1 /proc/sys/net/ipv4/* Variables:
7 Forward Packets between interfaces.
9 This variable is special, its change resets all configuration
10 parameters to their default state (RFC1122 for hosts, RFC1812
13 ip_default_ttl - INTEGER
14 Default value of TTL field (Time To Live) for outgoing (but not
15 forwarded) IP packets. Should be between 1 and 255 inclusive.
16 Default: 64 (as recommended by RFC1700)
18 ip_no_pmtu_disc - INTEGER
19 Disable Path MTU Discovery. If enabled in mode 1 and a
20 fragmentation-required ICMP is received, the PMTU to this
21 destination will be set to min_pmtu (see below). You will need
22 to raise min_pmtu to the smallest interface MTU on your system
23 manually if you want to avoid locally generated fragments.
25 In mode 2 incoming Path MTU Discovery messages will be
26 discarded. Outgoing frames are handled the same as in mode 1,
27 implicitly setting IP_PMTUDISC_DONT on every created socket.
29 Mode 3 is a hardend pmtu discover mode. The kernel will only
30 accept fragmentation-needed errors if the underlying protocol
31 can verify them besides a plain socket lookup. Current
32 protocols for which pmtu events will be honored are TCP, SCTP
33 and DCCP as they verify e.g. the sequence number or the
34 association. This mode should not be enabled globally but is
35 only intended to secure e.g. name servers in namespaces where
36 TCP path mtu must still work but path MTU information of other
37 protocols should be discarded. If enabled globally this mode
38 could break other protocols.
44 default 552 - minimum discovered Path MTU
46 ip_forward_use_pmtu - BOOLEAN
47 By default we don't trust protocol path MTUs while forwarding
48 because they could be easily forged and can lead to unwanted
49 fragmentation by the router.
50 You only need to enable this if you have user-space software
51 which tries to discover path mtus by itself and depends on the
52 kernel honoring this information. This is normally not the
59 fwmark_reflect - BOOLEAN
60 Controls the fwmark of kernel-generated IPv4 reply packets that are not
61 associated with a socket for example, TCP RSTs or ICMP echo replies).
62 If unset, these packets have a fwmark of zero. If set, they have the
63 fwmark of the packet they are replying to.
66 route/max_size - INTEGER
67 Maximum number of routes allowed in the kernel. Increase
68 this when using large numbers of interfaces and/or routes.
69 From linux kernel 3.6 onwards, this is deprecated for ipv4
70 as route cache is no longer used.
72 neigh/default/gc_thresh1 - INTEGER
73 Minimum number of entries to keep. Garbage collector will not
74 purge entries if there are fewer than this number.
77 neigh/default/gc_thresh2 - INTEGER
78 Threshold when garbage collector becomes more aggressive about
79 purging entries. Entries older than 5 seconds will be cleared
80 when over this number.
83 neigh/default/gc_thresh3 - INTEGER
84 Maximum number of neighbor entries allowed. Increase this
85 when using large numbers of interfaces and when communicating
86 with large numbers of directly-connected peers.
89 neigh/default/unres_qlen_bytes - INTEGER
90 The maximum number of bytes which may be used by packets
91 queued for each unresolved address by other network layers.
93 Setting negative value is meaningless and will return error.
94 Default: 65536 Bytes(64KB)
96 neigh/default/unres_qlen - INTEGER
97 The maximum number of packets which may be queued for each
98 unresolved address by other network layers.
99 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
100 Prior to linux 3.3, the default value is 3 which may cause
101 unexpected packet loss. The current default value is calculated
102 according to default value of unres_qlen_bytes and true size of
106 mtu_expires - INTEGER
107 Time, in seconds, that cached PMTU information is kept.
109 min_adv_mss - INTEGER
110 The advertised MSS depends on the first hop route MTU, but will
111 never be lower than this setting.
115 ipfrag_high_thresh - INTEGER
116 Maximum memory used to reassemble IP fragments. When
117 ipfrag_high_thresh bytes of memory is allocated for this purpose,
118 the fragment handler will toss packets until ipfrag_low_thresh
119 is reached. This also serves as a maximum limit to namespaces
120 different from the initial one.
122 ipfrag_low_thresh - INTEGER
123 Maximum memory used to reassemble IP fragments before the kernel
124 begins to remove incomplete fragment queues to free up resources.
125 The kernel still accepts new fragments for defragmentation.
127 ipfrag_time - INTEGER
128 Time in seconds to keep an IP fragment in memory.
130 ipfrag_max_dist - INTEGER
131 ipfrag_max_dist is a non-negative integer value which defines the
132 maximum "disorder" which is allowed among fragments which share a
133 common IP source address. Note that reordering of packets is
134 not unusual, but if a large number of fragments arrive from a source
135 IP address while a particular fragment queue remains incomplete, it
136 probably indicates that one or more fragments belonging to that queue
137 have been lost. When ipfrag_max_dist is positive, an additional check
138 is done on fragments before they are added to a reassembly queue - if
139 ipfrag_max_dist (or more) fragments have arrived from a particular IP
140 address between additions to any IP fragment queue using that source
141 address, it's presumed that one or more fragments in the queue are
142 lost. The existing fragment queue will be dropped, and a new one
143 started. An ipfrag_max_dist value of zero disables this check.
145 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
146 result in unnecessarily dropping fragment queues when normal
147 reordering of packets occurs, which could lead to poor application
148 performance. Using a very large value, e.g. 50000, increases the
149 likelihood of incorrectly reassembling IP fragments that originate
150 from different IP datagrams, which could result in data corruption.
155 inet_peer_threshold - INTEGER
156 The approximate size of the storage. Starting from this threshold
157 entries will be thrown aggressively. This threshold also determines
158 entries' time-to-live and time intervals between garbage collection
159 passes. More entries, less time-to-live, less GC interval.
161 inet_peer_minttl - INTEGER
162 Minimum time-to-live of entries. Should be enough to cover fragment
163 time-to-live on the reassembling side. This minimum time-to-live is
164 guaranteed if the pool size is less than inet_peer_threshold.
167 inet_peer_maxttl - INTEGER
168 Maximum time-to-live of entries. Unused entries will expire after
169 this period of time if there is no memory pressure on the pool (i.e.
170 when the number of entries in the pool is very small).
176 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
177 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
180 tcp_abort_on_overflow - BOOLEAN
181 If listening service is too slow to accept new connections,
182 reset them. Default state is FALSE. It means that if overflow
183 occurred due to a burst, connection will recover. Enable this
184 option _only_ if you are really sure that listening daemon
185 cannot be tuned to accept connections faster. Enabling this
186 option can harm clients of your server.
188 tcp_adv_win_scale - INTEGER
189 Count buffering overhead as bytes/2^tcp_adv_win_scale
190 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
192 Possible values are [-31, 31], inclusive.
195 tcp_allowed_congestion_control - STRING
196 Show/set the congestion control choices available to non-privileged
197 processes. The list is a subset of those listed in
198 tcp_available_congestion_control.
199 Default is "reno" and the default setting (tcp_congestion_control).
201 tcp_app_win - INTEGER
202 Reserve max(window/2^tcp_app_win, mss) of window for application
203 buffer. Value 0 is special, it means that nothing is reserved.
206 tcp_autocorking - BOOLEAN
207 Enable TCP auto corking :
208 When applications do consecutive small write()/sendmsg() system calls,
209 we try to coalesce these small writes as much as possible, to lower
210 total amount of sent packets. This is done if at least one prior
211 packet for the flow is waiting in Qdisc queues or device transmit
212 queue. Applications can still use TCP_CORK for optimal behavior
213 when they know how/when to uncork their sockets.
216 tcp_available_congestion_control - STRING
217 Shows the available congestion control choices that are registered.
218 More congestion control algorithms may be available as modules,
221 tcp_base_mss - INTEGER
222 The initial value of search_low to be used by the packetization layer
223 Path MTU discovery (MTU probing). If MTU probing is enabled,
224 this is the initial MSS used by the connection.
226 tcp_congestion_control - STRING
227 Set the congestion control algorithm to be used for new
228 connections. The algorithm "reno" is always available, but
229 additional choices may be available based on kernel configuration.
