| blob | 3c12d9a7ed00391d5c3f49ef80d7b1ae9abc40fe |
1 /proc/sys/net/ipv4/* Variables:
3 ip_forward - BOOLEAN
4 0 - disabled (default)
5 not 0 - enabled
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
11 for routers)
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 - BOOLEAN
19 Disable Path MTU Discovery.
20 default FALSE
22 min_pmtu - INTEGER
23 default 552 - minimum discovered Path MTU
25 route/max_size - INTEGER
26 Maximum number of routes allowed in the kernel. Increase
27 this when using large numbers of interfaces and/or routes.
29 neigh/default/gc_thresh1 - INTEGER
30 Minimum number of entries to keep. Garbage collector will not
31 purge entries if there are fewer than this number.
32 Default: 128
34 neigh/default/gc_thresh3 - INTEGER
35 Maximum number of neighbor entries allowed. Increase this
36 when using large numbers of interfaces and when communicating
37 with large numbers of directly-connected peers.
38 Default: 1024
40 neigh/default/unres_qlen_bytes - INTEGER
41 The maximum number of bytes which may be used by packets
42 queued for each unresolved address by other network layers.
43 (added in linux 3.3)
44 Setting negative value is meaningless and will return error.
45 Default: 65536 Bytes(64KB)
47 neigh/default/unres_qlen - INTEGER
48 The maximum number of packets which may be queued for each
49 unresolved address by other network layers.
50 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
51 Prior to linux 3.3, the default value is 3 which may cause
52 unexpected packet loss. The current default value is calculated
53 according to default value of unres_qlen_bytes and true size of
54 packet.
55 Default: 31
57 mtu_expires - INTEGER
58 Time, in seconds, that cached PMTU information is kept.
60 min_adv_mss - INTEGER
61 The advertised MSS depends on the first hop route MTU, but will
62 never be lower than this setting.
64 IP Fragmentation:
66 ipfrag_high_thresh - INTEGER
67 Maximum memory used to reassemble IP fragments. When
68 ipfrag_high_thresh bytes of memory is allocated for this purpose,
69 the fragment handler will toss packets until ipfrag_low_thresh
70 is reached.
72 ipfrag_low_thresh - INTEGER
73 See ipfrag_high_thresh
75 ipfrag_time - INTEGER
76 Time in seconds to keep an IP fragment in memory.
78 ipfrag_secret_interval - INTEGER
79 Regeneration interval (in seconds) of the hash secret (or lifetime
80 for the hash secret) for IP fragments.
81 Default: 600
83 ipfrag_max_dist - INTEGER
84 ipfrag_max_dist is a non-negative integer value which defines the
85 maximum "disorder" which is allowed among fragments which share a
86 common IP source address. Note that reordering of packets is
87 not unusual, but if a large number of fragments arrive from a source
88 IP address while a particular fragment queue remains incomplete, it
89 probably indicates that one or more fragments belonging to that queue
90 have been lost. When ipfrag_max_dist is positive, an additional check
91 is done on fragments before they are added to a reassembly queue - if
92 ipfrag_max_dist (or more) fragments have arrived from a particular IP
93 address between additions to any IP fragment queue using that source
94 address, it's presumed that one or more fragments in the queue are
95 lost. The existing fragment queue will be dropped, and a new one
96 started. An ipfrag_max_dist value of zero disables this check.
98 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
99 result in unnecessarily dropping fragment queues when normal
100 reordering of packets occurs, which could lead to poor application
101 performance. Using a very large value, e.g. 50000, increases the
102 likelihood of incorrectly reassembling IP fragments that originate
103 from different IP datagrams, which could result in data corruption.
104 Default: 64
106 INET peer storage:
108 inet_peer_threshold - INTEGER
109 The approximate size of the storage. Starting from this threshold
110 entries will be thrown aggressively. This threshold also determines
111 entries' time-to-live and time intervals between garbage collection
112 passes. More entries, less time-to-live, less GC interval.
114 inet_peer_minttl - INTEGER
115 Minimum time-to-live of entries. Should be enough to cover fragment
116 time-to-live on the reassembling side. This minimum time-to-live is
117 guaranteed if the pool size is less than inet_peer_threshold.
118 Measured in seconds.
120 inet_peer_maxttl - INTEGER
121 Maximum time-to-live of entries. Unused entries will expire after
122 this period of time if there is no memory pressure on the pool (i.e.
123 when the number of entries in the pool is very small).
124 Measured in seconds.
126 TCP variables:
128 somaxconn - INTEGER
129 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
130 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
131 for TCP sockets.
133 tcp_abort_on_overflow - BOOLEAN
134 If listening service is too slow to accept new connections,
135 reset them. Default state is FALSE. It means that if overflow
136 occurred due to a burst, connection will recover. Enable this
137 option _only_ if you are really sure that listening daemon
138 cannot be tuned to accept connections faster. Enabling this
139 option can harm clients of your server.
141 tcp_adv_win_scale - INTEGER
142 Count buffering overhead as bytes/2^tcp_adv_win_scale
143 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
144 if it is <= 0.
145 Possible values are [-31, 31], inclusive.
146 Default: 1
148 tcp_allowed_congestion_control - STRING
149 Show/set the congestion control choices available to non-privileged
150 processes. The list is a subset of those listed in
151 tcp_available_congestion_control.
152 Default is "reno" and the default setting (tcp_congestion_control).
154 tcp_app_win - INTEGER
155 Reserve max(window/2^tcp_app_win, mss) of window for application
156 buffer. Value 0 is special, it means that nothing is reserved.
157 Default: 31
159 tcp_available_congestion_control - STRING
160 Shows the available congestion control choices that are registered.
161 More congestion control algorithms may be available as modules,
162 but not loaded.
164 tcp_base_mss - INTEGER
165 The initial value of search_low to be used by the packetization layer
166 Path MTU discovery (MTU probing). If MTU probing is enabled,
167 this is the initial MSS used by the connection.
169 tcp_congestion_control - STRING
170 Set the congestion control algorithm to be used for new
171 connections. The algorithm "reno" is always available, but
172 additional choices may be available based on kernel configuration.
173 Default is set as part of kernel configuration.
174 For passive connections, the listener congestion control choice
175 is inherited.
176 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
178 tcp_dsack - BOOLEAN
179 Allows TCP to send "duplicate" SACKs.
181 tcp_early_retrans - INTEGER
182 Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold
183 for triggering fast retransmit when the amount of outstanding data is
184 small and when no previously unsent data can be transmitted (such
185 that limited transmit could be used). Also controls the use of
186 Tail loss probe (TLP) that converts RTOs occurring due to tail
187 losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01).
188 Possible values:
189 0 disables ER
190 1 enables ER
191 2 enables ER but delays fast recovery and fast retransmit
192 by a fourth of RTT. This mitigates connection falsely
193 recovers when network has a small degree of reordering
194 (less than 3 packets).
195 3 enables delayed ER and TLP.
196 4 enables TLP only.
197 Default: 3
199 tcp_ecn - INTEGER
200 Control use of Explicit Congestion Notification (ECN) by TCP.
201 ECN is used only when both ends of the TCP connection indicate
202 support for it. This feature is useful in avoiding losses due
203 to congestion by allowing supporting routers to signal
204 congestion before having to drop packets.
205 Possible values are:
206 0 Disable ECN. Neither initiate nor accept ECN.
