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 manually,
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 From linux kernel 6.3 onwards, this is deprecated for ipv6
160 as garbage collection manages cached route entries.
162 neigh/default/gc_thresh1 - INTEGER
163 Minimum number of entries to keep. Garbage collector will not
164 purge entries if there are fewer than this number.
168 neigh/default/gc_thresh2 - INTEGER
169 Threshold when garbage collector becomes more aggressive about
170 purging entries. Entries older than 5 seconds will be cleared
171 when over this number.
175 neigh/default/gc_thresh3 - INTEGER
176 Maximum number of non-PERMANENT neighbor entries allowed. Increase
177 this when using large numbers of interfaces and when communicating
178 with large numbers of directly-connected peers.
182 neigh/default/unres_qlen_bytes - INTEGER
183 The maximum number of bytes which may be used by packets
184 queued for each unresolved address by other network layers.
187 Setting negative value is meaningless and will return error.
189 Default: SK_WMEM_MAX, (same as net.core.wmem_default).
191 Exact value depends on architecture and kernel options,
192 but should be enough to allow queuing 256 packets
195 neigh/default/unres_qlen - INTEGER
196 The maximum number of packets which may be queued for each
197 unresolved address by other network layers.
199 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
201 Prior to linux 3.3, the default value is 3 which may cause
202 unexpected packet loss. The current default value is calculated
203 according to default value of unres_qlen_bytes and true size of
208 neigh/default/interval_probe_time_ms - INTEGER
209 The probe interval for neighbor entries with NTF_MANAGED flag,
214 mtu_expires - INTEGER
215 Time, in seconds, that cached PMTU information is kept.
217 min_adv_mss - INTEGER
218 The advertised MSS depends on the first hop route MTU, but will
219 never be lower than this setting.
221 fib_notify_on_flag_change - INTEGER
222 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
223 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
225 After installing a route to the kernel, user space receives an
226 acknowledgment, which means the route was installed in the kernel,
227 but not necessarily in hardware.
228 It is also possible for a route already installed in hardware to change
229 its action and therefore its flags. For example, a host route that is
230 trapping packets can be "promoted" to perform decapsulation following
231 the installation of an IPinIP/VXLAN tunnel.
232 The notifications will indicate to user-space the state of the route.
234 Default: 0 (Do not emit notifications.)
238 - 0 - Do not emit notifications.
239 - 1 - Emit notifications.
240 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
244 ipfrag_high_thresh - LONG INTEGER
245 Maximum memory used to reassemble IP fragments.
247 ipfrag_low_thresh - LONG INTEGER
248 (Obsolete since linux-4.17)
249 Maximum memory used to reassemble IP fragments before the kernel
250 begins to remove incomplete fragment queues to free up resources.
251 The kernel still accepts new fragments for defragmentation.
253 ipfrag_time - INTEGER
254 Time in seconds to keep an IP fragment in memory.
256 ipfrag_max_dist - INTEGER
257 ipfrag_max_dist is a non-negative integer value which defines the
258 maximum "disorder" which is allowed among fragments which share a
259 common IP source address. Note that reordering of packets is
260 not unusual, but if a large number of fragments arrive from a source
261 IP address while a particular fragment queue remains incomplete, it
262 probably indicates that one or more fragments belonging to that queue
263 have been lost. When ipfrag_max_dist is positive, an additional check
264 is done on fragments before they are added to a reassembly queue - if
265 ipfrag_max_dist (or more) fragments have arrived from a particular IP
266 address between additions to any IP fragment queue using that source
267 address, it's presumed that one or more fragments in the queue are
268 lost. The existing fragment queue will be dropped, and a new one
269 started. An ipfrag_max_dist value of zero disables this check.
271 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
272 result in unnecessarily dropping fragment queues when normal
273 reordering of packets occurs, which could lead to poor application
274 performance. Using a very large value, e.g. 50000, increases the
275 likelihood of incorrectly reassembling IP fragments that originate
276 from different IP datagrams, which could result in data corruption.
279 bc_forwarding - INTEGER
280 bc_forwarding enables the feature described in rfc1812#section-5.3.5.2
281 and rfc2644. It allows the router to forward directed broadcast.
282 To enable this feature, the 'all' entry and the input interface entry
289 inet_peer_threshold - INTEGER
290 The approximate size of the storage. Starting from this threshold
291 entries will be thrown aggressively. This threshold also determines
292 entries' time-to-live and time intervals between garbage collection
293 passes. More entries, less time-to-live, less GC interval.
295 inet_peer_minttl - INTEGER
296 Minimum time-to-live of entries. Should be enough to cover fragment
297 time-to-live on the reassembling side. This minimum time-to-live is
298 guaranteed if the pool size is less than inet_peer_threshold.
301 inet_peer_maxttl - INTEGER
302 Maximum time-to-live of entries. Unused entries will expire after
303 this period of time if there is no memory pressure on the pool (i.e.
304 when the number of entries in the pool is very small).
311 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
312 Defaults to 4096. (Was 128 before linux-5.4)
313 See also tcp_max_syn_backlog for additional tuning for TCP sockets.
315 tcp_abort_on_overflow - BOOLEAN
316 If listening service is too slow to accept new connections,
317 reset them. Default state is FALSE. It means that if overflow
318 occurred due to a burst, connection will recover. Enable this
319 option _only_ if you are really sure that listening daemon
320 cannot be tuned to accept connections faster. Enabling this
321 option can harm clients of your server.
323 tcp_adv_win_scale - INTEGER
324 Obsolete since linux-6.6
325 Count buffering overhead as bytes/2^tcp_adv_win_scale
326 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
329 Possible values are [-31, 31], inclusive.
333 tcp_allowed_congestion_control - STRING
334 Show/set the congestion control choices available to non-privileged
335 processes. The list is a subset of those listed in
336 tcp_available_congestion_control.
338 Default is "reno" and the default setting (tcp_congestion_control).
340 tcp_app_win - INTEGER
341 Reserve max(window/2^tcp_app_win, mss) of window for application
342 buffer. Value 0 is special, it means that nothing is reserved.
344 Possible values are [0, 31], inclusive.
348 tcp_autocorking - BOOLEAN
349 Enable TCP auto corking :
350 When applications do consecutive small write()/sendmsg() system calls,
351 we try to coalesce these small writes as much as possible, to lower
352 total amount of sent packets. This is done if at least one prior
353 packet for the flow is waiting in Qdisc queues or device transmit
354 queue. Applications can still use TCP_CORK for optimal behavior
355 when they know how/when to uncork their sockets.
359 tcp_available_congestion_control - STRING
360 Shows the available congestion control choices that are registered.
361 More congestion control algorithms may be available as modules,
364 tcp_base_mss - INTEGER
365 The initial value of search_low to be used by the packetization layer
366 Path MTU discovery (MTU probing). If MTU probing is enabled,
367 this is the initial MSS used by the connection.
369 tcp_mtu_probe_floor - INTEGER
370 If MTU probing is enabled this caps the minimum MSS used for search_low
375 tcp_min_snd_mss - INTEGER
376 TCP SYN and SYNACK messages usually advertise an ADVMSS option,
377 as described in RFC 1122 and RFC 6691.
379 If this ADVMSS option is smaller than tcp_min_snd_mss,
380 it is silently capped to tcp_min_snd_mss.
382 Default : 48 (at least 8 bytes of payload per segment)
384 tcp_congestion_control - STRING
385 Set the congestion control algorithm to be used for new
386 connections. The algorithm "reno" is always available, but
387 additional choices may be available based on kernel configuration.
388 Default is set as part of kernel configuration.
389 For passive connections, the listener congestion control choice
392 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
395 Allows TCP to send "duplicate" SACKs.
397 tcp_early_retrans - INTEGER
398 Tail loss probe (TLP) converts RTOs occurring due to tail
399 losses into fast recovery (draft-ietf-tcpm-rack). Note that
400 TLP requires RACK to function properly (see tcp_recovery below)
410 Control use of Explicit Congestion Notification (ECN) by TCP.
411 ECN is used only when both ends of the TCP connection indicate
412 support for it. This feature is useful in avoiding losses due
413 to congestion by allowing supporting routers to signal
414 congestion before having to drop packets.
418 = =====================================================
419 0 Disable ECN. Neither initiate nor accept ECN.
420 1 Enable ECN when requested by incoming connections and
421 also request ECN on outgoing connection attempts.
422 2 Enable ECN when requested by incoming connections
423 but do not request ECN on outgoing connections.
424 = =====================================================
428 tcp_ecn_fallback - BOOLEAN
429 If the kernel detects that ECN connection misbehaves, enable fall
430 back to non-ECN. Currently, this knob implements the fallback
431 from RFC3168, section 6.1.1.1., but we reserve that in future,
432 additional detection mechanisms could be implemented under this
433 knob. The value is not used, if tcp_ecn or per route (or congestion
434 control) ECN settings are disabled.
436 Default: 1 (fallback enabled)
439 This is a legacy option, it has no effect anymore.
441 tcp_fin_timeout - INTEGER
442 The length of time an orphaned (no longer referenced by any
443 application) connection will remain in the FIN_WAIT_2 state
444 before it is aborted at the local end. While a perfectly
445 valid "receive only" state for an un-orphaned connection, an
446 orphaned connection in FIN_WAIT_2 state could otherwise wait
447 forever for the remote to close its end of the connection.
454 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
455 F-RTO is an enhanced recovery algorithm for TCP retransmission
456 timeouts. It is particularly beneficial in networks where the
457 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
458 modification. It does not require any support from the peer.
460 By default it's enabled with a non-zero value. 0 disables F-RTO.
462 tcp_fwmark_accept - BOOLEAN
463 If set, incoming connections to listening sockets that do not have a
464 socket mark will set the mark of the accepting socket to the fwmark of
465 the incoming SYN packet. This will cause all packets on that connection
466 (starting from the first SYNACK) to be sent with that fwmark. The
467 listening socket's mark is unchanged. Listening sockets that already
468 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
473 tcp_invalid_ratelimit - INTEGER
474 Limit the maximal rate for sending duplicate acknowledgments
475 in response to incoming TCP packets that are for an existing
476 connection but that are invalid due to any of these reasons:
478 (a) out-of-window sequence number,
479 (b) out-of-window acknowledgment number, or
480 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
482 This can help mitigate simple "ack loop" DoS attacks, wherein
483 a buggy or malicious middlebox or man-in-the-middle can
484 rewrite TCP header fields in manner that causes each endpoint
485 to think that the other is sending invalid TCP segments, thus
486 causing each side to send an unterminating stream of duplicate
487 acknowledgments for invalid segments.
489 Using 0 disables rate-limiting of dupacks in response to
490 invalid segments; otherwise this value specifies the minimal
491 space between sending such dupacks, in milliseconds.
493 Default: 500 (milliseconds).
495 tcp_keepalive_time - INTEGER
496 How often TCP sends out keepalive messages when keepalive is enabled.
499 tcp_keepalive_probes - INTEGER
500 How many keepalive probes TCP sends out, until it decides that the
501 connection is broken. Default value: 9.
503 tcp_keepalive_intvl - INTEGER
504 How frequently the probes are send out. Multiplied by
505 tcp_keepalive_probes it is time to kill not responding connection,
506 after probes started. Default value: 75sec i.e. connection
507 will be aborted after ~11 minutes of retries.
509 tcp_l3mdev_accept - BOOLEAN
510 Enables child sockets to inherit the L3 master device index.
511 Enabling this option allows a "global" listen socket to work
512 across L3 master domains (e.g., VRFs) with connected sockets
513 derived from the listen socket to be bound to the L3 domain in
514 which the packets originated. Only valid when the kernel was
515 compiled with CONFIG_NET_L3_MASTER_DEV.
517 Default: 0 (disabled)
519 tcp_low_latency - BOOLEAN
520 This is a legacy option, it has no effect anymore.
522 tcp_max_orphans - INTEGER
523 Maximal number of TCP sockets not attached to any user file handle,
524 held by system. If this number is exceeded orphaned connections are
525 reset immediately and warning is printed. This limit exists
526 only to prevent simple DoS attacks, you _must_ not rely on this
527 or lower the limit artificially, but rather increase it
528 (probably, after increasing installed memory),
529 if network conditions require more than default value,
530 and tune network services to linger and kill such states
531 more aggressively. Let me to remind again: each orphan eats
532 up to ~64K of unswappable memory.
