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.
342 tcp_autocorking - BOOLEAN
343 Enable TCP auto corking :
344 When applications do consecutive small write()/sendmsg() system calls,
345 we try to coalesce these small writes as much as possible, to lower
346 total amount of sent packets. This is done if at least one prior
347 packet for the flow is waiting in Qdisc queues or device transmit
348 queue. Applications can still use TCP_CORK for optimal behavior
349 when they know how/when to uncork their sockets.
353 tcp_available_congestion_control - STRING
354 Shows the available congestion control choices that are registered.
355 More congestion control algorithms may be available as modules,
358 tcp_base_mss - INTEGER
359 The initial value of search_low to be used by the packetization layer
360 Path MTU discovery (MTU probing). If MTU probing is enabled,
361 this is the initial MSS used by the connection.
363 tcp_mtu_probe_floor - INTEGER
364 If MTU probing is enabled this caps the minimum MSS used for search_low
369 tcp_min_snd_mss - INTEGER
370 TCP SYN and SYNACK messages usually advertise an ADVMSS option,
371 as described in RFC 1122 and RFC 6691.
373 If this ADVMSS option is smaller than tcp_min_snd_mss,
374 it is silently capped to tcp_min_snd_mss.
376 Default : 48 (at least 8 bytes of payload per segment)
378 tcp_congestion_control - STRING
379 Set the congestion control algorithm to be used for new
380 connections. The algorithm "reno" is always available, but
381 additional choices may be available based on kernel configuration.
382 Default is set as part of kernel configuration.
383 For passive connections, the listener congestion control choice
386 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
389 Allows TCP to send "duplicate" SACKs.
391 tcp_early_retrans - INTEGER
392 Tail loss probe (TLP) converts RTOs occurring due to tail
393 losses into fast recovery (draft-ietf-tcpm-rack). Note that
394 TLP requires RACK to function properly (see tcp_recovery below)
404 Control use of Explicit Congestion Notification (ECN) by TCP.
405 ECN is used only when both ends of the TCP connection indicate
406 support for it. This feature is useful in avoiding losses due
407 to congestion by allowing supporting routers to signal
408 congestion before having to drop packets.
412 = =====================================================
413 0 Disable ECN. Neither initiate nor accept ECN.
414 1 Enable ECN when requested by incoming connections and
415 also request ECN on outgoing connection attempts.
416 2 Enable ECN when requested by incoming connections
417 but do not request ECN on outgoing connections.
418 = =====================================================
422 tcp_ecn_fallback - BOOLEAN
423 If the kernel detects that ECN connection misbehaves, enable fall
424 back to non-ECN. Currently, this knob implements the fallback
425 from RFC3168, section 6.1.1.1., but we reserve that in future,
426 additional detection mechanisms could be implemented under this
427 knob. The value is not used, if tcp_ecn or per route (or congestion
428 control) ECN settings are disabled.
430 Default: 1 (fallback enabled)
433 This is a legacy option, it has no effect anymore.
435 tcp_fin_timeout - INTEGER
436 The length of time an orphaned (no longer referenced by any
437 application) connection will remain in the FIN_WAIT_2 state
438 before it is aborted at the local end. While a perfectly
439 valid "receive only" state for an un-orphaned connection, an
440 orphaned connection in FIN_WAIT_2 state could otherwise wait
441 forever for the remote to close its end of the connection.
448 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
449 F-RTO is an enhanced recovery algorithm for TCP retransmission
450 timeouts. It is particularly beneficial in networks where the
451 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
452 modification. It does not require any support from the peer.
454 By default it's enabled with a non-zero value. 0 disables F-RTO.
456 tcp_fwmark_accept - BOOLEAN
457 If set, incoming connections to listening sockets that do not have a
458 socket mark will set the mark of the accepting socket to the fwmark of
459 the incoming SYN packet. This will cause all packets on that connection
460 (starting from the first SYNACK) to be sent with that fwmark. The
461 listening socket's mark is unchanged. Listening sockets that already
462 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
467 tcp_invalid_ratelimit - INTEGER
468 Limit the maximal rate for sending duplicate acknowledgments
469 in response to incoming TCP packets that are for an existing
470 connection but that are invalid due to any of these reasons:
472 (a) out-of-window sequence number,
473 (b) out-of-window acknowledgment number, or
474 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
476 This can help mitigate simple "ack loop" DoS attacks, wherein
477 a buggy or malicious middlebox or man-in-the-middle can
478 rewrite TCP header fields in manner that causes each endpoint
479 to think that the other is sending invalid TCP segments, thus
480 causing each side to send an unterminating stream of duplicate
481 acknowledgments for invalid segments.
483 Using 0 disables rate-limiting of dupacks in response to
484 invalid segments; otherwise this value specifies the minimal
485 space between sending such dupacks, in milliseconds.
487 Default: 500 (milliseconds).
489 tcp_keepalive_time - INTEGER
490 How often TCP sends out keepalive messages when keepalive is enabled.
493 tcp_keepalive_probes - INTEGER
494 How many keepalive probes TCP sends out, until it decides that the
495 connection is broken. Default value: 9.
497 tcp_keepalive_intvl - INTEGER
498 How frequently the probes are send out. Multiplied by
499 tcp_keepalive_probes it is time to kill not responding connection,
500 after probes started. Default value: 75sec i.e. connection
501 will be aborted after ~11 minutes of retries.
503 tcp_l3mdev_accept - BOOLEAN
504 Enables child sockets to inherit the L3 master device index.
505 Enabling this option allows a "global" listen socket to work
506 across L3 master domains (e.g., VRFs) with connected sockets
507 derived from the listen socket to be bound to the L3 domain in
508 which the packets originated. Only valid when the kernel was
509 compiled with CONFIG_NET_L3_MASTER_DEV.
511 Default: 0 (disabled)
513 tcp_low_latency - BOOLEAN
514 This is a legacy option, it has no effect anymore.
516 tcp_max_orphans - INTEGER
517 Maximal number of TCP sockets not attached to any user file handle,
518 held by system. If this number is exceeded orphaned connections are
519 reset immediately and warning is printed. This limit exists
520 only to prevent simple DoS attacks, you _must_ not rely on this
521 or lower the limit artificially, but rather increase it
522 (probably, after increasing installed memory),
523 if network conditions require more than default value,
524 and tune network services to linger and kill such states
525 more aggressively. Let me to remind again: each orphan eats
526 up to ~64K of unswappable memory.
528 tcp_max_syn_backlog - INTEGER
529 Maximal number of remembered connection requests (SYN_RECV),
530 which have not received an acknowledgment from connecting client.
532 This is a per-listener limit.
534 The minimal value is 128 for low memory machines, and it will
535 increase in proportion to the memory of machine.
537 If server suffers from overload, try increasing this number.
539 Remember to also check /proc/sys/net/core/somaxconn
540 A SYN_RECV request socket consumes about 304 bytes of memory.
542 tcp_max_tw_buckets - INTEGER
543 Maximal number of timewait sockets held by system simultaneously.
544 If this number is exceeded time-wait socket is immediately destroyed
545 and warning is printed. This limit exists only to prevent
546 simple DoS attacks, you _must_ not lower the limit artificially,
547 but rather increase it (probably, after increasing installed memory),
548 if network conditions require more than default value.
550 tcp_mem - vector of 3 INTEGERs: min, pressure, max
551 min: below this number of pages TCP is not bothered about its
554 pressure: when amount of memory allocated by TCP exceeds this number
555 of pages, TCP moderates its memory consumption and enters memory
556 pressure mode, which is exited when memory consumption falls
559 max: number of pages allowed for queueing by all TCP sockets.
561 Defaults are calculated at boot time from amount of available
564 tcp_min_rtt_wlen - INTEGER
565 The window length of the windowed min filter to track the minimum RTT.
566 A shorter window lets a flow more quickly pick up new (higher)
567 minimum RTT when it is moved to a longer path (e.g., due to traffic
568 engineering). A longer window makes the filter more resistant to RTT
569 inflations such as transient congestion. The unit is seconds.
571 Possible values: 0 - 86400 (1 day)
575 tcp_moderate_rcvbuf - BOOLEAN
576 If set, TCP performs receive buffer auto-tuning, attempting to
577 automatically size the buffer (no greater than tcp_rmem[2]) to
578 match the size required by the path for full throughput. Enabled by
581 tcp_mtu_probing - INTEGER
582 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
586 - 1 - Disabled by default, enabled when an ICMP black hole detected
587 - 2 - Always enabled, use initial MSS of tcp_base_mss.
589 tcp_probe_interval - UNSIGNED INTEGER
590 Controls how often to start TCP Packetization-Layer Path MTU
591 Discovery reprobe. The default is reprobing every 10 minutes as
594 tcp_probe_threshold - INTEGER
595 Controls when TCP Packetization-Layer Path MTU Discovery probing
596 will stop in respect to the width of search range in bytes. Default
599 tcp_no_metrics_save - BOOLEAN
600 By default, TCP saves various connection metrics in the route cache
601 when the connection closes, so that connections established in the
602 near future can use these to set initial conditions. Usually, this
603 increases overall performance, but may sometimes cause performance
604 degradation. If set, TCP will not cache metrics on closing
607 tcp_no_ssthresh_metrics_save - BOOLEAN
608 Controls whether TCP saves ssthresh metrics in the route cache.
610 Default is 1, which disables ssthresh metrics.
612 tcp_orphan_retries - INTEGER
613 This value influences the timeout of a locally closed TCP connection,
614 when RTO retransmissions remain unacknowledged.
615 See tcp_retries2 for more details.
617 The default value is 8.
619 If your machine is a loaded WEB server,
620 you should think about lowering this value, such sockets
621 may consume significant resources. Cf. tcp_max_orphans.
623 tcp_recovery - INTEGER
624 This value is a bitmap to enable various experimental loss recovery
627 ========= =============================================================
628 RACK: 0x1 enables the RACK loss detection for fast detection of lost
629 retransmissions and tail drops. It also subsumes and disables
630 RFC6675 recovery for SACK connections.
632 RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
634 RACK: 0x4 disables RACK's DUPACK threshold heuristic
635 ========= =============================================================
639 tcp_reflect_tos - BOOLEAN
640 For listening sockets, reuse the DSCP value of the initial SYN message
641 for outgoing packets. This allows to have both directions of a TCP
642 stream to use the same DSCP value, assuming DSCP remains unchanged for
643 the lifetime of the connection.
645 This options affects both IPv4 and IPv6.
647 Default: 0 (disabled)
649 tcp_reordering - INTEGER
650 Initial reordering level of packets in a TCP stream.
651 TCP stack can then dynamically adjust flow reordering level
652 between this initial value and tcp_max_reordering
656 tcp_max_reordering - INTEGER
657 Maximal reordering level of packets in a TCP stream.
658 300 is a fairly conservative value, but you might increase it
659 if paths are using per packet load balancing (like bonding rr mode)
663 tcp_retrans_collapse - BOOLEAN
664 Bug-to-bug compatibility with some broken printers.
665 On retransmit try to send bigger packets to work around bugs in
668 tcp_retries1 - INTEGER
669 This value influences the time, after which TCP decides, that
670 something is wrong due to unacknowledged RTO retransmissions,
671 and reports this suspicion to the network layer.
