time from amount of available memory.
1097 udp_rmem_min - INTEGER
1098 Minimal size of receive buffer used by UDP sockets in moderation.
1099 Each UDP socket is able to use the size for receiving data, even if
1100 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
1104 udp_wmem_min - INTEGER
1105 UDP does not have tx memory accounting and this tunable has no effect.
1110 raw_l3mdev_accept - BOOLEAN
1111 Enabling this option allows a "global" bound socket to work
1112 across L3 master domains (e.g., VRFs) with packets capable of
1113 being received regardless of the L3 domain in which they
1114 originated. Only valid when the kernel was compiled with
1115 CONFIG_NET_L3_MASTER_DEV.
1117 Default: 1 (enabled)
1122 cipso_cache_enable - BOOLEAN
1123 If set, enable additions to and lookups from the CIPSO label mapping
1124 cache. If unset, additions are ignored and lookups always result in a
1125 miss. However, regardless of the setting the cache is still
1126 invalidated when required when means you can safely toggle this on and
1127 off and the cache will always be "safe".
1131 cipso_cache_bucket_size - INTEGER
1132 The CIPSO label cache consists of a fixed size hash table with each
1133 hash bucket containing a number of cache entries. This variable limits
1134 the number of entries in each hash bucket; the larger the value is, the
1135 more CIPSO label mappings that can be cached. When the number of
1136 entries in a given hash bucket reaches this limit adding new entries
1137 causes the oldest entry in the bucket to be removed to make room.
1141 cipso_rbm_optfmt - BOOLEAN
1142 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
1143 the CIPSO draft specification (see Documentation/netlabel for details).
1144 This means that when set the CIPSO tag will be padded with empty
1145 categories in order to make the packet data 32-bit aligned.
1149 cipso_rbm_structvalid - BOOLEAN
1150 If set, do a very strict check of the CIPSO option when
1151 ip_options_compile() is called. If unset, relax the checks done during
1152 ip_options_compile(). Either way is "safe" as errors are caught else
1153 where in the CIPSO processing code but setting this to 0 (False) should
1154 result in less work (i.e. it should be faster) but could cause problems
1155 with other implementations that require strict checking.
1162 ip_local_port_range - 2 INTEGERS
1163 Defines the local port range that is used by TCP and UDP to
1164 choose the local port. The first number is the first, the
1165 second the last local port number.
1166 If possible, it is better these numbers have different parity
1167 (one even and one odd value).
1168 Must be greater than or equal to ip_unprivileged_port_start.
1169 The default values are 32768 and 60999 respectively.
1171 ip_local_reserved_ports - list of comma separated ranges
1172 Specify the ports which are reserved for known third-party
1173 applications. These ports will not be used by automatic port
1174 assignments (e.g. when calling connect() or bind() with port
1175 number 0). Explicit port allocation behavior is unchanged.
1177 The format used for both input and output is a comma separated
1178 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
1179 10). Writing to the file will clear all previously reserved
1180 ports and update the current list with the one given in the
1183 Note that ip_local_port_range and ip_local_reserved_ports
1184 settings are independent and both are considered by the kernel
1185 when determining which ports are available for automatic port
1188 You can reserve ports which are not in the current
1189 ip_local_port_range, e.g.::
1191 $ cat /proc/sys/net/ipv4/ip_local_port_range
1193 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
1196 although this is redundant. However such a setting is useful
1197 if later the port range is changed to a value that will
1198 include the reserved ports. Also keep in mind, that overlapping
1199 of these ranges may affect probability of selecting ephemeral
1200 ports which are right after block of reserved ports.
1204 ip_unprivileged_port_start - INTEGER
1205 This is a per-namespace sysctl. It defines the first
1206 unprivileged port in the network namespace. Privileged ports
1207 require root or CAP_NET_BIND_SERVICE in order to bind to them.
1208 To disable all privileged ports, set this to 0. They must not
1209 overlap with the ip_local_port_range.
1213 ip_nonlocal_bind - BOOLEAN
1214 If set, allows processes to bind() to non-local IP addresses,
1215 which can be quite useful - but may break some applications.
1219 ip_autobind_reuse - BOOLEAN
1220 By default, bind() does not select the ports automatically even if
1221 the new socket and all sockets bound to the port have SO_REUSEADDR.
1222 ip_autobind_reuse allows bind() to reuse the port and this is useful
1223 when you use bind()+connect(), but may break some applications.
1224 The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
1225 option should only be set by experts.
1228 ip_dynaddr - INTEGER
1229 If set non-zero, enables support for dynamic addresses.
1230 If set to a non-zero value larger than 1, a kernel log
1231 message will be printed when dynamic address rewriting
1236 ip_early_demux - BOOLEAN
1237 Optimize input packet processing down to one demux for
1238 certain kinds of local sockets. Currently we only do this
1239 for established TCP and connected UDP sockets.
1241 It may add an additional cost for pure routing workloads that
1242 reduces overall throughput, in such case you should disable it.
1246 ping_group_range - 2 INTEGERS
1247 Restrict ICMP_PROTO datagram sockets to users in the group range.
1248 The default is "1 0", meaning, that nobody (not even root) may
1249 create ping sockets. Setting it to "100 100" would grant permissions
1250 to the single group. "0 4294967295" would enable it for the world, "100
1251 4294967295" would enable it for the users, but not daemons.
1253 tcp_early_demux - BOOLEAN
1254 Enable early demux for established TCP sockets.
1258 udp_early_demux - BOOLEAN
1259 Enable early demux for connected UDP sockets. Disable this if
1260 your system could experience more unconnected load.
1264 icmp_echo_ignore_all - BOOLEAN
1265 If set non-zero, then the kernel will ignore all ICMP ECHO
1266 requests sent to it.
1270 icmp_echo_enable_probe - BOOLEAN
1271 If set to one, then the kernel will respond to RFC 8335 PROBE
1272 requests sent to it.
1276 icmp_echo_ignore_broadcasts - BOOLEAN
1277 If set non-zero, then the kernel will ignore all ICMP ECHO and
1278 TIMESTAMP requests sent to it via broadcast/multicast.
1282 icmp_ratelimit - INTEGER
1283 Limit the maximal rates for sending ICMP packets whose type matches
1284 icmp_ratemask (see below) to specific targets.
1285 0 to disable any limiting,
1286 otherwise the minimal space between responses in milliseconds.
1287 Note that another sysctl, icmp_msgs_per_sec limits the number
1288 of ICMP packets sent on all targets.
1292 icmp_msgs_per_sec - INTEGER
1293 Limit maximal number of ICMP packets sent per second from this host.
1294 Only messages whose type matches icmp_ratemask (see below) are
1295 controlled by this limit. For security reasons, the precise count
1296 of messages per second is randomized.
1300 icmp_msgs_burst - INTEGER
1301 icmp_msgs_per_sec controls number of ICMP packets sent per second,
1302 while icmp_msgs_burst controls the burst size of these packets.
1303 For security reasons, the precise burst size is randomized.
1307 icmp_ratemask - INTEGER
1308 Mask made of ICMP types for which rates are being limited.
1310 Significant bits: IHGFEDCBA9876543210
1312 Default mask: 0000001100000011000 (6168)
1314 Bit definitions (see include/linux/icmp.h):
1316 = =========================
1318 3 Destination Unreachable [1]_
1319 4 Source Quench [1]_
1322 B Time Exceeded [1]_
1323 C Parameter Problem [1]_
1328 H Address Mask Request
1329 I Address Mask Reply
1330 = =========================
1332 .. [1] These are rate limited by default (see default mask above)
1334 icmp_ignore_bogus_error_responses - BOOLEAN
1335 Some routers violate RFC1122 by sending bogus responses to broadcast
1336 frames. Such violations are normally logged via a kernel warning.
