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_l3mdev_accept - BOOLEAN
339 Enables child sockets to inherit the L3 master device index.
340 Enabling this option allows a "global" listen socket to work
341 across L3 master domains (e.g., VRFs) with connected sockets
342 derived from the listen socket to be bound to the L3 domain in
343 which the packets originated. Only valid when the kernel was
344 compiled with CONFIG_NET_L3_MASTER_DEV.
346 tcp_low_latency - BOOLEAN
347 If set, the TCP stack makes decisions that prefer lower
348 latency as opposed to higher throughput. By default, this
349 option is not set meaning that higher throughput is preferred.
350 An example of an application where this default should be
351 changed would be a Beowulf compute cluster.
354 tcp_max_orphans - INTEGER
355 Maximal number of TCP sockets not attached to any user file handle,
356 held by system. If this number is exceeded orphaned connections are
357 reset immediately and warning is printed. This limit exists
358 only to prevent simple DoS attacks, you _must_ not rely on this
359 or lower the limit artificially, but rather increase it
360 (probably, after increasing installed memory),
361 if network conditions require more than default value,
362 and tune network services to linger and kill such states
363 more aggressively. Let me to remind again: each orphan eats
364 up to ~64K of unswappable memory.
366 tcp_max_syn_backlog - INTEGER
367 Maximal number of remembered connection requests, which have not
368 received an acknowledgment from connecting client.
369 The minimal value is 128 for low memory machines, and it will
370 increase in proportion to the memory of machine.
371 If server suffers from overload, try increasing this number.
373 tcp_max_tw_buckets - INTEGER
374 Maximal number of timewait sockets held by system simultaneously.
375 If this number is exceeded time-wait socket is immediately destroyed
376 and warning is printed. This limit exists only to prevent
377 simple DoS attacks, you _must_ not lower the limit artificially,
378 but rather increase it (probably, after increasing installed memory),
379 if network conditions require more than default value.
381 tcp_mem - vector of 3 INTEGERs: min, pressure, max
382 min: below this number of pages TCP is not bothered about its
385 pressure: when amount of memory allocated by TCP exceeds this number
386 of pages, TCP moderates its memory consumption and enters memory
387 pressure mode, which is exited when memory consumption falls
390 max: number of pages allowed for queueing by all TCP sockets.
392 Defaults are calculated at boot time from amount of available
395 tcp_min_rtt_wlen - INTEGER
396 The window length of the windowed min filter to track the minimum RTT.
397 A shorter window lets a flow more quickly pick up new (higher)
398 minimum RTT when it is moved to a longer path (e.g., due to traffic
399 engineering). A longer window makes the filter more resistant to RTT
400 inflations such as transient congestion. The unit is seconds.
403 tcp_moderate_rcvbuf - BOOLEAN
404 If set, TCP performs receive buffer auto-tuning, attempting to
405 automatically size the buffer (no greater than tcp_rmem[2]) to
406 match the size required by the path for full throughput. Enabled by
409 tcp_mtu_probing - INTEGER
410 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
413 1 - Disabled by default, enabled when an ICMP black hole detected
414 2 - Always enabled, use initial MSS of tcp_base_mss.
416 tcp_probe_interval - INTEGER
417 Controls how often to start TCP Packetization-Layer Path MTU
418 Discovery reprobe. The default is reprobing every 10 minutes as
421 tcp_probe_threshold - INTEGER
422 Controls when TCP Packetization-Layer Path MTU Discovery probing
423 will stop in respect to the width of search range in bytes. Default
426 tcp_no_metrics_save - BOOLEAN
427 By default, TCP saves various connection metrics in the route cache
428 when the connection closes, so that connections established in the
429 near future can use these to set initial conditions. Usually, this
430 increases overall performance, but may sometimes cause performance
431 degradation. If set, TCP will not cache metrics on closing
434 tcp_orphan_retries - INTEGER
435 This value influences the timeout of a locally closed TCP connection,
436 when RTO retransmissions remain unacknowledged.
437 See tcp_retries2 for more details.
439 The default value is 8.
440 If your machine is a loaded WEB server,
441 you should think about lowering this value, such sockets
442 may consume significant resources. Cf. tcp_max_orphans.
444 tcp_recovery - INTEGER
445 This value is a bitmap to enable various experimental loss recovery
448 RACK: 0x1 enables the RACK loss detection for fast detection of lost
449 retransmissions and tail drops.
453 tcp_reordering - INTEGER
454 Initial reordering level of packets in a TCP stream.
455 TCP stack can then dynamically adjust flow reordering level
456 between this initial value and tcp_max_reordering
459 tcp_max_reordering - INTEGER
460 Maximal reordering level of packets in a TCP stream.
461 300 is a fairly conservative value, but you might increase it
462 if paths are using per packet load balancing (like bonding rr mode)
465 tcp_retrans_collapse - BOOLEAN
466 Bug-to-bug compatibility with some broken printers.
467 On retransmit try to send bigger packets to work around bugs in
470 tcp_retries1 - INTEGER
471 This value influences the time, after which TCP decides, that
472 something is wrong due to unacknowledged RTO retransmissions,
473 and reports this suspicion to the network layer.
474 See tcp_retries2 for more details.
476 RFC 1122 recommends at least 3 retransmissions, which is the
479 tcp_retries2 - INTEGER
480 This value influences the timeout of an alive TCP connection,
481 when RTO retransmissions remain unacknowledged.
482 Given a value of N, a hypothetical TCP connection following
483 exponential backoff with an initial RTO of TCP_RTO_MIN would
484 retransmit N times before killing the connection at the (N+1)th RTO.
486 The default value of 15 yields a hypothetical timeout of 924.6
487 seconds and is a lower bound for the effective timeout.
488 TCP will effectively time out at the first RTO which exceeds the
489 hypothetical timeout.
491 RFC 1122 recommends at least 100 seconds for the timeout,
492 which corresponds to a value of at least 8.
494 tcp_rfc1337 - BOOLEAN
495 If set, the TCP stack behaves conforming to RFC1337. If unset,
496 we are not conforming to RFC, but prevent TCP TIME_WAIT
500 tcp_rmem - vector of 3 INTEGERs: min, default, max
501 min: Minimal size of receive buffer used by TCP sockets.
502 It is guaranteed to each TCP socket, even under moderate memory
506 default: initial size of receive buffer used by TCP sockets.
507 This value overrides net.core.rmem_default used by other protocols.
508 Default: 87380 bytes. This value results in window of 65535 with
509 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
510 less for default tcp_app_win. See below about these variables.
512 max: maximal size of receive buffer allowed for automatically
513 selected receiver buffers for TCP socket. This value does not override
514 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
515 automatic tuning of that socket's receive buffer size, in which
516 case this value is ignored.