230 Default is set as part of kernel configuration.
231 For passive connections, the listener congestion control choice
233 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
236 Allows TCP to send "duplicate" SACKs.
238 tcp_early_retrans - INTEGER
239 Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold
240 for triggering fast retransmit when the amount of outstanding data is
241 small and when no previously unsent data can be transmitted (such
242 that limited transmit could be used). Also controls the use of
243 Tail loss probe (TLP) that converts RTOs occurring due to tail
244 losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01).
248 2 enables ER but delays fast recovery and fast retransmit
249 by a fourth of RTT. This mitigates connection falsely
250 recovers when network has a small degree of reordering
251 (less than 3 packets).
252 3 enables delayed ER and TLP.
257 Control use of Explicit Congestion Notification (ECN) by TCP.
258 ECN is used only when both ends of the TCP connection indicate
259 support for it. This feature is useful in avoiding losses due
260 to congestion by allowing supporting routers to signal
261 congestion before having to drop packets.
263 0 Disable ECN. Neither initiate nor accept ECN.
264 1 Enable ECN when requested by incoming connections and
265 also request ECN on outgoing connection attempts.
266 2 Enable ECN when requested by incoming connections
267 but do not request ECN on outgoing connections.
270 tcp_ecn_fallback - BOOLEAN
271 If the kernel detects that ECN connection misbehaves, enable fall
272 back to non-ECN. Currently, this knob implements the fallback
273 from RFC3168, section 6.1.1.1., but we reserve that in future,
274 additional detection mechanisms could be implemented under this
275 knob. The value is not used, if tcp_ecn or per route (or congestion
276 control) ECN settings are disabled.
277 Default: 1 (fallback enabled)
280 Enable FACK congestion avoidance and fast retransmission.
281 The value is not used, if tcp_sack is not enabled.
283 tcp_fin_timeout - INTEGER
284 The length of time an orphaned (no longer referenced by any
285 application) connection will remain in the FIN_WAIT_2 state
286 before it is aborted at the local end. While a perfectly
287 valid "receive only" state for an un-orphaned connection, an
288 orphaned connection in FIN_WAIT_2 state could otherwise wait
289 forever for the remote to close its end of the connection.
294 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
295 F-RTO is an enhanced recovery algorithm for TCP retransmission
296 timeouts. It is particularly beneficial in networks where the
297 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
298 modification. It does not require any support from the peer.
300 By default it's enabled with a non-zero value. 0 disables F-RTO.
302 tcp_invalid_ratelimit - INTEGER
303 Limit the maximal rate for sending duplicate acknowledgments
304 in response to incoming TCP packets that are for an existing
305 connection but that are invalid due to any of these reasons:
307 (a) out-of-window sequence number,
308 (b) out-of-window acknowledgment number, or
309 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
311 This can help mitigate simple "ack loop" DoS attacks, wherein
312 a buggy or malicious middlebox or man-in-the-middle can
313 rewrite TCP header fields in manner that causes each endpoint
314 to think that the other is sending invalid TCP segments, thus
315 causing each side to send an unterminating stream of duplicate
316 acknowledgments for invalid segments.
318 Using 0 disables rate-limiting of dupacks in response to
319 invalid segments; otherwise this value specifies the minimal
320 space between sending such dupacks, in milliseconds.
322 Default: 500 (milliseconds).
324 tcp_keepalive_time - INTEGER
325 How often TCP sends out keepalive messages when keepalive is enabled.
328 tcp_keepalive_probes - INTEGER
329 How many keepalive probes TCP sends out, until it decides that the
330 connection is broken. Default value: 9.
332 tcp_keepalive_intvl - INTEGER
333 How frequently the probes are send out. Multiplied by
334 tcp_keepalive_probes it is time to kill not responding connection,
335 after probes started. Default value: 75sec i.e. connection
336 will be aborted after ~11 minutes of retries.
338 tcp_low_latency - BOOLEAN
339 If set, the TCP stack makes decisions that prefer lower
340 latency as opposed to higher throughput. By default, this
341 option is not set meaning that higher throughput is preferred.
342 An example of an application where this default should be
343 changed would be a Beowulf compute cluster.
346 tcp_max_orphans - INTEGER
347 Maximal number of TCP sockets not attached to any user file handle,
348 held by system. If this number is exceeded orphaned connections are
349 reset immediately and warning is printed. This limit exists
350 only to prevent simple DoS attacks, you _must_ not rely on this
351 or lower the limit artificially, but rather increase it
352 (probably, after increasing installed memory),
353 if network conditions require more than default value,
354 and tune network services to linger and kill such states
355 more aggressively. Let me to remind again: each orphan eats
356 up to ~64K of unswappable memory.
358 tcp_max_syn_backlog - INTEGER
359 Maximal number of remembered connection requests, which have not
360 received an acknowledgment from connecting client.
361 The minimal value is 128 for low memory machines, and it will
362 increase in proportion to the memory of machine.
363 If server suffers from overload, try increasing this number.
365 tcp_max_tw_buckets - INTEGER
366 Maximal number of timewait sockets held by system simultaneously.
367 If this number is exceeded time-wait socket is immediately destroyed
368 and warning is printed. This limit exists only to prevent
369 simple DoS attacks, you _must_ not lower the limit artificially,
370 but rather increase it (probably, after increasing installed memory),
371 if network conditions require more than default value.
373 tcp_mem - vector of 3 INTEGERs: min, pressure, max
374 min: below this number of pages TCP is not bothered about its
377 pressure: when amount of memory allocated by TCP exceeds this number
378 of pages, TCP moderates its memory consumption and enters memory
379 pressure mode, which is exited when memory consumption falls
382 max: number of pages allowed for queueing by all TCP sockets.
384 Defaults are calculated at boot time from amount of available
387 tcp_min_rtt_wlen - INTEGER
388 The window length of the windowed min filter to track the minimum RTT.
389 A shorter window lets a flow more quickly pick up new (higher)
390 minimum RTT when it is moved to a longer path (e.g., due to traffic
391 engineering). A longer window makes the filter more resistant to RTT
392 inflations such as transient congestion. The unit is seconds.
395 tcp_moderate_rcvbuf - BOOLEAN
396 If set, TCP performs receive buffer auto-tuning, attempting to
397 automatically size the buffer (no greater than tcp_rmem[2]) to
398 match the size required by the path for full throughput. Enabled by
401 tcp_mtu_probing - INTEGER
402 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
405 1 - Disabled by default, enabled when an ICMP black hole detected
406 2 - Always enabled, use initial MSS of tcp_base_mss.
408 tcp_probe_interval - INTEGER
409 Controls how often to start TCP Packetization-Layer Path MTU
410 Discovery reprobe. The default is reprobing every 10 minutes as
413 tcp_probe_threshold - INTEGER
414 Controls when TCP Packetization-Layer Path MTU Discovery probing
415 will stop in respect to the width of search range in bytes. Default
418 tcp_no_metrics_save - BOOLEAN
419 By default, TCP saves various connection metrics in the route cache
420 when the connection closes, so that connections established in the
421 near future can use these to set initial conditions. Usually, this
422 increases overall performance, but may sometimes cause performance
423 degradation. If set, TCP will not cache metrics on closing
426 tcp_orphan_retries - INTEGER
427 This value influences the timeout of a locally closed TCP connection,
428 when RTO retransmissions remain unacknowledged.
429 See tcp_retries2 for more details.
431 The default value is 8.
432 If your machine is a loaded WEB server,
433 you should think about lowering this value, such sockets
434 may consume significant resources. Cf. tcp_max_orphans.
436 tcp_recovery - INTEGER
437 This value is a bitmap to enable various experimental loss recovery
440 RACK: 0x1 enables the RACK loss detection for fast detection of lost
441 retransmissions and tail drops.