672 See tcp_retries2 for more details.
674 RFC 1122 recommends at least 3 retransmissions, which is the
677 tcp_retries2 - INTEGER
678 This value influences the timeout of an alive TCP connection,
679 when RTO retransmissions remain unacknowledged.
680 Given a value of N, a hypothetical TCP connection following
681 exponential backoff with an initial RTO of TCP_RTO_MIN would
682 retransmit N times before killing the connection at the (N+1)th RTO.
684 The default value of 15 yields a hypothetical timeout of 924.6
685 seconds and is a lower bound for the effective timeout.
686 TCP will effectively time out at the first RTO which exceeds the
687 hypothetical timeout.
689 RFC 1122 recommends at least 100 seconds for the timeout,
690 which corresponds to a value of at least 8.
692 tcp_rfc1337 - BOOLEAN
693 If set, the TCP stack behaves conforming to RFC1337. If unset,
694 we are not conforming to RFC, but prevent TCP TIME_WAIT
699 tcp_rmem - vector of 3 INTEGERs: min, default, max
700 min: Minimal size of receive buffer used by TCP sockets.
701 It is guaranteed to each TCP socket, even under moderate memory
706 default: initial size of receive buffer used by TCP sockets.
707 This value overrides net.core.rmem_default used by other protocols.
708 Default: 131072 bytes.
709 This value results in initial window of 65535.
711 max: maximal size of receive buffer allowed for automatically
712 selected receiver buffers for TCP socket. This value does not override
713 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
714 automatic tuning of that socket's receive buffer size, in which
715 case this value is ignored.
716 Default: between 131072 and 6MB, depending on RAM size.
719 Enable select acknowledgments (SACKS).
721 tcp_comp_sack_delay_ns - LONG INTEGER
722 TCP tries to reduce number of SACK sent, using a timer
723 based on 5% of SRTT, capped by this sysctl, in nano seconds.
724 The default is 1ms, based on TSO autosizing period.
726 Default : 1,000,000 ns (1 ms)
728 tcp_comp_sack_slack_ns - LONG INTEGER
729 This sysctl control the slack used when arming the
730 timer used by SACK compression. This gives extra time
731 for small RTT flows, and reduces system overhead by allowing
732 opportunistic reduction of timer interrupts.
734 Default : 100,000 ns (100 us)
736 tcp_comp_sack_nr - INTEGER
737 Max number of SACK that can be compressed.
738 Using 0 disables SACK compression.
742 tcp_slow_start_after_idle - BOOLEAN
743 If set, provide RFC2861 behavior and time out the congestion
744 window after an idle period. An idle period is defined at
745 the current RTO. If unset, the congestion window will not
746 be timed out after an idle period.
751 Use the Host requirements interpretation of the TCP urgent pointer field.
752 Most hosts use the older BSD interpretation, so if you turn this on
753 Linux might not communicate correctly with them.
757 tcp_synack_retries - INTEGER
758 Number of times SYNACKs for a passive TCP connection attempt will
759 be retransmitted. Should not be higher than 255. Default value
760 is 5, which corresponds to 31seconds till the last retransmission
761 with the current initial RTO of 1second. With this the final timeout
762 for a passive TCP connection will happen after 63seconds.
764 tcp_syncookies - INTEGER
765 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
766 Send out syncookies when the syn backlog queue of a socket
767 overflows. This is to prevent against the common 'SYN flood attack'
770 Note, that syncookies is fallback facility.
771 It MUST NOT be used to help highly loaded servers to stand
772 against legal connection rate. If you see SYN flood warnings
773 in your logs, but investigation shows that they occur
774 because of overload with legal connections, you should tune
775 another parameters until this warning disappear.
776 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
778 syncookies seriously violate TCP protocol, do not allow
779 to use TCP extensions, can result in serious degradation
780 of some services (f.e. SMTP relaying), visible not by you,
781 but your clients and relays, contacting you. While you see
782 SYN flood warnings in logs not being really flooded, your server
783 is seriously misconfigured.
785 If you want to test which effects syncookies have to your
786 network connections you can set this knob to 2 to enable
787 unconditionally generation of syncookies.
789 tcp_migrate_req - BOOLEAN
790 The incoming connection is tied to a specific listening socket when
791 the initial SYN packet is received during the three-way handshake.
792 When a listener is closed, in-flight request sockets during the
793 handshake and established sockets in the accept queue are aborted.
795 If the listener has SO_REUSEPORT enabled, other listeners on the
796 same port should have been able to accept such connections. This
797 option makes it possible to migrate such child sockets to another
798 listener after close() or shutdown().
800 The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should
801 usually be used to define the policy to pick an alive listener.
802 Otherwise, the kernel will randomly pick an alive listener only if
803 this option is enabled.
805 Note that migration between listeners with different settings may
806 crash applications. Let's say migration happens from listener A to
807 B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from
808 the requests migrated from A. To avoid such a situation, cancel
809 migration by returning SK_DROP in the type of eBPF program, or
814 tcp_fastopen - INTEGER
815 Enable TCP Fast Open (RFC7413) to send and accept data in the opening
818 The client support is enabled by flag 0x1 (on by default). The client
819 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
820 rather than connect() to send data in SYN.
822 The server support is enabled by flag 0x2 (off by default). Then
823 either enable for all listeners with another flag (0x400) or
824 enable individual listeners via TCP_FASTOPEN socket option with
825 the option value being the length of the syn-data backlog.
827 The values (bitmap) are
829 ===== ======== ======================================================
830 0x1 (client) enables sending data in the opening SYN on the client.
831 0x2 (server) enables the server support, i.e., allowing data in
832 a SYN packet to be accepted and passed to the
833 application before 3-way handshake finishes.
834 0x4 (client) send data in the opening SYN regardless of cookie
835 availability and without a cookie option.
836 0x200 (server) accept data-in-SYN w/o any cookie option present.
837 0x400 (server) enable all listeners to support Fast Open by
838 default without explicit TCP_FASTOPEN socket option.
839 ===== ======== ======================================================
843 Note that additional client or server features are only
844 effective if the basic support (0x1 and 0x2) are enabled respectively.
846 tcp_fastopen_blackhole_timeout_sec - INTEGER
847 Initial time period in second to disable Fastopen on active TCP sockets
848 when a TFO firewall blackhole issue happens.
849 This time period will grow exponentially when more blackhole issues
850 get detected right after Fastopen is re-enabled and will reset to
851 initial value when the blackhole issue goes away.
852 0 to disable the blackhole detection.
854 By default, it is set to 0 (feature is disabled).
856 tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
857 The list consists of a primary key and an optional backup key. The
858 primary key is used for both creating and validating cookies, while the
859 optional backup key is only used for validating cookies. The purpose of
860 the backup key is to maximize TFO validation when keys are rotated.
862 A randomly chosen primary key may be configured by the kernel if
863 the tcp_fastopen sysctl is set to 0x400 (see above), or if the
864 TCP_FASTOPEN setsockopt() optname is set and a key has not been
865 previously configured via sysctl. If keys are configured via
866 setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
867 per-socket keys will be used instead of any keys that are specified via
870 A key is specified as 4 8-digit hexadecimal integers which are separated
871 by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
872 omitted. A primary and a backup key may be specified by separating them
873 by a comma. If only one key is specified, it becomes the primary key and
874 any previously configured backup keys are removed.
876 tcp_syn_retries - INTEGER
877 Number of times initial SYNs for an active TCP connection attempt
878 will be retransmitted. Should not be higher than 127. Default value
879 is 6, which corresponds to 63seconds till the last retransmission
880 with the current initial RTO of 1second. With this the final timeout
881 for an active TCP connection attempt will happen after 127seconds.
883 tcp_timestamps - INTEGER
884 Enable timestamps as defined in RFC1323.
887 - 1: Enable timestamps as defined in RFC1323 and use random offset for
888 each connection rather than only using the current time.
889 - 2: Like 1, but without random offsets.
893 tcp_min_tso_segs - INTEGER
894 Minimal number of segments per TSO frame.
896 Since linux-3.12, TCP does an automatic sizing of TSO frames,
897 depending on flow rate, instead of filling 64Kbytes packets.
898 For specific usages, it's possible to force TCP to build big
899 TSO frames. Note that TCP stack might split too big TSO packets
900 if available window is too small.
904 tcp_tso_rtt_log - INTEGER
905 Adjustment of TSO packet sizes based on min_rtt
907 Starting from linux-5.18, TCP autosizing can be tweaked
908 for flows having small RTT.
910 Old autosizing was splitting the pacing budget to send 1024 TSO
913 tso_packet_size = sk->sk_pacing_rate / 1024;
915 With the new mechanism, we increase this TSO sizing using:
917 distance = min_rtt_usec / (2^tcp_tso_rtt_log)
918 tso_packet_size += gso_max_size >> distance;
920 This means that flows between very close hosts can use bigger
921 TSO packets, reducing their cpu costs.
923 If you want to use the old autosizing, set this sysctl to 0.
925 Default: 9 (2^9 = 512 usec)
927 tcp_pacing_ss_ratio - INTEGER
928 sk->sk_pacing_rate is set by TCP stack using a ratio applied
929 to current rate. (current_rate = cwnd * mss / srtt)
930 If TCP is in slow start, tcp_pacing_ss_ratio is applied
931 to let TCP probe for bigger speeds, assuming cwnd can be
932 doubled every other RTT.
936 tcp_pacing_ca_ratio - INTEGER
937 sk->sk_pacing_rate is set by TCP stack using a ratio applied
938 to current rate. (current_rate = cwnd * mss / srtt)
939 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
940 is applied to conservatively probe for bigger throughput.
944 tcp_tso_win_divisor - INTEGER
945 This allows control over what percentage of the congestion window
946 can be consumed by a single TSO frame.
947 The setting of this parameter is a choice between burstiness and
948 building larger TSO frames.
952 tcp_tw_reuse - INTEGER
953 Enable reuse of TIME-WAIT sockets for new connections when it is
954 safe from protocol viewpoint.
958 - 2 - enable for loopback traffic only
960 It should not be changed without advice/request of technical
965 tcp_window_scaling - BOOLEAN
966 Enable window scaling as defined in RFC1323.
968 tcp_wmem - vector of 3 INTEGERs: min, default, max
969 min: Amount of memory reserved for send buffers for TCP sockets.
970 Each TCP socket has rights to use it due to fact of its birth.
974 default: initial size of send buffer used by TCP sockets. This
975 value overrides net.core.wmem_default used by other protocols.
977 It is usually lower than net.core.wmem_default.
981 max: Maximal amount of memory allowed for automatically tuned
982 send buffers for TCP sockets. This value does not override
983 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
984 automatic tuning of that socket's send buffer size, in which case
985 this value is ignored.
987 Default: between 64K and 4MB, depending on RAM size.
989 tcp_notsent_lowat - UNSIGNED INTEGER
990 A TCP socket can control the amount of unsent bytes in its write queue,
991 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
992 reports POLLOUT events if the amount of unsent bytes is below a per
993 socket value, and if the write queue is not full. sendmsg() will
994 also not add new buffers if the limit is hit.
996 This global variable controls the amount of unsent data for
997 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
998 to the global variable has immediate effect.