517 Default: between 87380B and 6MB, depending on RAM size.
520 Enable select acknowledgments (SACKS).
522 tcp_slow_start_after_idle - BOOLEAN
523 If set, provide RFC2861 behavior and time out the congestion
524 window after an idle period. An idle period is defined at
525 the current RTO. If unset, the congestion window will not
526 be timed out after an idle period.
530 Use the Host requirements interpretation of the TCP urgent pointer field.
531 Most hosts use the older BSD interpretation, so if you turn this on
532 Linux might not communicate correctly with them.
535 tcp_synack_retries - INTEGER
536 Number of times SYNACKs for a passive TCP connection attempt will
537 be retransmitted. Should not be higher than 255. Default value
538 is 5, which corresponds to 31seconds till the last retransmission
539 with the current initial RTO of 1second. With this the final timeout
540 for a passive TCP connection will happen after 63seconds.
542 tcp_syncookies - BOOLEAN
543 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
544 Send out syncookies when the syn backlog queue of a socket
545 overflows. This is to prevent against the common 'SYN flood attack'
548 Note, that syncookies is fallback facility.
549 It MUST NOT be used to help highly loaded servers to stand
550 against legal connection rate. If you see SYN flood warnings
551 in your logs, but investigation shows that they occur
552 because of overload with legal connections, you should tune
553 another parameters until this warning disappear.
554 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
556 syncookies seriously violate TCP protocol, do not allow
557 to use TCP extensions, can result in serious degradation
558 of some services (f.e. SMTP relaying), visible not by you,
559 but your clients and relays, contacting you. While you see
560 SYN flood warnings in logs not being really flooded, your server
561 is seriously misconfigured.
563 If you want to test which effects syncookies have to your
564 network connections you can set this knob to 2 to enable
565 unconditionally generation of syncookies.
567 tcp_fastopen - INTEGER
568 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
569 in the opening SYN packet. To use this feature, the client application
570 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
571 connect() to perform a TCP handshake automatically.
573 The values (bitmap) are
574 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN.
575 2: Enables TCP Fast Open on the server side, i.e., allowing data in
576 a SYN packet to be accepted and passed to the application before
577 3-way hand shake finishes.
578 4: Send data in the opening SYN regardless of cookie availability and
579 without a cookie option.
580 0x100: Accept SYN data w/o validating the cookie.
581 0x200: Accept data-in-SYN w/o any cookie option present.
582 0x400/0x800: Enable Fast Open on all listeners regardless of the
583 TCP_FASTOPEN socket option. The two different flags designate two
584 different ways of setting max_qlen without the TCP_FASTOPEN socket
589 Note that the client & server side Fast Open flags (1 and 2
590 respectively) must be also enabled before the rest of flags can take
593 See include/net/tcp.h and the code for more details.
595 tcp_syn_retries - INTEGER
596 Number of times initial SYNs for an active TCP connection attempt
597 will be retransmitted. Should not be higher than 127. Default value
598 is 6, which corresponds to 63seconds till the last retransmission
599 with the current initial RTO of 1second. With this the final timeout
600 for an active TCP connection attempt will happen after 127seconds.
602 tcp_timestamps - BOOLEAN
603 Enable timestamps as defined in RFC1323.
605 tcp_min_tso_segs - INTEGER
606 Minimal number of segments per TSO frame.
607 Since linux-3.12, TCP does an automatic sizing of TSO frames,
608 depending on flow rate, instead of filling 64Kbytes packets.
609 For specific usages, it's possible to force TCP to build big
610 TSO frames. Note that TCP stack might split too big TSO packets
611 if available window is too small.
614 tcp_pacing_ss_ratio - INTEGER
615 sk->sk_pacing_rate is set by TCP stack using a ratio applied
616 to current rate. (current_rate = cwnd * mss / srtt)
617 If TCP is in slow start, tcp_pacing_ss_ratio is applied
618 to let TCP probe for bigger speeds, assuming cwnd can be
619 doubled every other RTT.
622 tcp_pacing_ca_ratio - INTEGER
623 sk->sk_pacing_rate is set by TCP stack using a ratio applied
624 to current rate. (current_rate = cwnd * mss / srtt)
625 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
626 is applied to conservatively probe for bigger throughput.
629 tcp_tso_win_divisor - INTEGER
630 This allows control over what percentage of the congestion window
631 can be consumed by a single TSO frame.
632 The setting of this parameter is a choice between burstiness and
633 building larger TSO frames.
636 tcp_tw_recycle - BOOLEAN
637 Enable fast recycling TIME-WAIT sockets. Default value is 0.
638 It should not be changed without advice/request of technical
641 tcp_tw_reuse - BOOLEAN
642 Allow to reuse TIME-WAIT sockets for new connections when it is
643 safe from protocol viewpoint. Default value is 0.
644 It should not be changed without advice/request of technical
647 tcp_window_scaling - BOOLEAN
648 Enable window scaling as defined in RFC1323.
650 tcp_wmem - vector of 3 INTEGERs: min, default, max
651 min: Amount of memory reserved for send buffers for TCP sockets.
652 Each TCP socket has rights to use it due to fact of its birth.
655 default: initial size of send buffer used by TCP sockets. This
656 value overrides net.core.wmem_default used by other protocols.
657 It is usually lower than net.core.wmem_default.
660 max: Maximal amount of memory allowed for automatically tuned
661 send buffers for TCP sockets. This value does not override
662 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
663 automatic tuning of that socket's send buffer size, in which case
664 this value is ignored.
665 Default: between 64K and 4MB, depending on RAM size.
667 tcp_notsent_lowat - UNSIGNED INTEGER
668 A TCP socket can control the amount of unsent bytes in its write queue,
669 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
670 reports POLLOUT events if the amount of unsent bytes is below a per
671 socket value, and if the write queue is not full. sendmsg() will
672 also not add new buffers if the limit is hit.
674 This global variable controls the amount of unsent data for
675 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
676 to the global variable has immediate effect.
678 Default: UINT_MAX (0xFFFFFFFF)
680 tcp_workaround_signed_windows - BOOLEAN
681 If set, assume no receipt of a window scaling option means the
682 remote TCP is broken and treats the window as a signed quantity.
683 If unset, assume the remote TCP is not broken even if we do
684 not receive a window scaling option from them.
687 tcp_thin_linear_timeouts - BOOLEAN
688 Enable dynamic triggering of linear timeouts for thin streams.
689 If set, a check is performed upon retransmission by timeout to
690 determine if the stream is thin (less than 4 packets in flight).
691 As long as the stream is found to be thin, up to 6 linear
692 timeouts may be performed before exponential backoff mode is
693 initiated. This improves retransmission latency for
694 non-aggressive thin streams, often found to be time-dependent.