445 tcp_reordering - INTEGER
446 Initial reordering level of packets in a TCP stream.
447 TCP stack can then dynamically adjust flow reordering level
448 between this initial value and tcp_max_reordering
451 tcp_max_reordering - INTEGER
452 Maximal reordering level of packets in a TCP stream.
453 300 is a fairly conservative value, but you might increase it
454 if paths are using per packet load balancing (like bonding rr mode)
457 tcp_retrans_collapse - BOOLEAN
458 Bug-to-bug compatibility with some broken printers.
459 On retransmit try to send bigger packets to work around bugs in
462 tcp_retries1 - INTEGER
463 This value influences the time, after which TCP decides, that
464 something is wrong due to unacknowledged RTO retransmissions,
465 and reports this suspicion to the network layer.
466 See tcp_retries2 for more details.
468 RFC 1122 recommends at least 3 retransmissions, which is the
471 tcp_retries2 - INTEGER
472 This value influences the timeout of an alive TCP connection,
473 when RTO retransmissions remain unacknowledged.
474 Given a value of N, a hypothetical TCP connection following
475 exponential backoff with an initial RTO of TCP_RTO_MIN would
476 retransmit N times before killing the connection at the (N+1)th RTO.
478 The default value of 15 yields a hypothetical timeout of 924.6
479 seconds and is a lower bound for the effective timeout.
480 TCP will effectively time out at the first RTO which exceeds the
481 hypothetical timeout.
483 RFC 1122 recommends at least 100 seconds for the timeout,
484 which corresponds to a value of at least 8.
486 tcp_rfc1337 - BOOLEAN
487 If set, the TCP stack behaves conforming to RFC1337. If unset,
488 we are not conforming to RFC, but prevent TCP TIME_WAIT
492 tcp_rmem - vector of 3 INTEGERs: min, default, max
493 min: Minimal size of receive buffer used by TCP sockets.
494 It is guaranteed to each TCP socket, even under moderate memory
498 default: initial size of receive buffer used by TCP sockets.
499 This value overrides net.core.rmem_default used by other protocols.
500 Default: 87380 bytes. This value results in window of 65535 with
501 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
502 less for default tcp_app_win. See below about these variables.
504 max: maximal size of receive buffer allowed for automatically
505 selected receiver buffers for TCP socket. This value does not override
506 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
507 automatic tuning of that socket's receive buffer size, in which
508 case this value is ignored.
509 Default: between 87380B and 6MB, depending on RAM size.
512 Enable select acknowledgments (SACKS).
514 tcp_slow_start_after_idle - BOOLEAN
515 If set, provide RFC2861 behavior and time out the congestion
516 window after an idle period. An idle period is defined at
517 the current RTO. If unset, the congestion window will not
518 be timed out after an idle period.
522 Use the Host requirements interpretation of the TCP urgent pointer field.
523 Most hosts use the older BSD interpretation, so if you turn this on
524 Linux might not communicate correctly with them.
527 tcp_synack_retries - INTEGER
528 Number of times SYNACKs for a passive TCP connection attempt will
529 be retransmitted. Should not be higher than 255. Default value
530 is 5, which corresponds to 31seconds till the last retransmission
531 with the current initial RTO of 1second. With this the final timeout
532 for a passive TCP connection will happen after 63seconds.
534 tcp_syncookies - BOOLEAN
535 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
536 Send out syncookies when the syn backlog queue of a socket
537 overflows. This is to prevent against the common 'SYN flood attack'
540 Note, that syncookies is fallback facility.
541 It MUST NOT be used to help highly loaded servers to stand
542 against legal connection rate. If you see SYN flood warnings
543 in your logs, but investigation shows that they occur
544 because of overload with legal connections, you should tune
545 another parameters until this warning disappear.
546 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
548 syncookies seriously violate TCP protocol, do not allow
549 to use TCP extensions, can result in serious degradation
550 of some services (f.e. SMTP relaying), visible not by you,
551 but your clients and relays, contacting you. While you see
552 SYN flood warnings in logs not being really flooded, your server
553 is seriously misconfigured.
555 If you want to test which effects syncookies have to your
556 network connections you can set this knob to 2 to enable
557 unconditionally generation of syncookies.
559 tcp_fastopen - INTEGER
560 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
561 in the opening SYN packet. To use this feature, the client application
562 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
563 connect() to perform a TCP handshake automatically.
565 The values (bitmap) are
566 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN.
567 2: Enables TCP Fast Open on the server side, i.e., allowing data in
568 a SYN packet to be accepted and passed to the application before
569 3-way hand shake finishes.
570 4: Send data in the opening SYN regardless of cookie availability and
571 without a cookie option.
572 0x100: Accept SYN data w/o validating the cookie.
573 0x200: Accept data-in-SYN w/o any cookie option present.
574 0x400/0x800: Enable Fast Open on all listeners regardless of the
575 TCP_FASTOPEN socket option. The two different flags designate two
576 different ways of setting max_qlen without the TCP_FASTOPEN socket
581 Note that the client & server side Fast Open flags (1 and 2
582 respectively) must be also enabled before the rest of flags can take
585 See include/net/tcp.h and the code for more details.
587 tcp_fwmark_accept - BOOLEAN
588 If set, incoming connections to listening sockets that do not have a
589 socket mark will set the mark of the accepting socket to the fwmark of
590 the incoming SYN packet. This will cause all packets on that connection
591 (starting from the first SYNACK) to be sent with that fwmark. The
592 listening socket's mark is unchanged. Listening sockets that already
593 have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
597 tcp_syn_retries - INTEGER
598 Number of times initial SYNs for an active TCP connection attempt
599 will be retransmitted. Should not be higher than 255. Default value
600 is 6, which corresponds to 63seconds till the last retransmission
601 with the current initial RTO of 1second. With this the final timeout
602 for an active TCP connection attempt will happen after 127seconds.
604 tcp_timestamps - BOOLEAN
605 Enable timestamps as defined in RFC1323.
607 tcp_min_tso_segs - INTEGER
608 Minimal number of segments per TSO frame.
609 Since linux-3.12, TCP does an automatic sizing of TSO frames,
610 depending on flow rate, instead of filling 64Kbytes packets.
611 For specific usages, it's possible to force TCP to build big
612 TSO frames. Note that TCP stack might split too big TSO packets
613 if available window is too small.
616 tcp_pacing_ss_ratio - INTEGER
617 sk->sk_pacing_rate is set by TCP stack using a ratio applied
618 to current rate. (current_rate = cwnd * mss / srtt)
619 If TCP is in slow start, tcp_pacing_ss_ratio is applied
620 to let TCP probe for bigger speeds, assuming cwnd can be
621 doubled every other RTT.
624 tcp_pacing_ca_ratio - INTEGER
625 sk->sk_pacing_rate is set by TCP stack using a ratio applied
626 to current rate. (current_rate = cwnd * mss / srtt)
627 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
628 is applied to conservatively probe for bigger throughput.
631 tcp_tso_win_divisor - INTEGER
632 This allows control over what percentage of the congestion window
633 can be consumed by a single TSO frame.
634 The setting of this parameter is a choice between burstiness and
635 building larger TSO frames.
638 tcp_tw_recycle - BOOLEAN
639 Enable fast recycling TIME-WAIT sockets. Default value is 0.
640 It should not be changed without advice/request of technical
643 tcp_tw_reuse - BOOLEAN
644 Allow to reuse TIME-WAIT sockets for new connections when it is
645 safe from protocol viewpoint. Default value is 0.
646 It should not be changed without advice/request of technical
649 tcp_window_scaling - BOOLEAN
650 Enable window scaling as defined in RFC1323.
652 tcp_wmem - vector of 3 INTEGERs: min, default, max
653 min: Amount of memory reserved for send buffers for TCP sockets.
654 Each TCP socket has rights to use it due to fact of its birth.
657 default: initial size of send buffer used by TCP sockets. This
658 value overrides net.core.wmem_default used by other protocols.
659 It is usually lower than net.core.wmem_default.
662 max: Maximal amount of memory allowed for automatically tuned
663 send buffers for TCP sockets. This value does not override
664 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
665 automatic tuning of that socket's send buffer size, in which case
666 this value is ignored.
667 Default: between 64K and 4MB, depending on RAM size.
669 tcp_notsent_lowat - UNSIGNED INTEGER
670 A TCP socket can control the amount of unsent bytes in its write queue,
671 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
672 reports POLLOUT events if the amount of unsent bytes is below a per
673 socket value, and if the write queue is not full. sendmsg() will
674 also not add new buffers if the limit is hit.
