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 - BOOLEAN
19 Disable Path MTU Discovery.
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.
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.
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.
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
58 Time, in seconds, that cached PMTU information is kept.
61 The advertised MSS depends on the first hop route MTU, but will
62 never be lower than this setting.
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
72 ipfrag_low_thresh - INTEGER
73 See ipfrag_high_thresh
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.
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.
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.
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).
129 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
130 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
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),
145 Possible values are [-31, 31], inclusive.
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.
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,
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
176 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
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 occuring due to tail
187 losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01).
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.
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.
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.
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.
228 Enables Forward RTO-Recovery (F-RTO) defined in RFC4138.
229 F-RTO is an enhanced recovery algorithm for TCP retransmission
230 timeouts. It is particularly beneficial in wireless environments
231 where packet loss is typically due to random radio interference
232 rather than intermediate router congestion. F-RTO is sender-side
233 only modification. Therefore it does not require any support from
236 If set to 1, basic version is enabled. 2 enables SACK enhanced
237 F-RTO if flow uses SACK. The basic version can be used also when
238 SACK is in use though scenario(s) with it exists where F-RTO
239 interacts badly with the packet counting of the SACK enabled TCP
242 tcp_frto_response - INTEGER
243 When F-RTO has detected that a TCP retransmission timeout was
244 spurious (i.e, the timeout would have been avoided had TCP set a
245 longer retransmission timeout), TCP has several options what to do
246 next. Possible values are:
247 0 Rate halving based; a smooth and conservative response,
248 results in halved cwnd and ssthresh after one RTT
249 1 Very conservative response; not recommended because even
250 though being valid, it interacts poorly with the rest of
251 Linux TCP, halves cwnd and ssthresh immediately
252 2 Aggressive response; undoes congestion control measures
253 that are now known to be unnecessary (ignoring the
254 possibility of a lost retransmission that would require
255 TCP to be more cautious), cwnd and ssthresh are restored
256 to the values prior timeout
257 Default: 0 (rate halving based)
259 tcp_keepalive_time - INTEGER
260 How often TCP sends out keepalive messages when keepalive is enabled.
263 tcp_keepalive_probes - INTEGER
264 How many keepalive probes TCP sends out, until it decides that the
265 connection is broken. Default value: 9.
267 tcp_keepalive_intvl - INTEGER
268 How frequently the probes are send out. Multiplied by
269 tcp_keepalive_probes it is time to kill not responding connection,
270 after probes started. Default value: 75sec i.e. connection
271 will be aborted after ~11 minutes of retries.
273 tcp_low_latency - BOOLEAN
274 If set, the TCP stack makes decisions that prefer lower
275 latency as opposed to higher throughput. By default, this
276 option is not set meaning that higher throughput is preferred.
277 An example of an application where this default should be
278 changed would be a Beowulf compute cluster.
281 tcp_max_orphans - INTEGER
282 Maximal number of TCP sockets not attached to any user file handle,
283 held by system. If this number is exceeded orphaned connections are
284 reset immediately and warning is printed. This limit exists
285 only to prevent simple DoS attacks, you _must_ not rely on this
286 or lower the limit artificially, but rather increase it
287 (probably, after increasing installed memory),
288 if network conditions require more than default value,
289 and tune network services to linger and kill such states
290 more aggressively. Let me to remind again: each orphan eats
291 up to ~64K of unswappable memory.
293 tcp_max_ssthresh - INTEGER
294 Limited Slow-Start for TCP with large congestion windows (cwnd) defined in
295 RFC3742. Limited slow-start is a mechanism to limit growth of the cwnd
296 on the region where cwnd is larger than tcp_max_ssthresh. TCP increases cwnd
297 by at most tcp_max_ssthresh segments, and by at least tcp_max_ssthresh/2
298 segments per RTT when the cwnd is above tcp_max_ssthresh.
299 If TCP connection increased cwnd to thousands (or tens of thousands) segments,
300 and thousands of packets were being dropped during slow-start, you can set
301 tcp_max_ssthresh to improve performance for new TCP connection.