695 For more information on thin streams, see
696 Documentation/networking/tcp-thin.txt
699 tcp_thin_dupack - BOOLEAN
700 Enable dynamic triggering of retransmissions after one dupACK
701 for thin streams. If set, a check is performed upon reception
702 of a dupACK to determine if the stream is thin (less than 4
703 packets in flight). As long as the stream is found to be thin,
704 data is retransmitted on the first received dupACK. This
705 improves retransmission latency for non-aggressive thin
706 streams, often found to be time-dependent.
707 For more information on thin streams, see
708 Documentation/networking/tcp-thin.txt
711 tcp_limit_output_bytes - INTEGER
712 Controls TCP Small Queue limit per tcp socket.
713 TCP bulk sender tends to increase packets in flight until it
714 gets losses notifications. With SNDBUF autotuning, this can
715 result in a large amount of packets queued in qdisc/device
716 on the local machine, hurting latency of other flows, for
717 typical pfifo_fast qdiscs.
718 tcp_limit_output_bytes limits the number of bytes on qdisc
719 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
722 tcp_challenge_ack_limit - INTEGER
723 Limits number of Challenge ACK sent per second, as recommended
724 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
729 udp_mem - vector of 3 INTEGERs: min, pressure, max
730 Number of pages allowed for queueing by all UDP sockets.
732 min: Below this number of pages UDP is not bothered about its
733 memory appetite. When amount of memory allocated by UDP exceeds
734 this number, UDP starts to moderate memory usage.
736 pressure: This value was introduced to follow format of tcp_mem.
738 max: Number of pages allowed for queueing by all UDP sockets.
740 Default is calculated at boot time from amount of available memory.
742 udp_rmem_min - INTEGER
743 Minimal size of receive buffer used by UDP sockets in moderation.
744 Each UDP socket is able to use the size for receiving data, even if
745 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
748 udp_wmem_min - INTEGER
749 Minimal size of send buffer used by UDP sockets in moderation.
750 Each UDP socket is able to use the size for sending data, even if
751 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
756 cipso_cache_enable - BOOLEAN
757 If set, enable additions to and lookups from the CIPSO label mapping
758 cache. If unset, additions are ignored and lookups always result in a
759 miss. However, regardless of the setting the cache is still
760 invalidated when required when means you can safely toggle this on and
761 off and the cache will always be "safe".
764 cipso_cache_bucket_size - INTEGER
765 The CIPSO label cache consists of a fixed size hash table with each
766 hash bucket containing a number of cache entries. This variable limits
767 the number of entries in each hash bucket; the larger the value the
768 more CIPSO label mappings that can be cached. When the number of
769 entries in a given hash bucket reaches this limit adding new entries
770 causes the oldest entry in the bucket to be removed to make room.
773 cipso_rbm_optfmt - BOOLEAN
774 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
775 the CIPSO draft specification (see Documentation/netlabel for details).
776 This means that when set the CIPSO tag will be padded with empty
777 categories in order to make the packet data 32-bit aligned.
780 cipso_rbm_structvalid - BOOLEAN
781 If set, do a very strict check of the CIPSO option when
782 ip_options_compile() is called. If unset, relax the checks done during
783 ip_options_compile(). Either way is "safe" as errors are caught else
784 where in the CIPSO processing code but setting this to 0 (False) should
785 result in less work (i.e. it should be faster) but could cause problems
786 with other implementations that require strict checking.
791 ip_local_port_range - 2 INTEGERS
792 Defines the local port range that is used by TCP and UDP to
793 choose the local port. The first number is the first, the
794 second the last local port number.
795 If possible, it is better these numbers have different parity.
796 (one even and one odd values)
797 The default values are 32768 and 60999 respectively.
799 ip_local_reserved_ports - list of comma separated ranges
800 Specify the ports which are reserved for known third-party
801 applications. These ports will not be used by automatic port
802 assignments (e.g. when calling connect() or bind() with port
803 number 0). Explicit port allocation behavior is unchanged.
805 The format used for both input and output is a comma separated
806 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
807 10). Writing to the file will clear all previously reserved
808 ports and update the current list with the one given in the
811 Note that ip_local_port_range and ip_local_reserved_ports
812 settings are independent and both are considered by the kernel
813 when determining which ports are available for automatic port
816 You can reserve ports which are not in the current
817 ip_local_port_range, e.g.:
819 $ cat /proc/sys/net/ipv4/ip_local_port_range
821 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
824 although this is redundant. However such a setting is useful
825 if later the port range is changed to a value that will
826 include the reserved ports.
830 ip_nonlocal_bind - BOOLEAN
831 If set, allows processes to bind() to non-local IP addresses,
832 which can be quite useful - but may break some applications.
836 If set non-zero, enables support for dynamic addresses.
837 If set to a non-zero value larger than 1, a kernel log
838 message will be printed when dynamic address rewriting
842 ip_early_demux - BOOLEAN
843 Optimize input packet processing down to one demux for
844 certain kinds of local sockets. Currently we only do this
845 for established TCP sockets.
847 It may add an additional cost for pure routing workloads that
848 reduces overall throughput, in such case you should disable it.
851 icmp_echo_ignore_all - BOOLEAN
852 If set non-zero, then the kernel will ignore all ICMP ECHO
856 icmp_echo_ignore_broadcasts - BOOLEAN
857 If set non-zero, then the kernel will ignore all ICMP ECHO and
858 TIMESTAMP requests sent to it via broadcast/multicast.
861 icmp_ratelimit - INTEGER
862 Limit the maximal rates for sending ICMP packets whose type matches
863 icmp_ratemask (see below) to specific targets.
864 0 to disable any limiting,
865 otherwise the minimal space between responses in milliseconds.
866 Note that another sysctl, icmp_msgs_per_sec limits the number
867 of ICMP packets sent on all targets.
870 icmp_msgs_per_sec - INTEGER
871 Limit maximal number of ICMP packets sent per second from this host.
872 Only messages whose type matches icmp_ratemask (see below) are
873 controlled by this limit.
876 icmp_msgs_burst - INTEGER
877 icmp_msgs_per_sec controls number of ICMP packets sent per second,
878 while icmp_msgs_burst controls the burst size of these packets.
881 icmp_ratemask - INTEGER
882 Mask made of ICMP types for which rates are being limited.
883 Significant bits: IHGFEDCBA9876543210
884 Default mask: 0000001100000011000 (6168)
886 Bit definitions (see include/linux/icmp.h):
888 3 Destination Unreachable *
893 C Parameter Problem *
898 H Address Mask Request
901 * These are rate limited by default (see default mask above)
903 icmp_ignore_bogus_error_responses - BOOLEAN
904 Some routers violate RFC1122 by sending bogus responses to broadcast
905 frames. Such violations are normally logged via a kernel warning.
906 If this is set to TRUE, the kernel will not give such warnings, which
907 will avoid log file clutter.