1000 Default: UINT_MAX (0xFFFFFFFF)
1002 tcp_workaround_signed_windows - BOOLEAN
1003 If set, assume no receipt of a window scaling option means the
1004 remote TCP is broken and treats the window as a signed quantity.
1005 If unset, assume the remote TCP is not broken even if we do
1006 not receive a window scaling option from them.
1010 tcp_thin_linear_timeouts - BOOLEAN
1011 Enable dynamic triggering of linear timeouts for thin streams.
1012 If set, a check is performed upon retransmission by timeout to
1013 determine if the stream is thin (less than 4 packets in flight).
1014 As long as the stream is found to be thin, up to 6 linear
1015 timeouts may be performed before exponential backoff mode is
1016 initiated. This improves retransmission latency for
1017 non-aggressive thin streams, often found to be time-dependent.
1018 For more information on thin streams, see
1019 Documentation/networking/tcp-thin.rst
1023 tcp_limit_output_bytes - INTEGER
1024 Controls TCP Small Queue limit per tcp socket.
1025 TCP bulk sender tends to increase packets in flight until it
1026 gets losses notifications. With SNDBUF autotuning, this can
1027 result in a large amount of packets queued on the local machine
1028 (e.g.: qdiscs, CPU backlog, or device) hurting latency of other
1029 flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes
1030 limits the number of bytes on qdisc or device to reduce artificial
1031 RTT/cwnd and reduce bufferbloat.
1033 Default: 1048576 (16 * 65536)
1035 tcp_challenge_ack_limit - INTEGER
1036 Limits number of Challenge ACK sent per second, as recommended
1037 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
1038 Note that this per netns rate limit can allow some side channel
1039 attacks and probably should not be enabled.
1040 TCP stack implements per TCP socket limits anyway.
1041 Default: INT_MAX (unlimited)
1043 tcp_ehash_entries - INTEGER
1044 Show the number of hash buckets for TCP sockets in the current
1045 networking namespace.
1047 A negative value means the networking namespace does not own its
1048 hash buckets and shares the initial networking namespace's one.
1050 tcp_child_ehash_entries - INTEGER
1051 Control the number of hash buckets for TCP sockets in the child
1052 networking namespace, which must be set before clone() or unshare().
1054 If the value is not 0, the kernel uses a value rounded up to 2^n
1055 as the actual hash bucket size. 0 is a special value, meaning
1056 the child networking namespace will share the initial networking
1057 namespace's hash buckets.
1059 Note that the child will use the global one in case the kernel
1060 fails to allocate enough memory. In addition, the global hash
1061 buckets are spread over available NUMA nodes, but the allocation
1062 of the child hash table depends on the current process's NUMA
1063 policy, which could result in performance differences.
1065 Note also that the default value of tcp_max_tw_buckets and
1066 tcp_max_syn_backlog depend on the hash bucket size.
1068 Possible values: 0, 2^n (n: 0 - 24 (16Mi))
1072 tcp_plb_enabled - BOOLEAN
1073 If set and the underlying congestion control (e.g. DCTCP) supports
1074 and enables PLB feature, TCP PLB (Protective Load Balancing) is
1075 enabled. PLB is described in the following paper:
1076 https://doi.org/10.1145/3544216.3544226. Based on PLB parameters,
1077 upon sensing sustained congestion, TCP triggers a change in
1078 flow label field for outgoing IPv6 packets. A change in flow label
1079 field potentially changes the path of outgoing packets for switches
1080 that use ECMP/WCMP for routing.
1082 PLB changes socket txhash which results in a change in IPv6 Flow Label
1083 field, and currently no-op for IPv4 headers. It is possible
1084 to apply PLB for IPv4 with other network header fields (e.g. TCP
1085 or IPv4 options) or using encapsulation where outer header is used
1086 by switches to determine next hop. In either case, further host
1087 and switch side changes will be needed.
1089 When set, PLB assumes that congestion signal (e.g. ECN) is made
1090 available and used by congestion control module to estimate a
1091 congestion measure (e.g. ce_ratio). PLB needs a congestion measure to
1092 make repathing decisions.
1096 tcp_plb_idle_rehash_rounds - INTEGER
1097 Number of consecutive congested rounds (RTT) seen after which
1098 a rehash can be performed, given there are no packets in flight.
1099 This is referred to as M in PLB paper:
1100 https://doi.org/10.1145/3544216.3544226.
1102 Possible Values: 0 - 31
1106 tcp_plb_rehash_rounds - INTEGER
1107 Number of consecutive congested rounds (RTT) seen after which
1108 a forced rehash can be performed. Be careful when setting this
1109 parameter, as a small value increases the risk of retransmissions.
1110 This is referred to as N in PLB paper:
1111 https://doi.org/10.1145/3544216.3544226.
1113 Possible Values: 0 - 31
1117 tcp_plb_suspend_rto_sec - INTEGER
1118 Time, in seconds, to suspend PLB in event of an RTO. In order to avoid
1119 having PLB repath onto a connectivity "black hole", after an RTO a TCP
1120 connection suspends PLB repathing for a random duration between 1x and
1121 2x of this parameter. Randomness is added to avoid concurrent rehashing
1122 of multiple TCP connections. This should be set corresponding to the
1123 amount of time it takes to repair a failed link.
1125 Possible Values: 0 - 255
1129 tcp_plb_cong_thresh - INTEGER
1130 Fraction of packets marked with congestion over a round (RTT) to
1131 tag that round as congested. This is referred to as K in the PLB paper:
1132 https://doi.org/10.1145/3544216.3544226.
1134 The 0-1 fraction range is mapped to 0-256 range to avoid floating
1135 point operations. For example, 128 means that if at least 50% of
1136 the packets in a round were marked as congested then the round
1137 will be tagged as congested.
1139 Setting threshold to 0 means that PLB repaths every RTT regardless
1140 of congestion. This is not intended behavior for PLB and should be
1141 used only for experimentation purpose.
1143 Possible Values: 0 - 256
1150 udp_l3mdev_accept - BOOLEAN
1151 Enabling this option allows a "global" bound socket to work
1152 across L3 master domains (e.g., VRFs) with packets capable of
1153 being received regardless of the L3 domain in which they
1154 originated. Only valid when the kernel was compiled with
1155 CONFIG_NET_L3_MASTER_DEV.
1157 Default: 0 (disabled)
1159 udp_mem - vector of 3 INTEGERs: min, pressure, max
1160 Number of pages allowed for queueing by all UDP sockets.
1162 min: Number of pages allowed for queueing by all UDP sockets.
1164 pressure: This value was introduced to follow format of tcp_mem.
1166 max: This value was introduced to follow format of tcp_mem.
1168 Default is calculated at boot time from amount of available memory.
1170 udp_rmem_min - INTEGER
1171 Minimal size of receive buffer used by UDP sockets in moderation.
1172 Each UDP socket is able to use the size for receiving data, even if
1173 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
1177 udp_wmem_min - INTEGER
1178 UDP does not have tx memory accounting and this tunable has no effect.
1180 udp_hash_entries - INTEGER
1181 Show the number of hash buckets for UDP sockets in the current
1182 networking namespace.
1184 A negative value means the networking namespace does not own its
1185 hash buckets and shares the initial networking namespace's one.
1187 udp_child_ehash_entries - INTEGER
1188 Control the number of hash buckets for UDP sockets in the child
1189 networking namespace, which must be set before clone() or unshare().
1191 If the value is not 0, the kernel uses a value rounded up to 2^n
1192 as the actual hash bucket size. 0 is a special value, meaning
1193 the child networking namespace will share the initial networking
1194 namespace's hash buckets.
1196 Note that the child will use the global one in case the kernel
1197 fails to allocate enough memory. In addition, the global hash
1198 buckets are spread over available NUMA nodes, but the allocation
1199 of the child hash table depends on the current process's NUMA
1200 policy, which could result in performance differences.
1202 Possible values: 0, 2^n (n: 7 (128) - 16 (64K))
1210 raw_l3mdev_accept - BOOLEAN
1211 Enabling this option allows a "global" bound socket to work
1212 across L3 master domains (e.g., VRFs) with packets capable of
1213 being received regardless of the L3 domain in which they
1214 originated. Only valid when the kernel was compiled with
1215 CONFIG_NET_L3_MASTER_DEV.
1217 Default: 1 (enabled)
1222 cipso_cache_enable - BOOLEAN
1223 If set, enable additions to and lookups from the CIPSO label mapping
1224 cache. If unset, additions are ignored and lookups always result in a
1225 miss. However, regardless of the setting the cache is still
1226 invalidated when required when means you can safely toggle this on and
1227 off and the cache will always be "safe".
1231 cipso_cache_bucket_size - INTEGER
1232 The CIPSO label cache consists of a fixed size hash table with each
1233 hash bucket containing a number of cache entries. This variable limits
1234 the number of entries in each hash bucket; the larger the value is, the
1235 more CIPSO label mappings that can be cached. When the number of
1236 entries in a given hash bucket reaches this limit adding new entries
1237 causes the oldest entry in the bucket to be removed to make room.
1241 cipso_rbm_optfmt - BOOLEAN
1242 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
1243 the CIPSO draft specification (see Documentation/netlabel for details).
1244 This means that when set the CIPSO tag will be padded with empty
1245 categories in order to make the packet data 32-bit aligned.
1249 cipso_rbm_structvalid - BOOLEAN
1250 If set, do a very strict check of the CIPSO option when
1251 ip_options_compile() is called. If unset, relax the checks done during
1252 ip_options_compile(). Either way is "safe" as errors are caught else
1253 where in the CIPSO processing code but setting this to 0 (False) should
1254 result in less work (i.e. it should be faster) but could cause problems
1255 with other implementations that require strict checking.
1262 ip_local_port_range - 2 INTEGERS
1263 Defines the local port range that is used by TCP and UDP to
1264 choose the local port. The first number is the first, the
1265 second the last local port number.
1266 If possible, it is better these numbers have different parity
1267 (one even and one odd value).
1268 Must be greater than or equal to ip_unprivileged_port_start.
1269 The default values are 32768 and 60999 respectively.
1271 ip_local_reserved_ports - list of comma separated ranges
1272 Specify the ports which are reserved for known third-party
1273 applications. These ports will not be used by automatic port
1274 assignments (e.g. when calling connect() or bind() with port
1275 number 0). Explicit port allocation behavior is unchanged.
1277 The format used for both input and output is a comma separated
1278 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
1279 10). Writing to the file will clear all previously reserved
1280 ports and update the current list with the one given in the
1283 Note that ip_local_port_range and ip_local_reserved_ports
1284 settings are independent and both are considered by the kernel
1285 when determining which ports are available for automatic port
1288 You can reserve ports which are not in the current
1289 ip_local_port_range, e.g.::
1291 $ cat /proc/sys/net/ipv4/ip_local_port_range
1293 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
1296 although this is redundant. However such a setting is useful
1297 if later the port range is changed to a value that will
1298 include the reserved ports. Also keep in mind, that overlapping
1299 of these ranges may affect probability of selecting ephemeral
1300 ports which are right after block of reserved ports.
1304 ip_unprivileged_port_start - INTEGER
1305 This is a per-namespace sysctl. It defines the first
1306 unprivileged port in the network namespace. Privileged ports
1307 require root or CAP_NET_BIND_SERVICE in order to bind to them.