1337 If this is set to TRUE, the kernel will not give such warnings, which
1338 will avoid log file clutter.
1342 icmp_errors_use_inbound_ifaddr - BOOLEAN
1344 If zero, icmp error messages are sent with the primary address of
1345 the exiting interface.
1347 If non-zero, the message will be sent with the primary address of
1348 the interface that received the packet that caused the icmp error.
1349 This is the behaviour many network administrators will expect from
1350 a router. And it can make debugging complicated network layouts
1353 Note that if no primary address exists for the interface selected,
1354 then the primary address of the first non-loopback interface that
1355 has one will be used regardless of this setting.
1359 igmp_max_memberships - INTEGER
1360 Change the maximum number of multicast groups we can subscribe to.
1363 Theoretical maximum value is bounded by having to send a membership
1364 report in a single datagram (i.e. the report can't span multiple
1365 datagrams, or risk confusing the switch and leaving groups you don't
1368 The number of supported groups 'M' is bounded by the number of group
1369 report entries you can fit into a single datagram of 65535 bytes.
1371 M = 65536-sizeof (ip header)/(sizeof(Group record))
1373 Group records are variable length, with a minimum of 12 bytes.
1374 So net.ipv4.igmp_max_memberships should not be set higher than:
1376 (65536-24) / 12 = 5459
1378 The value 5459 assumes no IP header options, so in practice
1379 this number may be lower.
1381 igmp_max_msf - INTEGER
1382 Maximum number of addresses allowed in the source filter list for a
1388 Controls the IGMP query robustness variable (see RFC2236 8.1).
1390 Default: 2 (as specified by RFC2236 8.1)
1392 Minimum: 1 (as specified by RFC6636 4.5)
1394 force_igmp_version - INTEGER
1395 - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
1396 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
1397 Present timer expires.
1398 - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
1399 receive IGMPv2/v3 query.
1400 - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
1401 IGMPv1 query message. Will reply report if receive IGMPv3 query.
1402 - 3 - Enforce to use IGMP version 3. The same react with default 0.
1406 this is not the same with force_mld_version because IGMPv3 RFC3376
1407 Security Considerations does not have clear description that we could
1408 ignore other version messages completely as MLDv2 RFC3810. So make
1409 this value as default 0 is recommended.
1411 ``conf/interface/*``
1412 changes special settings per interface (where
1413 interface" is the name of your network interface)
1416 is special, changes the settings for all interfaces
1418 log_martians - BOOLEAN
1419 Log packets with impossible addresses to kernel log.
1420 log_martians for the interface will be enabled if at least one of
1421 conf/{all,interface}/log_martians is set to TRUE,
1422 it will be disabled otherwise
1424 accept_redirects - BOOLEAN
1425 Accept ICMP redirect messages.
1426 accept_redirects for the interface will be enabled if:
1428 - both conf/{all,interface}/accept_redirects are TRUE in the case
1429 forwarding for the interface is enabled
1433 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
1434 case forwarding for the interface is disabled
1436 accept_redirects for the interface will be disabled otherwise
1443 forwarding - BOOLEAN
1444 Enable IP forwarding on this interface. This controls whether packets
1445 received _on_ this interface can be forwarded.
1447 mc_forwarding - BOOLEAN
1448 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1449 and a multicast routing daemon is required.
1450 conf/all/mc_forwarding must also be set to TRUE to enable multicast
1451 routing for the interface
1454 Integer value used to differentiate the devices by the medium they
1455 are attached to. Two devices can have different id values when
1456 the broadcast packets are received only on one of them.
1457 The default value 0 means that the device is the only interface
1458 to its medium, value of -1 means that medium is not known.
1460 Currently, it is used to change the proxy_arp behavior:
1461 the proxy_arp feature is enabled for packets forwarded between
1462 two devices attached to different media.
1467 proxy_arp for the interface will be enabled if at least one of
1468 conf/{all,interface}/proxy_arp is set to TRUE,
1469 it will be disabled otherwise
1471 proxy_arp_pvlan - BOOLEAN
1472 Private VLAN proxy arp.
1474 Basically allow proxy arp replies back to the same interface
1475 (from which the ARP request/solicitation was received).
1477 This is done to support (ethernet) switch features, like RFC
1478 3069, where the individual ports are NOT allowed to
1479 communicate with each other, but they are allowed to talk to
1480 the upstream router. As described in RFC 3069, it is possible
1481 to allow these hosts to communicate through the upstream
1482 router by proxy_arp'ing. Don't need to be used together with
1485 This technology is known by different names:
1487 In RFC 3069 it is called VLAN Aggregation.
1488 Cisco and Allied Telesyn call it Private VLAN.
1489 Hewlett-Packard call it Source-Port filtering or port-isolation.
1490 Ericsson call it MAC-Forced Forwarding (RFC Draft).
1492 shared_media - BOOLEAN
1493 Send(router) or accept(host) RFC1620 shared media redirects.
1494 Overrides secure_redirects.
1496 shared_media for the interface will be enabled if at least one of
1497 conf/{all,interface}/shared_media is set to TRUE,
1498 it will be disabled otherwise
1502 secure_redirects - BOOLEAN
1503 Accept ICMP redirect messages only to gateways listed in the
1504 interface's current gateway list. Even if disabled, RFC1122 redirect
1507 Overridden by shared_media.
1509 secure_redirects for the interface will be enabled if at least one of
1510 conf/{all,interface}/secure_redirects is set to TRUE,
1511 it will be disabled otherwise
1515 send_redirects - BOOLEAN
1516 Send redirects, if router.
1518 send_redirects for the interface will be enabled if at least one of
1519 conf/{all,interface}/send_redirects is set to TRUE,
1520 it will be disabled otherwise
1524 bootp_relay - BOOLEAN
1525 Accept packets with source address 0.b.c.d destined
1526 not to this host as local ones. It is supposed, that
1527 BOOTP relay daemon will catch and forward such packets.
1528 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1533 Not Implemented Yet.
1535 accept_source_route - BOOLEAN
1536 Accept packets with SRR option.
1537 conf/all/accept_source_route must also be set to TRUE to accept packets
1538 with SRR option on the interface
1545 accept_local - BOOLEAN
1546 Accept packets with local source addresses. In combination with
1547 suitable routing, this can be used to direct packets between two
1548 local interfaces over the wire and have them accepted properly.
1551 route_localnet - BOOLEAN
1552 Do not consider loopback addresses as martian source or destination
1553 while routing. This enables the use of 127/8 for local routing purposes.
1558 - 0 - No source validation.
1559 - 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1560 Each incoming packet is tested against the FIB and if the interface
1561 is not the best reverse path the packet check will fail.
1562 By default failed packets are discarded.
1563 - 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1564 Each incoming packet's source address is also tested against the FIB
1565 and if the source address is not reachable via any interface
1566 the packet check will fail.
1568 Current recommended practice in RFC3704 is to enable strict mode
1569 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1570 or other complicated routing, then loose mode is recommended.
1572 The max value from conf/{all,interface}/rp_filter is used
1573 when doing source validation on the {interface}.