910 icmp_errors_use_inbound_ifaddr - BOOLEAN
912 If zero, icmp error messages are sent with the primary address of
913 the exiting interface.
915 If non-zero, the message will be sent with the primary address of
916 the interface that received the packet that caused the icmp error.
917 This is the behaviour network many administrators will expect from
918 a router. And it can make debugging complicated network layouts
921 Note that if no primary address exists for the interface selected,
922 then the primary address of the first non-loopback interface that
923 has one will be used regardless of this setting.
927 igmp_max_memberships - INTEGER
928 Change the maximum number of multicast groups we can subscribe to.
931 Theoretical maximum value is bounded by having to send a membership
932 report in a single datagram (i.e. the report can't span multiple
933 datagrams, or risk confusing the switch and leaving groups you don't
936 The number of supported groups 'M' is bounded by the number of group
937 report entries you can fit into a single datagram of 65535 bytes.
939 M = 65536-sizeof (ip header)/(sizeof(Group record))
941 Group records are variable length, with a minimum of 12 bytes.
942 So net.ipv4.igmp_max_memberships should not be set higher than:
944 (65536-24) / 12 = 5459
946 The value 5459 assumes no IP header options, so in practice
947 this number may be lower.
949 conf/interface/* changes special settings per interface (where
950 "interface" is the name of your network interface)
952 conf/all/* is special, changes the settings for all interfaces
955 Controls the IGMP query robustness variable (see RFC2236 8.1).
956 Default: 2 (as specified by RFC2236 8.1)
957 Minimum: 1 (as specified by RFC6636 4.5)
959 log_martians - BOOLEAN
960 Log packets with impossible addresses to kernel log.
961 log_martians for the interface will be enabled if at least one of
962 conf/{all,interface}/log_martians is set to TRUE,
963 it will be disabled otherwise
965 accept_redirects - BOOLEAN
966 Accept ICMP redirect messages.
967 accept_redirects for the interface will be enabled if:
968 - both conf/{all,interface}/accept_redirects are TRUE in the case
969 forwarding for the interface is enabled
971 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
972 case forwarding for the interface is disabled
973 accept_redirects for the interface will be disabled otherwise
978 Enable IP forwarding on this interface.
980 mc_forwarding - BOOLEAN
981 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
982 and a multicast routing daemon is required.
983 conf/all/mc_forwarding must also be set to TRUE to enable multicast
984 routing for the interface
987 Integer value used to differentiate the devices by the medium they
988 are attached to. Two devices can have different id values when
989 the broadcast packets are received only on one of them.
990 The default value 0 means that the device is the only interface
991 to its medium, value of -1 means that medium is not known.
993 Currently, it is used to change the proxy_arp behavior:
994 the proxy_arp feature is enabled for packets forwarded between
995 two devices attached to different media.
999 proxy_arp for the interface will be enabled if at least one of
1000 conf/{all,interface}/proxy_arp is set to TRUE,
1001 it will be disabled otherwise
1003 proxy_arp_pvlan - BOOLEAN
1004 Private VLAN proxy arp.
1005 Basically allow proxy arp replies back to the same interface
1006 (from which the ARP request/solicitation was received).
1008 This is done to support (ethernet) switch features, like RFC
1009 3069, where the individual ports are NOT allowed to
1010 communicate with each other, but they are allowed to talk to
1011 the upstream router. As described in RFC 3069, it is possible
1012 to allow these hosts to communicate through the upstream
1013 router by proxy_arp'ing. Don't need to be used together with
1016 This technology is known by different names:
1017 In RFC 3069 it is called VLAN Aggregation.
1018 Cisco and Allied Telesyn call it Private VLAN.
1019 Hewlett-Packard call it Source-Port filtering or port-isolation.
1020 Ericsson call it MAC-Forced Forwarding (RFC Draft).
1022 shared_media - BOOLEAN
1023 Send(router) or accept(host) RFC1620 shared media redirects.
1024 Overrides ip_secure_redirects.
1025 shared_media for the interface will be enabled if at least one of
1026 conf/{all,interface}/shared_media is set to TRUE,
1027 it will be disabled otherwise
1030 secure_redirects - BOOLEAN
1031 Accept ICMP redirect messages only for gateways,
1032 listed in default gateway list.
1033 secure_redirects for the interface will be enabled if at least one of
1034 conf/{all,interface}/secure_redirects is set to TRUE,
1035 it will be disabled otherwise
1038 send_redirects - BOOLEAN
1039 Send redirects, if router.
1040 send_redirects for the interface will be enabled if at least one of
1041 conf/{all,interface}/send_redirects is set to TRUE,
1042 it will be disabled otherwise
1045 bootp_relay - BOOLEAN
1046 Accept packets with source address 0.b.c.d destined
1047 not to this host as local ones. It is supposed, that
1048 BOOTP relay daemon will catch and forward such packets.
1049 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1052 Not Implemented Yet.
1054 accept_source_route - BOOLEAN
1055 Accept packets with SRR option.
1056 conf/all/accept_source_route must also be set to TRUE to accept packets
1057 with SRR option on the interface
1058 default TRUE (router)
1061 accept_local - BOOLEAN
1062 Accept packets with local source addresses. In combination with
1063 suitable routing, this can be used to direct packets between two
1064 local interfaces over the wire and have them accepted properly.
1067 route_localnet - BOOLEAN
1068 Do not consider loopback addresses as martian source or destination
1069 while routing. This enables the use of 127/8 for local routing purposes.
1073 0 - No source validation.
1074 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1075 Each incoming packet is tested against the FIB and if the interface
1076 is not the best reverse path the packet check will fail.
1077 By default failed packets are discarded.
1078 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1079 Each incoming packet's source address is also tested against the FIB
1080 and if the source address is not reachable via any interface
1081 the packet check will fail.
1083 Current recommended practice in RFC3704 is to enable strict mode
1084 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1085 or other complicated routing, then loose mode is recommended.
1087 The max value from conf/{all,interface}/rp_filter is used
1088 when doing source validation on the {interface}.
1090 Default value is 0. Note that some distributions enable it
1093 arp_filter - BOOLEAN
1094 1 - Allows you to have multiple network interfaces on the same
1095 subnet, and have the ARPs for each interface be answered
1096 based on whether or not the kernel would route a packet from
1097 the ARP'd IP out that interface (therefore you must use source
1098 based routing for this to work). In other words it allows control
1099 of which cards (usually 1) will respond to an arp request.
1101 0 - (default) The kernel can respond to arp requests with addresses
1102 from other interfaces. This may seem wrong but it usually makes
1103 sense, because it increases the chance of successful communication.
1104 IP addresses are owned by the complete host on Linux, not by
1105 particular interfaces. Only for more complex setups like load-
1106 balancing, does this behaviour cause problems.