534 tcp_max_syn_backlog - INTEGER
535 Maximal number of remembered connection requests (SYN_RECV),
536 which have not received an acknowledgment from connecting client.
538 This is a per-listener limit.
540 The minimal value is 128 for low memory machines, and it will
541 increase in proportion to the memory of machine.
543 If server suffers from overload, try increasing this number.
545 Remember to also check /proc/sys/net/core/somaxconn
546 A SYN_RECV request socket consumes about 304 bytes of memory.
548 tcp_max_tw_buckets - INTEGER
549 Maximal number of timewait sockets held by system simultaneously.
550 If this number is exceeded time-wait socket is immediately destroyed
551 and warning is printed. This limit exists only to prevent
552 simple DoS attacks, you _must_ not lower the limit artificially,
553 but rather increase it (probably, after increasing installed memory),
554 if network conditions require more than default value.
556 tcp_mem - vector of 3 INTEGERs: min, pressure, max
557 min: below this number of pages TCP is not bothered about its
560 pressure: when amount of memory allocated by TCP exceeds this number
561 of pages, TCP moderates its memory consumption and enters memory
562 pressure mode, which is exited when memory consumption falls
565 max: number of pages allowed for queueing by all TCP sockets.
567 Defaults are calculated at boot time from amount of available
570 tcp_min_rtt_wlen - INTEGER
571 The window length of the windowed min filter to track the minimum RTT.
572 A shorter window lets a flow more quickly pick up new (higher)
573 minimum RTT when it is moved to a longer path (e.g., due to traffic
574 engineering). A longer window makes the filter more resistant to RTT
575 inflations such as transient congestion. The unit is seconds.
577 Possible values: 0 - 86400 (1 day)
581 tcp_moderate_rcvbuf - BOOLEAN
582 If set, TCP performs receive buffer auto-tuning, attempting to
583 automatically size the buffer (no greater than tcp_rmem[2]) to
584 match the size required by the path for full throughput. Enabled by
587 tcp_mtu_probing - INTEGER
588 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
592 - 1 - Disabled by default, enabled when an ICMP black hole detected
593 - 2 - Always enabled, use initial MSS of tcp_base_mss.
595 tcp_probe_interval - UNSIGNED INTEGER
596 Controls how often to start TCP Packetization-Layer Path MTU
597 Discovery reprobe. The default is reprobing every 10 minutes as
600 tcp_probe_threshold - INTEGER
601 Controls when TCP Packetization-Layer Path MTU Discovery probing
602 will stop in respect to the width of search range in bytes. Default
605 tcp_no_metrics_save - BOOLEAN
606 By default, TCP saves various connection metrics in the route cache
607 when the connection closes, so that connections established in the
608 near future can use these to set initial conditions. Usually, this
609 increases overall performance, but may sometimes cause performance
610 degradation. If set, TCP will not cache metrics on closing
613 tcp_no_ssthresh_metrics_save - BOOLEAN
614 Controls whether TCP saves ssthresh metrics in the route cache.
616 Default is 1, which disables ssthresh metrics.
618 tcp_orphan_retries - INTEGER
619 This value influences the timeout of a locally closed TCP connection,
620 when RTO retransmissions remain unacknowledged.
621 See tcp_retries2 for more details.
623 The default value is 8.
625 If your machine is a loaded WEB server,
626 you should think about lowering this value, such sockets
627 may consume significant resources. Cf. tcp_max_orphans.
629 tcp_recovery - INTEGER
630 This value is a bitmap to enable various experimental loss recovery
633 ========= =============================================================
634 RACK: 0x1 enables the RACK loss detection for fast detection of lost
635 retransmissions and tail drops. It also subsumes and disables
636 RFC6675 recovery for SACK connections.
638 RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
640 RACK: 0x4 disables RACK's DUPACK threshold heuristic
641 ========= =============================================================
645 tcp_reflect_tos - BOOLEAN
646 For listening sockets, reuse the DSCP value of the initial SYN message
647 for outgoing packets. This allows to have both directions of a TCP
648 stream to use the same DSCP value, assuming DSCP remains unchanged for
649 the lifetime of the connection.
651 This options affects both IPv4 and IPv6.
653 Default: 0 (disabled)
655 tcp_reordering - INTEGER
656 Initial reordering level of packets in a TCP stream.
657 TCP stack can then dynamically adjust flow reordering level
658 between this initial value and tcp_max_reordering
662 tcp_max_reordering - INTEGER
663 Maximal reordering level of packets in a TCP stream.
664 300 is a fairly conservative value, but you might increase it
665 if paths are using per packet load balancing (like bonding rr mode)
669 tcp_retrans_collapse - BOOLEAN
670 Bug-to-bug compatibility with some broken printers.
671 On retransmit try to send bigger packets to work around bugs in
674 tcp_retries1 - INTEGER
675 This value influences the time, after which TCP decides, that
676 something is wrong due to unacknowledged RTO retransmissions,
677 and reports this suspicion to the network layer.
678 See tcp_retries2 for more details.
680 RFC 1122 recommends at least 3 retransmissions, which is the
683 tcp_retries2 - INTEGER
684 This value influences the timeout of an alive TCP connection,
685 when RTO retransmissions remain unacknowledged.
686 Given a value of N, a hypothetical TCP connection following
687 exponential backoff with an initial RTO of TCP_RTO_MIN would
688 retransmit N times before killing the connection at the (N+1)th RTO.
690 The default value of 15 yields a hypothetical timeout of 924.6
691 seconds and is a lower bound for the effective timeout.
692 TCP will effectively time out at the first RTO which exceeds the
693 hypothetical timeout.
695 RFC 1122 recommends at least 100 seconds for the timeout,
696 which corresponds to a value of at least 8.
698 tcp_rfc1337 - BOOLEAN
699 If set, the TCP stack behaves conforming to RFC1337. If unset,
700 we are not conforming to RFC, but prevent TCP TIME_WAIT
705 tcp_rmem - vector of 3 INTEGERs: min, default, max
706 min: Minimal size of receive buffer used by TCP sockets.
707 It is guaranteed to each TCP socket, even under moderate memory
712 default: initial size of receive buffer used by TCP sockets.
713 This value overrides net.core.rmem_default used by other protocols.
714 Default: 131072 bytes.
715 This value results in initial window of 65535.
717 max: maximal size of receive buffer allowed for automatically
718 selected receiver buffers for TCP socket. This value does not override
719 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
720 automatic tuning of that socket's receive buffer size, in which
721 case this value is ignored.
722 Default: between 131072 and 6MB, depending on RAM size.
725 Enable select acknowledgments (SACKS).
727 tcp_comp_sack_delay_ns - LONG INTEGER
728 TCP tries to reduce number of SACK sent, using a timer
729 based on 5% of SRTT, capped by this sysctl, in nano seconds.
730 The default is 1ms, based on TSO autosizing period.
732 Default : 1,000,000 ns (1 ms)
734 tcp_comp_sack_slack_ns - LONG INTEGER
735 This sysctl control the slack used when arming the
736 timer used by SACK compression. This gives extra time
737 for small RTT flows, and reduces system overhead by allowing
738 opportunistic reduction of timer interrupts.
740 Default : 100,000 ns (100 us)
742 tcp_comp_sack_nr - INTEGER
743 Max number of SACK that can be compressed.
744 Using 0 disables SACK compression.
748 tcp_slow_start_after_idle - BOOLEAN
749 If set, provide RFC2861 behavior and time out the congestion
750 window after an idle period. An idle period is defined at
751 the current RTO. If unset, the congestion window will not
752 be timed out after an idle period.
757 Use the Host requirements interpretation of the TCP urgent pointer field.
758 Most hosts use the older BSD interpretation, so if you turn this on
759 Linux might not communicate correctly with them.
763 tcp_synack_retries - INTEGER
764 Number of times SYNACKs for a passive TCP connection attempt will
765 be retransmitted. Should not be higher than 255. Default value
766 is 5, which corresponds to 31seconds till the last retransmission
767 with the current initial RTO of 1second. With this the final timeout
768 for a passive TCP connection will happen after 63seconds.
770 tcp_syncookies - INTEGER
771 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
772 Send out syncookies when the syn backlog queue of a socket
773 overflows. This is to prevent against the common 'SYN flood attack'
776 Note, that syncookies is fallback facility.
777 It MUST NOT be used to help highly loaded servers to stand
778 against legal connection rate. If you see SYN flood warnings
779 in your logs, but investigation shows that they occur
780 because of overload with legal connections, you should tune
781 another parameters until this warning disappear.
782 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
784 syncookies seriously violate TCP protocol, do not allow
785 to use TCP extensions, can result in serious degradation
786 of some services (f.e. SMTP relaying), visible not by you,
787 but your clients and relays, contacting you. While you see
788 SYN flood warnings in logs not being really flooded, your server
789 is seriously misconfigured.
791 If you want to test which effects syncookies have to your
792 network connections you can set this knob to 2 to enable
793 unconditionally generation of syncookies.
795 tcp_migrate_req - BOOLEAN
796 The incoming connection is tied to a specific listening socket when
797 the initial SYN packet is received during the three-way handshake.
798 When a listener is closed, in-flight request sockets during the
799 handshake and established sockets in the accept queue are aborted.
801 If the listener has SO_REUSEPORT enabled, other listeners on the
802 same port should have been able to accept such connections. This
803 option makes it possible to migrate such child sockets to another
804 listener after close() or shutdown().
806 The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should
807 usually be used to define the policy to pick an alive listener.
808 Otherwise, the kernel will randomly pick an alive listener only if
809 this option is enabled.
811 Note that migration between listeners with different settings may
812 crash applications. Let's say migration happens from listener A to
813 B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from
814 the requests migrated from A. To avoid such a situation, cancel
815 migration by returning SK_DROP in the type of eBPF program, or
820 tcp_fastopen - INTEGER
821 Enable TCP Fast Open (RFC7413) to send and accept data in the opening
824 The client support is enabled by flag 0x1 (on by default). The client
825 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
826 rather than connect() to send data in SYN.
828 The server support is enabled by flag 0x2 (off by default). Then
829 either enable for all listeners with another flag (0x400) or
830 enable individual listeners via TCP_FASTOPEN socket option with
831 the option value being the length of the syn-data backlog.
833 The values (bitmap) are
835 ===== ======== ======================================================
836 0x1 (client) enables sending data in the opening SYN on the client.
837 0x2 (server) enables the server support, i.e., allowing data in
838 a SYN packet to be accepted and passed to the
839 application before 3-way handshake finishes.
840 0x4 (client) send data in the opening SYN regardless of cookie
841 availability and without a cookie option.
842 0x200 (server) accept data-in-SYN w/o any cookie option present.
843 0x400 (server) enable all listeners to support Fast Open by
844 default without explicit TCP_FASTOPEN socket option.
845 ===== ======== ======================================================
849 Note that additional client or server features are only
850 effective if the basic support (0x1 and 0x2) are enabled respectively.
852 tcp_fastopen_blackhole_timeout_sec - INTEGER
853 Initial time period in second to disable Fastopen on active TCP sockets
854 when a TFO firewall blackhole issue happens.
855 This time period will grow exponentially when more blackhole issues
856 get detected right after Fastopen is re-enabled and will reset to
857 initial value when the blackhole issue goes away.
858 0 to disable the blackhole detection.
860 By default, it is set to 0 (feature is disabled).
862 tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
863 The list consists of a primary key and an optional backup key. The
864 primary key is used for both creating and validating cookies, while the
865 optional backup key is only used for validating cookies. The purpose of
866 the backup key is to maximize TFO validation when keys are rotated.
868 A randomly chosen primary key may be configured by the kernel if
869 the tcp_fastopen sysctl is set to 0x400 (see above), or if the
870 TCP_FASTOPEN setsockopt() optname is set and a key has not been
871 previously configured via sysctl. If keys are configured via
872 setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
873 per-socket keys will be used instead of any keys that are specified via
876 A key is specified as 4 8-digit hexadecimal integers which are separated
877 by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
878 omitted. A primary and a backup key may be specified by separating them
879 by a comma. If only one key is specified, it becomes the primary key and
880 any previously configured backup keys are removed.