304 tcp_max_syn_backlog - INTEGER
305 Maximal number of remembered connection requests, which have not
306 received an acknowledgment from connecting client.
307 The minimal value is 128 for low memory machines, and it will
308 increase in proportion to the memory of machine.
309 If server suffers from overload, try increasing this number.
311 tcp_max_tw_buckets - INTEGER
312 Maximal number of timewait sockets held by system simultaneously.
313 If this number is exceeded time-wait socket is immediately destroyed
314 and warning is printed. This limit exists only to prevent
315 simple DoS attacks, you _must_ not lower the limit artificially,
316 but rather increase it (probably, after increasing installed memory),
317 if network conditions require more than default value.
319 tcp_mem - vector of 3 INTEGERs: min, pressure, max
320 min: below this number of pages TCP is not bothered about its
323 pressure: when amount of memory allocated by TCP exceeds this number
324 of pages, TCP moderates its memory consumption and enters memory
325 pressure mode, which is exited when memory consumption falls
328 max: number of pages allowed for queueing by all TCP sockets.
330 Defaults are calculated at boot time from amount of available
333 tcp_moderate_rcvbuf - BOOLEAN
334 If set, TCP performs receive buffer auto-tuning, attempting to
335 automatically size the buffer (no greater than tcp_rmem[2]) to
336 match the size required by the path for full throughput. Enabled by
339 tcp_mtu_probing - INTEGER
340 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
343 1 - Disabled by default, enabled when an ICMP black hole detected
344 2 - Always enabled, use initial MSS of tcp_base_mss.
346 tcp_no_metrics_save - BOOLEAN
347 By default, TCP saves various connection metrics in the route cache
348 when the connection closes, so that connections established in the
349 near future can use these to set initial conditions. Usually, this
350 increases overall performance, but may sometimes cause performance
351 degradation. If set, TCP will not cache metrics on closing
354 tcp_orphan_retries - INTEGER
355 This value influences the timeout of a locally closed TCP connection,
356 when RTO retransmissions remain unacknowledged.
357 See tcp_retries2 for more details.
359 The default value is 8.
360 If your machine is a loaded WEB server,
361 you should think about lowering this value, such sockets
362 may consume significant resources. Cf. tcp_max_orphans.
364 tcp_reordering - INTEGER
365 Maximal reordering of packets in a TCP stream.
368 tcp_retrans_collapse - BOOLEAN
369 Bug-to-bug compatibility with some broken printers.
370 On retransmit try to send bigger packets to work around bugs in
373 tcp_retries1 - INTEGER
374 This value influences the time, after which TCP decides, that
375 something is wrong due to unacknowledged RTO retransmissions,
376 and reports this suspicion to the network layer.
377 See tcp_retries2 for more details.
379 RFC 1122 recommends at least 3 retransmissions, which is the
382 tcp_retries2 - INTEGER
383 This value influences the timeout of an alive TCP connection,
384 when RTO retransmissions remain unacknowledged.
385 Given a value of N, a hypothetical TCP connection following
386 exponential backoff with an initial RTO of TCP_RTO_MIN would
387 retransmit N times before killing the connection at the (N+1)th RTO.
389 The default value of 15 yields a hypothetical timeout of 924.6
390 seconds and is a lower bound for the effective timeout.
391 TCP will effectively time out at the first RTO which exceeds the
392 hypothetical timeout.
394 RFC 1122 recommends at least 100 seconds for the timeout,
395 which corresponds to a value of at least 8.
397 tcp_rfc1337 - BOOLEAN
398 If set, the TCP stack behaves conforming to RFC1337. If unset,
399 we are not conforming to RFC, but prevent TCP TIME_WAIT
403 tcp_rmem - vector of 3 INTEGERs: min, default, max
404 min: Minimal size of receive buffer used by TCP sockets.