1108 arp_filter for the interface will be enabled if at least one of
1109 conf/{all,interface}/arp_filter is set to TRUE,
1110 it will be disabled otherwise
1112 arp_announce - INTEGER
1113 Define different restriction levels for announcing the local
1114 source IP address from IP packets in ARP requests sent on
1116 0 - (default) Use any local address, configured on any interface
1117 1 - Try to avoid local addresses that are not in the target's
1118 subnet for this interface. This mode is useful when target
1119 hosts reachable via this interface require the source IP
1120 address in ARP requests to be part of their logical network
1121 configured on the receiving interface. When we generate the
1122 request we will check all our subnets that include the
1123 target IP and will preserve the source address if it is from
1124 such subnet. If there is no such subnet we select source
1125 address according to the rules for level 2.
1126 2 - Always use the best local address for this target.
1127 In this mode we ignore the source address in the IP packet
1128 and try to select local address that we prefer for talks with
1129 the target host. Such local address is selected by looking
1130 for primary IP addresses on all our subnets on the outgoing
1131 interface that include the target IP address. If no suitable
1132 local address is found we select the first local address
1133 we have on the outgoing interface or on all other interfaces,
1134 with the hope we will receive reply for our request and
1135 even sometimes no matter the source IP address we announce.
1137 The max value from conf/{all,interface}/arp_announce is used.
1139 Increasing the restriction level gives more chance for
1140 receiving answer from the resolved target while decreasing
1141 the level announces more valid sender's information.
1143 arp_ignore - INTEGER
1144 Define different modes for sending replies in response to
1145 received ARP requests that resolve local target IP addresses:
1146 0 - (default): reply for any local target IP address, configured
1148 1 - reply only if the target IP address is local address
1149 configured on the incoming interface
1150 2 - reply only if the target IP address is local address
1151 configured on the incoming interface and both with the
1152 sender's IP address are part from same subnet on this interface
1153 3 - do not reply for local addresses configured with scope host,
1154 only resolutions for global and link addresses are replied
1156 8 - do not reply for all local addresses
1158 The max value from conf/{all,interface}/arp_ignore is used
1159 when ARP request is received on the {interface}
1161 arp_notify - BOOLEAN
1162 Define mode for notification of address and device changes.
1163 0 - (default): do nothing
1164 1 - Generate gratuitous arp requests when device is brought up
1165 or hardware address changes.
1167 arp_accept - BOOLEAN
1168 Define behavior for gratuitous ARP frames who's IP is not
1169 already present in the ARP table:
1170 0 - don't create new entries in the ARP table
1171 1 - create new entries in the ARP table
1173 Both replies and requests type gratuitous arp will trigger the
1174 ARP table to be updated, if this setting is on.
1176 If the ARP table already contains the IP address of the
1177 gratuitous arp frame, the arp table will be updated regardless
1178 if this setting is on or off.
1180 mcast_solicit - INTEGER
1181 The maximum number of multicast probes in INCOMPLETE state,
1182 when the associated hardware address is unknown. Defaults
1185 ucast_solicit - INTEGER
1186 The maximum number of unicast probes in PROBE state, when
1187 the hardware address is being reconfirmed. Defaults to 3.
1189 app_solicit - INTEGER
1190 The maximum number of probes to send to the user space ARP daemon
1191 via netlink before dropping back to multicast probes (see
1192 mcast_resolicit). Defaults to 0.
1194 mcast_resolicit - INTEGER
1195 The maximum number of multicast probes after unicast and
1196 app probes in PROBE state. Defaults to 0.
1198 disable_policy - BOOLEAN
1199 Disable IPSEC policy (SPD) for this interface
1201 disable_xfrm - BOOLEAN
1202 Disable IPSEC encryption on this interface, whatever the policy
1204 igmpv2_unsolicited_report_interval - INTEGER
1205 The interval in milliseconds in which the next unsolicited
1206 IGMPv1 or IGMPv2 report retransmit will take place.
1207 Default: 10000 (10 seconds)
1209 igmpv3_unsolicited_report_interval - INTEGER
1210 The interval in milliseconds in which the next unsolicited
1211 IGMPv3 report retransmit will take place.
1212 Default: 1000 (1 seconds)
1214 promote_secondaries - BOOLEAN
1215 When a primary IP address is removed from this interface
1216 promote a corresponding secondary IP address instead of
1217 removing all the corresponding secondary IP addresses.
1219 drop_unicast_in_l2_multicast - BOOLEAN
1220 Drop any unicast IP packets that are received in link-layer
1221 multicast (or broadcast) frames.
1222 This behavior (for multicast) is actually a SHOULD in RFC
1223 1122, but is disabled by default for compatibility reasons.
1228 Allows you to write a number, which can be used as required.
1231 xfrm4_gc_thresh - INTEGER
1232 The threshold at which we will start garbage collecting for IPv4
1233 destination cache entries. At twice this value the system will
1234 refuse new allocations. The value must be set below the flowcache
1235 limit (4096 * number of online cpus) to take effect.
1237 igmp_link_local_mcast_reports - BOOLEAN
1238 Enable IGMP reports for link local multicast groups in the
1243 kuznet@ms2.inr.ac.ru
1249 delon.nicolas@wanadoo.fr
1254 /proc/sys/net/ipv6/* Variables:
1256 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1257 apply to IPv6 [XXX?].
1259 bindv6only - BOOLEAN
1260 Default value for IPV6_V6ONLY socket option,
1261 which restricts use of the IPv6 socket to IPv6 communication
1263 TRUE: disable IPv4-mapped address feature
1264 FALSE: enable IPv4-mapped address feature
1266 Default: FALSE (as specified in RFC3493)
1268 flowlabel_consistency - BOOLEAN
1269 Protect the consistency (and unicity) of flow label.
1270 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1276 auto_flowlabels - INTEGER
1277 Automatically generate flow labels based on a flow hash of the
1278 packet. This allows intermediate devices, such as routers, to
1279 identify packet flows for mechanisms like Equal Cost Multipath
1280 Routing (see RFC 6438).
1281 0: automatic flow labels are completely disabled
1282 1: automatic flow labels are enabled by default, they can be
1283 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1285 2: automatic flow labels are allowed, they may be enabled on a
1286 per socket basis using the IPV6_AUTOFLOWLABEL socket option
1287 3: automatic flow labels are enabled and enforced, they cannot
1288 be disabled by the socket option
1291 flowlabel_state_ranges - BOOLEAN
1292 Split the flow label number space into two ranges. 0-0x7FFFF is
1293 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1294 is reserved for stateless flow labels as described in RFC6437.