1575 Default value is 0. Note that some distributions enable it
1578 src_valid_mark - BOOLEAN
1579 - 0 - The fwmark of the packet is not included in reverse path
1580 route lookup. This allows for asymmetric routing configurations
1581 utilizing the fwmark in only one direction, e.g., transparent
1584 - 1 - The fwmark of the packet is included in reverse path route
1585 lookup. This permits rp_filter to function when the fwmark is
1586 used for routing traffic in both directions.
1588 This setting also affects the utilization of fmwark when
1589 performing source address selection for ICMP replies, or
1590 determining addresses stored for the IPOPT_TS_TSANDADDR and
1591 IPOPT_RR IP options.
1593 The max value from conf/{all,interface}/src_valid_mark is used.
1597 arp_filter - BOOLEAN
1598 - 1 - Allows you to have multiple network interfaces on the same
1599 subnet, and have the ARPs for each interface be answered
1600 based on whether or not the kernel would route a packet from
1601 the ARP'd IP out that interface (therefore you must use source
1602 based routing for this to work). In other words it allows control
1603 of which cards (usually 1) will respond to an arp request.
1605 - 0 - (default) The kernel can respond to arp requests with addresses
1606 from other interfaces. This may seem wrong but it usually makes
1607 sense, because it increases the chance of successful communication.
1608 IP addresses are owned by the complete host on Linux, not by
1609 particular interfaces. Only for more complex setups like load-
1610 balancing, does this behaviour cause problems.
1612 arp_filter for the interface will be enabled if at least one of
1613 conf/{all,interface}/arp_filter is set to TRUE,
1614 it will be disabled otherwise
1616 arp_announce - INTEGER
1617 Define different restriction levels for announcing the local
1618 source IP address from IP packets in ARP requests sent on
1621 - 0 - (default) Use any local address, configured on any interface
1622 - 1 - Try to avoid local addresses that are not in the target's
1623 subnet for this interface. This mode is useful when target
1624 hosts reachable via this interface require the source IP
1625 address in ARP requests to be part of their logical network
1626 configured on the receiving interface. When we generate the
1627 request we will check all our subnets that include the
1628 target IP and will preserve the source address if it is from
1629 such subnet. If there is no such subnet we select source
1630 address according to the rules for level 2.
1631 - 2 - Always use the best local address for this target.
1632 In this mode we ignore the source address in the IP packet
1633 and try to select local address that we prefer for talks with
1634 the target host. Such local address is selected by looking
1635 for primary IP addresses on all our subnets on the outgoing
1636 interface that include the target IP address. If no suitable
1637 local address is found we select the first local address
1638 we have on the outgoing interface or on all other interfaces,
1639 with the hope we will receive reply for our request and
1640 even sometimes no matter the source IP address we announce.
1642 The max value from conf/{all,interface}/arp_announce is used.
1644 Increasing the restriction level gives more chance for
1645 receiving answer from the resolved target while decreasing
1646 the level announces more valid sender's information.
1648 arp_ignore - INTEGER
1649 Define different modes for sending replies in response to
1650 received ARP requests that resolve local target IP addresses:
1652 - 0 - (default): reply for any local target IP address, configured
1654 - 1 - reply only if the target IP address is local address
1655 configured on the incoming interface
1656 - 2 - reply only if the target IP address is local address
1657 configured on the incoming interface and both with the
1658 sender's IP address are part from same subnet on this interface
1659 - 3 - do not reply for local addresses configured with scope host,
1660 only resolutions for global and link addresses are replied
1662 - 8 - do not reply for all local addresses
1664 The max value from conf/{all,interface}/arp_ignore is used
1665 when ARP request is received on the {interface}
1667 arp_notify - BOOLEAN
1668 Define mode for notification of address and device changes.
1670 == ==========================================================
1671 0 (default): do nothing
1672 1 Generate gratuitous arp requests when device is brought up
1673 or hardware address changes.
1674 == ==========================================================
1676 arp_accept - INTEGER
1677 Define behavior for accepting gratuitous ARP (garp) frames from devices
1678 that are not already present in the ARP table:
1680 - 0 - don't create new entries in the ARP table
1681 - 1 - create new entries in the ARP table
1682 - 2 - create new entries only if the source IP address is in the same
1683 subnet as an address configured on the interface that received the
1686 Both replies and requests type gratuitous arp will trigger the
1687 ARP table to be updated, if this setting is on.
1689 If the ARP table already contains the IP address of the
1690 gratuitous arp frame, the arp table will be updated regardless
1691 if this setting is on or off.
1693 arp_evict_nocarrier - BOOLEAN
1694 Clears the ARP cache on NOCARRIER events. This option is important for
1695 wireless devices where the ARP cache should not be cleared when roaming
1696 between access points on the same network. In most cases this should
1697 remain as the default (1).
1699 - 1 - (default): Clear the ARP cache on NOCARRIER events
1700 - 0 - Do not clear ARP cache on NOCARRIER events
1702 mcast_solicit - INTEGER
1703 The maximum number of multicast probes in INCOMPLETE state,
1704 when the associated hardware address is unknown. Defaults
1707 ucast_solicit - INTEGER
1708 The maximum number of unicast probes in PROBE state, when
1709 the hardware address is being reconfirmed. Defaults to 3.
1711 app_solicit - INTEGER
1712 The maximum number of probes to send to the user space ARP daemon
1713 via netlink before dropping back to multicast probes (see
1714 mcast_resolicit). Defaults to 0.
1716 mcast_resolicit - INTEGER
1717 The maximum number of multicast probes after unicast and
1718 app probes in PROBE state. Defaults to 0.
1720 disable_policy - BOOLEAN
1721 Disable IPSEC policy (SPD) for this interface
1723 disable_xfrm - BOOLEAN
1724 Disable IPSEC encryption on this interface, whatever the policy
1726 igmpv2_unsolicited_report_interval - INTEGER
1727 The interval in milliseconds in which the next unsolicited
1728 IGMPv1 or IGMPv2 report retransmit will take place.
1730 Default: 10000 (10 seconds)
1732 igmpv3_unsolicited_report_interval - INTEGER
1733 The interval in milliseconds in which the next unsolicited
1734 IGMPv3 report retransmit will take place.
1736 Default: 1000 (1 seconds)
1738 ignore_routes_with_linkdown - BOOLEAN
1739 Ignore routes whose link is down when performing a FIB lookup.
1741 promote_secondaries - BOOLEAN
1742 When a primary IP address is removed from this interface
1743 promote a corresponding secondary IP address instead of
1744 removing all the corresponding secondary IP addresses.
1746 drop_unicast_in_l2_multicast - BOOLEAN
1747 Drop any unicast IP packets that are received in link-layer
1748 multicast (or broadcast) frames.
1750 This behavior (for multicast) is actually a SHOULD in RFC
1751 1122, but is disabled by default for compatibility reasons.
1755 drop_gratuitous_arp - BOOLEAN
1756 Drop all gratuitous ARP frames, for example if there's a known
1757 good ARP proxy on the network and such frames need not be used
1758 (or in the case of 802.11, must not be used to prevent attacks.)
1764 Allows you to write a number, which can be used as required.
1768 xfrm4_gc_thresh - INTEGER
1769 (Obsolete since linux-4.14)
1770 The threshold at which we will start garbage collecting for IPv4
1771 destination cache entries. At twice this value the system will
1772 refuse new allocations.
1774 igmp_link_local_mcast_reports - BOOLEAN
1775 Enable IGMP reports for link local multicast groups in the
1781 kuznet@ms2.inr.ac.ru
1788 delon.nicolas@wanadoo.fr
1793 /proc/sys/net/ipv6/* Variables
1794 ==============================
1796 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1797 apply to IPv6 [XXX?].