882 tcp_syn_retries - INTEGER
883 Number of times initial SYNs for an active TCP connection attempt
884 will be retransmitted. Should not be higher than 127. Default value
885 is 6, which corresponds to 67seconds (with tcp_syn_linear_timeouts = 4)
886 till the last retransmission with the current initial RTO of 1second.
887 With this the final timeout for an active TCP connection attempt
888 will happen after 131seconds.
890 tcp_timestamps - INTEGER
891 Enable timestamps as defined in RFC1323.
894 - 1: Enable timestamps as defined in RFC1323 and use random offset for
895 each connection rather than only using the current time.
896 - 2: Like 1, but without random offsets.
900 tcp_min_tso_segs - INTEGER
901 Minimal number of segments per TSO frame.
903 Since linux-3.12, TCP does an automatic sizing of TSO frames,
904 depending on flow rate, instead of filling 64Kbytes packets.
905 For specific usages, it's possible to force TCP to build big
906 TSO frames. Note that TCP stack might split too big TSO packets
907 if available window is too small.
911 tcp_tso_rtt_log - INTEGER
912 Adjustment of TSO packet sizes based on min_rtt
914 Starting from linux-5.18, TCP autosizing can be tweaked
915 for flows having small RTT.
917 Old autosizing was splitting the pacing budget to send 1024 TSO
920 tso_packet_size = sk->sk_pacing_rate / 1024;
922 With the new mechanism, we increase this TSO sizing using:
924 distance = min_rtt_usec / (2^tcp_tso_rtt_log)
925 tso_packet_size += gso_max_size >> distance;
927 This means that flows between very close hosts can use bigger
928 TSO packets, reducing their cpu costs.
930 If you want to use the old autosizing, set this sysctl to 0.
932 Default: 9 (2^9 = 512 usec)
934 tcp_pacing_ss_ratio - INTEGER
935 sk->sk_pacing_rate is set by TCP stack using a ratio applied
936 to current rate. (current_rate = cwnd * mss / srtt)
937 If TCP is in slow start, tcp_pacing_ss_ratio is applied
938 to let TCP probe for bigger speeds, assuming cwnd can be
939 doubled every other RTT.
943 tcp_pacing_ca_ratio - INTEGER
944 sk->sk_pacing_rate is set by TCP stack using a ratio applied
945 to current rate. (current_rate = cwnd * mss / srtt)
946 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
947 is applied to conservatively probe for bigger throughput.
951 tcp_syn_linear_timeouts - INTEGER
952 The number of times for an active TCP connection to retransmit SYNs with
953 a linear backoff timeout before defaulting to an exponential backoff
954 timeout. This has no effect on SYNACK at the passive TCP side.
956 With an initial RTO of 1 and tcp_syn_linear_timeouts = 4 we would
957 expect SYN RTOs to be: 1, 1, 1, 1, 1, 2, 4, ... (4 linear timeouts,
958 and the first exponential backoff using 2^0 * initial_RTO).
961 tcp_tso_win_divisor - INTEGER
962 This allows control over what percentage of the congestion window
963 can be consumed by a single TSO frame.
964 The setting of this parameter is a choice between burstiness and
965 building larger TSO frames.
969 tcp_tw_reuse - INTEGER
970 Enable reuse of TIME-WAIT sockets for new connections when it is
971 safe from protocol viewpoint.
975 - 2 - enable for loopback traffic only
977 It should not be changed without advice/request of technical
982 tcp_window_scaling - BOOLEAN
983 Enable window scaling as defined in RFC1323.
985 tcp_shrink_window - BOOLEAN
986 This changes how the TCP receive window is calculated.
988 RFC 7323, section 2.4, says there are instances when a retracted
989 window can be offered, and that TCP implementations MUST ensure
990 that they handle a shrinking window, as specified in RFC 1122.
992 - 0 - Disabled. The window is never shrunk.
993 - 1 - Enabled. The window is shrunk when necessary to remain within
994 the memory limit set by autotuning (sk_rcvbuf).
995 This only occurs if a non-zero receive window
996 scaling factor is also in effect.
1000 tcp_wmem - vector of 3 INTEGERs: min, default, max
1001 min: Amount of memory reserved for send buffers for TCP sockets.
1002 Each TCP socket has rights to use it due to fact of its birth.
1006 default: initial size of send buffer used by TCP sockets. This
1007 value overrides net.core.wmem_default used by other protocols.
1009 It is usually lower than net.core.wmem_default.
1013 max: Maximal amount of memory allowed for automatically tuned
1014 send buffers for TCP sockets. This value does not override
1015 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
1016 automatic tuning of that socket's send buffer size, in which case
1017 this value is ignored.
1019 Default: between 64K and 4MB, depending on RAM size.
1021 tcp_notsent_lowat - UNSIGNED INTEGER
1022 A TCP socket can control the amount of unsent bytes in its write queue,
1023 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
1024 reports POLLOUT events if the amount of unsent bytes is below a per
1025 socket value, and if the write queue is not full. sendmsg() will
1026 also not add new buffers if the limit is hit.
1028 This global variable controls the amount of unsent data for
1029 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
1030 to the global variable has immediate effect.
1032 Default: UINT_MAX (0xFFFFFFFF)
1034 tcp_workaround_signed_windows - BOOLEAN
1035 If set, assume no receipt of a window scaling option means the
1036 remote TCP is broken and treats the window as a signed quantity.
1037 If unset, assume the remote TCP is not broken even if we do
1038 not receive a window scaling option from them.
1042 tcp_thin_linear_timeouts - BOOLEAN
1043 Enable dynamic triggering of linear timeouts for thin streams.
1044 If set, a check is performed upon retransmission by timeout to
1045 determine if the stream is thin (less than 4 packets in flight).
1046 As long as the stream is found to be thin, up to 6 linear
1047 timeouts may be performed before exponential backoff mode is
1048 initiated. This improves retransmission latency for
1049 non-aggressive thin streams, often found to be time-dependent.
1050 For more information on thin streams, see
1051 Documentation/networking/tcp-thin.rst
1055 tcp_limit_output_bytes - INTEGER
1056 Controls TCP Small Queue limit per tcp socket.
1057 TCP bulk sender tends to increase packets in flight until it
1058 gets losses notifications. With SNDBUF autotuning, this can
1059 result in a large amount of packets queued on the local machine
1060 (e.g.: qdiscs, CPU backlog, or device) hurting latency of other
1061 flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes
1062 limits the number of bytes on qdisc or device to reduce artificial
1063 RTT/cwnd and reduce bufferbloat.
1065 Default: 1048576 (16 * 65536)
1067 tcp_challenge_ack_limit - INTEGER
1068 Limits number of Challenge ACK sent per second, as recommended
1069 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
1070 Note that this per netns rate limit can allow some side channel
1071 attacks and probably should not be enabled.
1072 TCP stack implements per TCP socket limits anyway.
1073 Default: INT_MAX (unlimited)
1075 tcp_ehash_entries - INTEGER
1076 Show the number of hash buckets for TCP sockets in the current
1077 networking namespace.
1079 A negative value means the networking namespace does not own its
1080 hash buckets and shares the initial networking namespace's one.
1082 tcp_child_ehash_entries - INTEGER
1083 Control the number of hash buckets for TCP sockets in the child
1084 networking namespace, which must be set before clone() or unshare().
1086 If the value is not 0, the kernel uses a value rounded up to 2^n
1087 as the actual hash bucket size. 0 is a special value, meaning
1088 the child networking namespace will share the initial networking
1089 namespace's hash buckets.
1091 Note that the child will use the global one in case the kernel
1092 fails to allocate enough memory. In addition, the global hash
1093 buckets are spread over available NUMA nodes, but the allocation
1094 of the child hash table depends on the current process's NUMA
1095 policy, which could result in performance differences.
1097 Note also that the default value of tcp_max_tw_buckets and
1098 tcp_max_syn_backlog depend on the hash bucket size.
1100 Possible values: 0, 2^n (n: 0 - 24 (16Mi))
1104 tcp_plb_enabled - BOOLEAN
1105 If set and the underlying congestion control (e.g. DCTCP) supports
1106 and enables PLB feature, TCP PLB (Protective Load Balancing) is
1107 enabled. PLB is described in the following paper:
1108 https://doi.org/10.1145/3544216.3544226. Based on PLB parameters,
1109 upon sensing sustained congestion, TCP triggers a change in
1110 flow label field for outgoing IPv6 packets. A change in flow label
1111 field potentially changes the path of outgoing packets for switches
1112 that use ECMP/WCMP for routing.
1114 PLB changes socket txhash which results in a change in IPv6 Flow Label
1115 field, and currently no-op for IPv4 headers. It is possible
1116 to apply PLB for IPv4 with other network header fields (e.g. TCP
1117 or IPv4 options) or using encapsulation where outer header is used
1118 by switches to determine next hop. In either case, further host
1119 and switch side changes will be needed.
1121 When set, PLB assumes that congestion signal (e.g. ECN) is made
1122 available and used by congestion control module to estimate a
1123 congestion measure (e.g. ce_ratio). PLB needs a congestion measure to
1124 make repathing decisions.
1128 tcp_plb_idle_rehash_rounds - INTEGER
1129 Number of consecutive congested rounds (RTT) seen after which
1130 a rehash can be performed, given there are no packets in flight.
1131 This is referred to as M in PLB paper:
1132 https://doi.org/10.1145/3544216.3544226.
1134 Possible Values: 0 - 31
1138 tcp_plb_rehash_rounds - INTEGER
1139 Number of consecutive congested rounds (RTT) seen after which
1140 a forced rehash can be performed. Be careful when setting this
1141 parameter, as a small value increases the risk of retransmissions.
1142 This is referred to as N in PLB paper:
1143 https://doi.org/10.1145/3544216.3544226.
1145 Possible Values: 0 - 31
1149 tcp_plb_suspend_rto_sec - INTEGER
1150 Time, in seconds, to suspend PLB in event of an RTO. In order to avoid
1151 having PLB repath onto a connectivity "black hole", after an RTO a TCP
1152 connection suspends PLB repathing for a random duration between 1x and
1153 2x of this parameter. Randomness is added to avoid concurrent rehashing
1154 of multiple TCP connections. This should be set corresponding to the
1155 amount of time it takes to repair a failed link.
1157 Possible Values: 0 - 255
1161 tcp_plb_cong_thresh - INTEGER
1162 Fraction of packets marked with congestion over a round (RTT) to
1163 tag that round as congested. This is referred to as K in the PLB paper:
1164 https://doi.org/10.1145/3544216.3544226.
1166 The 0-1 fraction range is mapped to 0-256 range to avoid floating
1167 point operations. For example, 128 means that if at least 50% of
1168 the packets in a round were marked as congested then the round
1169 will be tagged as congested.
1171 Setting threshold to 0 means that PLB repaths every RTT regardless
1172 of congestion. This is not intended behavior for PLB and should be
1173 used only for experimentation purpose.
1175 Possible Values: 0 - 256
1182 udp_l3mdev_accept - BOOLEAN
1183 Enabling this option allows a "global" bound socket to work
1184 across L3 master domains (e.g., VRFs) with packets capable of
1185 being received regardless of the L3 domain in which they
1186 originated. Only valid when the kernel was compiled with
1187 CONFIG_NET_L3_MASTER_DEV.
1189 Default: 0 (disabled)
1191 udp_mem - vector of 3 INTEGERs: min, pressure, max
1192 Number of pages allowed for queueing by all UDP sockets.
1194 min: Number of pages allowed for queueing by all UDP sockets.
1196 pressure: This value was introduced to follow format of tcp_mem.
1198 max: This value was introduced to follow format of tcp_mem.
1200 Default is calculated at boot time from amount of available memory.
1202 udp_rmem_min - INTEGER
1203 Minimal size of receive buffer used by UDP sockets in moderation.
1204 Each UDP socket is able to use the size for receiving data, even if
1205 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
1209 udp_wmem_min - INTEGER
1210 UDP does not have tx memory accounting and this tunable has no effect.
1212 udp_hash_entries - INTEGER
1213 Show the number of hash buckets for UDP sockets in the current
1214 networking namespace.