1308 To disable all privileged ports, set this to 0. They must not
1309 overlap with the ip_local_port_range.
1313 ip_nonlocal_bind - BOOLEAN
1314 If set, allows processes to bind() to non-local IP addresses,
1315 which can be quite useful - but may break some applications.
1319 ip_autobind_reuse - BOOLEAN
1320 By default, bind() does not select the ports automatically even if
1321 the new socket and all sockets bound to the port have SO_REUSEADDR.
1322 ip_autobind_reuse allows bind() to reuse the port and this is useful
1323 when you use bind()+connect(), but may break some applications.
1324 The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
1325 option should only be set by experts.
1328 ip_dynaddr - INTEGER
1329 If set non-zero, enables support for dynamic addresses.
1330 If set to a non-zero value larger than 1, a kernel log
1331 message will be printed when dynamic address rewriting
1336 ip_early_demux - BOOLEAN
1337 Optimize input packet processing down to one demux for
1338 certain kinds of local sockets. Currently we only do this
1339 for established TCP and connected UDP sockets.
1341 It may add an additional cost for pure routing workloads that
1342 reduces overall throughput, in such case you should disable it.
1346 ping_group_range - 2 INTEGERS
1347 Restrict ICMP_PROTO datagram sockets to users in the group range.
1348 The default is "1 0", meaning, that nobody (not even root) may
1349 create ping sockets. Setting it to "100 100" would grant permissions
1350 to the single group. "0 4294967295" would enable it for the world, "100
1351 4294967295" would enable it for the users, but not daemons.
1353 tcp_early_demux - BOOLEAN
1354 Enable early demux for established TCP sockets.
1358 udp_early_demux - BOOLEAN
1359 Enable early demux for connected UDP sockets. Disable this if
1360 your system could experience more unconnected load.
1364 icmp_echo_ignore_all - BOOLEAN
1365 If set non-zero, then the kernel will ignore all ICMP ECHO
1366 requests sent to it.
1370 icmp_echo_enable_probe - BOOLEAN
1371 If set to one, then the kernel will respond to RFC 8335 PROBE
1372 requests sent to it.
1376 icmp_echo_ignore_broadcasts - BOOLEAN
1377 If set non-zero, then the kernel will ignore all ICMP ECHO and
1378 TIMESTAMP requests sent to it via broadcast/multicast.
1382 icmp_ratelimit - INTEGER
1383 Limit the maximal rates for sending ICMP packets whose type matches
1384 icmp_ratemask (see below) to specific targets.
1385 0 to disable any limiting,
1386 otherwise the minimal space between responses in milliseconds.
1387 Note that another sysctl, icmp_msgs_per_sec limits the number
1388 of ICMP packets sent on all targets.
1392 icmp_msgs_per_sec - INTEGER
1393 Limit maximal number of ICMP packets sent per second from this host.
1394 Only messages whose type matches icmp_ratemask (see below) are
1395 controlled by this limit. For security reasons, the precise count
1396 of messages per second is randomized.
1400 icmp_msgs_burst - INTEGER
1401 icmp_msgs_per_sec controls number of ICMP packets sent per second,
1402 while icmp_msgs_burst controls the burst size of these packets.
1403 For security reasons, the precise burst size is randomized.
1407 icmp_ratemask - INTEGER
1408 Mask made of ICMP types for which rates are being limited.
1410 Significant bits: IHGFEDCBA9876543210
1412 Default mask: 0000001100000011000 (6168)
1414 Bit definitions (see include/linux/icmp.h):
1416 = =========================
1418 3 Destination Unreachable [1]_
1419 4 Source Quench [1]_
1422 B Time Exceeded [1]_
1423 C Parameter Problem [1]_
1428 H Address Mask Request
1429 I Address Mask Reply
1430 = =========================
1432 .. [1] These are rate limited by default (see default mask above)
1434 icmp_ignore_bogus_error_responses - BOOLEAN
1435 Some routers violate RFC1122 by sending bogus responses to broadcast
1436 frames. Such violations are normally logged via a kernel warning.
1437 If this is set to TRUE, the kernel will not give such warnings, which
1438 will avoid log file clutter.
1442 icmp_errors_use_inbound_ifaddr - BOOLEAN
1444 If zero, icmp error messages are sent with the primary address of
1445 the exiting interface.
1447 If non-zero, the message will be sent with the primary address of
1448 the interface that received the packet that caused the icmp error.
1449 This is the behaviour many network administrators will expect from
1450 a router. And it can make debugging complicated network layouts
1453 Note that if no primary address exists for the interface selected,
1454 then the primary address of the first non-loopback interface that
1455 has one will be used regardless of this setting.
1459 igmp_max_memberships - INTEGER
1460 Change the maximum number of multicast groups we can subscribe to.
1463 Theoretical maximum value is bounded by having to send a membership
1464 report in a single datagram (i.e. the report can't span multiple
1465 datagrams, or risk confusing the switch and leaving groups you don't
1468 The number of supported groups 'M' is bounded by the number of group
1469 report entries you can fit into a single datagram of 65535 bytes.
1471 M = 65536-sizeof (ip header)/(sizeof(Group record))
1473 Group records are variable length, with a minimum of 12 bytes.
1474 So net.ipv4.igmp_max_memberships should not be set higher than:
1476 (65536-24) / 12 = 5459
1478 The value 5459 assumes no IP header options, so in practice
1479 this number may be lower.
1481 igmp_max_msf - INTEGER
1482 Maximum number of addresses allowed in the source filter list for a
1488 Controls the IGMP query robustness variable (see RFC2236 8.1).
1490 Default: 2 (as specified by RFC2236 8.1)
1492 Minimum: 1 (as specified by RFC6636 4.5)
1494 force_igmp_version - INTEGER
1495 - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
1496 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
1497 Present timer expires.
1498 - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
1499 receive IGMPv2/v3 query.
1500 - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
1501 IGMPv1 query message. Will reply report if receive IGMPv3 query.
1502 - 3 - Enforce to use IGMP version 3. The same react with default 0.
1506 this is not the same with force_mld_version because IGMPv3 RFC3376
1507 Security Considerations does not have clear description that we could
1508 ignore other version messages completely as MLDv2 RFC3810. So make
1509 this value as default 0 is recommended.
1511 ``conf/interface/*``
1512 changes special settings per interface (where
1513 interface" is the name of your network interface)
1516 is special, changes the settings for all interfaces
1518 log_martians - BOOLEAN
1519 Log packets with impossible addresses to kernel log.
1520 log_martians for the interface will be enabled if at least one of
1521 conf/{all,interface}/log_martians is set to TRUE,
1522 it will be disabled otherwise
1524 accept_redirects - BOOLEAN
1525 Accept ICMP redirect messages.
1526 accept_redirects for the interface will be enabled if:
1528 - both conf/{all,interface}/accept_redirects are TRUE in the case
1529 forwarding for the interface is enabled
1533 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
1534 case forwarding for the interface is disabled
1536 accept_redirects for the interface will be disabled otherwise
1543 forwarding - BOOLEAN
1544 Enable IP forwarding on this interface. This controls whether packets
1545 received _on_ this interface can be forwarded.
1547 mc_forwarding - BOOLEAN
1548 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1549 and a multicast routing daemon is required.
1550 conf/all/mc_forwarding must also be set to TRUE to enable multicast
1551 routing for the interface
1554 Integer value used to differentiate the devices by the medium they
1555 are attached to. Two devices can have different id values when
1556 the broadcast packets are received only on one of them.
1557 The default value 0 means that the device is the only interface
1558 to its medium, value of -1 means that medium is not known.
1560 Currently, it is used to change the proxy_arp behavior:
1561 the proxy_arp feature is enabled for packets forwarded between
1562 two devices attached to different media.
1567 proxy_arp for the interface will be enabled if at least one of
1568 conf/{all,interface}/proxy_arp is set to TRUE,
1569 it will be disabled otherwise
1571 proxy_arp_pvlan - BOOLEAN
1572 Private VLAN proxy arp.
1574 Basically allow proxy arp replies back to the same interface
1575 (from which the ARP request/solicitation was received).
1577 This is done to support (ethernet) switch features, like RFC
1578 3069, where the individual ports are NOT allowed to
1579 communicate with each other, but they are allowed to talk to
1580 the upstream router. As described in RFC 3069, it is possible
1581 to allow these hosts to communicate through the upstream
1582 router by proxy_arp'ing. Don't need to be used together with
1585 This technology is known by different names:
1587 In RFC 3069 it is called VLAN Aggregation.
1588 Cisco and Allied Telesyn call it Private VLAN.
1589 Hewlett-Packard call it Source-Port filtering or port-isolation.
1590 Ericsson call it MAC-Forced Forwarding (RFC Draft).
1592 shared_media - BOOLEAN
1593 Send(router) or accept(host) RFC1620 shared media redirects.
1594 Overrides secure_redirects.
1596 shared_media for the interface will be enabled if at least one of
1597 conf/{all,interface}/shared_media is set to TRUE,
1598 it will be disabled otherwise
1602 secure_redirects - BOOLEAN
1603 Accept ICMP redirect messages only to gateways listed in the
1604 interface's current gateway list. Even if disabled, RFC1122 redirect
1607 Overridden by shared_media.
1609 secure_redirects for the interface will be enabled if at least one of
1610 conf/{all,interface}/secure_redirects is set to TRUE,
1611 it will be disabled otherwise
1615 send_redirects - BOOLEAN
1616 Send redirects, if router.
1618 send_redirects for the interface will be enabled if at least one of
1619 conf/{all,interface}/send_redirects is set to TRUE,
1620 it will be disabled otherwise
1624 bootp_relay - BOOLEAN
1625 Accept packets with source address 0.b.c.d destined
1626 not to this host as local ones. It is supposed, that
1627 BOOTP relay daemon will catch and forward such packets.
1628 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1633 Not Implemented Yet.
1635 accept_source_route - BOOLEAN
1636 Accept packets with SRR option.
1637 conf/all/accept_source_route must also be set to TRUE to accept packets
1638 with SRR option on the interface
1645 accept_local - BOOLEAN
1646 Accept packets with local source addresses. In combination with
1647 suitable routing, this can be used to direct packets between two
1648 local interfaces over the wire and have them accepted properly.
1651 route_localnet - BOOLEAN
1652 Do not consider loopback addresses as martian source or destination
1653 while routing. This enables the use of 127/8 for local routing purposes.
1658 - 0 - No source validation.
1659 - 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1660 Each incoming packet is tested against the FIB and if the interface
1661 is not the best reverse path the packet check will fail.
1662 By default failed packets are discarded.
1663 - 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1664 Each incoming packet's source address is also tested against the FIB
1665 and if the source address is not reachable via any interface
1666 the packet check will fail.
1668 Current recommended practice in RFC3704 is to enable strict mode
1669 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1670 or other complicated routing, then loose mode is recommended.
1672 The max value from conf/{all,interface}/rp_filter is used
1673 when doing source validation on the {interface}.