1799 bindv6only - BOOLEAN
1800 Default value for IPV6_V6ONLY socket option,
1801 which restricts use of the IPv6 socket to IPv6 communication
1804 - TRUE: disable IPv4-mapped address feature
1805 - FALSE: enable IPv4-mapped address feature
1807 Default: FALSE (as specified in RFC3493)
1809 flowlabel_consistency - BOOLEAN
1810 Protect the consistency (and unicity) of flow label.
1811 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1819 auto_flowlabels - INTEGER
1820 Automatically generate flow labels based on a flow hash of the
1821 packet. This allows intermediate devices, such as routers, to
1822 identify packet flows for mechanisms like Equal Cost Multipath
1823 Routing (see RFC 6438).
1825 = ===========================================================
1826 0 automatic flow labels are completely disabled
1827 1 automatic flow labels are enabled by default, they can be
1828 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1830 2 automatic flow labels are allowed, they may be enabled on a
1831 per socket basis using the IPV6_AUTOFLOWLABEL socket option
1832 3 automatic flow labels are enabled and enforced, they cannot
1833 be disabled by the socket option
1834 = ===========================================================
1838 flowlabel_state_ranges - BOOLEAN
1839 Split the flow label number space into two ranges. 0-0x7FFFF is
1840 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1841 is reserved for stateless flow labels as described in RFC6437.
1848 flowlabel_reflect - INTEGER
1849 Control flow label reflection. Needed for Path MTU
1850 Discovery to work with Equal Cost Multipath Routing in anycast
1851 environments. See RFC 7690 and:
1852 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
1856 - 1: enabled for established flows
1858 Note that this prevents automatic flowlabel changes, as done
1859 in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
1860 and "tcp: Change txhash on every SYN and RTO retransmit"
1862 - 2: enabled for TCP RESET packets (no active listener)
1863 If set, a RST packet sent in response to a SYN packet on a closed
1864 port will reflect the incoming flow label.
1866 - 4: enabled for ICMPv6 echo reply messages.
1870 fib_multipath_hash_policy - INTEGER
1871 Controls which hash policy to use for multipath routes.
1873 Default: 0 (Layer 3)
1877 - 0 - Layer 3 (source and destination addresses plus flow label)
1878 - 1 - Layer 4 (standard 5-tuple)
1879 - 2 - Layer 3 or inner Layer 3 if present
1880 - 3 - Custom multipath hash. Fields used for multipath hash calculation
1881 are determined by fib_multipath_hash_fields sysctl
1883 fib_multipath_hash_fields - UNSIGNED INTEGER
1884 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
1885 fields used for multipath hash calculation are determined by this
1888 This value is a bitmask which enables various fields for multipath hash
1891 Possible fields are:
1893 ====== ============================
1894 0x0001 Source IP address
1895 0x0002 Destination IP address
1899 0x0020 Destination port
1900 0x0040 Inner source IP address
1901 0x0080 Inner destination IP address
1902 0x0100 Inner IP protocol
1903 0x0200 Inner Flow Label
1904 0x0400 Inner source port
1905 0x0800 Inner destination port
1906 ====== ============================
1908 Default: 0x0007 (source IP, destination IP and IP protocol)
1910 anycast_src_echo_reply - BOOLEAN
1911 Controls the use of anycast addresses as source addresses for ICMPv6
1919 idgen_delay - INTEGER
1920 Controls the delay in seconds after which time to retry
1921 privacy stable address generation if a DAD conflict is
1924 Default: 1 (as specified in RFC7217)
1926 idgen_retries - INTEGER
1927 Controls the number of retries to generate a stable privacy
1928 address if a DAD conflict is detected.
1930 Default: 3 (as specified in RFC7217)
1933 Controls the MLD query robustness variable (see RFC3810 9.1).
1935 Default: 2 (as specified by RFC3810 9.1)
1937 Minimum: 1 (as specified by RFC6636 4.5)
1939 max_dst_opts_number - INTEGER
1940 Maximum number of non-padding TLVs allowed in a Destination
1941 options extension header. If this value is less than zero
1942 then unknown options are disallowed and the number of known
1943 TLVs allowed is the absolute value of this number.
1947 max_hbh_opts_number - INTEGER
1948 Maximum number of non-padding TLVs allowed in a Hop-by-Hop
1949 options extension header. If this value is less than zero
1950 then unknown options are disallowed and the number of known
1951 TLVs allowed is the absolute value of this number.
1955 max_dst_opts_length - INTEGER
1956 Maximum length allowed for a Destination options extension
1959 Default: INT_MAX (unlimited)
1961 max_hbh_length - INTEGER
1962 Maximum length allowed for a Hop-by-Hop options extension
1965 Default: INT_MAX (unlimited)
1967 skip_notify_on_dev_down - BOOLEAN
1968 Controls whether an RTM_DELROUTE message is generated for routes
1969 removed when a device is taken down or deleted. IPv4 does not
1970 generate this message; IPv6 does by default. Setting this sysctl
1971 to true skips the message, making IPv4 and IPv6 on par in relying
1972 on userspace caches to track link events and evict routes.
1974 Default: false (generate message)
1976 nexthop_compat_mode - BOOLEAN
1977 New nexthop API provides a means for managing nexthops independent of
1978 prefixes. Backwards compatibilty with old route format is enabled by
1979 default which means route dumps and notifications contain the new
1980 nexthop attribute but also the full, expanded nexthop definition.
1981 Further, updates or deletes of a nexthop configuration generate route
1982 notifications for each fib entry using the nexthop. Once a system
1983 understands the new API, this sysctl can be disabled to achieve full
1984 performance benefits of the new API by disabling the nexthop expansion
1985 and extraneous notifications.
1986 Default: true (backward compat mode)
1988 fib_notify_on_flag_change - INTEGER
1989 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
1990 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
1992 After installing a route to the kernel, user space receives an
1993 acknowledgment, which means the route was installed in the kernel,
1994 but not necessarily in hardware.
1995 It is also possible for a route already installed in hardware to change
1996 its action and therefore its flags. For example, a host route that is
1997 trapping packets can be "promoted" to perform decapsulation following
1998 the installation of an IPinIP/VXLAN tunnel.
1999 The notifications will indicate to user-space the state of the route.
2001 Default: 0 (Do not emit notifications.)
2005 - 0 - Do not emit notifications.
2006 - 1 - Emit notifications.
2007 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
2010 Define the IOAM id of this node. Uses only 24 bits out of 32 in total.
2017 ioam6_id_wide - LONG INTEGER
2018 Define the wide IOAM id of this node. Uses only 56 bits out of 64 in
2019 total. Can be different from ioam6_id.
2022 Max: 0xFFFFFFFFFFFFFF
2024 Default: 0xFFFFFFFFFFFFFF
2028 ip6frag_high_thresh - INTEGER
2029 Maximum memory used to reassemble IPv6 fragments. When
2030 ip6frag_high_thresh bytes of memory is allocated for this purpose,
2031 the fragment handler will toss packets until ip6frag_low_thresh
2034 ip6frag_low_thresh - INTEGER
2035 See ip6frag_high_thresh
2037 ip6frag_time - INTEGER
2038 Time in seconds to keep an IPv6 fragment in memory.
2041 Change the interface-specific default settings.
2043 These settings would be used during creating new interfaces.
2047 Change all the interface-specific settings.
2049 [XXX: Other special features than forwarding?]