1299 anycast_src_echo_reply - BOOLEAN
1300 Controls the use of anycast addresses as source addresses for ICMPv6
1306 idgen_delay - INTEGER
1307 Controls the delay in seconds after which time to retry
1308 privacy stable address generation if a DAD conflict is
1310 Default: 1 (as specified in RFC7217)
1312 idgen_retries - INTEGER
1313 Controls the number of retries to generate a stable privacy
1314 address if a DAD conflict is detected.
1315 Default: 3 (as specified in RFC7217)
1318 Controls the MLD query robustness variable (see RFC3810 9.1).
1319 Default: 2 (as specified by RFC3810 9.1)
1320 Minimum: 1 (as specified by RFC6636 4.5)
1324 ip6frag_high_thresh - INTEGER
1325 Maximum memory used to reassemble IPv6 fragments. When
1326 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1327 the fragment handler will toss packets until ip6frag_low_thresh
1330 ip6frag_low_thresh - INTEGER
1331 See ip6frag_high_thresh
1333 ip6frag_time - INTEGER
1334 Time in seconds to keep an IPv6 fragment in memory.
1337 Change the interface-specific default settings.
1341 Change all the interface-specific settings.
1343 [XXX: Other special features than forwarding?]
1345 conf/all/forwarding - BOOLEAN
1346 Enable global IPv6 forwarding between all interfaces.
1348 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1349 to control which interfaces may forward packets and which not.
1351 This also sets all interfaces' Host/Router setting
1352 'forwarding' to the specified value. See below for details.
1354 This referred to as global forwarding.
1359 fwmark_reflect - BOOLEAN
1360 Controls the fwmark of kernel-generated IPv6 reply packets that are not
1361 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
1362 If unset, these packets have a fwmark of zero. If set, they have the
1363 fwmark of the packet they are replying to.
1367 Change special settings per interface.
1369 The functional behaviour for certain settings is different
1370 depending on whether local forwarding is enabled or not.
1373 Accept Router Advertisements; autoconfigure using them.
1375 It also determines whether or not to transmit Router
1376 Solicitations. If and only if the functional setting is to
1377 accept Router Advertisements, Router Solicitations will be
1380 Possible values are:
1381 0 Do not accept Router Advertisements.
1382 1 Accept Router Advertisements if forwarding is disabled.
1383 2 Overrule forwarding behaviour. Accept Router Advertisements
1384 even if forwarding is enabled.
1386 Functional default: enabled if local forwarding is disabled.
1387 disabled if local forwarding is enabled.
1389 accept_ra_defrtr - BOOLEAN
1390 Learn default router in Router Advertisement.
1392 Functional default: enabled if accept_ra is enabled.
1393 disabled if accept_ra is disabled.
1395 accept_ra_from_local - BOOLEAN
1396 Accept RA with source-address that is found on local machine
1397 if the RA is otherwise proper and able to be accepted.
1398 Default is to NOT accept these as it may be an un-intended
1402 enabled if accept_ra_from_local is enabled
1403 on a specific interface.
1404 disabled if accept_ra_from_local is disabled
1405 on a specific interface.
1407 accept_ra_min_hop_limit - INTEGER
1408 Minimum hop limit Information in Router Advertisement.
1410 Hop limit Information in Router Advertisement less than this
1411 variable shall be ignored.
1415 accept_ra_pinfo - BOOLEAN
1416 Learn Prefix Information in Router Advertisement.
1418 Functional default: enabled if accept_ra is enabled.
1419 disabled if accept_ra is disabled.
1421 accept_ra_rt_info_max_plen - INTEGER
1422 Maximum prefix length of Route Information in RA.
1424 Route Information w/ prefix larger than or equal to this
1425 variable shall be ignored.
1427 Functional default: 0 if accept_ra_rtr_pref is enabled.
1428 -1 if accept_ra_rtr_pref is disabled.
1430 accept_ra_rtr_pref - BOOLEAN
1431 Accept Router Preference in RA.
1433 Functional default: enabled if accept_ra is enabled.
1434 disabled if accept_ra is disabled.
1436 accept_ra_mtu - BOOLEAN
1437 Apply the MTU value specified in RA option 5 (RFC4861). If
1438 disabled, the MTU specified in the RA will be ignored.
1440 Functional default: enabled if accept_ra is enabled.
1441 disabled if accept_ra is disabled.
1443 accept_redirects - BOOLEAN
1446 Functional default: enabled if local forwarding is disabled.
1447 disabled if local forwarding is enabled.
1449 accept_source_route - INTEGER
1450 Accept source routing (routing extension header).
1452 >= 0: Accept only routing header type 2.
1453 < 0: Do not accept routing header.
1458 Autoconfigure addresses using Prefix Information in Router
1461 Functional default: enabled if accept_ra_pinfo is enabled.
1462 disabled if accept_ra_pinfo is disabled.
1464 dad_transmits - INTEGER
1465 The amount of Duplicate Address Detection probes to send.
1468 forwarding - INTEGER
1469 Configure interface-specific Host/Router behaviour.
1471 Note: It is recommended to have the same setting on all
1472 interfaces; mixed router/host scenarios are rather uncommon.
1474 Possible values are:
1475 0 Forwarding disabled
1476 1 Forwarding enabled
1480 By default, Host behaviour is assumed. This means:
1482 1. IsRouter flag is not set in Neighbour Advertisements.
1483 2. If accept_ra is TRUE (default), transmit Router
1485 3. If accept_ra is TRUE (default), accept Router
1486 Advertisements (and do autoconfiguration).
1487 4. If accept_redirects is TRUE (default), accept Redirects.
1491 If local forwarding is enabled, Router behaviour is assumed.
1492 This means exactly the reverse from the above:
1494 1. IsRouter flag is set in Neighbour Advertisements.
1495 2. Router Solicitations are not sent unless accept_ra is 2.
1496 3. Router Advertisements are ignored unless accept_ra is 2.
1497 4. Redirects are ignored.
1499 Default: 0 (disabled) if global forwarding is disabled (default),
1500 otherwise 1 (enabled).
1503 Default Hop Limit to set.
1507 Default Maximum Transfer Unit
1508 Default: 1280 (IPv6 required minimum)
1510 ip_nonlocal_bind - BOOLEAN
1511 If set, allows processes to bind() to non-local IPv6 addresses,
1512 which can be quite useful - but may break some applications.
1515 router_probe_interval - INTEGER
1516 Minimum interval (in seconds) between Router Probing described
1521 router_solicitation_delay - INTEGER
1522 Number of seconds to wait after interface is brought up
1523 before sending Router Solicitations.
1526 router_solicitation_interval - INTEGER
1527 Number of seconds to wait between Router Solicitations.
1530 router_solicitations - INTEGER
1531 Number of Router Solicitations to send until assuming no
1532 routers are present.