1216 A negative value means the networking namespace does not own its
1217 hash buckets and shares the initial networking namespace's one.
1219 udp_child_ehash_entries - INTEGER
1220 Control the number of hash buckets for UDP sockets in the child
1221 networking namespace, which must be set before clone() or unshare().
1223 If the value is not 0, the kernel uses a value rounded up to 2^n
1224 as the actual hash bucket size. 0 is a special value, meaning
1225 the child networking namespace will share the initial networking
1226 namespace's hash buckets.
1228 Note that the child will use the global one in case the kernel
1229 fails to allocate enough memory. In addition, the global hash
1230 buckets are spread over available NUMA nodes, but the allocation
1231 of the child hash table depends on the current process's NUMA
1232 policy, which could result in performance differences.
1234 Possible values: 0, 2^n (n: 7 (128) - 16 (64K))
1242 raw_l3mdev_accept - BOOLEAN
1243 Enabling this option allows a "global" bound socket to work
1244 across L3 master domains (e.g., VRFs) with packets capable of
1245 being received regardless of the L3 domain in which they
1246 originated. Only valid when the kernel was compiled with
1247 CONFIG_NET_L3_MASTER_DEV.
1249 Default: 1 (enabled)
1254 cipso_cache_enable - BOOLEAN
1255 If set, enable additions to and lookups from the CIPSO label mapping
1256 cache. If unset, additions are ignored and lookups always result in a
1257 miss. However, regardless of the setting the cache is still
1258 invalidated when required when means you can safely toggle this on and
1259 off and the cache will always be "safe".
1263 cipso_cache_bucket_size - INTEGER
1264 The CIPSO label cache consists of a fixed size hash table with each
1265 hash bucket containing a number of cache entries. This variable limits
1266 the number of entries in each hash bucket; the larger the value is, the
1267 more CIPSO label mappings that can be cached. When the number of
1268 entries in a given hash bucket reaches this limit adding new entries
1269 causes the oldest entry in the bucket to be removed to make room.
1273 cipso_rbm_optfmt - BOOLEAN
1274 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
1275 the CIPSO draft specification (see Documentation/netlabel for details).
1276 This means that when set the CIPSO tag will be padded with empty
1277 categories in order to make the packet data 32-bit aligned.
1281 cipso_rbm_structvalid - BOOLEAN
1282 If set, do a very strict check of the CIPSO option when
1283 ip_options_compile() is called. If unset, relax the checks done during
1284 ip_options_compile(). Either way is "safe" as errors are caught else
1285 where in the CIPSO processing code but setting this to 0 (False) should
1286 result in less work (i.e. it should be faster) but could cause problems
1287 with other implementations that require strict checking.
1294 ip_local_port_range - 2 INTEGERS
1295 Defines the local port range that is used by TCP and UDP to
1296 choose the local port. The first number is the first, the
1297 second the last local port number.
1298 If possible, it is better these numbers have different parity
1299 (one even and one odd value).
1300 Must be greater than or equal to ip_unprivileged_port_start.
1301 The default values are 32768 and 60999 respectively.
1303 ip_local_reserved_ports - list of comma separated ranges
1304 Specify the ports which are reserved for known third-party
1305 applications. These ports will not be used by automatic port
1306 assignments (e.g. when calling connect() or bind() with port
1307 number 0). Explicit port allocation behavior is unchanged.
1309 The format used for both input and output is a comma separated
1310 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
1311 10). Writing to the file will clear all previously reserved
1312 ports and update the current list with the one given in the
1315 Note that ip_local_port_range and ip_local_reserved_ports
1316 settings are independent and both are considered by the kernel
1317 when determining which ports are available for automatic port
1320 You can reserve ports which are not in the current
1321 ip_local_port_range, e.g.::
1323 $ cat /proc/sys/net/ipv4/ip_local_port_range
1325 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
1328 although this is redundant. However such a setting is useful
1329 if later the port range is changed to a value that will
1330 include the reserved ports. Also keep in mind, that overlapping
1331 of these ranges may affect probability of selecting ephemeral
1332 ports which are right after block of reserved ports.
1336 ip_unprivileged_port_start - INTEGER
1337 This is a per-namespace sysctl. It defines the first
1338 unprivileged port in the network namespace. Privileged ports
1339 require root or CAP_NET_BIND_SERVICE in order to bind to them.
1340 To disable all privileged ports, set this to 0. They must not
1341 overlap with the ip_local_port_range.
1345 ip_nonlocal_bind - BOOLEAN
1346 If set, allows processes to bind() to non-local IP addresses,
1347 which can be quite useful - but may break some applications.
1351 ip_autobind_reuse - BOOLEAN
1352 By default, bind() does not select the ports automatically even if
1353 the new socket and all sockets bound to the port have SO_REUSEADDR.
1354 ip_autobind_reuse allows bind() to reuse the port and this is useful
1355 when you use bind()+connect(), but may break some applications.
1356 The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
1357 option should only be set by experts.
1360 ip_dynaddr - INTEGER
1361 If set non-zero, enables support for dynamic addresses.
1362 If set to a non-zero value larger than 1, a kernel log
1363 message will be printed when dynamic address rewriting
1368 ip_early_demux - BOOLEAN
1369 Optimize input packet processing down to one demux for
1370 certain kinds of local sockets. Currently we only do this
1371 for established TCP and connected UDP sockets.
1373 It may add an additional cost for pure routing workloads that
1374 reduces overall throughput, in such case you should disable it.
1378 ping_group_range - 2 INTEGERS
1379 Restrict ICMP_PROTO datagram sockets to users in the group range.
1380 The default is "1 0", meaning, that nobody (not even root) may
1381 create ping sockets. Setting it to "100 100" would grant permissions
1382 to the single group. "0 4294967294" would enable it for the world, "100
1383 4294967294" would enable it for the users, but not daemons.
1385 tcp_early_demux - BOOLEAN
1386 Enable early demux for established TCP sockets.
1390 udp_early_demux - BOOLEAN
1391 Enable early demux for connected UDP sockets. Disable this if
1392 your system could experience more unconnected load.
1396 icmp_echo_ignore_all - BOOLEAN
1397 If set non-zero, then the kernel will ignore all ICMP ECHO
1398 requests sent to it.
1402 icmp_echo_enable_probe - BOOLEAN
1403 If set to one, then the kernel will respond to RFC 8335 PROBE
1404 requests sent to it.
1408 icmp_echo_ignore_broadcasts - BOOLEAN
1409 If set non-zero, then the kernel will ignore all ICMP ECHO and
1410 TIMESTAMP requests sent to it via broadcast/multicast.
1414 icmp_ratelimit - INTEGER
1415 Limit the maximal rates for sending ICMP packets whose type matches
1416 icmp_ratemask (see below) to specific targets.
1417 0 to disable any limiting,
1418 otherwise the minimal space between responses in milliseconds.
1419 Note that another sysctl, icmp_msgs_per_sec limits the number
1420 of ICMP packets sent on all targets.
1424 icmp_msgs_per_sec - INTEGER
1425 Limit maximal number of ICMP packets sent per second from this host.
1426 Only messages whose type matches icmp_ratemask (see below) are
1427 controlled by this limit. For security reasons, the precise count
1428 of messages per second is randomized.
1432 icmp_msgs_burst - INTEGER
1433 icmp_msgs_per_sec controls number of ICMP packets sent per second,
1434 while icmp_msgs_burst controls the burst size of these packets.
1435 For security reasons, the precise burst size is randomized.
1439 icmp_ratemask - INTEGER
1440 Mask made of ICMP types for which rates are being limited.
1442 Significant bits: IHGFEDCBA9876543210
1444 Default mask: 0000001100000011000 (6168)
1446 Bit definitions (see include/linux/icmp.h):
1448 = =========================
1450 3 Destination Unreachable [1]_
1451 4 Source Quench [1]_
1454 B Time Exceeded [1]_
1455 C Parameter Problem [1]_
1460 H Address Mask Request
1461 I Address Mask Reply
1462 = =========================
1464 .. [1] These are rate limited by default (see default mask above)
1466 icmp_ignore_bogus_error_responses - BOOLEAN
1467 Some routers violate RFC1122 by sending bogus responses to broadcast
1468 frames. Such violations are normally logged via a kernel warning.
1469 If this is set to TRUE, the kernel will not give such warnings, which
1470 will avoid log file clutter.
1474 icmp_errors_use_inbound_ifaddr - BOOLEAN
1476 If zero, icmp error messages are sent with the primary address of
1477 the exiting interface.
1479 If non-zero, the message will be sent with the primary address of
1480 the interface that received the packet that caused the icmp error.
1481 This is the behaviour many network administrators will expect from
1482 a router. And it can make debugging complicated network layouts
1485 Note that if no primary address exists for the interface selected,
1486 then the primary address of the first non-loopback interface that
1487 has one will be used regardless of this setting.
1491 igmp_max_memberships - INTEGER
1492 Change the maximum number of multicast groups we can subscribe to.
1495 Theoretical maximum value is bounded by having to send a membership
1496 report in a single datagram (i.e. the report can't span multiple
1497 datagrams, or risk confusing the switch and leaving groups you don't
1500 The number of supported groups 'M' is bounded by the number of group
1501 report entries you can fit into a single datagram of 65535 bytes.
1503 M = 65536-sizeof (ip header)/(sizeof(Group record))
1505 Group records are variable length, with a minimum of 12 bytes.
1506 So net.ipv4.igmp_max_memberships should not be set higher than:
1508 (65536-24) / 12 = 5459
1510 The value 5459 assumes no IP header options, so in practice
1511 this number may be lower.
1513 igmp_max_msf - INTEGER
1514 Maximum number of addresses allowed in the source filter list for a
1520 Controls the IGMP query robustness variable (see RFC2236 8.1).
1522 Default: 2 (as specified by RFC2236 8.1)
1524 Minimum: 1 (as specified by RFC6636 4.5)
1526 force_igmp_version - INTEGER
1527 - 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
1528 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
1529 Present timer expires.
1530 - 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
1531 receive IGMPv2/v3 query.
1532 - 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
1533 IGMPv1 query message. Will reply report if receive IGMPv3 query.
1534 - 3 - Enforce to use IGMP version 3. The same react with default 0.
1538 this is not the same with force_mld_version because IGMPv3 RFC3376
1539 Security Considerations does not have clear description that we could
1540 ignore other version messages completely as MLDv2 RFC3810. So make
1541 this value as default 0 is recommended.
1543 ``conf/interface/*``
1544 changes special settings per interface (where
1545 interface" is the name of your network interface)
1548 is special, changes the settings for all interfaces
1550 log_martians - BOOLEAN
1551 Log packets with impossible addresses to kernel log.
1552 log_martians for the interface will be enabled if at least one of
1553 conf/{all,interface}/log_martians is set to TRUE,
1554 it will be disabled otherwise
1556 accept_redirects - BOOLEAN
1557 Accept ICMP redirect messages.
1558 accept_redirects for the interface will be enabled if:
1560 - both conf/{all,interface}/accept_redirects are TRUE in the case
1561 forwarding for the interface is enabled
1565 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
1566 case forwarding for the interface is disabled
1568 accept_redirects for the interface will be disabled otherwise
1575 forwarding - BOOLEAN
1576 Enable IP forwarding on this interface. This controls whether packets
1577 received _on_ this interface can be forwarded.
1579 mc_forwarding - BOOLEAN
1580 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1581 and a multicast routing daemon is required.
1582 conf/all/mc_forwarding must also be set to TRUE to enable multicast
1583 routing for the interface
1586 Integer value used to differentiate the devices by the medium they
1587 are attached to. Two devices can have different id values when
1588 the broadcast packets are received only on one of them.
1589 The default value 0 means that the device is the only interface
1590 to its medium, value of -1 means that medium is not known.
1592 Currently, it is used to change the proxy_arp behavior:
1593 the proxy_arp feature is enabled for packets forwarded between
1594 two devices attached to different media.
1599 proxy_arp for the interface will be enabled if at least one of
1600 conf/{all,interface}/proxy_arp is set to TRUE,
1601 it will be disabled otherwise
1603 proxy_arp_pvlan - BOOLEAN
1604 Private VLAN proxy arp.
1606 Basically allow proxy arp replies back to the same interface
1607 (from which the ARP request/solicitation was received).