207 1 Enable ECN when requested by incoming connections and
208 also request ECN on outgoing connection attempts.
209 2 Enable ECN when requested by incoming connections
210 but do not request ECN on outgoing connections.
211 Default: 2
213 tcp_fack - BOOLEAN
214 Enable FACK congestion avoidance and fast retransmission.
215 The value is not used, if tcp_sack is not enabled.
217 tcp_fin_timeout - INTEGER
218 The length of time an orphaned (no longer referenced by any
219 application) connection will remain in the FIN_WAIT_2 state
220 before it is aborted at the local end. While a perfectly
221 valid "receive only" state for an un-orphaned connection, an
222 orphaned connection in FIN_WAIT_2 state could otherwise wait
223 forever for the remote to close its end of the connection.
224 Cf. tcp_max_orphans
225 Default: 60 seconds
227 tcp_frto - INTEGER
228 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
229 F-RTO is an enhanced recovery algorithm for TCP retransmission
230 timeouts. It is particularly beneficial in networks where the
231 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
232 modification. It does not require any support from the peer.
234 By default it's enabled with a non-zero value. 0 disables F-RTO.
236 tcp_keepalive_time - INTEGER
237 How often TCP sends out keepalive messages when keepalive is enabled.
238 Default: 2hours.
240 tcp_keepalive_probes - INTEGER
241 How many keepalive probes TCP sends out, until it decides that the
242 connection is broken. Default value: 9.
244 tcp_keepalive_intvl - INTEGER
245 How frequently the probes are send out. Multiplied by
246 tcp_keepalive_probes it is time to kill not responding connection,
247 after probes started. Default value: 75sec i.e. connection
248 will be aborted after ~11 minutes of retries.
250 tcp_low_latency - BOOLEAN
251 If set, the TCP stack makes decisions that prefer lower
252 latency as opposed to higher throughput. By default, this
253 option is not set meaning that higher throughput is preferred.
254 An example of an application where this default should be
255 changed would be a Beowulf compute cluster.
256 Default: 0
258 tcp_max_orphans - INTEGER
259 Maximal number of TCP sockets not attached to any user file handle,
260 held by system. If this number is exceeded orphaned connections are
261 reset immediately and warning is printed. This limit exists
262 only to prevent simple DoS attacks, you _must_ not rely on this
263 or lower the limit artificially, but rather increase it
264 (probably, after increasing installed memory),
265 if network conditions require more than default value,
266 and tune network services to linger and kill such states
267 more aggressively. Let me to remind again: each orphan eats
268 up to ~64K of unswappable memory.
270 tcp_max_syn_backlog - INTEGER
271 Maximal number of remembered connection requests, which have not
272 received an acknowledgment from connecting client.
273 The minimal value is 128 for low memory machines, and it will
274 increase in proportion to the memory of machine.
275 If server suffers from overload, try increasing this number.
277 tcp_max_tw_buckets - INTEGER
278 Maximal number of timewait sockets held by system simultaneously.
279 If this number is exceeded time-wait socket is immediately destroyed
280 and warning is printed. This limit exists only to prevent
281 simple DoS attacks, you _must_ not lower the limit artificially,
282 but rather increase it (probably, after increasing installed memory),
283 if network conditions require more than default value.
285 tcp_mem - vector of 3 INTEGERs: min, pressure, max
286 min: below this number of pages TCP is not bothered about its
287 memory appetite.
289 pressure: when amount of memory allocated by TCP exceeds this number
290 of pages, TCP moderates its memory consumption and enters memory
291 pressure mode, which is exited when memory consumption falls
292 under "min".
294 max: number of pages allowed for queueing by all TCP sockets.
296 Defaults are calculated at boot time from amount of available
297 memory.
299 tcp_moderate_rcvbuf - BOOLEAN
300 If set, TCP performs receive buffer auto-tuning, attempting to
301 automatically size the buffer (no greater than tcp_rmem[2]) to
302 match the size required by the path for full throughput. Enabled by
303 default.
305 tcp_mtu_probing - INTEGER
306 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
307 values:
308 0 - Disabled
309 1 - Disabled by default, enabled when an ICMP black hole detected
310 2 - Always enabled, use initial MSS of tcp_base_mss.
312 tcp_no_metrics_save - BOOLEAN
313 By default, TCP saves various connection metrics in the route cache
314 when the connection closes, so that connections established in the
315 near future can use these to set initial conditions. Usually, this
316 increases overall performance, but may sometimes cause performance
317 degradation. If set, TCP will not cache metrics on closing
318 connections.
320 tcp_orphan_retries - INTEGER
321 This value influences the timeout of a locally closed TCP connection,
322 when RTO retransmissions remain unacknowledged.
323 See tcp_retries2 for more details.
325 The default value is 8.
326 If your machine is a loaded WEB server,
327 you should think about lowering this value, such sockets
328 may consume significant resources. Cf. tcp_max_orphans.
330 tcp_reordering - INTEGER
331 Maximal reordering of packets in a TCP stream.
332 Default: 3
334 tcp_retrans_collapse - BOOLEAN
335 Bug-to-bug compatibility with some broken printers.
336 On retransmit try to send bigger packets to work around bugs in
337 certain TCP stacks.
339 tcp_retries1 - INTEGER
340 This value influences the time, after which TCP decides, that
341 something is wrong due to unacknowledged RTO retransmissions,
342 and reports this suspicion to the network layer.
343 See tcp_retries2 for more details.
345 RFC 1122 recommends at least 3 retransmissions, which is the
346 default.
348 tcp_retries2 - INTEGER
349 This value influences the timeout of an alive TCP connection,
350 when RTO retransmissions remain unacknowledged.
351 Given a value of N, a hypothetical TCP connection following
352 exponential backoff with an initial RTO of TCP_RTO_MIN would
353 retransmit N times before killing the connection at the (N+1)th RTO.
355 The default value of 15 yields a hypothetical timeout of 924.6
356 seconds and is a lower bound for the effective timeout.
357 TCP will effectively time out at the first RTO which exceeds the
358 hypothetical timeout.
360 RFC 1122 recommends at least 100 seconds for the timeout,
361 which corresponds to a value of at least 8.
363 tcp_rfc1337 - BOOLEAN
364 If set, the TCP stack behaves conforming to RFC1337. If unset,
365 we are not conforming to RFC, but prevent TCP TIME_WAIT
366 assassination.
367 Default: 0
369 tcp_rmem - vector of 3 INTEGERs: min, default, max
370 min: Minimal size of receive buffer used by TCP sockets.
371 It is guaranteed to each TCP socket, even under moderate memory
372 pressure.
373 Default: 1 page
375 default: initial size of receive buffer used by TCP sockets.
376 This value overrides net.core.rmem_default used by other protocols.
377 Default: 87380 bytes. This value results in window of 65535 with
378 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
379 less for default tcp_app_win. See below about these variables.
381 max: maximal size of receive buffer allowed for automatically
382 selected receiver buffers for TCP socket. This value does not override
383 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
384 automatic tuning of that socket's receive buffer size, in which
385 case this value is ignored.
386 Default: between 87380B and 6MB, depending on RAM size.
388 tcp_sack - BOOLEAN
389 Enable select acknowledgments (SACKS).
391 tcp_slow_start_after_idle - BOOLEAN
392 If set, provide RFC2861 behavior and time out the congestion
393 window after an idle period. An idle period is defined at
394 the current RTO. If unset, the congestion window will not
395 be timed out after an idle period.