2051 conf/all/disable_ipv6 - BOOLEAN
2052 Changing this value is same as changing ``conf/default/disable_ipv6``
2053 setting and also all per-interface ``disable_ipv6`` settings to the same
2056 Reading this value does not have any particular meaning. It does not say
2057 whether IPv6 support is enabled or disabled. Returned value can be 1
2058 also in the case when some interface has ``disable_ipv6`` set to 0 and
2059 has configured IPv6 addresses.
2061 conf/all/forwarding - BOOLEAN
2062 Enable global IPv6 forwarding between all interfaces.
2064 IPv4 and IPv6 work differently here; e.g. netfilter must be used
2065 to control which interfaces may forward packets and which not.
2067 This also sets all interfaces' Host/Router setting
2068 'forwarding' to the specified value. See below for details.
2070 This referred to as global forwarding.
2075 fwmark_reflect - BOOLEAN
2076 Controls the fwmark of kernel-generated IPv6 reply packets that are not
2077 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
2078 If unset, these packets have a fwmark of zero. If set, they have the
2079 fwmark of the packet they are replying to.
2083 ``conf/interface/*``:
2084 Change special settings per interface.
2086 The functional behaviour for certain settings is different
2087 depending on whether local forwarding is enabled or not.
2090 Accept Router Advertisements; autoconfigure using them.
2092 It also determines whether or not to transmit Router
2093 Solicitations. If and only if the functional setting is to
2094 accept Router Advertisements, Router Solicitations will be
2097 Possible values are:
2099 == ===========================================================
2100 0 Do not accept Router Advertisements.
2101 1 Accept Router Advertisements if forwarding is disabled.
2102 2 Overrule forwarding behaviour. Accept Router Advertisements
2103 even if forwarding is enabled.
2104 == ===========================================================
2108 - enabled if local forwarding is disabled.
2109 - disabled if local forwarding is enabled.
2111 accept_ra_defrtr - BOOLEAN
2112 Learn default router in Router Advertisement.
2116 - enabled if accept_ra is enabled.
2117 - disabled if accept_ra is disabled.
2119 ra_defrtr_metric - UNSIGNED INTEGER
2120 Route metric for default route learned in Router Advertisement. This value
2121 will be assigned as metric for the default route learned via IPv6 Router
2122 Advertisement. Takes affect only if accept_ra_defrtr is enabled.
2127 Default: IP6_RT_PRIO_USER i.e. 1024.
2129 accept_ra_from_local - BOOLEAN
2130 Accept RA with source-address that is found on local machine
2131 if the RA is otherwise proper and able to be accepted.
2133 Default is to NOT accept these as it may be an un-intended
2138 - enabled if accept_ra_from_local is enabled
2139 on a specific interface.
2140 - disabled if accept_ra_from_local is disabled
2141 on a specific interface.
2143 accept_ra_min_hop_limit - INTEGER
2144 Minimum hop limit Information in Router Advertisement.
2146 Hop limit Information in Router Advertisement less than this
2147 variable shall be ignored.
2151 accept_ra_pinfo - BOOLEAN
2152 Learn Prefix Information in Router Advertisement.
2156 - enabled if accept_ra is enabled.
2157 - disabled if accept_ra is disabled.
2159 accept_ra_rt_info_min_plen - INTEGER
2160 Minimum prefix length of Route Information in RA.
2162 Route Information w/ prefix smaller than this variable shall
2167 * 0 if accept_ra_rtr_pref is enabled.
2168 * -1 if accept_ra_rtr_pref is disabled.
2170 accept_ra_rt_info_max_plen - INTEGER
2171 Maximum prefix length of Route Information in RA.
2173 Route Information w/ prefix larger than this variable shall
2178 * 0 if accept_ra_rtr_pref is enabled.
2179 * -1 if accept_ra_rtr_pref is disabled.
2181 accept_ra_rtr_pref - BOOLEAN
2182 Accept Router Preference in RA.
2186 - enabled if accept_ra is enabled.
2187 - disabled if accept_ra is disabled.
2189 accept_ra_mtu - BOOLEAN
2190 Apply the MTU value specified in RA option 5 (RFC4861). If
2191 disabled, the MTU specified in the RA will be ignored.
2195 - enabled if accept_ra is enabled.
2196 - disabled if accept_ra is disabled.
2198 accept_redirects - BOOLEAN
2203 - enabled if local forwarding is disabled.
2204 - disabled if local forwarding is enabled.
2206 accept_source_route - INTEGER
2207 Accept source routing (routing extension header).
2209 - >= 0: Accept only routing header type 2.
2210 - < 0: Do not accept routing header.
2215 Autoconfigure addresses using Prefix Information in Router
2220 - enabled if accept_ra_pinfo is enabled.
2221 - disabled if accept_ra_pinfo is disabled.
2223 dad_transmits - INTEGER
2224 The amount of Duplicate Address Detection probes to send.
2228 forwarding - INTEGER
2229 Configure interface-specific Host/Router behaviour.
2233 It is recommended to have the same setting on all
2234 interfaces; mixed router/host scenarios are rather uncommon.
2236 Possible values are:
2238 - 0 Forwarding disabled
2239 - 1 Forwarding enabled
2243 By default, Host behaviour is assumed. This means:
2245 1. IsRouter flag is not set in Neighbour Advertisements.
2246 2. If accept_ra is TRUE (default), transmit Router
2248 3. If accept_ra is TRUE (default), accept Router
2249 Advertisements (and do autoconfiguration).
2250 4. If accept_redirects is TRUE (default), accept Redirects.
2254 If local forwarding is enabled, Router behaviour is assumed.
2255 This means exactly the reverse from the above:
2257 1. IsRouter flag is set in Neighbour Advertisements.
2258 2. Router Solicitations are not sent unless accept_ra is 2.
2259 3. Router Advertisements are ignored unless accept_ra is 2.
2260 4. Redirects are ignored.
2262 Default: 0 (disabled) if global forwarding is disabled (default),
2263 otherwise 1 (enabled).
2266 Default Hop Limit to set.
2271 Default Maximum Transfer Unit
2273 Default: 1280 (IPv6 required minimum)
2275 ip_nonlocal_bind - BOOLEAN
2276 If set, allows processes to bind() to non-local IPv6 addresses,
2277 which can be quite useful - but may break some applications.
2281 router_probe_interval - INTEGER
2282 Minimum interval (in seconds) between Router Probing described
2287 router_solicitation_delay - INTEGER
2288 Number of seconds to wait after interface is brought up
2289 before sending Router Solicitations.
2293 router_solicitation_interval - INTEGER
2294 Number of seconds to wait between Router Solicitations.
2298 router_solicitations - INTEGER
2299 Number of Router Solicitations to send until assuming no
2300 routers are present.
2304 use_oif_addrs_only - BOOLEAN
2305 When enabled, the candidate source addresses for destinations
2306 routed via this interface are restricted to the set of addresses
2307 configured on this interface (vis. RFC 6724, section 4).
2311 use_tempaddr - INTEGER
2312 Preference for Privacy Extensions (RFC3041).
2314 * <= 0 : disable Privacy Extensions
2315 * == 1 : enable Privacy Extensions, but prefer public
2316 addresses over temporary addresses.
2317 * > 1 : enable Privacy Extensions and prefer temporary
2318 addresses over public addresses.
2322 * 0 (for most devices)
2323 * -1 (for point-to-point devices and loopback devices)
2325 temp_valid_lft - INTEGER
2326 valid lifetime (in seconds) for temporary addresses.