1535 use_oif_addrs_only - BOOLEAN
1536 When enabled, the candidate source addresses for destinations
1537 routed via this interface are restricted to the set of addresses
1538 configured on this interface (vis. RFC 6724, section 4).
1542 use_tempaddr - INTEGER
1543 Preference for Privacy Extensions (RFC3041).
1544 <= 0 : disable Privacy Extensions
1545 == 1 : enable Privacy Extensions, but prefer public
1546 addresses over temporary addresses.
1547 > 1 : enable Privacy Extensions and prefer temporary
1548 addresses over public addresses.
1549 Default: 0 (for most devices)
1550 -1 (for point-to-point devices and loopback devices)
1552 temp_valid_lft - INTEGER
1553 valid lifetime (in seconds) for temporary addresses.
1554 Default: 604800 (7 days)
1556 temp_prefered_lft - INTEGER
1557 Preferred lifetime (in seconds) for temporary addresses.
1558 Default: 86400 (1 day)
1560 max_desync_factor - INTEGER
1561 Maximum value for DESYNC_FACTOR, which is a random value
1562 that ensures that clients don't synchronize with each
1563 other and generate new addresses at exactly the same time.
1564 value is in seconds.
1567 regen_max_retry - INTEGER
1568 Number of attempts before give up attempting to generate
1569 valid temporary addresses.
1572 max_addresses - INTEGER
1573 Maximum number of autoconfigured addresses per interface. Setting
1574 to zero disables the limitation. It is not recommended to set this
1575 value too large (or to zero) because it would be an easy way to
1576 crash the kernel by allowing too many addresses to be created.
1579 disable_ipv6 - BOOLEAN
1580 Disable IPv6 operation. If accept_dad is set to 2, this value
1581 will be dynamically set to TRUE if DAD fails for the link-local
1583 Default: FALSE (enable IPv6 operation)
1585 When this value is changed from 1 to 0 (IPv6 is being enabled),
1586 it will dynamically create a link-local address on the given
1587 interface and start Duplicate Address Detection, if necessary.
1589 When this value is changed from 0 to 1 (IPv6 is being disabled),
1590 it will dynamically delete all address on the given interface.
1592 accept_dad - INTEGER
1593 Whether to accept DAD (Duplicate Address Detection).
1595 1: Enable DAD (default)
1596 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1597 link-local address has been found.
1599 force_tllao - BOOLEAN
1600 Enable sending the target link-layer address option even when
1601 responding to a unicast neighbor solicitation.
1604 Quoting from RFC 2461, section 4.4, Target link-layer address:
1606 "The option MUST be included for multicast solicitations in order to
1607 avoid infinite Neighbor Solicitation "recursion" when the peer node
1608 does not have a cache entry to return a Neighbor Advertisements
1609 message. When responding to unicast solicitations, the option can be
1610 omitted since the sender of the solicitation has the correct link-
1611 layer address; otherwise it would not have be able to send the unicast
1612 solicitation in the first place. However, including the link-layer
1613 address in this case adds little overhead and eliminates a potential
1614 race condition where the sender deletes the cached link-layer address
1615 prior to receiving a response to a previous solicitation."
1617 ndisc_notify - BOOLEAN
1618 Define mode for notification of address and device changes.
1619 0 - (default): do nothing
1620 1 - Generate unsolicited neighbour advertisements when device is brought
1621 up or hardware address changes.
1623 mldv1_unsolicited_report_interval - INTEGER
1624 The interval in milliseconds in which the next unsolicited
1625 MLDv1 report retransmit will take place.
1626 Default: 10000 (10 seconds)
1628 mldv2_unsolicited_report_interval - INTEGER
1629 The interval in milliseconds in which the next unsolicited
1630 MLDv2 report retransmit will take place.
1631 Default: 1000 (1 second)
1633 force_mld_version - INTEGER
1634 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1635 1 - Enforce to use MLD version 1
1636 2 - Enforce to use MLD version 2
1638 suppress_frag_ndisc - INTEGER
1639 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1640 with IPv6 Neighbor Discovery) behavior:
1641 1 - (default) discard fragmented neighbor discovery packets
1642 0 - allow fragmented neighbor discovery packets
1644 optimistic_dad - BOOLEAN
1645 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
1646 0: disabled (default)
1649 use_optimistic - BOOLEAN
1650 If enabled, do not classify optimistic addresses as deprecated during
1651 source address selection. Preferred addresses will still be chosen
1652 before optimistic addresses, subject to other ranking in the source
1653 address selection algorithm.
1654 0: disabled (default)
1657 stable_secret - IPv6 address
1658 This IPv6 address will be used as a secret to generate IPv6
1659 addresses for link-local addresses and autoconfigured
1660 ones. All addresses generated after setting this secret will
1661 be stable privacy ones by default. This can be changed via the
1662 addrgenmode ip-link. conf/default/stable_secret is used as the
1663 secret for the namespace, the interface specific ones can
1664 overwrite that. Writes to conf/all/stable_secret are refused.
1666 It is recommended to generate this secret during installation
1667 of a system and keep it stable after that.
1669 By default the stable secret is unset.
1673 Limit the maximal rates for sending ICMPv6 packets.
1674 0 to disable any limiting,
1675 otherwise the minimal space between responses in milliseconds.
1678 xfrm6_gc_thresh - INTEGER
1679 The threshold at which we will start garbage collecting for IPv6
1680 destination cache entries. At twice this value the system will
1681 refuse new allocations. The value must be set below the flowcache
1682 limit (4096 * number of online cpus) to take effect.
1686 Pekka Savola <pekkas@netcore.fi>
1687 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1690 /proc/sys/net/bridge/* Variables:
1692 bridge-nf-call-arptables - BOOLEAN
1693 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1697 bridge-nf-call-iptables - BOOLEAN
1698 1 : pass bridged IPv4 traffic to iptables' chains.
1702 bridge-nf-call-ip6tables - BOOLEAN
1703 1 : pass bridged IPv6 traffic to ip6tables' chains.
1707 bridge-nf-filter-vlan-tagged - BOOLEAN
1708 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1712 bridge-nf-filter-pppoe-tagged - BOOLEAN
1713 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1717 bridge-nf-pass-vlan-input-dev - BOOLEAN
1718 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1719 interface on the bridge and set the netfilter input device to the vlan.
1720 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1721 target work with vlan-on-top-of-bridge interfaces. When no matching
1722 vlan interface is found, or this switch is off, the input device is
1723 set to the bridge interface.
1724 0: disable bridge netfilter vlan interface lookup.
1727 proc/sys/net/sctp/* Variables:
1729 addip_enable - BOOLEAN
1730 Enable or disable extension of Dynamic Address Reconfiguration
1731 (ADD-IP) functionality specified in RFC5061. This extension provides
1732 the ability to dynamically add and remove new addresses for the SCTP
1735 1: Enable extension.