1609 This is done to support (ethernet) switch features, like RFC
1610 3069, where the individual ports are NOT allowed to
1611 communicate with each other, but they are allowed to talk to
1612 the upstream router. As described in RFC 3069, it is possible
1613 to allow these hosts to communicate through the upstream
1614 router by proxy_arp'ing. Don't need to be used together with
1617 This technology is known by different names:
1619 In RFC 3069 it is called VLAN Aggregation.
1620 Cisco and Allied Telesyn call it Private VLAN.
1621 Hewlett-Packard call it Source-Port filtering or port-isolation.
1622 Ericsson call it MAC-Forced Forwarding (RFC Draft).
1624 proxy_delay - INTEGER
1625 Delay proxy response.
1627 Delay response to a neighbor solicitation when proxy_arp
1628 or proxy_ndp is enabled. A random value between [0, proxy_delay)
1629 will be chosen, setting to zero means reply with no delay.
1630 Value in jiffies. Defaults to 80.
1632 shared_media - BOOLEAN
1633 Send(router) or accept(host) RFC1620 shared media redirects.
1634 Overrides secure_redirects.
1636 shared_media for the interface will be enabled if at least one of
1637 conf/{all,interface}/shared_media is set to TRUE,
1638 it will be disabled otherwise
1642 secure_redirects - BOOLEAN
1643 Accept ICMP redirect messages only to gateways listed in the
1644 interface's current gateway list. Even if disabled, RFC1122 redirect
1647 Overridden by shared_media.
1649 secure_redirects for the interface will be enabled if at least one of
1650 conf/{all,interface}/secure_redirects is set to TRUE,
1651 it will be disabled otherwise
1655 send_redirects - BOOLEAN
1656 Send redirects, if router.
1658 send_redirects for the interface will be enabled if at least one of
1659 conf/{all,interface}/send_redirects is set to TRUE,
1660 it will be disabled otherwise
1664 bootp_relay - BOOLEAN
1665 Accept packets with source address 0.b.c.d destined
1666 not to this host as local ones. It is supposed, that
1667 BOOTP relay daemon will catch and forward such packets.
1668 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1673 Not Implemented Yet.
1675 accept_source_route - BOOLEAN
1676 Accept packets with SRR option.
1677 conf/all/accept_source_route must also be set to TRUE to accept packets
1678 with SRR option on the interface
1685 accept_local - BOOLEAN
1686 Accept packets with local source addresses. In combination with
1687 suitable routing, this can be used to direct packets between two
1688 local interfaces over the wire and have them accepted properly.
1691 route_localnet - BOOLEAN
1692 Do not consider loopback addresses as martian source or destination
1693 while routing. This enables the use of 127/8 for local routing purposes.
1698 - 0 - No source validation.
1699 - 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1700 Each incoming packet is tested against the FIB and if the interface
1701 is not the best reverse path the packet check will fail.
1702 By default failed packets are discarded.
1703 - 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1704 Each incoming packet's source address is also tested against the FIB
1705 and if the source address is not reachable via any interface
1706 the packet check will fail.
1708 Current recommended practice in RFC3704 is to enable strict mode
1709 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1710 or other complicated routing, then loose mode is recommended.
1712 The max value from conf/{all,interface}/rp_filter is used
1713 when doing source validation on the {interface}.
1715 Default value is 0. Note that some distributions enable it
1718 src_valid_mark - BOOLEAN
1719 - 0 - The fwmark of the packet is not included in reverse path
1720 route lookup. This allows for asymmetric routing configurations
1721 utilizing the fwmark in only one direction, e.g., transparent
1724 - 1 - The fwmark of the packet is included in reverse path route
1725 lookup. This permits rp_filter to function when the fwmark is
1726 used for routing traffic in both directions.
1728 This setting also affects the utilization of fmwark when
1729 performing source address selection for ICMP replies, or
1730 determining addresses stored for the IPOPT_TS_TSANDADDR and
1731 IPOPT_RR IP options.
1733 The max value from conf/{all,interface}/src_valid_mark is used.
1737 arp_filter - BOOLEAN
1738 - 1 - Allows you to have multiple network interfaces on the same
1739 subnet, and have the ARPs for each interface be answered
1740 based on whether or not the kernel would route a packet from
1741 the ARP'd IP out that interface (therefore you must use source
1742 based routing for this to work). In other words it allows control
1743 of which cards (usually 1) will respond to an arp request.
1745 - 0 - (default) The kernel can respond to arp requests with addresses
1746 from other interfaces. This may seem wrong but it usually makes
1747 sense, because it increases the chance of successful communication.
1748 IP addresses are owned by the complete host on Linux, not by
1749 particular interfaces. Only for more complex setups like load-
1750 balancing, does this behaviour cause problems.
1752 arp_filter for the interface will be enabled if at least one of
1753 conf/{all,interface}/arp_filter is set to TRUE,
1754 it will be disabled otherwise
1756 arp_announce - INTEGER
1757 Define different restriction levels for announcing the local
1758 source IP address from IP packets in ARP requests sent on
1761 - 0 - (default) Use any local address, configured on any interface
1762 - 1 - Try to avoid local addresses that are not in the target's
1763 subnet for this interface. This mode is useful when target
1764 hosts reachable via this interface require the source IP
1765 address in ARP requests to be part of their logical network
1766 configured on the receiving interface. When we generate the
1767 request we will check all our subnets that include the
1768 target IP and will preserve the source address if it is from
1769 such subnet. If there is no such subnet we select source
1770 address according to the rules for level 2.
1771 - 2 - Always use the best local address for this target.
1772 In this mode we ignore the source address in the IP packet
1773 and try to select local address that we prefer for talks with
1774 the target host. Such local address is selected by looking
1775 for primary IP addresses on all our subnets on the outgoing
1776 interface that include the target IP address. If no suitable
1777 local address is found we select the first local address
1778 we have on the outgoing interface or on all other interfaces,
1779 with the hope we will receive reply for our request and
1780 even sometimes no matter the source IP address we announce.
1782 The max value from conf/{all,interface}/arp_announce is used.
1784 Increasing the restriction level gives more chance for
1785 receiving answer from the resolved target while decreasing
1786 the level announces more valid sender's information.
1788 arp_ignore - INTEGER
1789 Define different modes for sending replies in response to
1790 received ARP requests that resolve local target IP addresses:
1792 - 0 - (default): reply for any local target IP address, configured
1794 - 1 - reply only if the target IP address is local address
1795 configured on the incoming interface
1796 - 2 - reply only if the target IP address is local address
1797 configured on the incoming interface and both with the
1798 sender's IP address are part from same subnet on this interface
1799 - 3 - do not reply for local addresses configured with scope host,
1800 only resolutions for global and link addresses are replied
1802 - 8 - do not reply for all local addresses
1804 The max value from conf/{all,interface}/arp_ignore is used
1805 when ARP request is received on the {interface}
1807 arp_notify - BOOLEAN
1808 Define mode for notification of address and device changes.
1810 == ==========================================================
1811 0 (default): do nothing
1812 1 Generate gratuitous arp requests when device is brought up
1813 or hardware address changes.
1814 == ==========================================================
1816 arp_accept - INTEGER
1817 Define behavior for accepting gratuitous ARP (garp) frames from devices
1818 that are not already present in the ARP table:
1820 - 0 - don't create new entries in the ARP table
1821 - 1 - create new entries in the ARP table
1822 - 2 - create new entries only if the source IP address is in the same
1823 subnet as an address configured on the interface that received the
1826 Both replies and requests type gratuitous arp will trigger the
1827 ARP table to be updated, if this setting is on.
1829 If the ARP table already contains the IP address of the
1830 gratuitous arp frame, the arp table will be updated regardless
1831 if this setting is on or off.
1833 arp_evict_nocarrier - BOOLEAN
1834 Clears the ARP cache on NOCARRIER events. This option is important for
1835 wireless devices where the ARP cache should not be cleared when roaming
1836 between access points on the same network. In most cases this should
1837 remain as the default (1).
1839 - 1 - (default): Clear the ARP cache on NOCARRIER events
1840 - 0 - Do not clear ARP cache on NOCARRIER events
1842 mcast_solicit - INTEGER
1843 The maximum number of multicast probes in INCOMPLETE state,
1844 when the associated hardware address is unknown. Defaults
1847 ucast_solicit - INTEGER
1848 The maximum number of unicast probes in PROBE state, when
1849 the hardware address is being reconfirmed. Defaults to 3.
1851 app_solicit - INTEGER
1852 The maximum number of probes to send to the user space ARP daemon
1853 via netlink before dropping back to multicast probes (see
1854 mcast_resolicit). Defaults to 0.
1856 mcast_resolicit - INTEGER
1857 The maximum number of multicast probes after unicast and
1858 app probes in PROBE state. Defaults to 0.
1860 disable_policy - BOOLEAN
1861 Disable IPSEC policy (SPD) for this interface
1863 disable_xfrm - BOOLEAN
1864 Disable IPSEC encryption on this interface, whatever the policy
1866 igmpv2_unsolicited_report_interval - INTEGER
1867 The interval in milliseconds in which the next unsolicited
1868 IGMPv1 or IGMPv2 report retransmit will take place.
1870 Default: 10000 (10 seconds)
1872 igmpv3_unsolicited_report_interval - INTEGER
1873 The interval in milliseconds in which the next unsolicited
1874 IGMPv3 report retransmit will take place.
1876 Default: 1000 (1 seconds)
1878 ignore_routes_with_linkdown - BOOLEAN
1879 Ignore routes whose link is down when performing a FIB lookup.
1881 promote_secondaries - BOOLEAN
1882 When a primary IP address is removed from this interface
1883 promote a corresponding secondary IP address instead of
1884 removing all the corresponding secondary IP addresses.
1886 drop_unicast_in_l2_multicast - BOOLEAN
1887 Drop any unicast IP packets that are received in link-layer
1888 multicast (or broadcast) frames.
1890 This behavior (for multicast) is actually a SHOULD in RFC
1891 1122, but is disabled by default for compatibility reasons.
1895 drop_gratuitous_arp - BOOLEAN
1896 Drop all gratuitous ARP frames, for example if there's a known
1897 good ARP proxy on the network and such frames need not be used
1898 (or in the case of 802.11, must not be used to prevent attacks.)
1904 Allows you to write a number, which can be used as required.
1908 xfrm4_gc_thresh - INTEGER
1909 (Obsolete since linux-4.14)
1910 The threshold at which we will start garbage collecting for IPv4
1911 destination cache entries. At twice this value the system will
1912 refuse new allocations.
1914 igmp_link_local_mcast_reports - BOOLEAN
1915 Enable IGMP reports for link local multicast groups in the
1921 kuznet@ms2.inr.ac.ru
1928 delon.nicolas@wanadoo.fr
1933 /proc/sys/net/ipv6/* Variables
1934 ==============================
1936 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1937 apply to IPv6 [XXX?].
1939 bindv6only - BOOLEAN
1940 Default value for IPV6_V6ONLY socket option,
1941 which restricts use of the IPv6 socket to IPv6 communication
1944 - TRUE: disable IPv4-mapped address feature
1945 - FALSE: enable IPv4-mapped address feature
1947 Default: FALSE (as specified in RFC3493)
1949 flowlabel_consistency - BOOLEAN
1950 Protect the consistency (and unicity) of flow label.
1951 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1959 auto_flowlabels - INTEGER
1960 Automatically generate flow labels based on a flow hash of the
1961 packet. This allows intermediate devices, such as routers, to
1962 identify packet flows for mechanisms like Equal Cost Multipath
1963 Routing (see RFC 6438).
1965 = ===========================================================
1966 0 automatic flow labels are completely disabled
1967 1 automatic flow labels are enabled by default, they can be
1968 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1970 2 automatic flow labels are allowed, they may be enabled on a
1971 per socket basis using the IPV6_AUTOFLOWLABEL socket option
1972 3 automatic flow labels are enabled and enforced, they cannot
1973 be disabled by the socket option
1974 = ===========================================================
1978 flowlabel_state_ranges - BOOLEAN
1979 Split the flow label number space into two ranges. 0-0x7FFFF is
1980 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1981 is reserved for stateless flow labels as described in RFC6437.