1675 Default value is 0. Note that some distributions enable it
1678 src_valid_mark - BOOLEAN
1679 - 0 - The fwmark of the packet is not included in reverse path
1680 route lookup. This allows for asymmetric routing configurations
1681 utilizing the fwmark in only one direction, e.g., transparent
1684 - 1 - The fwmark of the packet is included in reverse path route
1685 lookup. This permits rp_filter to function when the fwmark is
1686 used for routing traffic in both directions.
1688 This setting also affects the utilization of fmwark when
1689 performing source address selection for ICMP replies, or
1690 determining addresses stored for the IPOPT_TS_TSANDADDR and
1691 IPOPT_RR IP options.
1693 The max value from conf/{all,interface}/src_valid_mark is used.
1697 arp_filter - BOOLEAN
1698 - 1 - Allows you to have multiple network interfaces on the same
1699 subnet, and have the ARPs for each interface be answered
1700 based on whether or not the kernel would route a packet from
1701 the ARP'd IP out that interface (therefore you must use source
1702 based routing for this to work). In other words it allows control
1703 of which cards (usually 1) will respond to an arp request.
1705 - 0 - (default) The kernel can respond to arp requests with addresses
1706 from other interfaces. This may seem wrong but it usually makes
1707 sense, because it increases the chance of successful communication.
1708 IP addresses are owned by the complete host on Linux, not by
1709 particular interfaces. Only for more complex setups like load-
1710 balancing, does this behaviour cause problems.
1712 arp_filter for the interface will be enabled if at least one of
1713 conf/{all,interface}/arp_filter is set to TRUE,
1714 it will be disabled otherwise
1716 arp_announce - INTEGER
1717 Define different restriction levels for announcing the local
1718 source IP address from IP packets in ARP requests sent on
1721 - 0 - (default) Use any local address, configured on any interface
1722 - 1 - Try to avoid local addresses that are not in the target's
1723 subnet for this interface. This mode is useful when target
1724 hosts reachable via this interface require the source IP
1725 address in ARP requests to be part of their logical network
1726 configured on the receiving interface. When we generate the
1727 request we will check all our subnets that include the
1728 target IP and will preserve the source address if it is from
1729 such subnet. If there is no such subnet we select source
1730 address according to the rules for level 2.
1731 - 2 - Always use the best local address for this target.
1732 In this mode we ignore the source address in the IP packet
1733 and try to select local address that we prefer for talks with
1734 the target host. Such local address is selected by looking
1735 for primary IP addresses on all our subnets on the outgoing
1736 interface that include the target IP address. If no suitable
1737 local address is found we select the first local address
1738 we have on the outgoing interface or on all other interfaces,
1739 with the hope we will receive reply for our request and
1740 even sometimes no matter the source IP address we announce.
1742 The max value from conf/{all,interface}/arp_announce is used.
1744 Increasing the restriction level gives more chance for
1745 receiving answer from the resolved target while decreasing
1746 the level announces more valid sender's information.
1748 arp_ignore - INTEGER
1749 Define different modes for sending replies in response to
1750 received ARP requests that resolve local target IP addresses:
1752 - 0 - (default): reply for any local target IP address, configured
1754 - 1 - reply only if the target IP address is local address
1755 configured on the incoming interface
1756 - 2 - reply only if the target IP address is local address
1757 configured on the incoming interface and both with the
1758 sender's IP address are part from same subnet on this interface
1759 - 3 - do not reply for local addresses configured with scope host,
1760 only resolutions for global and link addresses are replied
1762 - 8 - do not reply for all local addresses
1764 The max value from conf/{all,interface}/arp_ignore is used
1765 when ARP request is received on the {interface}
1767 arp_notify - BOOLEAN
1768 Define mode for notification of address and device changes.
1770 == ==========================================================
1771 0 (default): do nothing
1772 1 Generate gratuitous arp requests when device is brought up
1773 or hardware address changes.
1774 == ==========================================================
1776 arp_accept - INTEGER
1777 Define behavior for accepting gratuitous ARP (garp) frames from devices
1778 that are not already present in the ARP table:
1780 - 0 - don't create new entries in the ARP table
1781 - 1 - create new entries in the ARP table
1782 - 2 - create new entries only if the source IP address is in the same
1783 subnet as an address configured on the interface that received the
1786 Both replies and requests type gratuitous arp will trigger the
1787 ARP table to be updated, if this setting is on.
1789 If the ARP table already contains the IP address of the
1790 gratuitous arp frame, the arp table will be updated regardless
1791 if this setting is on or off.
1793 arp_evict_nocarrier - BOOLEAN
1794 Clears the ARP cache on NOCARRIER events. This option is important for
1795 wireless devices where the ARP cache should not be cleared when roaming
1796 between access points on the same network. In most cases this should
1797 remain as the default (1).
1799 - 1 - (default): Clear the ARP cache on NOCARRIER events
1800 - 0 - Do not clear ARP cache on NOCARRIER events
1802 mcast_solicit - INTEGER
1803 The maximum number of multicast probes in INCOMPLETE state,
1804 when the associated hardware address is unknown. Defaults
1807 ucast_solicit - INTEGER
1808 The maximum number of unicast probes in PROBE state, when
1809 the hardware address is being reconfirmed. Defaults to 3.
1811 app_solicit - INTEGER
1812 The maximum number of probes to send to the user space ARP daemon
1813 via netlink before dropping back to multicast probes (see
1814 mcast_resolicit). Defaults to 0.
1816 mcast_resolicit - INTEGER
1817 The maximum number of multicast probes after unicast and
1818 app probes in PROBE state. Defaults to 0.
1820 disable_policy - BOOLEAN
1821 Disable IPSEC policy (SPD) for this interface
1823 disable_xfrm - BOOLEAN
1824 Disable IPSEC encryption on this interface, whatever the policy
1826 igmpv2_unsolicited_report_interval - INTEGER
1827 The interval in milliseconds in which the next unsolicited
1828 IGMPv1 or IGMPv2 report retransmit will take place.
1830 Default: 10000 (10 seconds)
1832 igmpv3_unsolicited_report_interval - INTEGER
1833 The interval in milliseconds in which the next unsolicited
1834 IGMPv3 report retransmit will take place.
1836 Default: 1000 (1 seconds)
1838 ignore_routes_with_linkdown - BOOLEAN
1839 Ignore routes whose link is down when performing a FIB lookup.
1841 promote_secondaries - BOOLEAN
1842 When a primary IP address is removed from this interface
1843 promote a corresponding secondary IP address instead of
1844 removing all the corresponding secondary IP addresses.
1846 drop_unicast_in_l2_multicast - BOOLEAN
1847 Drop any unicast IP packets that are received in link-layer
1848 multicast (or broadcast) frames.
1850 This behavior (for multicast) is actually a SHOULD in RFC
1851 1122, but is disabled by default for compatibility reasons.
1855 drop_gratuitous_arp - BOOLEAN
1856 Drop all gratuitous ARP frames, for example if there's a known
1857 good ARP proxy on the network and such frames need not be used
1858 (or in the case of 802.11, must not be used to prevent attacks.)
1864 Allows you to write a number, which can be used as required.
1868 xfrm4_gc_thresh - INTEGER
1869 (Obsolete since linux-4.14)
1870 The threshold at which we will start garbage collecting for IPv4
1871 destination cache entries. At twice this value the system will
1872 refuse new allocations.
1874 igmp_link_local_mcast_reports - BOOLEAN
1875 Enable IGMP reports for link local multicast groups in the
1881 kuznet@ms2.inr.ac.ru
1888 delon.nicolas@wanadoo.fr
1893 /proc/sys/net/ipv6/* Variables
1894 ==============================
1896 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1897 apply to IPv6 [XXX?].
1899 bindv6only - BOOLEAN
1900 Default value for IPV6_V6ONLY socket option,
1901 which restricts use of the IPv6 socket to IPv6 communication
1904 - TRUE: disable IPv4-mapped address feature
1905 - FALSE: enable IPv4-mapped address feature
1907 Default: FALSE (as specified in RFC3493)
1909 flowlabel_consistency - BOOLEAN
1910 Protect the consistency (and unicity) of flow label.
1911 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1919 auto_flowlabels - INTEGER
1920 Automatically generate flow labels based on a flow hash of the
1921 packet. This allows intermediate devices, such as routers, to
1922 identify packet flows for mechanisms like Equal Cost Multipath
1923 Routing (see RFC 6438).
1925 = ===========================================================
1926 0 automatic flow labels are completely disabled
1927 1 automatic flow labels are enabled by default, they can be
1928 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1930 2 automatic flow labels are allowed, they may be enabled on a
1931 per socket basis using the IPV6_AUTOFLOWLABEL socket option
1932 3 automatic flow labels are enabled and enforced, they cannot
1933 be disabled by the socket option
1934 = ===========================================================
1938 flowlabel_state_ranges - BOOLEAN
1939 Split the flow label number space into two ranges. 0-0x7FFFF is
1940 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1941 is reserved for stateless flow labels as described in RFC6437.
1948 flowlabel_reflect - INTEGER
1949 Control flow label reflection. Needed for Path MTU
1950 Discovery to work with Equal Cost Multipath Routing in anycast
1951 environments. See RFC 7690 and:
1952 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
1956 - 1: enabled for established flows
1958 Note that this prevents automatic flowlabel changes, as done
1959 in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
1960 and "tcp: Change txhash on every SYN and RTO retransmit"
1962 - 2: enabled for TCP RESET packets (no active listener)
1963 If set, a RST packet sent in response to a SYN packet on a closed
1964 port will reflect the incoming flow label.
1966 - 4: enabled for ICMPv6 echo reply messages.
1970 fib_multipath_hash_policy - INTEGER
1971 Controls which hash policy to use for multipath routes.
1973 Default: 0 (Layer 3)
1977 - 0 - Layer 3 (source and destination addresses plus flow label)
1978 - 1 - Layer 4 (standard 5-tuple)
1979 - 2 - Layer 3 or inner Layer 3 if present
1980 - 3 - Custom multipath hash. Fields used for multipath hash calculation
1981 are determined by fib_multipath_hash_fields sysctl
1983 fib_multipath_hash_fields - UNSIGNED INTEGER
1984 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
1985 fields used for multipath hash calculation are determined by this
1988 This value is a bitmask which enables various fields for multipath hash
1991 Possible fields are:
1993 ====== ============================
1994 0x0001 Source IP address
1995 0x0002 Destination IP address
1999 0x0020 Destination port
2000 0x0040 Inner source IP address
2001 0x0080 Inner destination IP address
2002 0x0100 Inner IP protocol
2003 0x0200 Inner Flow Label
2004 0x0400 Inner source port
2005 0x0800 Inner destination port
2006 ====== ============================
2008 Default: 0x0007 (source IP, destination IP and IP protocol)
2010 anycast_src_echo_reply - BOOLEAN
2011 Controls the use of anycast addresses as source addresses for ICMPv6
2019 idgen_delay - INTEGER
2020 Controls the delay in seconds after which time to retry
2021 privacy stable address generation if a DAD conflict is
2024 Default: 1 (as specified in RFC7217)
2026 idgen_retries - INTEGER
2027 Controls the number of retries to generate a stable privacy
2028 address if a DAD conflict is detected.