405 It is guaranteed to each TCP socket, even under moderate memory
409 default: initial size of receive buffer used by TCP sockets.
410 This value overrides net.core.rmem_default used by other protocols.
411 Default: 87380 bytes. This value results in window of 65535 with
412 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
413 less for default tcp_app_win. See below about these variables.
415 max: maximal size of receive buffer allowed for automatically
416 selected receiver buffers for TCP socket. This value does not override
417 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
418 automatic tuning of that socket's receive buffer size, in which
419 case this value is ignored.
420 Default: between 87380B and 6MB, depending on RAM size.
423 Enable select acknowledgments (SACKS).
425 tcp_slow_start_after_idle - BOOLEAN
426 If set, provide RFC2861 behavior and time out the congestion
427 window after an idle period. An idle period is defined at
428 the current RTO. If unset, the congestion window will not
429 be timed out after an idle period.
433 Use the Host requirements interpretation of the TCP urgent pointer field.
434 Most hosts use the older BSD interpretation, so if you turn this on
435 Linux might not communicate correctly with them.
438 tcp_synack_retries - INTEGER
439 Number of times SYNACKs for a passive TCP connection attempt will
440 be retransmitted. Should not be higher than 255. Default value
441 is 5, which corresponds to 31seconds till the last retransmission
442 with the current initial RTO of 1second. With this the final timeout
443 for a passive TCP connection will happen after 63seconds.
445 tcp_syncookies - BOOLEAN
446 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES
447 Send out syncookies when the syn backlog queue of a socket
448 overflows. This is to prevent against the common 'SYN flood attack'
451 Note, that syncookies is fallback facility.
452 It MUST NOT be used to help highly loaded servers to stand
453 against legal connection rate. If you see SYN flood warnings
454 in your logs, but investigation shows that they occur
455 because of overload with legal connections, you should tune
456 another parameters until this warning disappear.
457 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
459 syncookies seriously violate TCP protocol, do not allow
460 to use TCP extensions, can result in serious degradation
461 of some services (f.e. SMTP relaying), visible not by you,
462 but your clients and relays, contacting you. While you see
463 SYN flood warnings in logs not being really flooded, your server
464 is seriously misconfigured.
466 tcp_fastopen - INTEGER
467 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
468 in the opening SYN packet. To use this feature, the client application
469 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
470 connect() to perform a TCP handshake automatically.
472 The values (bitmap) are
473 1: Enables sending data in the opening SYN on the client.
474 2: Enables TCP Fast Open on the server side, i.e., allowing data in
475 a SYN packet to be accepted and passed to the application before
476 3-way hand shake finishes.
477 4: Send data in the opening SYN regardless of cookie availability and
478 without a cookie option.
479 0x100: Accept SYN data w/o validating the cookie.
480 0x200: Accept data-in-SYN w/o any cookie option present.
481 0x400/0x800: Enable Fast Open on all listeners regardless of the
482 TCP_FASTOPEN socket option. The two different flags designate two
483 different ways of setting max_qlen without the TCP_FASTOPEN socket
488 Note that the client & server side Fast Open flags (1 and 2
489 respectively) must be also enabled before the rest of flags can take
492 See include/net/tcp.h and the code for more details.
494 tcp_syn_retries - INTEGER
495 Number of times initial SYNs for an active TCP connection attempt
496 will be retransmitted. Should not be higher than 255. Default value
497 is 6, which corresponds to 63seconds till the last retransmission
498 with the current initial RTO of 1second. With this the final timeout
499 for an active TCP connection attempt will happen after 127seconds.
501 tcp_timestamps - BOOLEAN
502 Enable timestamps as defined in RFC1323.
504 tcp_tso_win_divisor - INTEGER
505 This allows control over what percentage of the congestion window
506 can be consumed by a single TSO frame.
507 The setting of this parameter is a choice between burstiness and
508 building larger TSO frames.