1988 flowlabel_reflect - INTEGER
1989 Control flow label reflection. Needed for Path MTU
1990 Discovery to work with Equal Cost Multipath Routing in anycast
1991 environments. See RFC 7690 and:
1992 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
1996 - 1: enabled for established flows
1998 Note that this prevents automatic flowlabel changes, as done
1999 in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
2000 and "tcp: Change txhash on every SYN and RTO retransmit"
2002 - 2: enabled for TCP RESET packets (no active listener)
2003 If set, a RST packet sent in response to a SYN packet on a closed
2004 port will reflect the incoming flow label.
2006 - 4: enabled for ICMPv6 echo reply messages.
2010 fib_multipath_hash_policy - INTEGER
2011 Controls which hash policy to use for multipath routes.
2013 Default: 0 (Layer 3)
2017 - 0 - Layer 3 (source and destination addresses plus flow label)
2018 - 1 - Layer 4 (standard 5-tuple)
2019 - 2 - Layer 3 or inner Layer 3 if present
2020 - 3 - Custom multipath hash. Fields used for multipath hash calculation
2021 are determined by fib_multipath_hash_fields sysctl
2023 fib_multipath_hash_fields - UNSIGNED INTEGER
2024 When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
2025 fields used for multipath hash calculation are determined by this
2028 This value is a bitmask which enables various fields for multipath hash
2031 Possible fields are:
2033 ====== ============================
2034 0x0001 Source IP address
2035 0x0002 Destination IP address
2039 0x0020 Destination port
2040 0x0040 Inner source IP address
2041 0x0080 Inner destination IP address
2042 0x0100 Inner IP protocol
2043 0x0200 Inner Flow Label
2044 0x0400 Inner source port
2045 0x0800 Inner destination port
2046 ====== ============================
2048 Default: 0x0007 (source IP, destination IP and IP protocol)
2050 anycast_src_echo_reply - BOOLEAN
2051 Controls the use of anycast addresses as source addresses for ICMPv6
2059 idgen_delay - INTEGER
2060 Controls the delay in seconds after which time to retry
2061 privacy stable address generation if a DAD conflict is
2064 Default: 1 (as specified in RFC7217)
2066 idgen_retries - INTEGER
2067 Controls the number of retries to generate a stable privacy
2068 address if a DAD conflict is detected.
2070 Default: 3 (as specified in RFC7217)
2073 Controls the MLD query robustness variable (see RFC3810 9.1).
2075 Default: 2 (as specified by RFC3810 9.1)
2077 Minimum: 1 (as specified by RFC6636 4.5)
2079 max_dst_opts_number - INTEGER
2080 Maximum number of non-padding TLVs allowed in a Destination
2081 options extension header. If this value is less than zero
2082 then unknown options are disallowed and the number of known
2083 TLVs allowed is the absolute value of this number.
2087 max_hbh_opts_number - INTEGER
2088 Maximum number of non-padding TLVs allowed in a Hop-by-Hop
2089 options extension header. If this value is less than zero
2090 then unknown options are disallowed and the number of known
2091 TLVs allowed is the absolute value of this number.
2095 max_dst_opts_length - INTEGER
2096 Maximum length allowed for a Destination options extension
2099 Default: INT_MAX (unlimited)
2101 max_hbh_length - INTEGER
2102 Maximum length allowed for a Hop-by-Hop options extension
2105 Default: INT_MAX (unlimited)
2107 skip_notify_on_dev_down - BOOLEAN
2108 Controls whether an RTM_DELROUTE message is generated for routes
2109 removed when a device is taken down or deleted. IPv4 does not
2110 generate this message; IPv6 does by default. Setting this sysctl
2111 to true skips the message, making IPv4 and IPv6 on par in relying
2112 on userspace caches to track link events and evict routes.
2114 Default: false (generate message)
2116 nexthop_compat_mode - BOOLEAN
2117 New nexthop API provides a means for managing nexthops independent of
2118 prefixes. Backwards compatibility with old route format is enabled by
2119 default which means route dumps and notifications contain the new
2120 nexthop attribute but also the full, expanded nexthop definition.
2121 Further, updates or deletes of a nexthop configuration generate route
2122 notifications for each fib entry using the nexthop. Once a system
2123 understands the new API, this sysctl can be disabled to achieve full
2124 performance benefits of the new API by disabling the nexthop expansion
2125 and extraneous notifications.
2126 Default: true (backward compat mode)
2128 fib_notify_on_flag_change - INTEGER
2129 Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
2130 RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
2132 After installing a route to the kernel, user space receives an
2133 acknowledgment, which means the route was installed in the kernel,
2134 but not necessarily in hardware.
2135 It is also possible for a route already installed in hardware to change
2136 its action and therefore its flags. For example, a host route that is
2137 trapping packets can be "promoted" to perform decapsulation following
2138 the installation of an IPinIP/VXLAN tunnel.
2139 The notifications will indicate to user-space the state of the route.
2141 Default: 0 (Do not emit notifications.)
2145 - 0 - Do not emit notifications.
2146 - 1 - Emit notifications.
2147 - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
2150 Define the IOAM id of this node. Uses only 24 bits out of 32 in total.
2157 ioam6_id_wide - LONG INTEGER
2158 Define the wide IOAM id of this node. Uses only 56 bits out of 64 in
2159 total. Can be different from ioam6_id.
2162 Max: 0xFFFFFFFFFFFFFF
2164 Default: 0xFFFFFFFFFFFFFF
2168 ip6frag_high_thresh - INTEGER
2169 Maximum memory used to reassemble IPv6 fragments. When
2170 ip6frag_high_thresh bytes of memory is allocated for this purpose,
2171 the fragment handler will toss packets until ip6frag_low_thresh
2174 ip6frag_low_thresh - INTEGER
2175 See ip6frag_high_thresh
2177 ip6frag_time - INTEGER
2178 Time in seconds to keep an IPv6 fragment in memory.
2181 Change the interface-specific default settings.
2183 These settings would be used during creating new interfaces.
2187 Change all the interface-specific settings.
2189 [XXX: Other special features than forwarding?]
2191 conf/all/disable_ipv6 - BOOLEAN
2192 Changing this value is same as changing ``conf/default/disable_ipv6``
2193 setting and also all per-interface ``disable_ipv6`` settings to the same
2196 Reading this value does not have any particular meaning. It does not say
2197 whether IPv6 support is enabled or disabled. Returned value can be 1
2198 also in the case when some interface has ``disable_ipv6`` set to 0 and
2199 has configured IPv6 addresses.
2201 conf/all/forwarding - BOOLEAN
2202 Enable global IPv6 forwarding between all interfaces.
2204 IPv4 and IPv6 work differently here; e.g. netfilter must be used
2205 to control which interfaces may forward packets and which not.
2207 This also sets all interfaces' Host/Router setting
2208 'forwarding' to the specified value. See below for details.
2210 This referred to as global forwarding.
2215 fwmark_reflect - BOOLEAN
2216 Controls the fwmark of kernel-generated IPv6 reply packets that are not
2217 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
2218 If unset, these packets have a fwmark of zero. If set, they have the
2219 fwmark of the packet they are replying to.
2223 ``conf/interface/*``:
2224 Change special settings per interface.
2226 The functional behaviour for certain settings is different
2227 depending on whether local forwarding is enabled or not.
2230 Accept Router Advertisements; autoconfigure using them.
2232 It also determines whether or not to transmit Router
2233 Solicitations. If and only if the functional setting is to
2234 accept Router Advertisements, Router Solicitations will be
2237 Possible values are:
2239 == ===========================================================
2240 0 Do not accept Router Advertisements.
2241 1 Accept Router Advertisements if forwarding is disabled.
2242 2 Overrule forwarding behaviour. Accept Router Advertisements
2243 even if forwarding is enabled.
2244 == ===========================================================
2248 - enabled if local forwarding is disabled.
2249 - disabled if local forwarding is enabled.
2251 accept_ra_defrtr - BOOLEAN
2252 Learn default router in Router Advertisement.
2256 - enabled if accept_ra is enabled.
2257 - disabled if accept_ra is disabled.
2259 ra_defrtr_metric - UNSIGNED INTEGER
2260 Route metric for default route learned in Router Advertisement. This value
2261 will be assigned as metric for the default route learned via IPv6 Router
2262 Advertisement. Takes affect only if accept_ra_defrtr is enabled.
2267 Default: IP6_RT_PRIO_USER i.e. 1024.
2269 accept_ra_from_local - BOOLEAN
2270 Accept RA with source-address that is found on local machine
2271 if the RA is otherwise proper and able to be accepted.
2273 Default is to NOT accept these as it may be an un-intended
2278 - enabled if accept_ra_from_local is enabled
2279 on a specific interface.
2280 - disabled if accept_ra_from_local is disabled
2281 on a specific interface.
2283 accept_ra_min_hop_limit - INTEGER
2284 Minimum hop limit Information in Router Advertisement.
2286 Hop limit Information in Router Advertisement less than this
2287 variable shall be ignored.
2291 accept_ra_min_lft - INTEGER
2292 Minimum acceptable lifetime value in Router Advertisement.
2294 RA sections with a lifetime less than this value shall be
2295 ignored. Zero lifetimes stay unaffected.
2299 accept_ra_pinfo - BOOLEAN
2300 Learn Prefix Information in Router Advertisement.
2304 - enabled if accept_ra is enabled.
2305 - disabled if accept_ra is disabled.
2307 accept_ra_rt_info_min_plen - INTEGER
2308 Minimum prefix length of Route Information in RA.
2310 Route Information w/ prefix smaller than this variable shall
2315 * 0 if accept_ra_rtr_pref is enabled.
2316 * -1 if accept_ra_rtr_pref is disabled.
2318 accept_ra_rt_info_max_plen - INTEGER
2319 Maximum prefix length of Route Information in RA.
2321 Route Information w/ prefix larger than this variable shall
2326 * 0 if accept_ra_rtr_pref is enabled.
2327 * -1 if accept_ra_rtr_pref is disabled.
2329 accept_ra_rtr_pref - BOOLEAN
2330 Accept Router Preference in RA.
2334 - enabled if accept_ra is enabled.
2335 - disabled if accept_ra is disabled.
2337 accept_ra_mtu - BOOLEAN
2338 Apply the MTU value specified in RA option 5 (RFC4861). If
2339 disabled, the MTU specified in the RA will be ignored.
2343 - enabled if accept_ra is enabled.
2344 - disabled if accept_ra is disabled.
2346 accept_redirects - BOOLEAN
2351 - enabled if local forwarding is disabled.
2352 - disabled if local forwarding is enabled.
2354 accept_source_route - INTEGER
2355 Accept source routing (routing extension header).
2357 - >= 0: Accept only routing header type 2.
2358 - < 0: Do not accept routing header.
2363 Autoconfigure addresses using Prefix Information in Router
2368 - enabled if accept_ra_pinfo is enabled.
2369 - disabled if accept_ra_pinfo is disabled.
2371 dad_transmits - INTEGER
2372 The amount of Duplicate Address Detection probes to send.
2376 forwarding - INTEGER
2377 Configure interface-specific Host/Router behaviour.
2381 It is recommended to have the same setting on all
2382 interfaces; mixed router/host scenarios are rather uncommon.
2384 Possible values are:
2386 - 0 Forwarding disabled
2387 - 1 Forwarding enabled
2391 By default, Host behaviour is assumed. This means:
2393 1. IsRouter flag is not set in Neighbour Advertisements.
2394 2. If accept_ra is TRUE (default), transmit Router
2396 3. If accept_ra is TRUE (default), accept Router
2397 Advertisements (and do autoconfiguration).
2398 4. If accept_redirects is TRUE (default), accept Redirects.
2402 If local forwarding is enabled, Router behaviour is assumed.
2403 This means exactly the reverse from the above:
2405 1. IsRouter flag is set in Neighbour Advertisements.
2406 2. Router Solicitations are not sent unless accept_ra is 2.
2407 3. Router Advertisements are ignored unless accept_ra is 2.
2408 4. Redirects are ignored.
2410 Default: 0 (disabled) if global forwarding is disabled (default),
2411 otherwise 1 (enabled).