2328 Default: 172800 (2 days)
2330 temp_prefered_lft - INTEGER
2331 Preferred lifetime (in seconds) for temporary addresses.
2333 Default: 86400 (1 day)
2335 keep_addr_on_down - INTEGER
2336 Keep all IPv6 addresses on an interface down event. If set static
2337 global addresses with no expiration time are not flushed.
2340 * 0 : system default
2343 Default: 0 (addresses are removed)
2345 max_desync_factor - INTEGER
2346 Maximum value for DESYNC_FACTOR, which is a random value
2347 that ensures that clients don't synchronize with each
2348 other and generate new addresses at exactly the same time.
2349 value is in seconds.
2353 regen_max_retry - INTEGER
2354 Number of attempts before give up attempting to generate
2355 valid temporary addresses.
2359 max_addresses - INTEGER
2360 Maximum number of autoconfigured addresses per interface. Setting
2361 to zero disables the limitation. It is not recommended to set this
2362 value too large (or to zero) because it would be an easy way to
2363 crash the kernel by allowing too many addresses to be created.
2367 disable_ipv6 - BOOLEAN
2368 Disable IPv6 operation. If accept_dad is set to 2, this value
2369 will be dynamically set to TRUE if DAD fails for the link-local
2372 Default: FALSE (enable IPv6 operation)
2374 When this value is changed from 1 to 0 (IPv6 is being enabled),
2375 it will dynamically create a link-local address on the given
2376 interface and start Duplicate Address Detection, if necessary.
2378 When this value is changed from 0 to 1 (IPv6 is being disabled),
2379 it will dynamically delete all addresses and routes on the given
2380 interface. From now on it will not possible to add addresses/routes
2381 to the selected interface.
2383 accept_dad - INTEGER
2384 Whether to accept DAD (Duplicate Address Detection).
2386 == ==============================================================
2388 1 Enable DAD (default)
2389 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate
2390 link-local address has been found.
2391 == ==============================================================
2393 DAD operation and mode on a given interface will be selected according
2394 to the maximum value of conf/{all,interface}/accept_dad.
2396 force_tllao - BOOLEAN
2397 Enable sending the target link-layer address option even when
2398 responding to a unicast neighbor solicitation.
2402 Quoting from RFC 2461, section 4.4, Target link-layer address:
2404 "The option MUST be included for multicast solicitations in order to
2405 avoid infinite Neighbor Solicitation "recursion" when the peer node
2406 does not have a cache entry to return a Neighbor Advertisements
2407 message. When responding to unicast solicitations, the option can be
2408 omitted since the sender of the solicitation has the correct link-
2409 layer address; otherwise it would not have be able to send the unicast
2410 solicitation in the first place. However, including the link-layer
2411 address in this case adds little overhead and eliminates a potential
2412 race condition where the sender deletes the cached link-layer address
2413 prior to receiving a response to a previous solicitation."
2415 ndisc_notify - BOOLEAN
2416 Define mode for notification of address and device changes.
2418 * 0 - (default): do nothing
2419 * 1 - Generate unsolicited neighbour advertisements when device is brought
2420 up or hardware address changes.
2422 ndisc_tclass - INTEGER
2423 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
2424 Discovery (Router Solicitation, Router Advertisement, Neighbor
2425 Solicitation, Neighbor Advertisement, Redirect) messages.
2426 These 8 bits can be interpreted as 6 high order bits holding the DSCP
2427 value and 2 low order bits representing ECN (which you probably want
2432 ndisc_evict_nocarrier - BOOLEAN
2433 Clears the neighbor discovery table on NOCARRIER events. This option is
2434 important for wireless devices where the neighbor discovery cache should
2435 not be cleared when roaming between access points on the same network.
2436 In most cases this should remain as the default (1).
2438 - 1 - (default): Clear neighbor discover cache on NOCARRIER events.
2439 - 0 - Do not clear neighbor discovery cache on NOCARRIER events.
2441 mldv1_unsolicited_report_interval - INTEGER
2442 The interval in milliseconds in which the next unsolicited
2443 MLDv1 report retransmit will take place.
2445 Default: 10000 (10 seconds)
2447 mldv2_unsolicited_report_interval - INTEGER
2448 The interval in milliseconds in which the next unsolicited
2449 MLDv2 report retransmit will take place.
2451 Default: 1000 (1 second)
2453 force_mld_version - INTEGER
2454 * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
2455 * 1 - Enforce to use MLD version 1
2456 * 2 - Enforce to use MLD version 2
2458 suppress_frag_ndisc - INTEGER
2459 Control RFC 6980 (Security Implications of IPv6 Fragmentation
2460 with IPv6 Neighbor Discovery) behavior:
2462 * 1 - (default) discard fragmented neighbor discovery packets
2463 * 0 - allow fragmented neighbor discovery packets
2465 optimistic_dad - BOOLEAN
2466 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
2468 * 0: disabled (default)
2471 Optimistic Duplicate Address Detection for the interface will be enabled
2472 if at least one of conf/{all,interface}/optimistic_dad is set to 1,
2473 it will be disabled otherwise.
2475 use_optimistic - BOOLEAN
2476 If enabled, do not classify optimistic addresses as deprecated during
2477 source address selection. Preferred addresses will still be chosen
2478 before optimistic addresses, subject to other ranking in the source
2479 address selection algorithm.
2481 * 0: disabled (default)
2484 This will be enabled if at least one of
2485 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
2487 stable_secret - IPv6 address
2488 This IPv6 address will be used as a secret to generate IPv6
2489 addresses for link-local addresses and autoconfigured
2490 ones. All addresses generated after setting this secret will
2491 be stable privacy ones by default. This can be changed via the
2492 addrgenmode ip-link. conf/default/stable_secret is used as the
2493 secret for the namespace, the interface specific ones can
2494 overwrite that. Writes to conf/all/stable_secret are refused.
2496 It is recommended to generate this secret during installation
2497 of a system and keep it stable after that.
2499 By default the stable secret is unset.
2501 addr_gen_mode - INTEGER
2502 Defines how link-local and autoconf addresses are generated.
2504 = =================================================================
2505 0 generate address based on EUI64 (default)
2506 1 do no generate a link-local address, use EUI64 for addresses
2507 generated from autoconf
2508 2 generate stable privacy addresses, using the secret from
2509 stable_secret (RFC7217)
2510 3 generate stable privacy addresses, using a random secret if unset
2511 = =================================================================
2513 drop_unicast_in_l2_multicast - BOOLEAN
2514 Drop any unicast IPv6 packets that are received in link-layer
2515 multicast (or broadcast) frames.
2517 By default this is turned off.
2519 drop_unsolicited_na - BOOLEAN
2520 Drop all unsolicited neighbor advertisements, for example if there's
2521 a known good NA proxy on the network and such frames need not be used
2522 (or in the case of 802.11, must not be used to prevent attacks.)
2524 By default this is turned off.
2526 accept_untracked_na - INTEGER
2527 Define behavior for accepting neighbor advertisements from devices that
2528 are absent in the neighbor cache:
2530 - 0 - (default) Do not accept unsolicited and untracked neighbor
2533 - 1 - Add a new neighbor cache entry in STALE state for routers on
2534 receiving a neighbor advertisement (either solicited or unsolicited)
2535 with target link-layer address option specified if no neighbor entry
2536 is already present for the advertised IPv6 address. Without this knob,
2537 NAs received for untracked addresses (absent in neighbor cache) are
2540 This is as per router-side behavior documented in RFC9131.
2542 This has lower precedence than drop_unsolicited_na.
2544 This will optimize the return path for the initial off-link
2545 communication that is initiated by a directly connected host, by
2546 ensuring that the first-hop router which turns on this setting doesn't
2547 have to buffer the initial return packets to do neighbor-solicitation.