2030 Default: 3 (as specified in RFC7217)
2033 Controls the MLD query robustness variable (see RFC3810 9.1).
2035 Default: 2 (as specified by RFC3810 9.1)
2037 Minimum: 1 (as specified by RFC6636 4.5)
2039 max_dst_opts_number - INTEGER
2040 Maximum number of non-padding TLVs allowed in a Destination
2041 options extension header. If this value is less than zero
2042 then unknown options are disallowed and the number of known
2043 TLVs allowed is the absolute value of this number.
2047 max_hbh_opts_number - INTEGER
2048 Maximum number of non-padding TLVs allowed in a Hop-by-Hop
2049 options extension header. If this value is less than zero
2050 then unknown options are disallowed and the number of known
2051 TLVs allowed is the absolute value of this number.
2055 max_dst_opts_length - INTEGER
2056 Maximum length allowed for a Destination options extension
2059 Default: INT_MAX (unlimited)
2061 max_hbh_length - INTEGER
2062 Maximum length allowed for a Hop-by-Hop options extension
2065 Default: INT_MAX (unlimited)
2067 skip_notify_on_dev_down - BOOLEAN
2068 Controls whether an RTM_DELROUTE message is generated for routes
2069 removed when a device is taken down or deleted. IPv4 does not
2070 generate this message; IPv6 does by default. Setting this sysctl
2071 to true skips the message, making IPv4 and IPv6 on par in relying
2072 on userspace caches to track link events and evict routes.
2074 Default: false (generate message)
2076 nexthop_compat_mode - BOOLEAN
2077 New nexthop API provides a means for managing nexthops independent of
2078 prefixes. Backwards compatibilty with old route format is enabled by
2079 default which means route dumps and notifications contain the new
2080 nexthop attribute but also the full, expanded nexthop definition.
2081 Further, updates or deletes of a nexthop configuration generate route
2082 notifications for each fib entry using the nexthop. Once a system
2083 understands the new API, this sysctl can be disabled to achieve full
2084 performance benefits of the new API by disabling the nexthop expansion
2085 and extraneous notifications.
2086 Default: true (backward compat mode)
2088 fib_notify_on_flag_change - INTEGER
2089 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
2090 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
2092 After installing a route to the kernel, user space receives an
2093 acknowledgment, which means the route was installed in the kernel,
2094 but not necessarily in hardware.
2095 It is also possible for a route already installed in hardware to change
2096 its action and therefore its flags. For example, a host route that is
2097 trapping packets can be "promoted" to perform decapsulation following
2098 the installation of an IPinIP/VXLAN tunnel.
2099 The notifications will indicate to user-space the state of the route.
2101 Default: 0 (Do not emit notifications.)
2105 - 0 - Do not emit notifications.
2106 - 1 - Emit notifications.
2107 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
2110 Define the IOAM id of this node. Uses only 24 bits out of 32 in total.
2117 ioam6_id_wide - LONG INTEGER
2118 Define the wide IOAM id of this node. Uses only 56 bits out of 64 in
2119 total. Can be different from ioam6_id.
2122 Max: 0xFFFFFFFFFFFFFF
2124 Default: 0xFFFFFFFFFFFFFF
2128 ip6frag_high_thresh - INTEGER
2129 Maximum memory used to reassemble IPv6 fragments. When
2130 ip6frag_high_thresh bytes of memory is allocated for this purpose,
2131 the fragment handler will toss packets until ip6frag_low_thresh
2134 ip6frag_low_thresh - INTEGER
2135 See ip6frag_high_thresh
2137 ip6frag_time - INTEGER
2138 Time in seconds to keep an IPv6 fragment in memory.
2141 Change the interface-specific default settings.
2143 These settings would be used during creating new interfaces.
2147 Change all the interface-specific settings.
2149 [XXX: Other special features than forwarding?]
2151 conf/all/disable_ipv6 - BOOLEAN
2152 Changing this value is same as changing ``conf/default/disable_ipv6``
2153 setting and also all per-interface ``disable_ipv6`` settings to the same
2156 Reading this value does not have any particular meaning. It does not say
2157 whether IPv6 support is enabled or disabled. Returned value can be 1
2158 also in the case when some interface has ``disable_ipv6`` set to 0 and
2159 has configured IPv6 addresses.
2161 conf/all/forwarding - BOOLEAN
2162 Enable global IPv6 forwarding between all interfaces.
2164 IPv4 and IPv6 work differently here; e.g. netfilter must be used
2165 to control which interfaces may forward packets and which not.
2167 This also sets all interfaces' Host/Router setting
2168 'forwarding' to the specified value. See below for details.
2170 This referred to as global forwarding.
2175 fwmark_reflect - BOOLEAN
2176 Controls the fwmark of kernel-generated IPv6 reply packets that are not
2177 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
2178 If unset, these packets have a fwmark of zero. If set, they have the
2179 fwmark of the packet they are replying to.
2183 ``conf/interface/*``:
2184 Change special settings per interface.
2186 The functional behaviour for certain settings is different
2187 depending on whether local forwarding is enabled or not.
2190 Accept Router Advertisements; autoconfigure using them.
2192 It also determines whether or not to transmit Router
2193 Solicitations. If and only if the functional setting is to
2194 accept Router Advertisements, Router Solicitations will be
2197 Possible values are:
2199 == ===========================================================
2200 0 Do not accept Router Advertisements.
2201 1 Accept Router Advertisements if forwarding is disabled.
2202 2 Overrule forwarding behaviour. Accept Router Advertisements
2203 even if forwarding is enabled.
2204 == ===========================================================
2208 - enabled if local forwarding is disabled.
2209 - disabled if local forwarding is enabled.
2211 accept_ra_defrtr - BOOLEAN
2212 Learn default router in Router Advertisement.
2216 - enabled if accept_ra is enabled.
2217 - disabled if accept_ra is disabled.
2219 ra_defrtr_metric - UNSIGNED INTEGER
2220 Route metric for default route learned in Router Advertisement. This value
2221 will be assigned as metric for the default route learned via IPv6 Router
2222 Advertisement. Takes affect only if accept_ra_defrtr is enabled.
2227 Default: IP6_RT_PRIO_USER i.e. 1024.
2229 accept_ra_from_local - BOOLEAN
2230 Accept RA with source-address that is found on local machine
2231 if the RA is otherwise proper and able to be accepted.
2233 Default is to NOT accept these as it may be an un-intended
2238 - enabled if accept_ra_from_local is enabled
2239 on a specific interface.
2240 - disabled if accept_ra_from_local is disabled
2241 on a specific interface.
2243 accept_ra_min_hop_limit - INTEGER
2244 Minimum hop limit Information in Router Advertisement.
2246 Hop limit Information in Router Advertisement less than this
2247 variable shall be ignored.
2251 accept_ra_pinfo - BOOLEAN
2252 Learn Prefix Information in Router Advertisement.
2256 - enabled if accept_ra is enabled.
2257 - disabled if accept_ra is disabled.
2259 accept_ra_rt_info_min_plen - INTEGER
2260 Minimum prefix length of Route Information in RA.
2262 Route Information w/ prefix smaller than this variable shall
2267 * 0 if accept_ra_rtr_pref is enabled.
2268 * -1 if accept_ra_rtr_pref is disabled.
2270 accept_ra_rt_info_max_plen - INTEGER
2271 Maximum prefix length of Route Information in RA.
2273 Route Information w/ prefix larger than this variable shall
2278 * 0 if accept_ra_rtr_pref is enabled.
2279 * -1 if accept_ra_rtr_pref is disabled.
2281 accept_ra_rtr_pref - BOOLEAN
2282 Accept Router Preference in RA.
2286 - enabled if accept_ra is enabled.
2287 - disabled if accept_ra is disabled.
2289 accept_ra_mtu - BOOLEAN
2290 Apply the MTU value specified in RA option 5 (RFC4861). If
2291 disabled, the MTU specified in the RA will be ignored.
2295 - enabled if accept_ra is enabled.
2296 - disabled if accept_ra is disabled.
2298 accept_redirects - BOOLEAN
2303 - enabled if local forwarding is disabled.
2304 - disabled if local forwarding is enabled.
2306 accept_source_route - INTEGER
2307 Accept source routing (routing extension header).
2309 - >= 0: Accept only routing header type 2.
2310 - < 0: Do not accept routing header.
2315 Autoconfigure addresses using Prefix Information in Router
2320 - enabled if accept_ra_pinfo is enabled.
2321 - disabled if accept_ra_pinfo is disabled.
2323 dad_transmits - INTEGER
2324 The amount of Duplicate Address Detection probes to send.
2328 forwarding - INTEGER
2329 Configure interface-specific Host/Router behaviour.
2333 It is recommended to have the same setting on all
2334 interfaces; mixed router/host scenarios are rather uncommon.
2336 Possible values are:
2338 - 0 Forwarding disabled
2339 - 1 Forwarding enabled
2343 By default, Host behaviour is assumed. This means:
2345 1. IsRouter flag is not set in Neighbour Advertisements.
2346 2. If accept_ra is TRUE (default), transmit Router
2348 3. If accept_ra is TRUE (default), accept Router
2349 Advertisements (and do autoconfiguration).
2350 4. If accept_redirects is TRUE (default), accept Redirects.
2354 If local forwarding is enabled, Router behaviour is assumed.
2355 This means exactly the reverse from the above:
2357 1. IsRouter flag is set in Neighbour Advertisements.
2358 2. Router Solicitations are not sent unless accept_ra is 2.
2359 3. Router Advertisements are ignored unless accept_ra is 2.
2360 4. Redirects are ignored.
2362 Default: 0 (disabled) if global forwarding is disabled (default),
2363 otherwise 1 (enabled).
2366 Default Hop Limit to set.
2371 Default Maximum Transfer Unit
2373 Default: 1280 (IPv6 required minimum)
2375 ip_nonlocal_bind - BOOLEAN
2376 If set, allows processes to bind() to non-local IPv6 addresses,
2377 which can be quite useful - but may break some applications.
2381 router_probe_interval - INTEGER
2382 Minimum interval (in seconds) between Router Probing described
2387 router_solicitation_delay - INTEGER
2388 Number of seconds to wait after interface is brought up
2389 before sending Router Solicitations.
2393 router_solicitation_interval - INTEGER
2394 Number of seconds to wait between Router Solicitations.
2398 router_solicitations - INTEGER
2399 Number of Router Solicitations to send until assuming no
2400 routers are present.
2404 use_oif_addrs_only - BOOLEAN
2405 When enabled, the candidate source addresses for destinations
2406 routed via this interface are restricted to the set of addresses
2407 configured on this interface (vis. RFC 6724, section 4).
2411 use_tempaddr - INTEGER
2412 Preference for Privacy Extensions (RFC3041).
2414 * <= 0 : disable Privacy Extensions
2415 * == 1 : enable Privacy Extensions, but prefer public
2416 addresses over temporary addresses.
2417 * > 1 : enable Privacy Extensions and prefer temporary
2418 addresses over public addresses.
2422 * 0 (for most devices)
2423 * -1 (for point-to-point devices and loopback devices)
2425 temp_valid_lft - INTEGER
2426 valid lifetime (in seconds) for temporary addresses.
2428 Default: 172800 (2 days)
2430 temp_prefered_lft - INTEGER
2431 Preferred lifetime (in seconds) for temporary addresses.
2433 Default: 86400 (1 day)
2435 keep_addr_on_down - INTEGER
2436 Keep all IPv6 addresses on an interface down event. If set static
2437 global addresses with no expiration time are not flushed.