2414 Default Hop Limit to set.
2419 Default Maximum Transfer Unit
2421 Default: 1280 (IPv6 required minimum)
2423 ip_nonlocal_bind - BOOLEAN
2424 If set, allows processes to bind() to non-local IPv6 addresses,
2425 which can be quite useful - but may break some applications.
2429 router_probe_interval - INTEGER
2430 Minimum interval (in seconds) between Router Probing described
2435 router_solicitation_delay - INTEGER
2436 Number of seconds to wait after interface is brought up
2437 before sending Router Solicitations.
2441 router_solicitation_interval - INTEGER
2442 Number of seconds to wait between Router Solicitations.
2446 router_solicitations - INTEGER
2447 Number of Router Solicitations to send until assuming no
2448 routers are present.
2452 use_oif_addrs_only - BOOLEAN
2453 When enabled, the candidate source addresses for destinations
2454 routed via this interface are restricted to the set of addresses
2455 configured on this interface (vis. RFC 6724, section 4).
2459 use_tempaddr - INTEGER
2460 Preference for Privacy Extensions (RFC3041).
2462 * <= 0 : disable Privacy Extensions
2463 * == 1 : enable Privacy Extensions, but prefer public
2464 addresses over temporary addresses.
2465 * > 1 : enable Privacy Extensions and prefer temporary
2466 addresses over public addresses.
2470 * 0 (for most devices)
2471 * -1 (for point-to-point devices and loopback devices)
2473 temp_valid_lft - INTEGER
2474 valid lifetime (in seconds) for temporary addresses.
2476 Default: 172800 (2 days)
2478 temp_prefered_lft - INTEGER
2479 Preferred lifetime (in seconds) for temporary addresses.
2481 Default: 86400 (1 day)
2483 keep_addr_on_down - INTEGER
2484 Keep all IPv6 addresses on an interface down event. If set static
2485 global addresses with no expiration time are not flushed.
2488 * 0 : system default
2491 Default: 0 (addresses are removed)
2493 max_desync_factor - INTEGER
2494 Maximum value for DESYNC_FACTOR, which is a random value
2495 that ensures that clients don't synchronize with each
2496 other and generate new addresses at exactly the same time.
2497 value is in seconds.
2501 regen_max_retry - INTEGER
2502 Number of attempts before give up attempting to generate
2503 valid temporary addresses.
2507 max_addresses - INTEGER
2508 Maximum number of autoconfigured addresses per interface. Setting
2509 to zero disables the limitation. It is not recommended to set this
2510 value too large (or to zero) because it would be an easy way to
2511 crash the kernel by allowing too many addresses to be created.
2515 disable_ipv6 - BOOLEAN
2516 Disable IPv6 operation. If accept_dad is set to 2, this value
2517 will be dynamically set to TRUE if DAD fails for the link-local
2520 Default: FALSE (enable IPv6 operation)
2522 When this value is changed from 1 to 0 (IPv6 is being enabled),
2523 it will dynamically create a link-local address on the given
2524 interface and start Duplicate Address Detection, if necessary.
2526 When this value is changed from 0 to 1 (IPv6 is being disabled),
2527 it will dynamically delete all addresses and routes on the given
2528 interface. From now on it will not possible to add addresses/routes
2529 to the selected interface.
2531 accept_dad - INTEGER
2532 Whether to accept DAD (Duplicate Address Detection).
2534 == ==============================================================
2536 1 Enable DAD (default)
2537 2 Enable DAD, and disable IPv6 operation if MAC-based duplicate
2538 link-local address has been found.
2539 == ==============================================================
2541 DAD operation and mode on a given interface will be selected according
2542 to the maximum value of conf/{all,interface}/accept_dad.
2544 force_tllao - BOOLEAN
2545 Enable sending the target link-layer address option even when
2546 responding to a unicast neighbor solicitation.
2550 Quoting from RFC 2461, section 4.4, Target link-layer address:
2552 "The option MUST be included for multicast solicitations in order to
2553 avoid infinite Neighbor Solicitation "recursion" when the peer node
2554 does not have a cache entry to return a Neighbor Advertisements
2555 message. When responding to unicast solicitations, the option can be
2556 omitted since the sender of the solicitation has the correct link-
2557 layer address; otherwise it would not have be able to send the unicast
2558 solicitation in the first place. However, including the link-layer
2559 address in this case adds little overhead and eliminates a potential
2560 race condition where the sender deletes the cached link-layer address
2561 prior to receiving a response to a previous solicitation."
2563 ndisc_notify - BOOLEAN
2564 Define mode for notification of address and device changes.
2566 * 0 - (default): do nothing
2567 * 1 - Generate unsolicited neighbour advertisements when device is brought
2568 up or hardware address changes.
2570 ndisc_tclass - INTEGER
2571 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
2572 Discovery (Router Solicitation, Router Advertisement, Neighbor
2573 Solicitation, Neighbor Advertisement, Redirect) messages.
2574 These 8 bits can be interpreted as 6 high order bits holding the DSCP
2575 value and 2 low order bits representing ECN (which you probably want
2580 ndisc_evict_nocarrier - BOOLEAN
2581 Clears the neighbor discovery table on NOCARRIER events. This option is
2582 important for wireless devices where the neighbor discovery cache should
2583 not be cleared when roaming between access points on the same network.
2584 In most cases this should remain as the default (1).
2586 - 1 - (default): Clear neighbor discover cache on NOCARRIER events.
2587 - 0 - Do not clear neighbor discovery cache on NOCARRIER events.
2589 mldv1_unsolicited_report_interval - INTEGER
2590 The interval in milliseconds in which the next unsolicited
2591 MLDv1 report retransmit will take place.
2593 Default: 10000 (10 seconds)
2595 mldv2_unsolicited_report_interval - INTEGER
2596 The interval in milliseconds in which the next unsolicited
2597 MLDv2 report retransmit will take place.
2599 Default: 1000 (1 second)
2601 force_mld_version - INTEGER
2602 * 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
2603 * 1 - Enforce to use MLD version 1
2604 * 2 - Enforce to use MLD version 2
2606 suppress_frag_ndisc - INTEGER
2607 Control RFC 6980 (Security Implications of IPv6 Fragmentation
2608 with IPv6 Neighbor Discovery) behavior:
2610 * 1 - (default) discard fragmented neighbor discovery packets
2611 * 0 - allow fragmented neighbor discovery packets
2613 optimistic_dad - BOOLEAN
2614 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
2616 * 0: disabled (default)
2619 Optimistic Duplicate Address Detection for the interface will be enabled
2620 if at least one of conf/{all,interface}/optimistic_dad is set to 1,
2621 it will be disabled otherwise.
2623 use_optimistic - BOOLEAN
2624 If enabled, do not classify optimistic addresses as deprecated during
2625 source address selection. Preferred addresses will still be chosen
2626 before optimistic addresses, subject to other ranking in the source
2627 address selection algorithm.
2629 * 0: disabled (default)
2632 This will be enabled if at least one of
2633 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
2635 stable_secret - IPv6 address
2636 This IPv6 address will be used as a secret to generate IPv6
2637 addresses for link-local addresses and autoconfigured
2638 ones. All addresses generated after setting this secret will
2639 be stable privacy ones by default. This can be changed via the
2640 addrgenmode ip-link. conf/default/stable_secret is used as the
2641 secret for the namespace, the interface specific ones can
2642 overwrite that. Writes to conf/all/stable_secret are refused.
2644 It is recommended to generate this secret during installation
2645 of a system and keep it stable after that.
2647 By default the stable secret is unset.
2649 addr_gen_mode - INTEGER
2650 Defines how link-local and autoconf addresses are generated.
2652 = =================================================================
2653 0 generate address based on EUI64 (default)
2654 1 do no generate a link-local address, use EUI64 for addresses
2655 generated from autoconf
2656 2 generate stable privacy addresses, using the secret from
2657 stable_secret (RFC7217)
2658 3 generate stable privacy addresses, using a random secret if unset
2659 = =================================================================
2661 drop_unicast_in_l2_multicast - BOOLEAN
2662 Drop any unicast IPv6 packets that are received in link-layer
2663 multicast (or broadcast) frames.
2665 By default this is turned off.
2667 drop_unsolicited_na - BOOLEAN
2668 Drop all unsolicited neighbor advertisements, for example if there's
2669 a known good NA proxy on the network and such frames need not be used
2670 (or in the case of 802.11, must not be used to prevent attacks.)
2672 By default this is turned off.
2674 accept_untracked_na - INTEGER
2675 Define behavior for accepting neighbor advertisements from devices that
2676 are absent in the neighbor cache:
2678 - 0 - (default) Do not accept unsolicited and untracked neighbor
2681 - 1 - Add a new neighbor cache entry in STALE state for routers on
2682 receiving a neighbor advertisement (either solicited or unsolicited)
2683 with target link-layer address option specified if no neighbor entry
2684 is already present for the advertised IPv6 address. Without this knob,
2685 NAs received for untracked addresses (absent in neighbor cache) are
2688 This is as per router-side behavior documented in RFC9131.
2690 This has lower precedence than drop_unsolicited_na.
2692 This will optimize the return path for the initial off-link
2693 communication that is initiated by a directly connected host, by
2694 ensuring that the first-hop router which turns on this setting doesn't
2695 have to buffer the initial return packets to do neighbor-solicitation.
2696 The prerequisite is that the host is configured to send unsolicited
2697 neighbor advertisements on interface bringup. This setting should be
2698 used in conjunction with the ndisc_notify setting on the host to
2699 satisfy this prerequisite.
2701 - 2 - Extend option (1) to add a new neighbor cache entry only if the
2702 source IP address is in the same subnet as an address configured on
2703 the interface that received the neighbor advertisement.
2705 enhanced_dad - BOOLEAN
2706 Include a nonce option in the IPv6 neighbor solicitation messages used for
2707 duplicate address detection per RFC7527. A received DAD NS will only signal
2708 a duplicate address if the nonce is different. This avoids any false
2709 detection of duplicates due to loopback of the NS messages that we send.
2710 The nonce option will be sent on an interface unless both of
2711 conf/{all,interface}/enhanced_dad are set to FALSE.
2719 Limit the maximal rates for sending ICMPv6 messages.
2721 0 to disable any limiting,
2722 otherwise the minimal space between responses in milliseconds.
2726 ratemask - list of comma separated ranges
2727 For ICMPv6 message types matching the ranges in the ratemask, limit
2728 the sending of the message according to ratelimit parameter.
2730 The format used for both input and output is a comma separated
2731 list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
2732 129). Writing to the file will clear all previous ranges of ICMPv6
2733 message types and update the current list with the input.
2735 Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
2736 for numerical values of ICMPv6 message types, e.g. echo request is 128
2737 and echo reply is 129.
2739 Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
2741 echo_ignore_all - BOOLEAN
2742 If set non-zero, then the kernel will ignore all ICMP ECHO
2743 requests sent to it over the IPv6 protocol.
2747 echo_ignore_multicast - BOOLEAN
2748 If set non-zero, then the kernel will ignore all ICMP ECHO
2749 requests sent to it over the IPv6 protocol via multicast.
2753 echo_ignore_anycast - BOOLEAN
2754 If set non-zero, then the kernel will ignore all ICMP ECHO
2755 requests sent to it over the IPv6 protocol destined to anycast address.
2759 error_anycast_as_unicast - BOOLEAN
2760 If set to 1, then the kernel will respond with ICMP Errors
2761 resulting from requests sent to it over the IPv6 protocol destined
2762 to anycast address essentially treating anycast as unicast.
2766 xfrm6_gc_thresh - INTEGER
2767 (Obsolete since linux-4.14)
2768 The threshold at which we will start garbage collecting for IPv6
2769 destination cache entries. At twice this value the system will
2770 refuse new allocations.
2774 Pekka Savola <pekkas@netcore.fi>
2775 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
2778 /proc/sys/net/bridge/* Variables:
2779 =================================
2781 bridge-nf-call-arptables - BOOLEAN
2782 - 1 : pass bridged ARP traffic to arptables' FORWARD chain.
2787 bridge-nf-call-iptables - BOOLEAN
2788 - 1 : pass bridged IPv4 traffic to iptables' chains.