2548 The prerequisite is that the host is configured to send unsolicited
2549 neighbor advertisements on interface bringup. This setting should be
2550 used in conjunction with the ndisc_notify setting on the host to
2551 satisfy this prerequisite.
2553 - 2 - Extend option (1) to add a new neighbor cache entry only if the
2554 source IP address is in the same subnet as an address configured on
2555 the interface that received the neighbor advertisement.
2557 enhanced_dad - BOOLEAN
2558 Include a nonce option in the IPv6 neighbor solicitation messages used for
2559 duplicate address detection per RFC7527. A received DAD NS will only signal
2560 a duplicate address if the nonce is different. This avoids any false
2561 detection of duplicates due to loopback of the NS messages that we send.
2562 The nonce option will be sent on an interface unless both of
2563 conf/{all,interface}/enhanced_dad are set to FALSE.
2571 Limit the maximal rates for sending ICMPv6 messages.
2573 0 to disable any limiting,
2574 otherwise the minimal space between responses in milliseconds.
2578 ratemask - list of comma separated ranges
2579 For ICMPv6 message types matching the ranges in the ratemask, limit
2580 the sending of the message according to ratelimit parameter.
2582 The format used for both input and output is a comma separated
2583 list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
2584 129). Writing to the file will clear all previous ranges of ICMPv6
2585 message types and update the current list with the input.
2587 Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
2588 for numerical values of ICMPv6 message types, e.g. echo request is 128
2589 and echo reply is 129.
2591 Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
2593 echo_ignore_all - BOOLEAN
2594 If set non-zero, then the kernel will ignore all ICMP ECHO
2595 requests sent to it over the IPv6 protocol.
2599 echo_ignore_multicast - BOOLEAN
2600 If set non-zero, then the kernel will ignore all ICMP ECHO
2601 requests sent to it over the IPv6 protocol via multicast.
2605 echo_ignore_anycast - BOOLEAN
2606 If set non-zero, then the kernel will ignore all ICMP ECHO
2607 requests sent to it over the IPv6 protocol destined to anycast address.
2611 xfrm6_gc_thresh - INTEGER
2612 (Obsolete since linux-4.14)
2613 The threshold at which we will start garbage collecting for IPv6
2614 destination cache entries. At twice this value the system will
2615 refuse new allocations.
2619 Pekka Savola <pekkas@netcore.fi>
2620 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
2623 /proc/sys/net/bridge/* Variables:
2624 =================================
2626 bridge-nf-call-arptables - BOOLEAN
2627 - 1 : pass bridged ARP traffic to arptables' FORWARD chain.
2632 bridge-nf-call-iptables - BOOLEAN
2633 - 1 : pass bridged IPv4 traffic to iptables' chains.
2638 bridge-nf-call-ip6tables - BOOLEAN
2639 - 1 : pass bridged IPv6 traffic to ip6tables' chains.
2644 bridge-nf-filter-vlan-tagged - BOOLEAN
2645 - 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
2650 bridge-nf-filter-pppoe-tagged - BOOLEAN
2651 - 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
2656 bridge-nf-pass-vlan-input-dev - BOOLEAN
2657 - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
2658 interface on the bridge and set the netfilter input device to the
2659 vlan. This allows use of e.g. "iptables -i br0.1" and makes the
2660 REDIRECT target work with vlan-on-top-of-bridge interfaces. When no
2661 matching vlan interface is found, or this switch is off, the input
2662 device is set to the bridge interface.
2664 - 0: disable bridge netfilter vlan interface lookup.
2668 ``proc/sys/net/sctp/*`` Variables:
2669 ==================================
2671 addip_enable - BOOLEAN
2672 Enable or disable extension of Dynamic Address Reconfiguration
2673 (ADD-IP) functionality specified in RFC5061. This extension provides
2674 the ability to dynamically add and remove new addresses for the SCTP
2677 1: Enable extension.
2679 0: Disable extension.
2684 Enable or disable pf (pf is short for potentially failed) state. A value
2685 of pf_retrans > path_max_retrans also disables pf state. That is, one of
2686 both pf_enable and pf_retrans > path_max_retrans can disable pf state.
2687 Since pf_retrans and path_max_retrans can be changed by userspace
2688 application, sometimes user expects to disable pf state by the value of
2689 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
2690 or path_max_retrans is changed by the user application, this pf state is
2691 enabled. As such, it is necessary to add this to dynamically enable
2692 and disable pf state. See:
2693 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
2703 Unset or enable/disable pf (pf is short for potentially failed) state
2704 exposure. Applications can control the exposure of the PF path state
2705 in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO
2706 sockopt. When it's unset, no SCTP_PEER_ADDR_CHANGE event with
2707 SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info
2708 can be got via SCTP_GET_PEER_ADDR_INFO sockopt; When it's enabled,
2709 a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming
2710 SCTP_PF state and a SCTP_PF-state transport info can be got via
2711 SCTP_GET_PEER_ADDR_INFO sockopt; When it's diabled, no
2712 SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when
2713 trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO
2716 0: Unset pf state exposure, Compatible with old applications.
2718 1: Disable pf state exposure.
2720 2: Enable pf state exposure.
2724 addip_noauth_enable - BOOLEAN
2725 Dynamic Address Reconfiguration (ADD-IP) requires the use of
2726 authentication to protect the operations of adding or removing new
2727 addresses. This requirement is mandated so that unauthorized hosts
2728 would not be able to hijack associations. However, older
2729 implementations may not have implemented this requirement while
2730 allowing the ADD-IP extension. For reasons of interoperability,
2731 we provide this variable to control the enforcement of the
2732 authentication requirement.
2734 == ===============================================================
2735 1 Allow ADD-IP extension to be used without authentication. This
2736 should only be set in a closed environment for interoperability
2737 with older implementations.
2739 0 Enforce the authentication requirement
2740 == ===============================================================
2744 auth_enable - BOOLEAN
2745 Enable or disable Authenticated Chunks extension. This extension
2746 provides the ability to send and receive authenticated chunks and is
2747 required for secure operation of Dynamic Address Reconfiguration
2750 - 1: Enable this extension.
2751 - 0: Disable this extension.
2755 prsctp_enable - BOOLEAN
2756 Enable or disable the Partial Reliability extension (RFC3758) which
2757 is used to notify peers that a given DATA should no longer be expected.
2759 - 1: Enable extension
2765 The limit of the number of new packets that can be initially sent. It
2766 controls how bursty the generated traffic can be.
2770 association_max_retrans - INTEGER
2771 Set the maximum number for retransmissions that an association can
2772 attempt deciding that the remote end is unreachable. If this value
2773 is exceeded, the association is terminated.
2777 max_init_retransmits - INTEGER
2778 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
2779 that an association will attempt before declaring the destination
2780 unreachable and terminating.
2784 path_max_retrans - INTEGER
2785 The maximum number of retransmissions that will be attempted on a given
2786 path. Once this threshold is exceeded, the path is considered
2787 unreachable, and new traffic will use a different path when the
2788 association is multihomed.