396 Default: 1
398 tcp_stdurg - BOOLEAN
399 Use the Host requirements interpretation of the TCP urgent pointer field.
400 Most hosts use the older BSD interpretation, so if you turn this on
401 Linux might not communicate correctly with them.
402 Default: FALSE
404 tcp_synack_retries - INTEGER
405 Number of times SYNACKs for a passive TCP connection attempt will
406 be retransmitted. Should not be higher than 255. Default value
407 is 5, which corresponds to 31seconds till the last retransmission
408 with the current initial RTO of 1second. With this the final timeout
409 for a passive TCP connection will happen after 63seconds.
411 tcp_syncookies - BOOLEAN
412 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
413 Send out syncookies when the syn backlog queue of a socket
414 overflows. This is to prevent against the common 'SYN flood attack'
415 Default: 1
417 Note, that syncookies is fallback facility.
418 It MUST NOT be used to help highly loaded servers to stand
419 against legal connection rate. If you see SYN flood warnings
420 in your logs, but investigation shows that they occur
421 because of overload with legal connections, you should tune
422 another parameters until this warning disappear.
423 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
425 syncookies seriously violate TCP protocol, do not allow
426 to use TCP extensions, can result in serious degradation
427 of some services (f.e. SMTP relaying), visible not by you,
428 but your clients and relays, contacting you. While you see
429 SYN flood warnings in logs not being really flooded, your server
430 is seriously misconfigured.
432 If you want to test which effects syncookies have to your
433 network connections you can set this knob to 2 to enable
434 unconditionally generation of syncookies.
436 tcp_fastopen - INTEGER
437 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
438 in the opening SYN packet. To use this feature, the client application
439 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
440 connect() to perform a TCP handshake automatically.
442 The values (bitmap) are
443 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN.
444 2: Enables TCP Fast Open on the server side, i.e., allowing data in
445 a SYN packet to be accepted and passed to the application before
446 3-way hand shake finishes.
447 4: Send data in the opening SYN regardless of cookie availability and
448 without a cookie option.
449 0x100: Accept SYN data w/o validating the cookie.
450 0x200: Accept data-in-SYN w/o any cookie option present.
451 0x400/0x800: Enable Fast Open on all listeners regardless of the
452 TCP_FASTOPEN socket option. The two different flags designate two
453 different ways of setting max_qlen without the TCP_FASTOPEN socket
454 option.
456 Default: 1
458 Note that the client & server side Fast Open flags (1 and 2
459 respectively) must be also enabled before the rest of flags can take
460 effect.
462 See include/net/tcp.h and the code for more details.
464 tcp_syn_retries - INTEGER
465 Number of times initial SYNs for an active TCP connection attempt
466 will be retransmitted. Should not be higher than 255. Default value
467 is 6, which corresponds to 63seconds till the last retransmission
468 with the current initial RTO of 1second. With this the final timeout
469 for an active TCP connection attempt will happen after 127seconds.
471 tcp_timestamps - BOOLEAN
472 Enable timestamps as defined in RFC1323.
474 tcp_min_tso_segs - INTEGER
475 Minimal number of segments per TSO frame.
476 Since linux-3.12, TCP does an automatic sizing of TSO frames,
477 depending on flow rate, instead of filling 64Kbytes packets.
478 For specific usages, it's possible to force TCP to build big
479 TSO frames. Note that TCP stack might split too big TSO packets
480 if available window is too small.
481 Default: 2
483 tcp_tso_win_divisor - INTEGER
484 This allows control over what percentage of the congestion window
485 can be consumed by a single TSO frame.
486 The setting of this parameter is a choice between burstiness and
487 building larger TSO frames.
488 Default: 3
490 tcp_tw_recycle - BOOLEAN
491 Enable fast recycling TIME-WAIT sockets. Default value is 0.
492 It should not be changed without advice/request of technical
493 experts.
495 tcp_tw_reuse - BOOLEAN
496 Allow to reuse TIME-WAIT sockets for new connections when it is
497 safe from protocol viewpoint. Default value is 0.
498 It should not be changed without advice/request of technical
499 experts.
501 tcp_window_scaling - BOOLEAN
502 Enable window scaling as defined in RFC1323.
504 tcp_wmem - vector of 3 INTEGERs: min, default, max
505 min: Amount of memory reserved for send buffers for TCP sockets.
506 Each TCP socket has rights to use it due to fact of its birth.
507 Default: 1 page
509 default: initial size of send buffer used by TCP sockets. This
510 value overrides net.core.wmem_default used by other protocols.
511 It is usually lower than net.core.wmem_default.
512 Default: 16K
514 max: Maximal amount of memory allowed for automatically tuned
515 send buffers for TCP sockets. This value does not override
516 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
517 automatic tuning of that socket's send buffer size, in which case
518 this value is ignored.
519 Default: between 64K and 4MB, depending on RAM size.
521 tcp_notsent_lowat - UNSIGNED INTEGER
522 A TCP socket can control the amount of unsent bytes in its write queue,
523 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
524 reports POLLOUT events if the amount of unsent bytes is below a per
525 socket value, and if the write queue is not full. sendmsg() will
526 also not add new buffers if the limit is hit.
528 This global variable controls the amount of unsent data for
529 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
530 to the global variable has immediate effect.
532 Default: UINT_MAX (0xFFFFFFFF)
534 tcp_workaround_signed_windows - BOOLEAN
535 If set, assume no receipt of a window scaling option means the
536 remote TCP is broken and treats the window as a signed quantity.
537 If unset, assume the remote TCP is not broken even if we do
538 not receive a window scaling option from them.
539 Default: 0
541 tcp_dma_copybreak - INTEGER
542 Lower limit, in bytes, of the size of socket reads that will be
543 offloaded to a DMA copy engine, if one is present in the system
544 and CONFIG_NET_DMA is enabled.
545 Default: 4096
547 tcp_thin_linear_timeouts - BOOLEAN
548 Enable dynamic triggering of linear timeouts for thin streams.
549 If set, a check is performed upon retransmission by timeout to
550 determine if the stream is thin (less than 4 packets in flight).
551 As long as the stream is found to be thin, up to 6 linear
552 timeouts may be performed before exponential backoff mode is
553 initiated. This improves retransmission latency for
554 non-aggressive thin streams, often found to be time-dependent.
555 For more information on thin streams, see
556 Documentation/networking/tcp-thin.txt
557 Default: 0
559 tcp_thin_dupack - BOOLEAN
560 Enable dynamic triggering of retransmissions after one dupACK
561 for thin streams. If set, a check is performed upon reception
562 of a dupACK to determine if the stream is thin (less than 4
563 packets in flight). As long as the stream is found to be thin,
564 data is retransmitted on the first received dupACK. This
565 improves retransmission latency for non-aggressive thin
566 streams, often found to be time-dependent.
567 For more information on thin streams, see
568 Documentation/networking/tcp-thin.txt
569 Default: 0
571 tcp_limit_output_bytes - INTEGER
572 Controls TCP Small Queue limit per tcp socket.
573 TCP bulk sender tends to increase packets in flight until it
574 gets losses notifications. With SNDBUF autotuning, this can
575 result in a large amount of packets queued in qdisc/device
576 on the local machine, hurting latency of other flows, for
577 typical pfifo_fast qdiscs.