2793 bridge-nf-call-ip6tables - BOOLEAN
2794 - 1 : pass bridged IPv6 traffic to ip6tables' chains.
2799 bridge-nf-filter-vlan-tagged - BOOLEAN
2800 - 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
2805 bridge-nf-filter-pppoe-tagged - BOOLEAN
2806 - 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
2811 bridge-nf-pass-vlan-input-dev - BOOLEAN
2812 - 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
2813 interface on the bridge and set the netfilter input device to the
2814 vlan. This allows use of e.g. "iptables -i br0.1" and makes the
2815 REDIRECT target work with vlan-on-top-of-bridge interfaces. When no
2816 matching vlan interface is found, or this switch is off, the input
2817 device is set to the bridge interface.
2819 - 0: disable bridge netfilter vlan interface lookup.
2823 ``proc/sys/net/sctp/*`` Variables:
2824 ==================================
2826 addip_enable - BOOLEAN
2827 Enable or disable extension of Dynamic Address Reconfiguration
2828 (ADD-IP) functionality specified in RFC5061. This extension provides
2829 the ability to dynamically add and remove new addresses for the SCTP
2832 1: Enable extension.
2834 0: Disable extension.
2839 Enable or disable pf (pf is short for potentially failed) state. A value
2840 of pf_retrans > path_max_retrans also disables pf state. That is, one of
2841 both pf_enable and pf_retrans > path_max_retrans can disable pf state.
2842 Since pf_retrans and path_max_retrans can be changed by userspace
2843 application, sometimes user expects to disable pf state by the value of
2844 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
2845 or path_max_retrans is changed by the user application, this pf state is
2846 enabled. As such, it is necessary to add this to dynamically enable
2847 and disable pf state. See:
2848 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
2858 Unset or enable/disable pf (pf is short for potentially failed) state
2859 exposure. Applications can control the exposure of the PF path state
2860 in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO
2861 sockopt. When it's unset, no SCTP_PEER_ADDR_CHANGE event with
2862 SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info
2863 can be got via SCTP_GET_PEER_ADDR_INFO sockopt; When it's enabled,
2864 a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming
2865 SCTP_PF state and a SCTP_PF-state transport info can be got via
2866 SCTP_GET_PEER_ADDR_INFO sockopt; When it's disabled, no
2867 SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when
2868 trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO
2871 0: Unset pf state exposure, Compatible with old applications.
2873 1: Disable pf state exposure.
2875 2: Enable pf state exposure.
2879 addip_noauth_enable - BOOLEAN
2880 Dynamic Address Reconfiguration (ADD-IP) requires the use of
2881 authentication to protect the operations of adding or removing new
2882 addresses. This requirement is mandated so that unauthorized hosts
2883 would not be able to hijack associations. However, older
2884 implementations may not have implemented this requirement while
2885 allowing the ADD-IP extension. For reasons of interoperability,
2886 we provide this variable to control the enforcement of the
2887 authentication requirement.
2889 == ===============================================================
2890 1 Allow ADD-IP extension to be used without authentication. This
2891 should only be set in a closed environment for interoperability
2892 with older implementations.
2894 0 Enforce the authentication requirement
2895 == ===============================================================
2899 auth_enable - BOOLEAN
2900 Enable or disable Authenticated Chunks extension. This extension
2901 provides the ability to send and receive authenticated chunks and is
2902 required for secure operation of Dynamic Address Reconfiguration
2905 - 1: Enable this extension.
2906 - 0: Disable this extension.
2910 prsctp_enable - BOOLEAN
2911 Enable or disable the Partial Reliability extension (RFC3758) which
2912 is used to notify peers that a given DATA should no longer be expected.
2914 - 1: Enable extension
2920 The limit of the number of new packets that can be initially sent. It
2921 controls how bursty the generated traffic can be.
2925 association_max_retrans - INTEGER
2926 Set the maximum number for retransmissions that an association can
2927 attempt deciding that the remote end is unreachable. If this value
2928 is exceeded, the association is terminated.
2932 max_init_retransmits - INTEGER
2933 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
2934 that an association will attempt before declaring the destination
2935 unreachable and terminating.
2939 path_max_retrans - INTEGER
2940 The maximum number of retransmissions that will be attempted on a given
2941 path. Once this threshold is exceeded, the path is considered
2942 unreachable, and new traffic will use a different path when the
2943 association is multihomed.
2947 pf_retrans - INTEGER
2948 The number of retransmissions that will be attempted on a given path
2949 before traffic is redirected to an alternate transport (should one
2950 exist). Note this is distinct from path_max_retrans, as a path that
2951 passes the pf_retrans threshold can still be used. Its only
2952 deprioritized when a transmission path is selected by the stack. This
2953 setting is primarily used to enable fast failover mechanisms without
2954 having to reduce path_max_retrans to a very low value. See:
2955 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
2956 for details. Note also that a value of pf_retrans > path_max_retrans
2957 disables this feature. Since both pf_retrans and path_max_retrans can
2958 be changed by userspace application, a variable pf_enable is used to
2963 ps_retrans - INTEGER
2964 Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
2965 from section-5 "Primary Path Switchover" in rfc7829. The primary path
2966 will be changed to another active path when the path error counter on
2967 the old primary path exceeds PSMR, so that "the SCTP sender is allowed
2968 to continue data transmission on a new working path even when the old
2969 primary destination address becomes active again". Note this feature
2970 is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
2971 and its value can't be less than 'pf_retrans' when changing by sysctl.
2975 rto_initial - INTEGER
2976 The initial round trip timeout value in milliseconds that will be used
2977 in calculating round trip times. This is the initial time interval
2978 for retransmissions.
2983 The maximum value (in milliseconds) of the round trip timeout. This
2984 is the largest time interval that can elapse between retransmissions.
2989 The minimum value (in milliseconds) of the round trip timeout. This
2990 is the smallest time interval the can elapse between retransmissions.
2994 hb_interval - INTEGER
2995 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
2996 are sent at the specified interval on idle paths to probe the state of
2997 a given path between 2 associations.
3001 sack_timeout - INTEGER
3002 The amount of time (in milliseconds) that the implementation will wait
3007 valid_cookie_life - INTEGER
3008 The default lifetime of the SCTP cookie (in milliseconds). The cookie
3009 is used during association establishment.
3013 cookie_preserve_enable - BOOLEAN
3014 Enable or disable the ability to extend the lifetime of the SCTP cookie
3015 that is used during the establishment phase of SCTP association
3017 - 1: Enable cookie lifetime extension.
3022 cookie_hmac_alg - STRING
3023 Select the hmac algorithm used when generating the cookie value sent by
3024 a listening sctp socket to a connecting client in the INIT-ACK chunk.
3031 Ability to assign md5 or sha1 as the selected alg is predicated on the
3032 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
3033 CONFIG_CRYPTO_SHA1).
3035 Default: Dependent on configuration. MD5 if available, else SHA1 if
3036 available, else none.
3038 rcvbuf_policy - INTEGER
3039 Determines if the receive buffer is attributed to the socket or to
3040 association. SCTP supports the capability to create multiple
3041 associations on a single socket. When using this capability, it is
3042 possible that a single stalled association that's buffering a lot
3043 of data may block other associations from delivering their data by
3044 consuming all of the receive buffer space. To work around this,
3045 the rcvbuf_policy could be set to attribute the receiver buffer space
3046 to each association instead of the socket. This prevents the described
3049 - 1: rcvbuf space is per association
3050 - 0: rcvbuf space is per socket
3054 sndbuf_policy - INTEGER
3055 Similar to rcvbuf_policy above, this applies to send buffer space.
3057 - 1: Send buffer is tracked per association
3058 - 0: Send buffer is tracked per socket.
3062 sctp_mem - vector of 3 INTEGERs: min, pressure, max
3063 Number of pages allowed for queueing by all SCTP sockets.
3065 min: Below this number of pages SCTP is not bothered about its
3066 memory appetite. When amount of memory allocated by SCTP exceeds
3067 this number, SCTP starts to moderate memory usage.
3069 pressure: This value was introduced to follow format of tcp_mem.
3071 max: Number of pages allowed for queueing by all SCTP sockets.
3073 Default is calculated at boot time from amount of available memory.
3075 sctp_rmem - vector of 3 INTEGERs: min, default, max
3076 Only the first value ("min") is used, "default" and "max" are
3079 min: Minimal size of receive buffer used by SCTP socket.
3080 It is guaranteed to each SCTP socket (but not association) even
3081 under moderate memory pressure.
3085 sctp_wmem - vector of 3 INTEGERs: min, default, max
3086 Only the first value ("min") is used, "default" and "max" are
3089 min: Minimum size of send buffer that can be used by SCTP sockets.
3090 It is guaranteed to each SCTP socket (but not association) even
3091 under moderate memory pressure.
3095 addr_scope_policy - INTEGER
3096 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
3098 - 0 - Disable IPv4 address scoping
3099 - 1 - Enable IPv4 address scoping
3100 - 2 - Follow draft but allow IPv4 private addresses
3101 - 3 - Follow draft but allow IPv4 link local addresses
3106 The listening port for the local UDP tunneling sock. Normally it's
3107 using the IANA-assigned UDP port number 9899 (sctp-tunneling).
3109 This UDP sock is used for processing the incoming UDP-encapsulated
3110 SCTP packets (from RFC6951), and shared by all applications in the
3111 same net namespace. This UDP sock will be closed when the value is
3114 The value will also be used to set the src port of the UDP header
3115 for the outgoing UDP-encapsulated SCTP packets. For the dest port,
3116 please refer to 'encap_port' below.
3120 encap_port - INTEGER
3121 The default remote UDP encapsulation port.
3123 This value is used to set the dest port of the UDP header for the
3124 outgoing UDP-encapsulated SCTP packets by default. Users can also
3125 change the value for each sock/asoc/transport by using setsockopt.
3126 For further information, please refer to RFC6951.
3128 Note that when connecting to a remote server, the client should set
3129 this to the port that the UDP tunneling sock on the peer server is
3130 listening to and the local UDP tunneling sock on the client also
3131 must be started. On the server, it would get the encap_port from
3132 the incoming packet's source port.
3136 plpmtud_probe_interval - INTEGER
3137 The time interval (in milliseconds) for the PLPMTUD probe timer,
3138 which is configured to expire after this period to receive an
3139 acknowledgment to a probe packet. This is also the time interval
3140 between the probes for the current pmtu when the probe search
3143 PLPMTUD will be disabled when 0 is set, and other values for it
3148 reconf_enable - BOOLEAN
3149 Enable or disable extension of Stream Reconfiguration functionality
3150 specified in RFC6525. This extension provides the ability to "reset"
3151 a stream, and it includes the Parameters of "Outgoing/Incoming SSN
3152 Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams".
3154 - 1: Enable extension.
3155 - 0: Disable extension.
3159 intl_enable - BOOLEAN
3160 Enable or disable extension of User Message Interleaving functionality
3161 specified in RFC8260. This extension allows the interleaving of user
3162 messages sent on different streams. With this feature enabled, I-DATA
3163 chunk will replace DATA chunk to carry user messages if also supported
3164 by the peer. Note that to use this feature, one needs to set this option
3165 to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2
3166 and SCTP_INTERLEAVING_SUPPORTED to 1.
3168 - 1: Enable extension.
3169 - 0: Disable extension.
3173 ecn_enable - BOOLEAN
3174 Control use of Explicit Congestion Notification (ECN) by SCTP.
3175 Like in TCP, ECN is used only when both ends of the SCTP connection
3176 indicate support for it. This feature is useful in avoiding losses
3177 due to congestion by allowing supporting routers to signal congestion
3178 before having to drop packets.
3185 l3mdev_accept - BOOLEAN
3186 Enabling this option allows a "global" bound socket to work
3187 across L3 master domains (e.g., VRFs) with packets capable of
3188 being received regardless of the L3 domain in which they
3189 originated. Only valid when the kernel was compiled with
3190 CONFIG_NET_L3_MASTER_DEV.
3192 Default: 1 (enabled)
3195 ``/proc/sys/net/core/*``
3196 ========================
3198 Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
3201 ``/proc/sys/net/unix/*``
3202 ========================
3204 max_dgram_qlen - INTEGER
3205 The maximum length of dgram socket receive queue