2440 * 0 : system default
2443 Default: 0 (addresses are removed)
2445 max_desync_factor - INTEGER
2446 Maximum value for DESYNC_FACTOR, which is a random value
2447 that ensures that clients don't synchronize with each
2448 other and generate new addresses at exactly the same time.
2449 value is in seconds.
2453 regen_max_retry - INTEGER
2454 Number of attempts before give up attempting to generate
2455 valid temporary addresses.
2459 max_addresses - INTEGER
2460 Maximum number of autoconfigured addresses per interface. Setting
2461 to zero disables the limitation. It is not recommended to set this
2462 value too large (or to zero) because it would be an easy way to
2463 crash the kernel by allowing too many addresses to be created.
2467 disable_ipv6 - BOOLEAN
2468 Disable IPv6 operation. If accept_dad is set to 2, this value
2469 will be dynamically set to TRUE if DAD fails for the link-local
2472 Default: FALSE (enable IPv6 operation)
2474 When this value is changed from 1 to 0 (IPv6 is being enabled),
2475 it will dynamically create a link-local address on the given
2476 interface and start Duplicate Address Detection, if necessary.
2478 When this value is changed from 0 to 1 (IPv6 is being disabled),
2479 it will dynamically delete all addresses and routes on the given
2480 interface. From now on it will not possible to add addresses/routes
2481 to the selected interface.
2483 accept_dad - INTEGER
2484 Whether to accept DAD (Duplicate Address Detection).
2486 == ==============================================================
2488 1 Enable DAD (default)
2489 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate
2490 link-local address has been found.
2491 == ==============================================================
2493 DAD operation and mode on a given interface will be selected according
2494 to the maximum value of conf/{all,interface}/accept_dad.
2496 force_tllao - BOOLEAN
2497 Enable sending the target link-layer address option even when
2498 responding to a unicast neighbor solicitation.
2502 Quoting from RFC 2461, section 4.4, Target link-layer address:
2504 "The option MUST be included for multicast solicitations in order to
2505 avoid infinite Neighbor Solicitation "recursion" when the peer node
2506 does not have a cache entry to return a Neighbor Advertisements
2507 message. When responding to unicast solicitations, the option can be
2508 omitted since the sender of the solicitation has the correct link-
2509 layer address; otherwise it would not have be able to send the unicast
2510 solicitation in the first place. However, including the link-layer
2511 address in this case adds little overhead and eliminates a potential
2512 race condition where the sender deletes the cached link-layer address
2513 prior to receiving a response to a previous solicitation."
2515 ndisc_notify - BOOLEAN
2516 Define mode for notification of address and device changes.
2518 * 0 - (default): do nothing
2519 * 1 - Generate unsolicited neighbour advertisements when device is brought
2520 up or hardware address changes.
2522 ndisc_tclass - INTEGER
2523 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
2524 Discovery (Router Solicitation, Router Advertisement, Neighbor
2525 Solicitation, Neighbor Advertisement, Redirect) messages.
2526 These 8 bits can be interpreted as 6 high order bits holding the DSCP
2527 value and 2 low order bits representing ECN (which you probably want
2532 ndisc_evict_nocarrier - BOOLEAN
2533 Clears the neighbor discovery table on NOCARRIER events. This option is
2534 important for wireless devices where the neighbor discovery cache should
2535 not be cleared when roaming between access points on the same network.
2536 In most cases this should remain as the default (1).
2538 - 1 - (default): Clear neighbor discover cache on NOCARRIER events.
2539 - 0 - Do not clear neighbor discovery cache on NOCARRIER events.
2541 mldv1_unsolicited_report_interval - INTEGER
2542 The interval in milliseconds in which the next unsolicited
2543 MLDv1 report retransmit will take place.
2545 Default: 10000 (10 seconds)
2547 mldv2_unsolicited_report_interval - INTEGER
2548 The interval in milliseconds in which the next unsolicited
2549 MLDv2 report retransmit will take place.
2551 Default: 1000 (1 second)
2553 force_mld_version - INTEGER
2554 * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
2555 * 1 - Enforce to use MLD version 1
2556 * 2 - Enforce to use MLD version 2
2558 suppress_frag_ndisc - INTEGER
2559 Control RFC 6980 (Security Implications of IPv6 Fragmentation
2560 with IPv6 Neighbor Discovery) behavior:
2562 * 1 - (default) discard fragmented neighbor discovery packets
2563 * 0 - allow fragmented neighbor discovery packets
2565 optimistic_dad - BOOLEAN
2566 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
2568 * 0: disabled (default)
2571 Optimistic Duplicate Address Detection for the interface will be enabled
2572 if at least one of conf/{all,interface}/optimistic_dad is set to 1,
2573 it will be disabled otherwise.
2575 use_optimistic - BOOLEAN
2576 If enabled, do not classify optimistic addresses as deprecated during
2577 source address selection. Preferred addresses will still be chosen
2578 before optimistic addresses, subject to other ranking in the source
2579 address selection algorithm.
2581 * 0: disabled (default)
2584 This will be enabled if at least one of
2585 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
2587 stable_secret - IPv6 address
2588 This IPv6 address will be used as a secret to generate IPv6
2589 addresses for link-local addresses and autoconfigured
2590 ones. All addresses generated after setting this secret will
2591 be stable privacy ones by default. This can be changed via the
2592 addrgenmode ip-link. conf/default/stable_secret is used as the
2593 secret for the namespace, the interface specific ones can
2594 overwrite that. Writes to conf/all/stable_secret are refused.
2596 It is recommended to generate this secret during installation
2597 of a system and keep it stable after that.
2599 By default the stable secret is unset.
2601 addr_gen_mode - INTEGER
2602 Defines how link-local and autoconf addresses are generated.
2604 = =================================================================
2605 0 generate address based on EUI64 (default)
2606 1 do no generate a link-local address, use EUI64 for addresses
2607 generated from autoconf
2608 2 generate stable privacy addresses, using the secret from
2609 stable_secret (RFC7217)
2610 3 generate stable privacy addresses, using a random secret if unset
2611 = =================================================================
2613 drop_unicast_in_l2_multicast - BOOLEAN
2614 Drop any unicast IPv6 packets that are received in link-layer
2615 multicast (or broadcast) frames.
2617 By default this is turned off.
2619 drop_unsolicited_na - BOOLEAN
2620 Drop all unsolicited neighbor advertisements, for example if there's
2621 a known good NA proxy on the network and such frames need not be used
2622 (or in the case of 802.11, must not be used to prevent attacks.)
2624 By default this is turned off.
2626 accept_untracked_na - INTEGER
2627 Define behavior for accepting neighbor advertisements from devices that
2628 are absent in the neighbor cache:
2630 - 0 - (default) Do not accept unsolicited and untracked neighbor
2633 - 1 - Add a new neighbor cache entry in STALE state for routers on
2634 receiving a neighbor advertisement (either solicited or unsolicited)
2635 with target link-layer address option specified if no neighbor entry
2636 is already present for the advertised IPv6 address. Without this knob,
2637 NAs received for untracked addresses (absent in neighbor cache) are
2640 This is as per router-side behavior documented in RFC9131.
2642 This has lower precedence than drop_unsolicited_na.
2644 This will optimize the return path for the initial off-link
2645 communication that is initiated by a directly connected host, by
2646 ensuring that the first-hop router which turns on this setting doesn't
2647 have to buffer the initial return packets to do neighbor-solicitation.
2648 The prerequisite is that the host is configured to send unsolicited
2649 neighbor advertisements on interface bringup. This setting should be
2650 used in conjunction with the ndisc_notify setting on the host to
2651 satisfy this prerequisite.
2653 - 2 - Extend option (1) to add a new neighbor cache entry only if the
2654 source IP address is in the same subnet as an address configured on
2655 the interface that received the neighbor advertisement.
2657 enhanced_dad - BOOLEAN
2658 Include a nonce option in the IPv6 neighbor solicitation messages used for
2659 duplicate address detection per RFC7527. A received DAD NS will only signal
2660 a duplicate address if the nonce is different. This avoids any false
2661 detection of duplicates due to loopback of the NS messages that we send.
2662 The nonce option will be sent on an interface unless both of
2663 conf/{all,interface}/enhanced_dad are set to FALSE.
2671 Limit the maximal rates for sending ICMPv6 messages.
2673 0 to disable any limiting,
2674 otherwise the minimal space between responses in milliseconds.
2678 ratemask - list of comma separated ranges
2679 For ICMPv6 message types matching the ranges in the ratemask, limit
2680 the sending of the message according to ratelimit parameter.
2682 The format used for both input and output is a comma separated
2683 list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
2684 129). Writing to the file will clear all previous ranges of ICMPv6
2685 message types and update the current list with the input.
2687 Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
2688 for numerical values of ICMPv6 message types, e.g. echo request is 128
2689 and echo reply is 129.
2691 Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
2693 echo_ignore_all - BOOLEAN
2694 If set non-zero, then the kernel will ignore all ICMP ECHO
2695 requests sent to it over the IPv6 protocol.
2699 echo_ignore_multicast - BOOLEAN
2700 If set non-zero, then the kernel will ignore all ICMP ECHO
2701 requests sent to it over the IPv6 protocol via multicast.
2705 echo_ignore_anycast - BOOLEAN
2706 If set non-zero, then the kernel will ignore all ICMP ECHO
2707 requests sent to it over the IPv6 protocol destined to anycast address.
2711 xfrm6_gc_thresh - INTEGER
2712 (Obsolete since linux-4.14)
2713 The threshold at which we will start garbage collecting for IPv6
2714 destination cache entries. At twice this value the system will
2715 refuse new allocations.
2719 Pekka Savola <pekkas@netcore.fi>
2720 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
2723 /proc/sys/net/bridge/* Variables:
2724 =================================
2726 bridge-nf-call-arptables - BOOLEAN
2727 - 1 : pass bridged ARP traffic to arptables' FORWARD chain.
2732 bridge-nf-call-iptables - BOOLEAN
2733 - 1 : pass bridged IPv4 traffic to iptables' chains.
2738 bridge-nf-call-ip6tables - BOOLEAN
2739 - 1 : pass bridged IPv6 traffic to ip6tables' chains.
2744 bridge-nf-filter-vlan-tagged - BOOLEAN
2745 - 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
2750 bridge-nf-filter-pppoe-tagged - BOOLEAN
2751 - 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
2756 bridge-nf-pass-vlan-input-dev - BOOLEAN
2757 - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
2758 interface on the bridge and set the netfilter input device to the
2759 vlan. This allows use of e.g. "iptables -i br0.1" and makes the
2760 REDIRECT target work with vlan-on-top-of-bridge interfaces. When no
2761 matching vlan interface is found, or this switch is off, the input
2762 device is set to the bridge interface.
2764 - 0: disable bridge netfilter vlan interface lookup.
2768 ``proc/sys/net/sctp/*`` Variables:
2769 ==================================
2771 addip_enable - BOOLEAN
2772 Enable or disable extension of Dynamic Address Reconfiguration
2773 (ADD-IP) functionality specified in RFC5061. This extension provides
2774 the ability to dynamically add and remove new addresses for the SCTP
2777 1: Enable extension.