676 This global variable controls the amount of unsent data for
677 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
678 to the global variable has immediate effect.
680 Default: UINT_MAX (0xFFFFFFFF)
682 tcp_workaround_signed_windows - BOOLEAN
683 If set, assume no receipt of a window scaling option means the
684 remote TCP is broken and treats the window as a signed quantity.
685 If unset, assume the remote TCP is not broken even if we do
686 not receive a window scaling option from them.
689 tcp_thin_linear_timeouts - BOOLEAN
690 Enable dynamic triggering of linear timeouts for thin streams.
691 If set, a check is performed upon retransmission by timeout to
692 determine if the stream is thin (less than 4 packets in flight).
693 As long as the stream is found to be thin, up to 6 linear
694 timeouts may be performed before exponential backoff mode is
695 initiated. This improves retransmission latency for
696 non-aggressive thin streams, often found to be time-dependent.
697 For more information on thin streams, see
698 Documentation/networking/tcp-thin.txt
701 tcp_thin_dupack - BOOLEAN
702 Enable dynamic triggering of retransmissions after one dupACK
703 for thin streams. If set, a check is performed upon reception
704 of a dupACK to determine if the stream is thin (less than 4
705 packets in flight). As long as the stream is found to be thin,
706 data is retransmitted on the first received dupACK. This
707 improves retransmission latency for non-aggressive thin
708 streams, often found to be time-dependent.
709 For more information on thin streams, see
710 Documentation/networking/tcp-thin.txt
713 tcp_limit_output_bytes - INTEGER
714 Controls TCP Small Queue limit per tcp socket.
715 TCP bulk sender tends to increase packets in flight until it
716 gets losses notifications. With SNDBUF autotuning, this can
717 result in a large amount of packets queued in qdisc/device
718 on the local machine, hurting latency of other flows, for
719 typical pfifo_fast qdiscs.
720 tcp_limit_output_bytes limits the number of bytes on qdisc
721 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
724 tcp_challenge_ack_limit - INTEGER
725 Limits number of Challenge ACK sent per second, as recommended
726 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
731 udp_mem - vector of 3 INTEGERs: min, pressure, max
732 Number of pages allowed for queueing by all UDP sockets.
734 min: Below this number of pages UDP is not bothered about its
735 memory appetite. When amount of memory allocated by UDP exceeds
736 this number, UDP starts to moderate memory usage.
738 pressure: This value was introduced to follow format of tcp_mem.
740 max: Number of pages allowed for queueing by all UDP sockets.
742 Default is calculated at boot time from amount of available memory.
744 udp_rmem_min - INTEGER
745 Minimal size of receive buffer used by UDP sockets in moderation.
746 Each UDP socket is able to use the size for receiving data, even if
747 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
750 udp_wmem_min - INTEGER
751 Minimal size of send buffer used by UDP sockets in moderation.
752 Each UDP socket is able to use the size for sending data, even if
753 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
758 cipso_cache_enable - BOOLEAN
759 If set, enable additions to and lookups from the CIPSO label mapping
760 cache. If unset, additions are ignored and lookups always result in a
761 miss. However, regardless of the setting the cache is still
762 invalidated when required when means you can safely toggle this on and
763 off and the cache will always be "safe".
766 cipso_cache_bucket_size - INTEGER
767 The CIPSO label cache consists of a fixed size hash table with each
768 hash bucket containing a number of cache entries. This variable limits
769 the number of entries in each hash bucket; the larger the value the
770 more CIPSO label mappings that can be cached. When the number of
771 entries in a given hash bucket reaches this limit adding new entries
772 causes the oldest entry in the bucket to be removed to make room.
775 cipso_rbm_optfmt - BOOLEAN
776 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
777 the CIPSO draft specification (see Documentation/netlabel for details).
778 This means that when set the CIPSO tag will be padded with empty
779 categories in order to make the packet data 32-bit aligned.
782 cipso_rbm_structvalid - BOOLEAN
783 If set, do a very strict check of the CIPSO option when
784 ip_options_compile() is called. If unset, relax the checks done during
785 ip_options_compile(). Either way is "safe" as errors are caught else
786 where in the CIPSO processing code but setting this to 0 (False) should
787 result in less work (i.e. it should be faster) but could cause problems
788 with other implementations that require strict checking.
793 ip_local_port_range - 2 INTEGERS
794 Defines the local port range that is used by TCP and UDP to
795 choose the local port. The first number is the first, the
796 second the last local port number.
797 If possible, it is better these numbers have different parity.
798 (one even and one odd values)
799 The default values are 32768 and 60999 respectively.
801 ip_local_reserved_ports - list of comma separated ranges
802 Specify the ports which are reserved for known third-party
803 applications. These ports will not be used by automatic port
804 assignments (e.g. when calling connect() or bind() with port
805 number 0). Explicit port allocation behavior is unchanged.
807 The format used for both input and output is a comma separated
808 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
809 10). Writing to the file will clear all previously reserved
810 ports and update the current list with the one given in the
813 Note that ip_local_port_range and ip_local_reserved_ports
814 settings are independent and both are considered by the kernel
815 when determining which ports are available for automatic port
818 You can reserve ports which are not in the current
819 ip_local_port_range, e.g.:
821 $ cat /proc/sys/net/ipv4/ip_local_port_range
823 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
826 although this is redundant. However such a setting is useful
827 if later the port range is changed to a value that will
828 include the reserved ports.
832 ip_nonlocal_bind - BOOLEAN
833 If set, allows processes to bind() to non-local IP addresses,
834 which can be quite useful - but may break some applications.
838 If set non-zero, enables support for dynamic addresses.
839 If set to a non-zero value larger than 1, a kernel log
840 message will be printed when dynamic address rewriting
844 ip_early_demux - BOOLEAN
845 Optimize input packet processing down to one demux for
846 certain kinds of local sockets. Currently we only do this
847 for established TCP sockets.
849 It may add an additional cost for pure routing workloads that
850 reduces overall throughput, in such case you should disable it.
853 icmp_echo_ignore_all - BOOLEAN
854 If set non-zero, then the kernel will ignore all ICMP ECHO
858 icmp_echo_ignore_broadcasts - BOOLEAN
859 If set non-zero, then the kernel will ignore all ICMP ECHO and
860 TIMESTAMP requests sent to it via broadcast/multicast.
863 icmp_ratelimit - INTEGER
864 Limit the maximal rates for sending ICMP packets whose type matches
865 icmp_ratemask (see below) to specific targets.
866 0 to disable any limiting,
867 otherwise the minimal space between responses in milliseconds.
868 Note that another sysctl, icmp_msgs_per_sec limits the number
869 of ICMP packets sent on all targets.
872 icmp_msgs_per_sec - INTEGER
873 Limit maximal number of ICMP packets sent per second from this host.
874 Only messages whose type matches icmp_ratemask (see below) are
875 controlled by this limit.
878 icmp_msgs_burst - INTEGER
879 icmp_msgs_per_sec controls number of ICMP packets sent per second,
880 while icmp_msgs_burst controls the burst size of these packets.
883 icmp_ratemask - INTEGER
884 Mask made of ICMP types for which rates are being limited.
885 Significant bits: IHGFEDCBA9876543210
886 Default mask: 0000001100000011000 (6168)
888 Bit definitions (see include/linux/icmp.h):
890 3 Destination Unreachable *
895 C Parameter Problem *
900 H Address Mask Request
903 * These are rate limited by default (see default mask above)
905 icmp_ignore_bogus_error_responses - BOOLEAN
906 Some routers violate RFC1122 by sending bogus responses to broadcast
907 frames. Such violations are normally logged via a kernel warning.
908 If this is set to TRUE, the kernel will not give such warnings, which
909 will avoid log file clutter.
912 icmp_errors_use_inbound_ifaddr - BOOLEAN
914 If zero, icmp error messages are sent with the primary address of
915 the exiting interface.
917 If non-zero, the message will be sent with the primary address of
918 the interface that received the packet that caused the icmp error.