578 tcp_limit_output_bytes limits the number of bytes on qdisc
579 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
580 Default: 131072
582 tcp_challenge_ack_limit - INTEGER
583 Limits number of Challenge ACK sent per second, as recommended
584 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
585 Default: 100
587 UDP variables:
589 udp_mem - vector of 3 INTEGERs: min, pressure, max
590 Number of pages allowed for queueing by all UDP sockets.
592 min: Below this number of pages UDP is not bothered about its
593 memory appetite. When amount of memory allocated by UDP exceeds
594 this number, UDP starts to moderate memory usage.
596 pressure: This value was introduced to follow format of tcp_mem.
598 max: Number of pages allowed for queueing by all UDP sockets.
600 Default is calculated at boot time from amount of available memory.
602 udp_rmem_min - INTEGER
603 Minimal size of receive buffer used by UDP sockets in moderation.
604 Each UDP socket is able to use the size for receiving data, even if
605 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
606 Default: 1 page
608 udp_wmem_min - INTEGER
609 Minimal size of send buffer used by UDP sockets in moderation.
610 Each UDP socket is able to use the size for sending data, even if
611 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
612 Default: 1 page
614 CIPSOv4 Variables:
616 cipso_cache_enable - BOOLEAN
617 If set, enable additions to and lookups from the CIPSO label mapping
618 cache. If unset, additions are ignored and lookups always result in a
619 miss. However, regardless of the setting the cache is still
620 invalidated when required when means you can safely toggle this on and
621 off and the cache will always be "safe".
622 Default: 1
624 cipso_cache_bucket_size - INTEGER
625 The CIPSO label cache consists of a fixed size hash table with each
626 hash bucket containing a number of cache entries. This variable limits
627 the number of entries in each hash bucket; the larger the value the
628 more CIPSO label mappings that can be cached. When the number of
629 entries in a given hash bucket reaches this limit adding new entries
630 causes the oldest entry in the bucket to be removed to make room.
631 Default: 10
633 cipso_rbm_optfmt - BOOLEAN
634 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
635 the CIPSO draft specification (see Documentation/netlabel for details).
636 This means that when set the CIPSO tag will be padded with empty
637 categories in order to make the packet data 32-bit aligned.
638 Default: 0
640 cipso_rbm_structvalid - BOOLEAN
641 If set, do a very strict check of the CIPSO option when
642 ip_options_compile() is called. If unset, relax the checks done during
643 ip_options_compile(). Either way is "safe" as errors are caught else
644 where in the CIPSO processing code but setting this to 0 (False) should
645 result in less work (i.e. it should be faster) but could cause problems
646 with other implementations that require strict checking.
647 Default: 0
649 IP Variables:
651 ip_local_port_range - 2 INTEGERS
652 Defines the local port range that is used by TCP and UDP to
653 choose the local port. The first number is the first, the
654 second the last local port number. The default values are
655 32768 and 61000 respectively.
657 ip_local_reserved_ports - list of comma separated ranges
658 Specify the ports which are reserved for known third-party
659 applications. These ports will not be used by automatic port
660 assignments (e.g. when calling connect() or bind() with port
661 number 0). Explicit port allocation behavior is unchanged.
663 The format used for both input and output is a comma separated
664 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
665 10). Writing to the file will clear all previously reserved
666 ports and update the current list with the one given in the
667 input.
669 Note that ip_local_port_range and ip_local_reserved_ports
670 settings are independent and both are considered by the kernel
671 when determining which ports are available for automatic port
672 assignments.
674 You can reserve ports which are not in the current
675 ip_local_port_range, e.g.:
677 $ cat /proc/sys/net/ipv4/ip_local_port_range
678 32000 61000
679 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
680 8080,9148
682 although this is redundant. However such a setting is useful
683 if later the port range is changed to a value that will
684 include the reserved ports.
686 Default: Empty
688 ip_nonlocal_bind - BOOLEAN
689 If set, allows processes to bind() to non-local IP addresses,
690 which can be quite useful - but may break some applications.
691 Default: 0
693 ip_dynaddr - BOOLEAN
694 If set non-zero, enables support for dynamic addresses.
695 If set to a non-zero value larger than 1, a kernel log
696 message will be printed when dynamic address rewriting
697 occurs.
698 Default: 0
700 ip_early_demux - BOOLEAN
701 Optimize input packet processing down to one demux for
702 certain kinds of local sockets. Currently we only do this
703 for established TCP sockets.
705 It may add an additional cost for pure routing workloads that
706 reduces overall throughput, in such case you should disable it.
707 Default: 1
709 icmp_echo_ignore_all - BOOLEAN
710 If set non-zero, then the kernel will ignore all ICMP ECHO
711 requests sent to it.
712 Default: 0
714 icmp_echo_ignore_broadcasts - BOOLEAN
715 If set non-zero, then the kernel will ignore all ICMP ECHO and
716 TIMESTAMP requests sent to it via broadcast/multicast.
717 Default: 1
719 icmp_ratelimit - INTEGER
720 Limit the maximal rates for sending ICMP packets whose type matches
721 icmp_ratemask (see below) to specific targets.
722 0 to disable any limiting,
723 otherwise the minimal space between responses in milliseconds.
724 Default: 1000
726 icmp_ratemask - INTEGER
727 Mask made of ICMP types for which rates are being limited.
728 Significant bits: IHGFEDCBA9876543210
729 Default mask: 0000001100000011000 (6168)
731 Bit definitions (see include/linux/icmp.h):
732 0 Echo Reply
733 3 Destination Unreachable *
734 4 Source Quench *
735 5 Redirect
736 8 Echo Request
737 B Time Exceeded *
738 C Parameter Problem *
739 D Timestamp Request
740 E Timestamp Reply
741 F Info Request
742 G Info Reply
743 H Address Mask Request
744 I Address Mask Reply
746 * These are rate limited by default (see default mask above)
748 icmp_ignore_bogus_error_responses - BOOLEAN
749 Some routers violate RFC1122 by sending bogus responses to broadcast
750 frames. Such violations are normally logged via a kernel warning.
751 If this is set to TRUE, the kernel will not give such warnings, which
752 will avoid log file clutter.
753 Default: 1
755 icmp_errors_use_inbound_ifaddr - BOOLEAN
757 If zero, icmp error messages are sent with the primary address of
758 the exiting interface.
760 If non-zero, the message will be sent with the primary address of
761 the interface that received the packet that caused the icmp error.
762 This is the behaviour network many administrators will expect from
763 a router. And it can make debugging complicated network layouts
764 much easier.
766 Note that if no primary address exists for the interface selected,
767 then the primary address of the first non-loopback interface that
768 has one will be used regardless of this setting.
770 Default: 0
772 igmp_max_memberships - INTEGER
773 Change the maximum number of multicast groups we can subscribe to.
774 Default: 20
776 Theoretical maximum value is bounded by having to send a membership
777 report in a single datagram (i.e. the report can't span multiple
778 datagrams, or risk confusing the switch and leaving groups you don't
779 intend to).
781 The number of supported groups 'M' is bounded by the number of group
782 report entries you can fit into a single datagram of 65535 bytes.