511 tcp_tw_recycle - BOOLEAN
512 Enable fast recycling TIME-WAIT sockets. Default value is 0.
513 It should not be changed without advice/request of technical
516 tcp_tw_reuse - BOOLEAN
517 Allow to reuse TIME-WAIT sockets for new connections when it is
518 safe from protocol viewpoint. Default value is 0.
519 It should not be changed without advice/request of technical
522 tcp_window_scaling - BOOLEAN
523 Enable window scaling as defined in RFC1323.
525 tcp_wmem - vector of 3 INTEGERs: min, default, max
526 min: Amount of memory reserved for send buffers for TCP sockets.
527 Each TCP socket has rights to use it due to fact of its birth.
530 default: initial size of send buffer used by TCP sockets. This
531 value overrides net.core.wmem_default used by other protocols.
532 It is usually lower than net.core.wmem_default.
535 max: Maximal amount of memory allowed for automatically tuned
536 send buffers for TCP sockets. This value does not override
537 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
538 automatic tuning of that socket's send buffer size, in which case
539 this value is ignored.
540 Default: between 64K and 4MB, depending on RAM size.
542 tcp_workaround_signed_windows - BOOLEAN
543 If set, assume no receipt of a window scaling option means the
544 remote TCP is broken and treats the window as a signed quantity.
545 If unset, assume the remote TCP is not broken even if we do
546 not receive a window scaling option from them.
549 tcp_dma_copybreak - INTEGER
550 Lower limit, in bytes, of the size of socket reads that will be
551 offloaded to a DMA copy engine, if one is present in the system
552 and CONFIG_NET_DMA is enabled.
555 tcp_thin_linear_timeouts - BOOLEAN
556 Enable dynamic triggering of linear timeouts for thin streams.
557 If set, a check is performed upon retransmission by timeout to
558 determine if the stream is thin (less than 4 packets in flight).
559 As long as the stream is found to be thin, up to 6 linear
560 timeouts may be performed before exponential backoff mode is
561 initiated. This improves retransmission latency for
562 non-aggressive thin streams, often found to be time-dependent.
563 For more information on thin streams, see
564 Documentation/networking/tcp-thin.txt
567 tcp_thin_dupack - BOOLEAN
568 Enable dynamic triggering of retransmissions after one dupACK
569 for thin streams. If set, a check is performed upon reception
570 of a dupACK to determine if the stream is thin (less than 4
571 packets in flight). As long as the stream is found to be thin,
572 data is retransmitted on the first received dupACK. This
573 improves retransmission latency for non-aggressive thin
574 streams, often found to be time-dependent.
575 For more information on thin streams, see
576 Documentation/networking/tcp-thin.txt
579 tcp_limit_output_bytes - INTEGER
580 Controls TCP Small Queue limit per tcp socket.
581 TCP bulk sender tends to increase packets in flight until it
582 gets losses notifications. With SNDBUF autotuning, this can
583 result in a large amount of packets queued in qdisc/device
584 on the local machine, hurting latency of other flows, for
585 typical pfifo_fast qdiscs.
586 tcp_limit_output_bytes limits the number of bytes on qdisc
587 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
588 Note: For GSO/TSO enabled flows, we try to have at least two
589 packets in flight. Reducing tcp_limit_output_bytes might also
590 reduce the size of individual GSO packet (64KB being the max)
593 tcp_challenge_ack_limit - INTEGER
594 Limits number of Challenge ACK sent per second, as recommended
595 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
600 udp_mem - vector of 3 INTEGERs: min, pressure, max
601 Number of pages allowed for queueing by all UDP sockets.
603 min: Below this number of pages UDP is not bothered about its
604 memory appetite. When amount of memory allocated by UDP exceeds
605 this number, UDP starts to moderate memory usage.
607 pressure: This value was introduced to follow format of tcp_mem.
609 max: Number of pages allowed for queueing by all UDP sockets.