919 This is the behaviour network many administrators will expect from
920 a router. And it can make debugging complicated network layouts
923 Note that if no primary address exists for the interface selected,
924 then the primary address of the first non-loopback interface that
925 has one will be used regardless of this setting.
929 igmp_max_memberships - INTEGER
930 Change the maximum number of multicast groups we can subscribe to.
933 Theoretical maximum value is bounded by having to send a membership
934 report in a single datagram (i.e. the report can't span multiple
935 datagrams, or risk confusing the switch and leaving groups you don't
938 The number of supported groups 'M' is bounded by the number of group
939 report entries you can fit into a single datagram of 65535 bytes.
941 M = 65536-sizeof (ip header)/(sizeof(Group record))
943 Group records are variable length, with a minimum of 12 bytes.
944 So net.ipv4.igmp_max_memberships should not be set higher than:
946 (65536-24) / 12 = 5459
948 The value 5459 assumes no IP header options, so in practice
949 this number may be lower.
951 conf/interface/* changes special settings per interface (where
952 "interface" is the name of your network interface)
954 conf/all/* is special, changes the settings for all interfaces
957 Controls the IGMP query robustness variable (see RFC2236 8.1).
958 Default: 2 (as specified by RFC2236 8.1)
959 Minimum: 1 (as specified by RFC6636 4.5)
961 log_martians - BOOLEAN
962 Log packets with impossible addresses to kernel log.
963 log_martians for the interface will be enabled if at least one of
964 conf/{all,interface}/log_martians is set to TRUE,
965 it will be disabled otherwise
967 accept_redirects - BOOLEAN
968 Accept ICMP redirect messages.
969 accept_redirects for the interface will be enabled if:
970 - both conf/{all,interface}/accept_redirects are TRUE in the case
971 forwarding for the interface is enabled
973 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
974 case forwarding for the interface is disabled
975 accept_redirects for the interface will be disabled otherwise
980 Enable IP forwarding on this interface.
982 mc_forwarding - BOOLEAN
983 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
984 and a multicast routing daemon is required.
985 conf/all/mc_forwarding must also be set to TRUE to enable multicast
986 routing for the interface
989 Integer value used to differentiate the devices by the medium they
990 are attached to. Two devices can have different id values when
991 the broadcast packets are received only on one of them.
992 The default value 0 means that the device is the only interface
993 to its medium, value of -1 means that medium is not known.
995 Currently, it is used to change the proxy_arp behavior:
996 the proxy_arp feature is enabled for packets forwarded between
997 two devices attached to different media.
1001 proxy_arp for the interface will be enabled if at least one of
1002 conf/{all,interface}/proxy_arp is set to TRUE,
1003 it will be disabled otherwise
1005 proxy_arp_pvlan - BOOLEAN
1006 Private VLAN proxy arp.
1007 Basically allow proxy arp replies back to the same interface
1008 (from which the ARP request/solicitation was received).
1010 This is done to support (ethernet) switch features, like RFC
1011 3069, where the individual ports are NOT allowed to
1012 communicate with each other, but they are allowed to talk to
1013 the upstream router. As described in RFC 3069, it is possible
1014 to allow these hosts to communicate through the upstream
1015 router by proxy_arp'ing. Don't need to be used together with
1018 This technology is known by different names:
1019 In RFC 3069 it is called VLAN Aggregation.
1020 Cisco and Allied Telesyn call it Private VLAN.
1021 Hewlett-Packard call it Source-Port filtering or port-isolation.
1022 Ericsson call it MAC-Forced Forwarding (RFC Draft).
1024 shared_media - BOOLEAN
1025 Send(router) or accept(host) RFC1620 shared media redirects.
1026 Overrides ip_secure_redirects.
1027 shared_media for the interface will be enabled if at least one of
1028 conf/{all,interface}/shared_media is set to TRUE,
1029 it will be disabled otherwise
1032 secure_redirects - BOOLEAN
1033 Accept ICMP redirect messages only for gateways,
1034 listed in default gateway list.
1035 secure_redirects for the interface will be enabled if at least one of
1036 conf/{all,interface}/secure_redirects is set to TRUE,
1037 it will be disabled otherwise
1040 send_redirects - BOOLEAN
1041 Send redirects, if router.
1042 send_redirects for the interface will be enabled if at least one of
1043 conf/{all,interface}/send_redirects is set to TRUE,
1044 it will be disabled otherwise
1047 bootp_relay - BOOLEAN
1048 Accept packets with source address 0.b.c.d destined
1049 not to this host as local ones. It is supposed, that
1050 BOOTP relay daemon will catch and forward such packets.
1051 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1054 Not Implemented Yet.
1056 accept_source_route - BOOLEAN
1057 Accept packets with SRR option.
1058 conf/all/accept_source_route must also be set to TRUE to accept packets
1059 with SRR option on the interface
1060 default TRUE (router)
1063 accept_local - BOOLEAN
1064 Accept packets with local source addresses. In combination with
1065 suitable routing, this can be used to direct packets between two
1066 local interfaces over the wire and have them accepted properly.
1069 route_localnet - BOOLEAN
1070 Do not consider loopback addresses as martian source or destination
1071 while routing. This enables the use of 127/8 for local routing purposes.
1075 0 - No source validation.
1076 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1077 Each incoming packet is tested against the FIB and if the interface
1078 is not the best reverse path the packet check will fail.
1079 By default failed packets are discarded.
1080 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1081 Each incoming packet's source address is also tested against the FIB
1082 and if the source address is not reachable via any interface
1083 the packet check will fail.
1085 Current recommended practice in RFC3704 is to enable strict mode
1086 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1087 or other complicated routing, then loose mode is recommended.
1089 The max value from conf/{all,interface}/rp_filter is used
1090 when doing source validation on the {interface}.
1092 Default value is 0. Note that some distributions enable it
1095 arp_filter - BOOLEAN
1096 1 - Allows you to have multiple network interfaces on the same
1097 subnet, and have the ARPs for each interface be answered
1098 based on whether or not the kernel would route a packet from
1099 the ARP'd IP out that interface (therefore you must use source
1100 based routing for this to work). In other words it allows control
1101 of which cards (usually 1) will respond to an arp request.
1103 0 - (default) The kernel can respond to arp requests with addresses
1104 from other interfaces. This may seem wrong but it usually makes
1105 sense, because it increases the chance of successful communication.
1106 IP addresses are owned by the complete host on Linux, not by
1107 particular interfaces. Only for more complex setups like load-
1108 balancing, does this behaviour cause problems.
1110 arp_filter for the interface will be enabled if at least one of
1111 conf/{all,interface}/arp_filter is set to TRUE,
1112 it will be disabled otherwise
1114 arp_announce - INTEGER
1115 Define different restriction levels for announcing the local
1116 source IP address from IP packets in ARP requests sent on
1118 0 - (default) Use any local address, configured on any interface
1119 1 - Try to avoid local addresses that are not in the target's
1120 subnet for this interface. This mode is useful when target
1121 hosts reachable via this interface require the source IP
1122 address in ARP requests to be part of their logical network
1123 configured on the receiving interface. When we generate the
1124 request we will check all our subnets that include the
1125 target IP and will preserve the source address if it is from
1126 such subnet. If there is no such subnet we select source
1127 address according to the rules for level 2.
1128 2 - Always use the best local address for this target.
1129 In this mode we ignore the source address in the IP packet
1130 and try to select local address that we prefer for talks with
1131 the target host. Such local address is selected by looking
1132 for primary IP addresses on all our subnets on the outgoing
1133 interface that include the target IP address. If no suitable
1134 local address is found we select the first local address
1135 we have on the outgoing interface or on all other interfaces,
1136 with the hope we will receive reply for our request and
1137 even sometimes no matter the source IP address we announce.
1139 The max value from conf/{all,interface}/arp_announce is used.
1141 Increasing the restriction level gives more chance for
1142 receiving answer from the resolved target while decreasing
1143 the level announces more valid sender's information.