784 M = 65536-sizeof (ip header)/(sizeof(Group record))
786 Group records are variable length, with a minimum of 12 bytes.
787 So net.ipv4.igmp_max_memberships should not be set higher than:
789 (65536-24) / 12 = 5459
791 The value 5459 assumes no IP header options, so in practice
792 this number may be lower.
794 conf/interface/* changes special settings per interface (where
795 "interface" is the name of your network interface)
797 conf/all/* is special, changes the settings for all interfaces
799 log_martians - BOOLEAN
800 Log packets with impossible addresses to kernel log.
801 log_martians for the interface will be enabled if at least one of
802 conf/{all,interface}/log_martians is set to TRUE,
803 it will be disabled otherwise
805 accept_redirects - BOOLEAN
806 Accept ICMP redirect messages.
807 accept_redirects for the interface will be enabled if:
808 - both conf/{all,interface}/accept_redirects are TRUE in the case
809 forwarding for the interface is enabled
810 or
811 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
812 case forwarding for the interface is disabled
813 accept_redirects for the interface will be disabled otherwise
814 default TRUE (host)
815 FALSE (router)
817 forwarding - BOOLEAN
818 Enable IP forwarding on this interface.
820 mc_forwarding - BOOLEAN
821 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
822 and a multicast routing daemon is required.
823 conf/all/mc_forwarding must also be set to TRUE to enable multicast
824 routing for the interface
826 medium_id - INTEGER
827 Integer value used to differentiate the devices by the medium they
828 are attached to. Two devices can have different id values when
829 the broadcast packets are received only on one of them.
830 The default value 0 means that the device is the only interface
831 to its medium, value of -1 means that medium is not known.
833 Currently, it is used to change the proxy_arp behavior:
834 the proxy_arp feature is enabled for packets forwarded between
835 two devices attached to different media.
837 proxy_arp - BOOLEAN
838 Do proxy arp.
839 proxy_arp for the interface will be enabled if at least one of
840 conf/{all,interface}/proxy_arp is set to TRUE,
841 it will be disabled otherwise
843 proxy_arp_pvlan - BOOLEAN
844 Private VLAN proxy arp.
845 Basically allow proxy arp replies back to the same interface
846 (from which the ARP request/solicitation was received).
848 This is done to support (ethernet) switch features, like RFC
849 3069, where the individual ports are NOT allowed to
850 communicate with each other, but they are allowed to talk to
851 the upstream router. As described in RFC 3069, it is possible
852 to allow these hosts to communicate through the upstream
853 router by proxy_arp'ing. Don't need to be used together with
854 proxy_arp.
856 This technology is known by different names:
857 In RFC 3069 it is called VLAN Aggregation.
858 Cisco and Allied Telesyn call it Private VLAN.
859 Hewlett-Packard call it Source-Port filtering or port-isolation.
860 Ericsson call it MAC-Forced Forwarding (RFC Draft).
862 shared_media - BOOLEAN
863 Send(router) or accept(host) RFC1620 shared media redirects.
864 Overrides ip_secure_redirects.
865 shared_media for the interface will be enabled if at least one of
866 conf/{all,interface}/shared_media is set to TRUE,
867 it will be disabled otherwise
868 default TRUE
870 secure_redirects - BOOLEAN
871 Accept ICMP redirect messages only for gateways,
872 listed in default gateway list.
873 secure_redirects for the interface will be enabled if at least one of
874 conf/{all,interface}/secure_redirects is set to TRUE,
875 it will be disabled otherwise
876 default TRUE
878 send_redirects - BOOLEAN
879 Send redirects, if router.
880 send_redirects for the interface will be enabled if at least one of
881 conf/{all,interface}/send_redirects is set to TRUE,
882 it will be disabled otherwise
883 Default: TRUE
885 bootp_relay - BOOLEAN
886 Accept packets with source address 0.b.c.d destined
887 not to this host as local ones. It is supposed, that
888 BOOTP relay daemon will catch and forward such packets.
889 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
890 for the interface
891 default FALSE
892 Not Implemented Yet.
894 accept_source_route - BOOLEAN
895 Accept packets with SRR option.
896 conf/all/accept_source_route must also be set to TRUE to accept packets
897 with SRR option on the interface
898 default TRUE (router)
899 FALSE (host)
901 accept_local - BOOLEAN
902 Accept packets with local source addresses. In combination
903 with suitable routing, this can be used to direct packets
904 between two local interfaces over the wire and have them
905 accepted properly.
907 rp_filter must be set to a non-zero value in order for
908 accept_local to have an effect.
910 default FALSE
912 route_localnet - BOOLEAN
913 Do not consider loopback addresses as martian source or destination
914 while routing. This enables the use of 127/8 for local routing purposes.
915 default FALSE
917 rp_filter - INTEGER
918 0 - No source validation.
919 1 - Strict mode as defined in RFC3704 Strict Reverse Path
920 Each incoming packet is tested against the FIB and if the interface
921 is not the best reverse path the packet check will fail.
922 By default failed packets are discarded.
923 2 - Loose mode as defined in RFC3704 Loose Reverse Path
924 Each incoming packet's source address is also tested against the FIB
925 and if the source address is not reachable via any interface
926 the packet check will fail.
928 Current recommended practice in RFC3704 is to enable strict mode
929 to prevent IP spoofing from DDos attacks. If using asymmetric routing
930 or other complicated routing, then loose mode is recommended.
932 The max value from conf/{all,interface}/rp_filter is used
933 when doing source validation on the {interface}.
935 Default value is 0. Note that some distributions enable it
936 in startup scripts.
938 arp_filter - BOOLEAN
939 1 - Allows you to have multiple network interfaces on the same
940 subnet, and have the ARPs for each interface be answered
941 based on whether or not the kernel would route a packet from
942 the ARP'd IP out that interface (therefore you must use source
943 based routing for this to work). In other words it allows control
944 of which cards (usually 1) will respond to an arp request.
946 0 - (default) The kernel can respond to arp requests with addresses
947 from other interfaces. This may seem wrong but it usually makes
948 sense, because it increases the chance of successful communication.
949 IP addresses are owned by the complete host on Linux, not by
950 particular interfaces. Only for more complex setups like load-
951 balancing, does this behaviour cause problems.
953 arp_filter for the interface will be enabled if at least one of
954 conf/{all,interface}/arp_filter is set to TRUE,
955 it will be disabled otherwise
957 arp_announce - INTEGER
958 Define different restriction levels for announcing the local
959 source IP address from IP packets in ARP requests sent on
960 interface:
961 0 - (default) Use any local address, configured on any interface
962 1 - Try to avoid local addresses that are not in the target's
963 subnet for this interface. This mode is useful when target
964 hosts reachable via this interface require the source IP
965 address in ARP requests to be part of their logical network
966 configured on the receiving interface. When we generate the
967 request we will check all our subnets that include the
968 target IP and will preserve the source address if it is from
969 such subnet. If there is no such subnet we select source
970 address according to the rules for level 2.
971 2 - Always use the best local address for this target.
972 In this mode we ignore the source address in the IP packet
973 and try to select local address that we prefer for talks with
974 the target host. Such local address is selected by looking
975 for primary IP addresses on all our subnets on the outgoing
976 interface that include the target IP address. If no suitable
977 local address is found we select the first local address
978 we have on the outgoing interface or on all other interfaces,
979 with the hope we will receive reply for our request and
980 even sometimes no matter the source IP address we announce.