2779 0: Disable extension.
2784 Enable or disable pf (pf is short for potentially failed) state. A value
2785 of pf_retrans > path_max_retrans also disables pf state. That is, one of
2786 both pf_enable and pf_retrans > path_max_retrans can disable pf state.
2787 Since pf_retrans and path_max_retrans can be changed by userspace
2788 application, sometimes user expects to disable pf state by the value of
2789 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
2790 or path_max_retrans is changed by the user application, this pf state is
2791 enabled. As such, it is necessary to add this to dynamically enable
2792 and disable pf state. See:
2793 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
2803 Unset or enable/disable pf (pf is short for potentially failed) state
2804 exposure. Applications can control the exposure of the PF path state
2805 in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO
2806 sockopt. When it's unset, no SCTP_PEER_ADDR_CHANGE event with
2807 SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info
2808 can be got via SCTP_GET_PEER_ADDR_INFO sockopt; When it's enabled,
2809 a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming
2810 SCTP_PF state and a SCTP_PF-state transport info can be got via
2811 SCTP_GET_PEER_ADDR_INFO sockopt; When it's diabled, no
2812 SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when
2813 trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO
2816 0: Unset pf state exposure, Compatible with old applications.
2818 1: Disable pf state exposure.
2820 2: Enable pf state exposure.
2824 addip_noauth_enable - BOOLEAN
2825 Dynamic Address Reconfiguration (ADD-IP) requires the use of
2826 authentication to protect the operations of adding or removing new
2827 addresses. This requirement is mandated so that unauthorized hosts
2828 would not be able to hijack associations. However, older
2829 implementations may not have implemented this requirement while
2830 allowing the ADD-IP extension. For reasons of interoperability,
2831 we provide this variable to control the enforcement of the
2832 authentication requirement.
2834 == ===============================================================
2835 1 Allow ADD-IP extension to be used without authentication. This
2836 should only be set in a closed environment for interoperability
2837 with older implementations.
2839 0 Enforce the authentication requirement
2840 == ===============================================================
2844 auth_enable - BOOLEAN
2845 Enable or disable Authenticated Chunks extension. This extension
2846 provides the ability to send and receive authenticated chunks and is
2847 required for secure operation of Dynamic Address Reconfiguration
2850 - 1: Enable this extension.
2851 - 0: Disable this extension.
2855 prsctp_enable - BOOLEAN
2856 Enable or disable the Partial Reliability extension (RFC3758) which
2857 is used to notify peers that a given DATA should no longer be expected.
2859 - 1: Enable extension
2865 The limit of the number of new packets that can be initially sent. It
2866 controls how bursty the generated traffic can be.
2870 association_max_retrans - INTEGER
2871 Set the maximum number for retransmissions that an association can
2872 attempt deciding that the remote end is unreachable. If this value
2873 is exceeded, the association is terminated.
2877 max_init_retransmits - INTEGER
2878 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
2879 that an association will attempt before declaring the destination
2880 unreachable and terminating.
2884 path_max_retrans - INTEGER
2885 The maximum number of retransmissions that will be attempted on a given
2886 path. Once this threshold is exceeded, the path is considered
2887 unreachable, and new traffic will use a different path when the
2888 association is multihomed.
2892 pf_retrans - INTEGER
2893 The number of retransmissions that will be attempted on a given path
2894 before traffic is redirected to an alternate transport (should one
2895 exist). Note this is distinct from path_max_retrans, as a path that
2896 passes the pf_retrans threshold can still be used. Its only
2897 deprioritized when a transmission path is selected by the stack. This
2898 setting is primarily used to enable fast failover mechanisms without
2899 having to reduce path_max_retrans to a very low value. See:
2900 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
2901 for details. Note also that a value of pf_retrans > path_max_retrans
2902 disables this feature. Since both pf_retrans and path_max_retrans can
2903 be changed by userspace application, a variable pf_enable is used to
2908 ps_retrans - INTEGER
2909 Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
2910 from section-5 "Primary Path Switchover" in rfc7829. The primary path
2911 will be changed to another active path when the path error counter on
2912 the old primary path exceeds PSMR, so that "the SCTP sender is allowed
2913 to continue data transmission on a new working path even when the old
2914 primary destination address becomes active again". Note this feature
2915 is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
2916 and its value can't be less than 'pf_retrans' when changing by sysctl.
2920 rto_initial - INTEGER
2921 The initial round trip timeout value in milliseconds that will be used
2922 in calculating round trip times. This is the initial time interval
2923 for retransmissions.
2928 The maximum value (in milliseconds) of the round trip timeout. This
2929 is the largest time interval that can elapse between retransmissions.
2934 The minimum value (in milliseconds) of the round trip timeout. This
2935 is the smallest time interval the can elapse between retransmissions.
2939 hb_interval - INTEGER
2940 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
2941 are sent at the specified interval on idle paths to probe the state of
2942 a given path between 2 associations.
2946 sack_timeout - INTEGER
2947 The amount of time (in milliseconds) that the implementation will wait
2952 valid_cookie_life - INTEGER
2953 The default lifetime of the SCTP cookie (in milliseconds). The cookie
2954 is used during association establishment.
2958 cookie_preserve_enable - BOOLEAN
2959 Enable or disable the ability to extend the lifetime of the SCTP cookie
2960 that is used during the establishment phase of SCTP association
2962 - 1: Enable cookie lifetime extension.
2967 cookie_hmac_alg - STRING
2968 Select the hmac algorithm used when generating the cookie value sent by
2969 a listening sctp socket to a connecting client in the INIT-ACK chunk.
2976 Ability to assign md5 or sha1 as the selected alg is predicated on the
2977 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
2978 CONFIG_CRYPTO_SHA1).
2980 Default: Dependent on configuration. MD5 if available, else SHA1 if
2981 available, else none.
2983 rcvbuf_policy - INTEGER
2984 Determines if the receive buffer is attributed to the socket or to
2985 association. SCTP supports the capability to create multiple
2986 associations on a single socket. When using this capability, it is
2987 possible that a single stalled association that's buffering a lot
2988 of data may block other associations from delivering their data by
2989 consuming all of the receive buffer space. To work around this,
2990 the rcvbuf_policy could be set to attribute the receiver buffer space
2991 to each association instead of the socket. This prevents the described
2994 - 1: rcvbuf space is per association
2995 - 0: rcvbuf space is per socket
2999 sndbuf_policy - INTEGER
3000 Similar to rcvbuf_policy above, this applies to send buffer space.
3002 - 1: Send buffer is tracked per association
3003 - 0: Send buffer is tracked per socket.
3007 sctp_mem - vector of 3 INTEGERs: min, pressure, max
3008 Number of pages allowed for queueing by all SCTP sockets.
3010 min: Below this number of pages SCTP is not bothered about its
3011 memory appetite. When amount of memory allocated by SCTP exceeds
3012 this number, SCTP starts to moderate memory usage.
3014 pressure: This value was introduced to follow format of tcp_mem.
3016 max: Number of pages allowed for queueing by all SCTP sockets.
3018 Default is calculated at boot time from amount of available memory.
3020 sctp_rmem - vector of 3 INTEGERs: min, default, max
3021 Only the first value ("min") is used, "default" and "max" are
3024 min: Minimal size of receive buffer used by SCTP socket.
3025 It is guaranteed to each SCTP socket (but not association) even
3026 under moderate memory pressure.
3030 sctp_wmem - vector of 3 INTEGERs: min, default, max
3031 Only the first value ("min") is used, "default" and "max" are
3034 min: Minimum size of send buffer that can be used by SCTP sockets.
3035 It is guaranteed to each SCTP socket (but not association) even
3036 under moderate memory pressure.
3040 addr_scope_policy - INTEGER
3041 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
3043 - 0 - Disable IPv4 address scoping
3044 - 1 - Enable IPv4 address scoping
3045 - 2 - Follow draft but allow IPv4 private addresses
3046 - 3 - Follow draft but allow IPv4 link local addresses
3051 The listening port for the local UDP tunneling sock. Normally it's
3052 using the IANA-assigned UDP port number 9899 (sctp-tunneling).
3054 This UDP sock is used for processing the incoming UDP-encapsulated
3055 SCTP packets (from RFC6951), and shared by all applications in the
3056 same net namespace. This UDP sock will be closed when the value is
3059 The value will also be used to set the src port of the UDP header
3060 for the outgoing UDP-encapsulated SCTP packets. For the dest port,
3061 please refer to 'encap_port' below.
3065 encap_port - INTEGER
3066 The default remote UDP encapsulation port.
3068 This value is used to set the dest port of the UDP header for the
3069 outgoing UDP-encapsulated SCTP packets by default. Users can also
3070 change the value for each sock/asoc/transport by using setsockopt.
3071 For further information, please refer to RFC6951.
3073 Note that when connecting to a remote server, the client should set
3074 this to the port that the UDP tunneling sock on the peer server is
3075 listening to and the local UDP tunneling sock on the client also
3076 must be started. On the server, it would get the encap_port from
3077 the incoming packet's source port.
3081 plpmtud_probe_interval - INTEGER
3082 The time interval (in milliseconds) for the PLPMTUD probe timer,
3083 which is configured to expire after this period to receive an
3084 acknowledgment to a probe packet. This is also the time interval
3085 between the probes for the current pmtu when the probe search
3088 PLPMTUD will be disabled when 0 is set, and other values for it
3093 reconf_enable - BOOLEAN
3094 Enable or disable extension of Stream Reconfiguration functionality
3095 specified in RFC6525. This extension provides the ability to "reset"
3096 a stream, and it includes the Parameters of "Outgoing/Incoming SSN
3097 Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams".
3099 - 1: Enable extension.
3100 - 0: Disable extension.
3104 intl_enable - BOOLEAN
3105 Enable or disable extension of User Message Interleaving functionality
3106 specified in RFC8260. This extension allows the interleaving of user
3107 messages sent on different streams. With this feature enabled, I-DATA
3108 chunk will replace DATA chunk to carry user messages if also supported
3109 by the peer. Note that to use this feature, one needs to set this option
3110 to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2
3111 and SCTP_INTERLEAVING_SUPPORTED to 1.
3113 - 1: Enable extension.
3114 - 0: Disable extension.
3118 ecn_enable - BOOLEAN
3119 Control use of Explicit Congestion Notification (ECN) by SCTP.
3120 Like in TCP, ECN is used only when both ends of the SCTP connection
3121 indicate support for it. This feature is useful in avoiding losses
3122 due to congestion by allowing supporting routers to signal congestion
3123 before having to drop packets.
3130 l3mdev_accept - BOOLEAN
3131 Enabling this option allows a "global" bound socket to work
3132 across L3 master domains (e.g., VRFs) with packets capable of
3133 being received regardless of the L3 domain in which they
3134 originated. Only valid when the kernel was compiled with
3135 CONFIG_NET_L3_MASTER_DEV.
3137 Default: 1 (enabled)
3140 ``/proc/sys/net/core/*``
3141 ========================
3143 Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
3146 ``/proc/sys/net/unix/*``
3147 ========================
3149 max_dgram_qlen - INTEGER
3150 The maximum length of dgram socket receive queue