1145 arp_ignore - INTEGER
1146 Define different modes for sending replies in response to
1147 received ARP requests that resolve local target IP addresses:
1148 0 - (default): reply for any local target IP address, configured
1150 1 - reply only if the target IP address is local address
1151 configured on the incoming interface
1152 2 - reply only if the target IP address is local address
1153 configured on the incoming interface and both with the
1154 sender's IP address are part from same subnet on this interface
1155 3 - do not reply for local addresses configured with scope host,
1156 only resolutions for global and link addresses are replied
1158 8 - do not reply for all local addresses
1160 The max value from conf/{all,interface}/arp_ignore is used
1161 when ARP request is received on the {interface}
1163 arp_notify - BOOLEAN
1164 Define mode for notification of address and device changes.
1165 0 - (default): do nothing
1166 1 - Generate gratuitous arp requests when device is brought up
1167 or hardware address changes.
1169 arp_accept - BOOLEAN
1170 Define behavior for gratuitous ARP frames who's IP is not
1171 already present in the ARP table:
1172 0 - don't create new entries in the ARP table
1173 1 - create new entries in the ARP table
1175 Both replies and requests type gratuitous arp will trigger the
1176 ARP table to be updated, if this setting is on.
1178 If the ARP table already contains the IP address of the
1179 gratuitous arp frame, the arp table will be updated regardless
1180 if this setting is on or off.
1182 mcast_solicit - INTEGER
1183 The maximum number of multicast probes in INCOMPLETE state,
1184 when the associated hardware address is unknown. Defaults
1187 ucast_solicit - INTEGER
1188 The maximum number of unicast probes in PROBE state, when
1189 the hardware address is being reconfirmed. Defaults to 3.
1191 app_solicit - INTEGER
1192 The maximum number of probes to send to the user space ARP daemon
1193 via netlink before dropping back to multicast probes (see
1194 mcast_resolicit). Defaults to 0.
1196 mcast_resolicit - INTEGER
1197 The maximum number of multicast probes after unicast and
1198 app probes in PROBE state. Defaults to 0.
1200 disable_policy - BOOLEAN
1201 Disable IPSEC policy (SPD) for this interface
1203 disable_xfrm - BOOLEAN
1204 Disable IPSEC encryption on this interface, whatever the policy
1206 igmpv2_unsolicited_report_interval - INTEGER
1207 The interval in milliseconds in which the next unsolicited
1208 IGMPv1 or IGMPv2 report retransmit will take place.
1209 Default: 10000 (10 seconds)
1211 igmpv3_unsolicited_report_interval - INTEGER
1212 The interval in milliseconds in which the next unsolicited
1213 IGMPv3 report retransmit will take place.
1214 Default: 1000 (1 seconds)
1216 promote_secondaries - BOOLEAN
1217 When a primary IP address is removed from this interface
1218 promote a corresponding secondary IP address instead of
1219 removing all the corresponding secondary IP addresses.
1223 Allows you to write a number, which can be used as required.
1226 xfrm4_gc_thresh - INTEGER
1227 The threshold at which we will start garbage collecting for IPv4
1228 destination cache entries. At twice this value the system will
1229 refuse new allocations. The value must be set below the flowcache
1230 limit (4096 * number of online cpus) to take effect.
1232 igmp_link_local_mcast_reports - BOOLEAN
1233 Enable IGMP reports for link local multicast groups in the
1238 kuznet@ms2.inr.ac.ru
1244 delon.nicolas@wanadoo.fr
1249 /proc/sys/net/ipv6/* Variables:
1251 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1252 apply to IPv6 [XXX?].
1254 bindv6only - BOOLEAN
1255 Default value for IPV6_V6ONLY socket option,
1256 which restricts use of the IPv6 socket to IPv6 communication
1258 TRUE: disable IPv4-mapped address feature
1259 FALSE: enable IPv4-mapped address feature
1261 Default: FALSE (as specified in RFC3493)
1263 flowlabel_consistency - BOOLEAN
1264 Protect the consistency (and unicity) of flow label.
1265 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1271 auto_flowlabels - INTEGER
1272 Automatically generate flow labels based on a flow hash of the
1273 packet. This allows intermediate devices, such as routers, to
1274 identify packet flows for mechanisms like Equal Cost Multipath
1275 Routing (see RFC 6438).
1276 0: automatic flow labels are completely disabled
1277 1: automatic flow labels are enabled by default, they can be
1278 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1280 2: automatic flow labels are allowed, they may be enabled on a
1281 per socket basis using the IPV6_AUTOFLOWLABEL socket option
1282 3: automatic flow labels are enabled and enforced, they cannot
1283 be disabled by the socket option
1286 flowlabel_state_ranges - BOOLEAN
1287 Split the flow label number space into two ranges. 0-0x7FFFF is
1288 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1289 is reserved for stateless flow labels as described in RFC6437.
1294 anycast_src_echo_reply - BOOLEAN
1295 Controls the use of anycast addresses as source addresses for ICMPv6
1301 idgen_delay - INTEGER
1302 Controls the delay in seconds after which time to retry
1303 privacy stable address generation if a DAD conflict is
1305 Default: 1 (as specified in RFC7217)
1307 idgen_retries - INTEGER
1308 Controls the number of retries to generate a stable privacy
1309 address if a DAD conflict is detected.
1310 Default: 3 (as specified in RFC7217)
1313 Controls the MLD query robustness variable (see RFC3810 9.1).
1314 Default: 2 (as specified by RFC3810 9.1)
1315 Minimum: 1 (as specified by RFC6636 4.5)
1319 ip6frag_high_thresh - INTEGER
1320 Maximum memory used to reassemble IPv6 fragments. When
1321 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1322 the fragment handler will toss packets until ip6frag_low_thresh
1325 ip6frag_low_thresh - INTEGER
1326 See ip6frag_high_thresh
1328 ip6frag_time - INTEGER
1329 Time in seconds to keep an IPv6 fragment in memory.
1332 Change the interface-specific default settings.
1336 Change all the interface-specific settings.
1338 [XXX: Other special features than forwarding?]
1340 conf/all/forwarding - BOOLEAN
1341 Enable global IPv6 forwarding between all interfaces.
1343 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1344 to control which interfaces may forward packets and which not.
1346 This also sets all interfaces' Host/Router setting
1347 'forwarding' to the specified value. See below for details.
1349 This referred to as global forwarding.
1354 fwmark_reflect - BOOLEAN
1355 Controls the fwmark of kernel-generated IPv6 reply packets that are not
1356 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
1357 If unset, these packets have a fwmark of zero. If set, they have the
1358 fwmark of the packet they are replying to.
1362 Change special settings per interface.
1364 The functional behaviour for certain settings is different
1365 depending on whether local forwarding is enabled or not.
1368 Accept Router Advertisements; autoconfigure using them.
1370 It also determines whether or not to transmit Router
1371 Solicitations. If and only if the functional setting is to
1372 accept Router Advertisements, Router Solicitations will be
1375 Possible values are:
1376 0 Do not accept Router Advertisements.
1377 1 Accept Router Advertisements if forwarding is disabled.
1378 2 Overrule forwarding behaviour. Accept Router Advertisements
1379 even if forwarding is enabled.
1381 Functional default: enabled if local forwarding is disabled.
1382 disabled if local forwarding is enabled.
1384 accept_ra_defrtr - BOOLEAN
1385 Learn default router in Router Advertisement.
1387 Functional default: enabled if accept_ra is enabled.
1388 disabled if accept_ra is disabled.
1390 accept_ra_from_local - BOOLEAN
1391 Accept RA with source-address that is found on local machine
1392 if the RA is otherwise proper and able to be accepted.
1393 Default is to NOT accept these as it may be an un-intended
1397 enabled if accept_ra_from_local is enabled
1398 on a specific interface.
1399 disabled if accept_ra_from_local is disabled
1400 on a specific interface.
1402 accept_ra_min_hop_limit - INTEGER
1403 Minimum hop limit Information in Router Advertisement.
1405 Hop limit Information in Router Advertisement less than this
1406 variable shall be ignored.
1410 accept_ra_pinfo - BOOLEAN
1411 Learn Prefix Information in Router Advertisement.
1413 Functional default: enabled if accept_ra is enabled.
1414 disabled if accept_ra is disabled.