982 The max value from conf/{all,interface}/arp_announce is used.
984 Increasing the restriction level gives more chance for
985 receiving answer from the resolved target while decreasing
986 the level announces more valid sender's information.
988 arp_ignore - INTEGER
989 Define different modes for sending replies in response to
990 received ARP requests that resolve local target IP addresses:
991 0 - (default): reply for any local target IP address, configured
992 on any interface
993 1 - reply only if the target IP address is local address
994 configured on the incoming interface
995 2 - reply only if the target IP address is local address
996 configured on the incoming interface and both with the
997 sender's IP address are part from same subnet on this interface
998 3 - do not reply for local addresses configured with scope host,
999 only resolutions for global and link addresses are replied
1000 4-7 - reserved
1001 8 - do not reply for all local addresses
1003 The max value from conf/{all,interface}/arp_ignore is used
1004 when ARP request is received on the {interface}
1006 arp_notify - BOOLEAN
1007 Define mode for notification of address and device changes.
1008 0 - (default): do nothing
1009 1 - Generate gratuitous arp requests when device is brought up
1010 or hardware address changes.
1012 arp_accept - BOOLEAN
1013 Define behavior for gratuitous ARP frames who's IP is not
1014 already present in the ARP table:
1015 0 - don't create new entries in the ARP table
1016 1 - create new entries in the ARP table
1018 Both replies and requests type gratuitous arp will trigger the
1019 ARP table to be updated, if this setting is on.
1021 If the ARP table already contains the IP address of the
1022 gratuitous arp frame, the arp table will be updated regardless
1023 if this setting is on or off.
1026 app_solicit - INTEGER
1027 The maximum number of probes to send to the user space ARP daemon
1028 via netlink before dropping back to multicast probes (see
1029 mcast_solicit). Defaults to 0.
1031 disable_policy - BOOLEAN
1032 Disable IPSEC policy (SPD) for this interface
1034 disable_xfrm - BOOLEAN
1035 Disable IPSEC encryption on this interface, whatever the policy
1037 igmpv2_unsolicited_report_interval - INTEGER
1038 The interval in milliseconds in which the next unsolicited
1039 IGMPv1 or IGMPv2 report retransmit will take place.
1040 Default: 10000 (10 seconds)
1042 igmpv3_unsolicited_report_interval - INTEGER
1043 The interval in milliseconds in which the next unsolicited
1044 IGMPv3 report retransmit will take place.
1045 Default: 1000 (1 seconds)
1047 tag - INTEGER
1048 Allows you to write a number, which can be used as required.
1049 Default value is 0.
1051 Alexey Kuznetsov.
1052 kuznet@ms2.inr.ac.ru
1054 Updated by:
1055 Andi Kleen
1056 ak@muc.de
1057 Nicolas Delon
1058 delon.nicolas@wanadoo.fr
1063 /proc/sys/net/ipv6/* Variables:
1065 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1066 apply to IPv6 [XXX?].
1068 bindv6only - BOOLEAN
1069 Default value for IPV6_V6ONLY socket option,
1070 which restricts use of the IPv6 socket to IPv6 communication
1071 only.
1072 TRUE: disable IPv4-mapped address feature
1073 FALSE: enable IPv4-mapped address feature
1075 Default: FALSE (as specified in RFC3493)
1077 IPv6 Fragmentation:
1079 ip6frag_high_thresh - INTEGER
1080 Maximum memory used to reassemble IPv6 fragments. When
1081 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1082 the fragment handler will toss packets until ip6frag_low_thresh
1083 is reached.
1085 ip6frag_low_thresh - INTEGER
1086 See ip6frag_high_thresh
1088 ip6frag_time - INTEGER
1089 Time in seconds to keep an IPv6 fragment in memory.
1091 ip6frag_secret_interval - INTEGER
1092 Regeneration interval (in seconds) of the hash secret (or lifetime
1093 for the hash secret) for IPv6 fragments.
1094 Default: 600
1096 conf/default/*:
1097 Change the interface-specific default settings.
1100 conf/all/*:
1101 Change all the interface-specific settings.
1103 [XXX: Other special features than forwarding?]
1105 conf/all/forwarding - BOOLEAN
1106 Enable global IPv6 forwarding between all interfaces.
1108 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1109 to control which interfaces may forward packets and which not.
1111 This also sets all interfaces' Host/Router setting
1112 'forwarding' to the specified value. See below for details.
1114 This referred to as global forwarding.
1116 proxy_ndp - BOOLEAN
1117 Do proxy ndp.
1119 conf/interface/*:
1120 Change special settings per interface.
1122 The functional behaviour for certain settings is different
1123 depending on whether local forwarding is enabled or not.
1125 accept_ra - INTEGER
1126 Accept Router Advertisements; autoconfigure using them.
1128 It also determines whether or not to transmit Router
1129 Solicitations. If and only if the functional setting is to
1130 accept Router Advertisements, Router Solicitations will be
1131 transmitted.
1133 Possible values are:
1134 0 Do not accept Router Advertisements.
1135 1 Accept Router Advertisements if forwarding is disabled.
1136 2 Overrule forwarding behaviour. Accept Router Advertisements
1137 even if forwarding is enabled.
1139 Functional default: enabled if local forwarding is disabled.
1140 disabled if local forwarding is enabled.
1142 accept_ra_defrtr - BOOLEAN
1143 Learn default router in Router Advertisement.
1145 Functional default: enabled if accept_ra is enabled.
1146 disabled if accept_ra is disabled.
1148 accept_ra_pinfo - BOOLEAN
1149 Learn Prefix Information in Router Advertisement.
1151 Functional default: enabled if accept_ra is enabled.
1152 disabled if accept_ra is disabled.
1154 accept_ra_rt_info_max_plen - INTEGER
1155 Maximum prefix length of Route Information in RA.
1157 Route Information w/ prefix larger than or equal to this
1158 variable shall be ignored.
1160 Functional default: 0 if accept_ra_rtr_pref is enabled.
1161 -1 if accept_ra_rtr_pref is disabled.
1163 accept_ra_rtr_pref - BOOLEAN
1164 Accept Router Preference in RA.
1166 Functional default: enabled if accept_ra is enabled.
1167 disabled if accept_ra is disabled.
1169 accept_redirects - BOOLEAN
1170 Accept Redirects.
1172 Functional default: enabled if local forwarding is disabled.
1173 disabled if local forwarding is enabled.
1175 accept_source_route - INTEGER
1176 Accept source routing (routing extension header).
1178 >= 0: Accept only routing header type 2.
1179 < 0: Do not accept routing header.
1181 Default: 0
1183 autoconf - BOOLEAN
1184 Autoconfigure addresses using Prefix Information in Router
1185 Advertisements.
1187 Functional default: enabled if accept_ra_pinfo is enabled.
1188 disabled if accept_ra_pinfo is disabled.
1190 dad_transmits - INTEGER
1191 The amount of Duplicate Address Detection probes to send.
1192 Default: 1
1194 forwarding - INTEGER
1195 Configure interface-specific Host/Router behaviour.
1197 Note: It is recommended to have the same setting on all
1198 interfaces; mixed router/host scenarios are rather uncommon.
1200 Possible values are:
1201 0 Forwarding disabled
1202 1 Forwarding enabled
1204 FALSE (0):
1206 By default, Host behaviour is assumed. This means:
1208 1. IsRouter flag is not set in Neighbour Advertisements.