2792 pf_retrans - INTEGER
2793 The number of retransmissions that will be attempted on a given path
2794 before traffic is redirected to an alternate transport (should one
2795 exist). Note this is distinct from path_max_retrans, as a path that
2796 passes the pf_retrans threshold can still be used. Its only
2797 deprioritized when a transmission path is selected by the stack. This
2798 setting is primarily used to enable fast failover mechanisms without
2799 having to reduce path_max_retrans to a very low value. See:
2800 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
2801 for details. Note also that a value of pf_retrans > path_max_retrans
2802 disables this feature. Since both pf_retrans and path_max_retrans can
2803 be changed by userspace application, a variable pf_enable is used to
2808 ps_retrans - INTEGER
2809 Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
2810 from section-5 "Primary Path Switchover" in rfc7829. The primary path
2811 will be changed to another active path when the path error counter on
2812 the old primary path exceeds PSMR, so that "the SCTP sender is allowed
2813 to continue data transmission on a new working path even when the old
2814 primary destination address becomes active again". Note this feature
2815 is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
2816 and its value can't be less than 'pf_retrans' when changing by sysctl.
2820 rto_initial - INTEGER
2821 The initial round trip timeout value in milliseconds that will be used
2822 in calculating round trip times. This is the initial time interval
2823 for retransmissions.
2828 The maximum value (in milliseconds) of the round trip timeout. This
2829 is the largest time interval that can elapse between retransmissions.
2834 The minimum value (in milliseconds) of the round trip timeout. This
2835 is the smallest time interval the can elapse between retransmissions.
2839 hb_interval - INTEGER
2840 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
2841 are sent at the specified interval on idle paths to probe the state of
2842 a given path between 2 associations.
2846 sack_timeout - INTEGER
2847 The amount of time (in milliseconds) that the implementation will wait
2852 valid_cookie_life - INTEGER
2853 The default lifetime of the SCTP cookie (in milliseconds). The cookie
2854 is used during association establishment.
2858 cookie_preserve_enable - BOOLEAN
2859 Enable or disable the ability to extend the lifetime of the SCTP cookie
2860 that is used during the establishment phase of SCTP association
2862 - 1: Enable cookie lifetime extension.
2867 cookie_hmac_alg - STRING
2868 Select the hmac algorithm used when generating the cookie value sent by
2869 a listening sctp socket to a connecting client in the INIT-ACK chunk.
2876 Ability to assign md5 or sha1 as the selected alg is predicated on the
2877 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
2878 CONFIG_CRYPTO_SHA1).
2880 Default: Dependent on configuration. MD5 if available, else SHA1 if
2881 available, else none.
2883 rcvbuf_policy - INTEGER
2884 Determines if the receive buffer is attributed to the socket or to
2885 association. SCTP supports the capability to create multiple
2886 associations on a single socket. When using this capability, it is
2887 possible that a single stalled association that's buffering a lot
2888 of data may block other associations from delivering their data by
2889 consuming all of the receive buffer space. To work around this,
2890 the rcvbuf_policy could be set to attribute the receiver buffer space
2891 to each association instead of the socket. This prevents the described
2894 - 1: rcvbuf space is per association
2895 - 0: rcvbuf space is per socket
2899 sndbuf_policy - INTEGER
2900 Similar to rcvbuf_policy above, this applies to send buffer space.
2902 - 1: Send buffer is tracked per association
2903 - 0: Send buffer is tracked per socket.
2907 sctp_mem - vector of 3 INTEGERs: min, pressure, max
2908 Number of pages allowed for queueing by all SCTP sockets.
2910 min: Below this number of pages SCTP is not bothered about its
2911 memory appetite. When amount of memory allocated by SCTP exceeds
2912 this number, SCTP starts to moderate memory usage.
2914 pressure: This value was introduced to follow format of tcp_mem.
2916 max: Number of pages allowed for queueing by all SCTP sockets.
2918 Default is calculated at boot time from amount of available memory.
2920 sctp_rmem - vector of 3 INTEGERs: min, default, max
2921 Only the first value ("min") is used, "default" and "max" are
2924 min: Minimal size of receive buffer used by SCTP socket.
2925 It is guaranteed to each SCTP socket (but not association) even
2926 under moderate memory pressure.
2930 sctp_wmem - vector of 3 INTEGERs: min, default, max
2931 Only the first value ("min") is used, "default" and "max" are
2934 min: Minimum size of send buffer that can be used by SCTP sockets.
2935 It is guaranteed to each SCTP socket (but not association) even
2936 under moderate memory pressure.
2940 addr_scope_policy - INTEGER
2941 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
2943 - 0 - Disable IPv4 address scoping
2944 - 1 - Enable IPv4 address scoping
2945 - 2 - Follow draft but allow IPv4 private addresses
2946 - 3 - Follow draft but allow IPv4 link local addresses
2951 The listening port for the local UDP tunneling sock. Normally it's
2952 using the IANA-assigned UDP port number 9899 (sctp-tunneling).
2954 This UDP sock is used for processing the incoming UDP-encapsulated
2955 SCTP packets (from RFC6951), and shared by all applications in the
2956 same net namespace. This UDP sock will be closed when the value is
2959 The value will also be used to set the src port of the UDP header
2960 for the outgoing UDP-encapsulated SCTP packets. For the dest port,
2961 please refer to 'encap_port' below.
2965 encap_port - INTEGER
2966 The default remote UDP encapsulation port.
2968 This value is used to set the dest port of the UDP header for the
2969 outgoing UDP-encapsulated SCTP packets by default. Users can also
2970 change the value for each sock/asoc/transport by using setsockopt.
2971 For further information, please refer to RFC6951.
2973 Note that when connecting to a remote server, the client should set
2974 this to the port that the UDP tunneling sock on the peer server is
2975 listening to and the local UDP tunneling sock on the client also
2976 must be started. On the server, it would get the encap_port from
2977 the incoming packet's source port.
2981 plpmtud_probe_interval - INTEGER
2982 The time interval (in milliseconds) for the PLPMTUD probe timer,
2983 which is configured to expire after this period to receive an
2984 acknowledgment to a probe packet. This is also the time interval
2985 between the probes for the current pmtu when the probe search
2988 PLPMTUD will be disabled when 0 is set, and other values for it
2993 reconf_enable - BOOLEAN
2994 Enable or disable extension of Stream Reconfiguration functionality
2995 specified in RFC6525. This extension provides the ability to "reset"
2996 a stream, and it includes the Parameters of "Outgoing/Incoming SSN
2997 Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams".
2999 - 1: Enable extension.
3000 - 0: Disable extension.
3004 intl_enable - BOOLEAN
3005 Enable or disable extension of User Message Interleaving functionality
3006 specified in RFC8260. This extension allows the interleaving of user
3007 messages sent on different streams. With this feature enabled, I-DATA
3008 chunk will replace DATA chunk to carry user messages if also supported
3009 by the peer. Note that to use this feature, one needs to set this option
3010 to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2
3011 and SCTP_INTERLEAVING_SUPPORTED to 1.
3013 - 1: Enable extension.
3014 - 0: Disable extension.
3018 ecn_enable - BOOLEAN
3019 Control use of Explicit Congestion Notification (ECN) by SCTP.
3020 Like in TCP, ECN is used only when both ends of the SCTP connection
3021 indicate support for it. This feature is useful in avoiding losses
3022 due to congestion by allowing supporting routers to signal congestion
3023 before having to drop packets.
3031 ``/proc/sys/net/core/*``
3032 ========================
3034 Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
3037 ``/proc/sys/net/unix/*``
3038 ========================
3040 max_dgram_qlen - INTEGER
3041 The maximum length of dgram socket receive queue