611 Default is calculated at boot time from amount of available memory.
613 udp_rmem_min - INTEGER
614 Minimal size of receive buffer used by UDP sockets in moderation.
615 Each UDP socket is able to use the size for receiving data, even if
616 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
619 udp_wmem_min - INTEGER
620 Minimal size of send buffer used by UDP sockets in moderation.
621 Each UDP socket is able to use the size for sending data, even if
622 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
627 cipso_cache_enable - BOOLEAN
628 If set, enable additions to and lookups from the CIPSO label mapping
629 cache. If unset, additions are ignored and lookups always result in a
630 miss. However, regardless of the setting the cache is still
631 invalidated when required when means you can safely toggle this on and
632 off and the cache will always be "safe".
635 cipso_cache_bucket_size - INTEGER
636 The CIPSO label cache consists of a fixed size hash table with each
637 hash bucket containing a number of cache entries. This variable limits
638 the number of entries in each hash bucket; the larger the value the
639 more CIPSO label mappings that can be cached. When the number of
640 entries in a given hash bucket reaches this limit adding new entries
641 causes the oldest entry in the bucket to be removed to make room.
644 cipso_rbm_optfmt - BOOLEAN
645 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
646 the CIPSO draft specification (see Documentation/netlabel for details).
647 This means that when set the CIPSO tag will be padded with empty
648 categories in order to make the packet data 32-bit aligned.
651 cipso_rbm_structvalid - BOOLEAN
652 If set, do a very strict check of the CIPSO option when
653 ip_options_compile() is called. If unset, relax the checks done during
654 ip_options_compile(). Either way is "safe" as errors are caught else
655 where in the CIPSO processing code but setting this to 0 (False) should
656 result in less work (i.e. it should be faster) but could cause problems
657 with other implementations that require strict checking.
662 ip_local_port_range - 2 INTEGERS
663 Defines the local port range that is used by TCP and UDP to
664 choose the local port. The first number is the first, the
665 second the last local port number. The default values are
666 32768 and 61000 respectively.
668 ip_local_reserved_ports - list of comma separated ranges
669 Specify the ports which are reserved for known third-party
670 applications. These ports will not be used by automatic port
671 assignments (e.g. when calling connect() or bind() with port
672 number 0). Explicit port allocation behavior is unchanged.
674 The format used for both input and output is a comma separated
675 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
676 10). Writing to the file will clear all previously reserved
677 ports and update the current list with the one given in the
680 Note that ip_local_port_range and ip_local_reserved_ports
681 settings are independent and both are considered by the kernel
682 when determining which ports are available for automatic port
685 You can reserve ports which are not in the current
686 ip_local_port_range, e.g.:
688 $ cat /proc/sys/net/ipv4/ip_local_port_range
690 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
693 although this is redundant. However such a setting is useful
694 if later the port range is changed to a value that will
695 include the reserved ports.
699 ip_nonlocal_bind - BOOLEAN
700 If set, allows processes to bind() to non-local IP addresses,
701 which can be quite useful - but may break some applications.
705 If set non-zero, enables support for dynamic addresses.
706 If set to a non-zero value larger than 1, a kernel log
707 message will be printed when dynamic address rewriting
711 icmp_echo_ignore_all - BOOLEAN
712 If set non-zero, then the kernel will ignore all ICMP ECHO
716 icmp_echo_ignore_broadcasts - BOOLEAN
717 If set non-zero, then the kernel will ignore all ICMP ECHO and
718 TIMESTAMP requests sent to it via broadcast/multicast.
721 icmp_ratelimit - INTEGER
722 Limit the maximal rates for sending ICMP packets whose type matches
723 icmp_ratemask (see below) to specific targets.
724 0 to disable any limiting,
725 otherwise the minimal space between responses in milliseconds.
728 icmp_ratemask - INTEGER
729 Mask made of ICMP types for which rates are being limited.
730 Significant bits: IHGFEDCBA9876543210
731 Default mask: 0000001100000011000 (6168)
733 Bit definitions (see include/linux/icmp.h):
735 3 Destination Unreachable *
740 C Parameter Problem *
745 H Address Mask Request
748 * These are rate limited by default (see default mask above)
750 icmp_ignore_bogus_error_responses - BOOLEAN
751 Some routers violate RFC1122 by sending bogus responses to broadcast
752 frames. Such violations are normally logged via a kernel warning.