1209 2. If accept_ra is TRUE (default), transmit Router
1210 Solicitations.
1211 3. If accept_ra is TRUE (default), accept Router
1212 Advertisements (and do autoconfiguration).
1213 4. If accept_redirects is TRUE (default), accept Redirects.
1215 TRUE (1):
1217 If local forwarding is enabled, Router behaviour is assumed.
1218 This means exactly the reverse from the above:
1220 1. IsRouter flag is set in Neighbour Advertisements.
1221 2. Router Solicitations are not sent unless accept_ra is 2.
1222 3. Router Advertisements are ignored unless accept_ra is 2.
1223 4. Redirects are ignored.
1225 Default: 0 (disabled) if global forwarding is disabled (default),
1226 otherwise 1 (enabled).
1228 hop_limit - INTEGER
1229 Default Hop Limit to set.
1230 Default: 64
1232 mtu - INTEGER
1233 Default Maximum Transfer Unit
1234 Default: 1280 (IPv6 required minimum)
1236 router_probe_interval - INTEGER
1237 Minimum interval (in seconds) between Router Probing described
1238 in RFC4191.
1240 Default: 60
1242 router_solicitation_delay - INTEGER
1243 Number of seconds to wait after interface is brought up
1244 before sending Router Solicitations.
1245 Default: 1
1247 router_solicitation_interval - INTEGER
1248 Number of seconds to wait between Router Solicitations.
1249 Default: 4
1251 router_solicitations - INTEGER
1252 Number of Router Solicitations to send until assuming no
1253 routers are present.
1254 Default: 3
1256 use_tempaddr - INTEGER
1257 Preference for Privacy Extensions (RFC3041).
1258 <= 0 : disable Privacy Extensions
1259 == 1 : enable Privacy Extensions, but prefer public
1260 addresses over temporary addresses.
1261 > 1 : enable Privacy Extensions and prefer temporary
1262 addresses over public addresses.
1263 Default: 0 (for most devices)
1264 -1 (for point-to-point devices and loopback devices)
1266 temp_valid_lft - INTEGER
1267 valid lifetime (in seconds) for temporary addresses.
1268 Default: 604800 (7 days)
1270 temp_prefered_lft - INTEGER
1271 Preferred lifetime (in seconds) for temporary addresses.
1272 Default: 86400 (1 day)
1274 max_desync_factor - INTEGER
1275 Maximum value for DESYNC_FACTOR, which is a random value
1276 that ensures that clients don't synchronize with each
1277 other and generate new addresses at exactly the same time.
1278 value is in seconds.
1279 Default: 600
1281 regen_max_retry - INTEGER
1282 Number of attempts before give up attempting to generate
1283 valid temporary addresses.
1284 Default: 5
1286 max_addresses - INTEGER
1287 Maximum number of autoconfigured addresses per interface. Setting
1288 to zero disables the limitation. It is not recommended to set this
1289 value too large (or to zero) because it would be an easy way to
1290 crash the kernel by allowing too many addresses to be created.
1291 Default: 16
1293 disable_ipv6 - BOOLEAN
1294 Disable IPv6 operation. If accept_dad is set to 2, this value
1295 will be dynamically set to TRUE if DAD fails for the link-local
1296 address.
1297 Default: FALSE (enable IPv6 operation)
1299 When this value is changed from 1 to 0 (IPv6 is being enabled),
1300 it will dynamically create a link-local address on the given
1301 interface and start Duplicate Address Detection, if necessary.
1303 When this value is changed from 0 to 1 (IPv6 is being disabled),
1304 it will dynamically delete all address on the given interface.
1306 accept_dad - INTEGER
1307 Whether to accept DAD (Duplicate Address Detection).
1308 0: Disable DAD
1309 1: Enable DAD (default)
1310 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1311 link-local address has been found.
1313 force_tllao - BOOLEAN
1314 Enable sending the target link-layer address option even when
1315 responding to a unicast neighbor solicitation.
1316 Default: FALSE
1318 Quoting from RFC 2461, section 4.4, Target link-layer address:
1320 "The option MUST be included for multicast solicitations in order to
1321 avoid infinite Neighbor Solicitation "recursion" when the peer node
1322 does not have a cache entry to return a Neighbor Advertisements
1323 message. When responding to unicast solicitations, the option can be
1324 omitted since the sender of the solicitation has the correct link-
1325 layer address; otherwise it would not have be able to send the unicast
1326 solicitation in the first place. However, including the link-layer
1327 address in this case adds little overhead and eliminates a potential
1328 race condition where the sender deletes the cached link-layer address
1329 prior to receiving a response to a previous solicitation."
1331 ndisc_notify - BOOLEAN
1332 Define mode for notification of address and device changes.
1333 0 - (default): do nothing
1334 1 - Generate unsolicited neighbour advertisements when device is brought
1335 up or hardware address changes.
1337 mldv1_unsolicited_report_interval - INTEGER
1338 The interval in milliseconds in which the next unsolicited
1339 MLDv1 report retransmit will take place.
1340 Default: 10000 (10 seconds)
1342 mldv2_unsolicited_report_interval - INTEGER
1343 The interval in milliseconds in which the next unsolicited
1344 MLDv2 report retransmit will take place.
1345 Default: 1000 (1 second)
1347 force_mld_version - INTEGER
1348 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1349 1 - Enforce to use MLD version 1
1350 2 - Enforce to use MLD version 2
1352 suppress_frag_ndisc - INTEGER
1353 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1354 with IPv6 Neighbor Discovery) behavior:
1355 1 - (default) discard fragmented neighbor discovery packets
1356 0 - allow fragmented neighbor discovery packets
1358 icmp/*:
1359 ratelimit - INTEGER
1360 Limit the maximal rates for sending ICMPv6 packets.
1361 0 to disable any limiting,
1362 otherwise the minimal space between responses in milliseconds.
1363 Default: 1000
1366 IPv6 Update by:
1367 Pekka Savola <pekkas@netcore.fi>
1368 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1371 /proc/sys/net/bridge/* Variables:
1373 bridge-nf-call-arptables - BOOLEAN
1374 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1375 0 : disable this.
1376 Default: 1
1378 bridge-nf-call-iptables - BOOLEAN
1379 1 : pass bridged IPv4 traffic to iptables' chains.
1380 0 : disable this.
1381 Default: 1
1383 bridge-nf-call-ip6tables - BOOLEAN
1384 1 : pass bridged IPv6 traffic to ip6tables' chains.
1385 0 : disable this.
1386 Default: 1
1388 bridge-nf-filter-vlan-tagged - BOOLEAN
1389 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1390 0 : disable this.
1391 Default: 0
1393 bridge-nf-filter-pppoe-tagged - BOOLEAN
1394 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1395 0 : disable this.
1396 Default: 0
1398 bridge-nf-pass-vlan-input-dev - BOOLEAN
1399 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1400 interface on the bridge and set the netfilter input device to the vlan.
1401 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1402 target work with vlan-on-top-of-bridge interfaces. When no matching
1403 vlan interface is found, or this switch is off, the input device is
1404 set to the bridge interface.
1405 0: disable bridge netfilter vlan interface lookup.