1416 accept_ra_rt_info_max_plen - INTEGER
1417 Maximum prefix length of Route Information in RA.
1419 Route Information w/ prefix larger than or equal to this
1420 variable shall be ignored.
1422 Functional default: 0 if accept_ra_rtr_pref is enabled.
1423 -1 if accept_ra_rtr_pref is disabled.
1425 accept_ra_rtr_pref - BOOLEAN
1426 Accept Router Preference in RA.
1428 Functional default: enabled if accept_ra is enabled.
1429 disabled if accept_ra is disabled.
1431 accept_ra_mtu - BOOLEAN
1432 Apply the MTU value specified in RA option 5 (RFC4861). If
1433 disabled, the MTU specified in the RA will be ignored.
1435 Functional default: enabled if accept_ra is enabled.
1436 disabled if accept_ra is disabled.
1438 accept_redirects - BOOLEAN
1441 Functional default: enabled if local forwarding is disabled.
1442 disabled if local forwarding is enabled.
1444 accept_source_route - INTEGER
1445 Accept source routing (routing extension header).
1447 >= 0: Accept only routing header type 2.
1448 < 0: Do not accept routing header.
1453 Autoconfigure addresses using Prefix Information in Router
1456 Functional default: enabled if accept_ra_pinfo is enabled.
1457 disabled if accept_ra_pinfo is disabled.
1459 dad_transmits - INTEGER
1460 The amount of Duplicate Address Detection probes to send.
1463 forwarding - INTEGER
1464 Configure interface-specific Host/Router behaviour.
1466 Note: It is recommended to have the same setting on all
1467 interfaces; mixed router/host scenarios are rather uncommon.
1469 Possible values are:
1470 0 Forwarding disabled
1471 1 Forwarding enabled
1475 By default, Host behaviour is assumed. This means:
1477 1. IsRouter flag is not set in Neighbour Advertisements.
1478 2. If accept_ra is TRUE (default), transmit Router
1480 3. If accept_ra is TRUE (default), accept Router
1481 Advertisements (and do autoconfiguration).
1482 4. If accept_redirects is TRUE (default), accept Redirects.
1486 If local forwarding is enabled, Router behaviour is assumed.
1487 This means exactly the reverse from the above:
1489 1. IsRouter flag is set in Neighbour Advertisements.
1490 2. Router Solicitations are not sent unless accept_ra is 2.
1491 3. Router Advertisements are ignored unless accept_ra is 2.
1492 4. Redirects are ignored.
1494 Default: 0 (disabled) if global forwarding is disabled (default),
1495 otherwise 1 (enabled).
1498 Default Hop Limit to set.
1502 Default Maximum Transfer Unit
1503 Default: 1280 (IPv6 required minimum)
1505 ip_nonlocal_bind - BOOLEAN
1506 If set, allows processes to bind() to non-local IPv6 addresses,
1507 which can be quite useful - but may break some applications.
1510 router_probe_interval - INTEGER
1511 Minimum interval (in seconds) between Router Probing described
1516 router_solicitation_delay - INTEGER
1517 Number of seconds to wait after interface is brought up
1518 before sending Router Solicitations.
1521 router_solicitation_interval - INTEGER
1522 Number of seconds to wait between Router Solicitations.
1525 router_solicitations - INTEGER
1526 Number of Router Solicitations to send until assuming no
1527 routers are present.
1530 use_oif_addrs_only - BOOLEAN
1531 When enabled, the candidate source addresses for destinations
1532 routed via this interface are restricted to the set of addresses
1533 configured on this interface (vis. RFC 6724, section 4).
1537 use_tempaddr - INTEGER
1538 Preference for Privacy Extensions (RFC3041).
1539 <= 0 : disable Privacy Extensions
1540 == 1 : enable Privacy Extensions, but prefer public
1541 addresses over temporary addresses.
1542 > 1 : enable Privacy Extensions and prefer temporary
1543 addresses over public addresses.
1544 Default: 0 (for most devices)
1545 -1 (for point-to-point devices and loopback devices)
1547 temp_valid_lft - INTEGER
1548 valid lifetime (in seconds) for temporary addresses.
1549 Default: 604800 (7 days)
1551 temp_prefered_lft - INTEGER
1552 Preferred lifetime (in seconds) for temporary addresses.
1553 Default: 86400 (1 day)
1555 max_desync_factor - INTEGER
1556 Maximum value for DESYNC_FACTOR, which is a random value
1557 that ensures that clients don't synchronize with each
1558 other and generate new addresses at exactly the same time.
1559 value is in seconds.
1562 regen_max_retry - INTEGER
1563 Number of attempts before give up attempting to generate
1564 valid temporary addresses.
1567 max_addresses - INTEGER
1568 Maximum number of autoconfigured addresses per interface. Setting
1569 to zero disables the limitation. It is not recommended to set this
1570 value too large (or to zero) because it would be an easy way to
1571 crash the kernel by allowing too many addresses to be created.
1574 disable_ipv6 - BOOLEAN
1575 Disable IPv6 operation. If accept_dad is set to 2, this value
1576 will be dynamically set to TRUE if DAD fails for the link-local
1578 Default: FALSE (enable IPv6 operation)
1580 When this value is changed from 1 to 0 (IPv6 is being enabled),
1581 it will dynamically create a link-local address on the given
1582 interface and start Duplicate Address Detection, if necessary.
1584 When this value is changed from 0 to 1 (IPv6 is being disabled),
1585 it will dynamically delete all address on the given interface.
1587 accept_dad - INTEGER
1588 Whether to accept DAD (Duplicate Address Detection).
1590 1: Enable DAD (default)
1591 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1592 link-local address has been found.
1594 force_tllao - BOOLEAN
1595 Enable sending the target link-layer address option even when
1596 responding to a unicast neighbor solicitation.
1599 Quoting from RFC 2461, section 4.4, Target link-layer address:
1601 "The option MUST be included for multicast solicitations in order to
1602 avoid infinite Neighbor Solicitation "recursion" when the peer node
1603 does not have a cache entry to return a Neighbor Advertisements
1604 message. When responding to unicast solicitations, the option can be
1605 omitted since the sender of the solicitation has the correct link-
1606 layer address; otherwise it would not have be able to send the unicast
1607 solicitation in the first place. However, including the link-layer
1608 address in this case adds little overhead and eliminates a potential
1609 race condition where the sender deletes the cached link-layer address
1610 prior to receiving a response to a previous solicitation."
1612 ndisc_notify - BOOLEAN
1613 Define mode for notification of address and device changes.
1614 0 - (default): do nothing
1615 1 - Generate unsolicited neighbour advertisements when device is brought
1616 up or hardware address changes.
1618 mldv1_unsolicited_report_interval - INTEGER
1619 The interval in milliseconds in which the next unsolicited
1620 MLDv1 report retransmit will take place.
1621 Default: 10000 (10 seconds)
1623 mldv2_unsolicited_report_interval - INTEGER
1624 The interval in milliseconds in which the next unsolicited
1625 MLDv2 report retransmit will take place.
1626 Default: 1000 (1 second)
1628 force_mld_version - INTEGER
1629 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1630 1 - Enforce to use MLD version 1
1631 2 - Enforce to use MLD version 2
1633 suppress_frag_ndisc - INTEGER
1634 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1635 with IPv6 Neighbor Discovery) behavior:
1636 1 - (default) discard fragmented neighbor discovery packets
1637 0 - allow fragmented neighbor discovery packets
1639 optimistic_dad - BOOLEAN
1640 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
1641 0: disabled (default)
1644 use_optimistic - BOOLEAN
1645 If enabled, do not classify optimistic addresses as deprecated during
1646 source address selection. Preferred addresses will still be chosen
1647 before optimistic addresses, subject to other ranking in the source
1648 address selection algorithm.
1649 0: disabled (default)
1652 stable_secret - IPv6 address
1653 This IPv6 address will be used as a secret to generate IPv6
1654 addresses for link-local addresses and autoconfigured
1655 ones. All addresses generated after setting this secret will
1656 be stable privacy ones by default. This can be changed via the
1657 addrgenmode ip-link. conf/default/stable_secret is used as the
1658 secret for the namespace, the interface specific ones can
1659 overwrite that. Writes to conf/all/stable_secret are refused.