1737 0: Disable extension.
1742 Enable or disable pf (pf is short for potentially failed) state. A value
1743 of pf_retrans > path_max_retrans also disables pf state. That is, one of
1744 both pf_enable and pf_retrans > path_max_retrans can disable pf state.
1745 Since pf_retrans and path_max_retrans can be changed by userspace
1746 application, sometimes user expects to disable pf state by the value of
1747 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
1748 or path_max_retrans is changed by the user application, this pf state is
1749 enabled. As such, it is necessary to add this to dynamically enable
1750 and disable pf state. See:
1751 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
1760 addip_noauth_enable - BOOLEAN
1761 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1762 authentication to protect the operations of adding or removing new
1763 addresses. This requirement is mandated so that unauthorized hosts
1764 would not be able to hijack associations. However, older
1765 implementations may not have implemented this requirement while
1766 allowing the ADD-IP extension. For reasons of interoperability,
1767 we provide this variable to control the enforcement of the
1768 authentication requirement.
1770 1: Allow ADD-IP extension to be used without authentication. This
1771 should only be set in a closed environment for interoperability
1772 with older implementations.
1774 0: Enforce the authentication requirement
1778 auth_enable - BOOLEAN
1779 Enable or disable Authenticated Chunks extension. This extension
1780 provides the ability to send and receive authenticated chunks and is
1781 required for secure operation of Dynamic Address Reconfiguration
1784 1: Enable this extension.
1785 0: Disable this extension.
1789 prsctp_enable - BOOLEAN
1790 Enable or disable the Partial Reliability extension (RFC3758) which
1791 is used to notify peers that a given DATA should no longer be expected.
1799 The limit of the number of new packets that can be initially sent. It
1800 controls how bursty the generated traffic can be.
1804 association_max_retrans - INTEGER
1805 Set the maximum number for retransmissions that an association can
1806 attempt deciding that the remote end is unreachable. If this value
1807 is exceeded, the association is terminated.
1811 max_init_retransmits - INTEGER
1812 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1813 that an association will attempt before declaring the destination
1814 unreachable and terminating.
1818 path_max_retrans - INTEGER
1819 The maximum number of retransmissions that will be attempted on a given
1820 path. Once this threshold is exceeded, the path is considered
1821 unreachable, and new traffic will use a different path when the
1822 association is multihomed.
1826 pf_retrans - INTEGER
1827 The number of retransmissions that will be attempted on a given path
1828 before traffic is redirected to an alternate transport (should one
1829 exist). Note this is distinct from path_max_retrans, as a path that
1830 passes the pf_retrans threshold can still be used. Its only
1831 deprioritized when a transmission path is selected by the stack. This
1832 setting is primarily used to enable fast failover mechanisms without
1833 having to reduce path_max_retrans to a very low value. See:
1834 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1835 for details. Note also that a value of pf_retrans > path_max_retrans
1836 disables this feature. Since both pf_retrans and path_max_retrans can
1837 be changed by userspace application, a variable pf_enable is used to
1842 rto_initial - INTEGER
1843 The initial round trip timeout value in milliseconds that will be used
1844 in calculating round trip times. This is the initial time interval
1845 for retransmissions.
1850 The maximum value (in milliseconds) of the round trip timeout. This
1851 is the largest time interval that can elapse between retransmissions.
1856 The minimum value (in milliseconds) of the round trip timeout. This
1857 is the smallest time interval the can elapse between retransmissions.
1861 hb_interval - INTEGER
1862 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1863 are sent at the specified interval on idle paths to probe the state of
1864 a given path between 2 associations.
1868 sack_timeout - INTEGER
1869 The amount of time (in milliseconds) that the implementation will wait
1874 valid_cookie_life - INTEGER
1875 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1876 is used during association establishment.
1880 cookie_preserve_enable - BOOLEAN
1881 Enable or disable the ability to extend the lifetime of the SCTP cookie
1882 that is used during the establishment phase of SCTP association
1884 1: Enable cookie lifetime extension.
1889 cookie_hmac_alg - STRING
1890 Select the hmac algorithm used when generating the cookie value sent by
1891 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1896 Ability to assign md5 or sha1 as the selected alg is predicated on the
1897 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1898 CONFIG_CRYPTO_SHA1).
1900 Default: Dependent on configuration. MD5 if available, else SHA1 if
1901 available, else none.
1903 rcvbuf_policy - INTEGER
1904 Determines if the receive buffer is attributed to the socket or to
1905 association. SCTP supports the capability to create multiple
1906 associations on a single socket. When using this capability, it is
1907 possible that a single stalled association that's buffering a lot
1908 of data may block other associations from delivering their data by
1909 consuming all of the receive buffer space. To work around this,
1910 the rcvbuf_policy could be set to attribute the receiver buffer space
1911 to each association instead of the socket. This prevents the described
1914 1: rcvbuf space is per association
1915 0: rcvbuf space is per socket
1919 sndbuf_policy - INTEGER
1920 Similar to rcvbuf_policy above, this applies to send buffer space.
1922 1: Send buffer is tracked per association
1923 0: Send buffer is tracked per socket.
1927 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1928 Number of pages allowed for queueing by all SCTP sockets.
1930 min: Below this number of pages SCTP is not bothered about its
1931 memory appetite. When amount of memory allocated by SCTP exceeds
1932 this number, SCTP starts to moderate memory usage.
1934 pressure: This value was introduced to follow format of tcp_mem.
1936 max: Number of pages allowed for queueing by all SCTP sockets.
1938 Default is calculated at boot time from amount of available memory.
1940 sctp_rmem - vector of 3 INTEGERs: min, default, max
1941 Only the first value ("min") is used, "default" and "max" are
1944 min: Minimal size of receive buffer used by SCTP socket.
1945 It is guaranteed to each SCTP socket (but not association) even
1946 under moderate memory pressure.
1950 sctp_wmem - vector of 3 INTEGERs: min, default, max
1951 Currently this tunable has no effect.
1953 addr_scope_policy - INTEGER
1954 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1956 0 - Disable IPv4 address scoping
1957 1 - Enable IPv4 address scoping
1958 2 - Follow draft but allow IPv4 private addresses
1959 3 - Follow draft but allow IPv4 link local addresses
1964 /proc/sys/net/core/*
1965 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1968 /proc/sys/net/unix/*
1969 max_dgram_qlen - INTEGER
1970 The maximum length of dgram socket receive queue
1977 /proc/sys/net/irda/*
1978 fast_poll_increase FIXME
1979 warn_noreply_time FIXME
1980 discovery_slots FIXME
1983 discovery_timeout FIXME
1984 lap_keepalive_time FIXME
1985 max_noreply_time FIXME
1986 max_tx_data_size FIXME
1988 min_tx_turn_time FIXME