1406 Default: 0
1408 proc/sys/net/sctp/* Variables:
1410 addip_enable - BOOLEAN
1411 Enable or disable extension of Dynamic Address Reconfiguration
1412 (ADD-IP) functionality specified in RFC5061. This extension provides
1413 the ability to dynamically add and remove new addresses for the SCTP
1414 associations.
1416 1: Enable extension.
1418 0: Disable extension.
1420 Default: 0
1422 addip_noauth_enable - BOOLEAN
1423 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1424 authentication to protect the operations of adding or removing new
1425 addresses. This requirement is mandated so that unauthorized hosts
1426 would not be able to hijack associations. However, older
1427 implementations may not have implemented this requirement while
1428 allowing the ADD-IP extension. For reasons of interoperability,
1429 we provide this variable to control the enforcement of the
1430 authentication requirement.
1432 1: Allow ADD-IP extension to be used without authentication. This
1433 should only be set in a closed environment for interoperability
1434 with older implementations.
1436 0: Enforce the authentication requirement
1438 Default: 0
1440 auth_enable - BOOLEAN
1441 Enable or disable Authenticated Chunks extension. This extension
1442 provides the ability to send and receive authenticated chunks and is
1443 required for secure operation of Dynamic Address Reconfiguration
1444 (ADD-IP) extension.
1446 1: Enable this extension.
1447 0: Disable this extension.
1449 Default: 0
1451 prsctp_enable - BOOLEAN
1452 Enable or disable the Partial Reliability extension (RFC3758) which
1453 is used to notify peers that a given DATA should no longer be expected.
1455 1: Enable extension
1456 0: Disable
1458 Default: 1
1460 max_burst - INTEGER
1461 The limit of the number of new packets that can be initially sent. It
1462 controls how bursty the generated traffic can be.
1464 Default: 4
1466 association_max_retrans - INTEGER
1467 Set the maximum number for retransmissions that an association can
1468 attempt deciding that the remote end is unreachable. If this value
1469 is exceeded, the association is terminated.
1471 Default: 10
1473 max_init_retransmits - INTEGER
1474 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1475 that an association will attempt before declaring the destination
1476 unreachable and terminating.
1478 Default: 8
1480 path_max_retrans - INTEGER
1481 The maximum number of retransmissions that will be attempted on a given
1482 path. Once this threshold is exceeded, the path is considered
1483 unreachable, and new traffic will use a different path when the
1484 association is multihomed.
1486 Default: 5
1488 pf_retrans - INTEGER
1489 The number of retransmissions that will be attempted on a given path
1490 before traffic is redirected to an alternate transport (should one
1491 exist). Note this is distinct from path_max_retrans, as a path that
1492 passes the pf_retrans threshold can still be used. Its only
1493 deprioritized when a transmission path is selected by the stack. This
1494 setting is primarily used to enable fast failover mechanisms without
1495 having to reduce path_max_retrans to a very low value. See:
1496 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1497 for details. Note also that a value of pf_retrans > path_max_retrans
1498 disables this feature
1500 Default: 0
1502 rto_initial - INTEGER
1503 The initial round trip timeout value in milliseconds that will be used
1504 in calculating round trip times. This is the initial time interval
1505 for retransmissions.
1507 Default: 3000
1509 rto_max - INTEGER
1510 The maximum value (in milliseconds) of the round trip timeout. This
1511 is the largest time interval that can elapse between retransmissions.
1513 Default: 60000
1515 rto_min - INTEGER
1516 The minimum value (in milliseconds) of the round trip timeout. This
1517 is the smallest time interval the can elapse between retransmissions.
1519 Default: 1000
1521 hb_interval - INTEGER
1522 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1523 are sent at the specified interval on idle paths to probe the state of
1524 a given path between 2 associations.
1526 Default: 30000
1528 sack_timeout - INTEGER
1529 The amount of time (in milliseconds) that the implementation will wait
1530 to send a SACK.
1532 Default: 200
1534 valid_cookie_life - INTEGER
1535 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1536 is used during association establishment.
1538 Default: 60000
1540 cookie_preserve_enable - BOOLEAN
1541 Enable or disable the ability to extend the lifetime of the SCTP cookie
1542 that is used during the establishment phase of SCTP association
1544 1: Enable cookie lifetime extension.
1545 0: Disable
1547 Default: 1
1549 cookie_hmac_alg - STRING
1550 Select the hmac algorithm used when generating the cookie value sent by
1551 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1552 Valid values are:
1553 * md5
1554 * sha1
1555 * none
1556 Ability to assign md5 or sha1 as the selected alg is predicated on the
1557 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1558 CONFIG_CRYPTO_SHA1).
1560 Default: Dependent on configuration. MD5 if available, else SHA1 if
1561 available, else none.
1563 rcvbuf_policy - INTEGER
1564 Determines if the receive buffer is attributed to the socket or to
1565 association. SCTP supports the capability to create multiple
1566 associations on a single socket. When using this capability, it is
1567 possible that a single stalled association that's buffering a lot
1568 of data may block other associations from delivering their data by
1569 consuming all of the receive buffer space. To work around this,
1570 the rcvbuf_policy could be set to attribute the receiver buffer space
1571 to each association instead of the socket. This prevents the described
1572 blocking.
1574 1: rcvbuf space is per association
1575 0: rcvbuf space is per socket
1577 Default: 0
1579 sndbuf_policy - INTEGER
1580 Similar to rcvbuf_policy above, this applies to send buffer space.
1582 1: Send buffer is tracked per association
1583 0: Send buffer is tracked per socket.
1585 Default: 0
1587 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1588 Number of pages allowed for queueing by all SCTP sockets.
1590 min: Below this number of pages SCTP is not bothered about its
1591 memory appetite. When amount of memory allocated by SCTP exceeds
1592 this number, SCTP starts to moderate memory usage.
1594 pressure: This value was introduced to follow format of tcp_mem.
1596 max: Number of pages allowed for queueing by all SCTP sockets.
1598 Default is calculated at boot time from amount of available memory.
1600 sctp_rmem - vector of 3 INTEGERs: min, default, max
1601 Only the first value ("min") is used, "default" and "max" are
1602 ignored.
1604 min: Minimal size of receive buffer used by SCTP socket.
1605 It is guaranteed to each SCTP socket (but not association) even
1606 under moderate memory pressure.
1608 Default: 1 page
1610 sctp_wmem - vector of 3 INTEGERs: min, default, max
1611 Currently this tunable has no effect.
1613 addr_scope_policy - INTEGER
1614 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1616 0 - Disable IPv4 address scoping
1617 1 - Enable IPv4 address scoping
1618 2 - Follow draft but allow IPv4 private addresses
1619 3 - Follow draft but allow IPv4 link local addresses
1621 Default: 1
1624 /proc/sys/net/core/*
1625 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1628 /proc/sys/net/unix/*
1629 max_dgram_qlen - INTEGER
1630 The maximum length of dgram socket receive queue
1632 Default: 10
1635 UNDOCUMENTED:
1637 /proc/sys/net/irda/*
1638 fast_poll_increase FIXME
1639 warn_noreply_time FIXME
1640 discovery_slots FIXME
1641 slot_timeout FIXME
1642 max_baud_rate FIXME
1643 discovery_timeout FIXME
1644 lap_keepalive_time FIXME
1645 max_noreply_time FIXME
1646 max_tx_data_size FIXME
1647 max_tx_window FIXME
1648 min_tx_turn_time FIXME