1661 It is recommended to generate this secret during installation
1662 of a system and keep it stable after that.
1664 By default the stable secret is unset.
1668 Limit the maximal rates for sending ICMPv6 packets.
1669 0 to disable any limiting,
1670 otherwise the minimal space between responses in milliseconds.
1673 xfrm6_gc_thresh - INTEGER
1674 The threshold at which we will start garbage collecting for IPv6
1675 destination cache entries. At twice this value the system will
1676 refuse new allocations. The value must be set below the flowcache
1677 limit (4096 * number of online cpus) to take effect.
1681 Pekka Savola <pekkas@netcore.fi>
1682 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1685 /proc/sys/net/bridge/* Variables:
1687 bridge-nf-call-arptables - BOOLEAN
1688 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1692 bridge-nf-call-iptables - BOOLEAN
1693 1 : pass bridged IPv4 traffic to iptables' chains.
1697 bridge-nf-call-ip6tables - BOOLEAN
1698 1 : pass bridged IPv6 traffic to ip6tables' chains.
1702 bridge-nf-filter-vlan-tagged - BOOLEAN
1703 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1707 bridge-nf-filter-pppoe-tagged - BOOLEAN
1708 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1712 bridge-nf-pass-vlan-input-dev - BOOLEAN
1713 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1714 interface on the bridge and set the netfilter input device to the vlan.
1715 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1716 target work with vlan-on-top-of-bridge interfaces. When no matching
1717 vlan interface is found, or this switch is off, the input device is
1718 set to the bridge interface.
1719 0: disable bridge netfilter vlan interface lookup.
1722 proc/sys/net/sctp/* Variables:
1724 addip_enable - BOOLEAN
1725 Enable or disable extension of Dynamic Address Reconfiguration
1726 (ADD-IP) functionality specified in RFC5061. This extension provides
1727 the ability to dynamically add and remove new addresses for the SCTP
1730 1: Enable extension.
1732 0: Disable extension.
1736 addip_noauth_enable - BOOLEAN
1737 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1738 authentication to protect the operations of adding or removing new
1739 addresses. This requirement is mandated so that unauthorized hosts
1740 would not be able to hijack associations. However, older
1741 implementations may not have implemented this requirement while
1742 allowing the ADD-IP extension. For reasons of interoperability,
1743 we provide this variable to control the enforcement of the
1744 authentication requirement.
1746 1: Allow ADD-IP extension to be used without authentication. This
1747 should only be set in a closed environment for interoperability
1748 with older implementations.
1750 0: Enforce the authentication requirement
1754 auth_enable - BOOLEAN
1755 Enable or disable Authenticated Chunks extension. This extension
1756 provides the ability to send and receive authenticated chunks and is
1757 required for secure operation of Dynamic Address Reconfiguration
1760 1: Enable this extension.
1761 0: Disable this extension.
1765 prsctp_enable - BOOLEAN
1766 Enable or disable the Partial Reliability extension (RFC3758) which
1767 is used to notify peers that a given DATA should no longer be expected.
1775 The limit of the number of new packets that can be initially sent. It
1776 controls how bursty the generated traffic can be.
1780 association_max_retrans - INTEGER
1781 Set the maximum number for retransmissions that an association can
1782 attempt deciding that the remote end is unreachable. If this value
1783 is exceeded, the association is terminated.
1787 max_init_retransmits - INTEGER
1788 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1789 that an association will attempt before declaring the destination
1790 unreachable and terminating.
1794 path_max_retrans - INTEGER
1795 The maximum number of retransmissions that will be attempted on a given
1796 path. Once this threshold is exceeded, the path is considered
1797 unreachable, and new traffic will use a different path when the
1798 association is multihomed.
1802 pf_retrans - INTEGER
1803 The number of retransmissions that will be attempted on a given path
1804 before traffic is redirected to an alternate transport (should one
1805 exist). Note this is distinct from path_max_retrans, as a path that
1806 passes the pf_retrans threshold can still be used. Its only
1807 deprioritized when a transmission path is selected by the stack. This
1808 setting is primarily used to enable fast failover mechanisms without
1809 having to reduce path_max_retrans to a very low value. See:
1810 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1811 for details. Note also that a value of pf_retrans > path_max_retrans
1812 disables this feature
1816 rto_initial - INTEGER
1817 The initial round trip timeout value in milliseconds that will be used
1818 in calculating round trip times. This is the initial time interval
1819 for retransmissions.
1824 The maximum value (in milliseconds) of the round trip timeout. This
1825 is the largest time interval that can elapse between retransmissions.
1830 The minimum value (in milliseconds) of the round trip timeout. This
1831 is the smallest time interval the can elapse between retransmissions.
1835 hb_interval - INTEGER
1836 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1837 are sent at the specified interval on idle paths to probe the state of
1838 a given path between 2 associations.
1842 sack_timeout - INTEGER
1843 The amount of time (in milliseconds) that the implementation will wait
1848 valid_cookie_life - INTEGER
1849 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1850 is used during association establishment.
1854 cookie_preserve_enable - BOOLEAN
1855 Enable or disable the ability to extend the lifetime of the SCTP cookie
1856 that is used during the establishment phase of SCTP association
1858 1: Enable cookie lifetime extension.
1863 cookie_hmac_alg - STRING
1864 Select the hmac algorithm used when generating the cookie value sent by
1865 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1870 Ability to assign md5 or sha1 as the selected alg is predicated on the
1871 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1872 CONFIG_CRYPTO_SHA1).
1874 Default: Dependent on configuration. MD5 if available, else SHA1 if
1875 available, else none.
1877 rcvbuf_policy - INTEGER
1878 Determines if the receive buffer is attributed to the socket or to
1879 association. SCTP supports the capability to create multiple
1880 associations on a single socket. When using this capability, it is
1881 possible that a single stalled association that's buffering a lot
1882 of data may block other associations from delivering their data by
1883 consuming all of the receive buffer space. To work around this,
1884 the rcvbuf_policy could be set to attribute the receiver buffer space
1885 to each association instead of the socket. This prevents the described
1888 1: rcvbuf space is per association
1889 0: rcvbuf space is per socket
1893 sndbuf_policy - INTEGER
1894 Similar to rcvbuf_policy above, this applies to send buffer space.
1896 1: Send buffer is tracked per association
1897 0: Send buffer is tracked per socket.
1901 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1902 Number of pages allowed for queueing by all SCTP sockets.
1904 min: Below this number of pages SCTP is not bothered about its
1905 memory appetite. When amount of memory allocated by SCTP exceeds
1906 this number, SCTP starts to moderate memory usage.
1908 pressure: This value was introduced to follow format of tcp_mem.
1910 max: Number of pages allowed for queueing by all SCTP sockets.
1912 Default is calculated at boot time from amount of available memory.
1914 sctp_rmem - vector of 3 INTEGERs: min, default, max
1915 Only the first value ("min") is used, "default" and "max" are
1918 min: Minimal size of receive buffer used by SCTP socket.
1919 It is guaranteed to each SCTP socket (but not association) even
1920 under moderate memory pressure.
1924 sctp_wmem - vector of 3 INTEGERs: min, default, max
1925 Currently this tunable has no effect.
1927 addr_scope_policy - INTEGER
1928 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1930 0 - Disable IPv4 address scoping
1931 1 - Enable IPv4 address scoping
1932 2 - Follow draft but allow IPv4 private addresses
1933 3 - Follow draft but allow IPv4 link local addresses
1938 /proc/sys/net/core/*
1939 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1942 /proc/sys/net/unix/*
1943 max_dgram_qlen - INTEGER
1944 The maximum length of dgram socket receive queue
1951 /proc/sys/net/irda/*
1952 fast_poll_increase FIXME
1953 warn_noreply_time FIXME
1954 discovery_slots FIXME
1957 discovery_timeout FIXME
1958 lap_keepalive_time FIXME
1959 max_noreply_time FIXME
1960 max_tx_data_size FIXME
1962 min_tx_turn_time FIXME