753 If this is set to TRUE, the kernel will not give such warnings, which
754 will avoid log file clutter.
757 icmp_errors_use_inbound_ifaddr - BOOLEAN
759 If zero, icmp error messages are sent with the primary address of
760 the exiting interface.
762 If non-zero, the message will be sent with the primary address of
763 the interface that received the packet that caused the icmp error.
764 This is the behaviour network many administrators will expect from
765 a router. And it can make debugging complicated network layouts
768 Note that if no primary address exists for the interface selected,
769 then the primary address of the first non-loopback interface that
770 has one will be used regardless of this setting.
774 igmp_max_memberships - INTEGER
775 Change the maximum number of multicast groups we can subscribe to.
778 Theoretical maximum value is bounded by having to send a membership
779 report in a single datagram (i.e. the report can't span multiple
780 datagrams, or risk confusing the switch and leaving groups you don't
783 The number of supported groups 'M' is bounded by the number of group
784 report entries you can fit into a single datagram of 65535 bytes.
786 M = 65536-sizeof (ip header)/(sizeof(Group record))
788 Group records are variable length, with a minimum of 12 bytes.
789 So net.ipv4.igmp_max_memberships should not be set higher than:
791 (65536-24) / 12 = 5459
793 The value 5459 assumes no IP header options, so in practice
794 this number may be lower.
796 conf/interface/* changes special settings per interface (where
797 "interface" is the name of your network interface)
799 conf/all/* is special, changes the settings for all interfaces
801 log_martians - BOOLEAN
802 Log packets with impossible addresses to kernel log.
803 log_martians for the interface will be enabled if at least one of
804 conf/{all,interface}/log_martians is set to TRUE,
805 it will be disabled otherwise
807 accept_redirects - BOOLEAN
808 Accept ICMP redirect messages.
809 accept_redirects for the interface will be enabled if:
810 - both conf/{all,interface}/accept_redirects are TRUE in the case
811 forwarding for the interface is enabled
813 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
814 case forwarding for the interface is disabled
815 accept_redirects for the interface will be disabled otherwise
820 Enable IP forwarding on this interface.
822 mc_forwarding - BOOLEAN
823 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
824 and a multicast routing daemon is required.
825 conf/all/mc_forwarding must also be set to TRUE to enable multicast
826 routing for the interface
829 Integer value used to differentiate the devices by the medium they
830 are attached to. Two devices can have different id values when
831 the broadcast packets are received only on one of them.
832 The default value 0 means that the device is the only interface
833 to its medium, value of -1 means that medium is not known.
835 Currently, it is used to change the proxy_arp behavior:
836 the proxy_arp feature is enabled for packets forwarded between
837 two devices attached to different media.
841 proxy_arp for the interface will be enabled if at least one of
842 conf/{all,interface}/proxy_arp is set to TRUE,
843 it will be disabled otherwise
845 proxy_arp_pvlan - BOOLEAN
846 Private VLAN proxy arp.
847 Basically allow proxy arp replies back to the same interface
848 (from which the ARP request/solicitation was received).
850 This is done to support (ethernet) switch features, like RFC
851 3069, where the individual ports are NOT allowed to
852 communicate with each other, but they are allowed to talk to
853 the upstream router. As described in RFC 3069, it is possible
854 to allow these hosts to communicate through the upstream
855 router by proxy_arp'ing. Don't need to be used together with
858 This technology is known by different names:
859 In RFC 3069 it is called VLAN Aggregation.
860 Cisco and Allied Telesyn call it Private VLAN.
861 Hewlett-Packard call it Source-Port filtering or port-isolation.
862 Ericsson call it MAC-Forced Forwarding (RFC Draft).
864 shared_media - BOOLEAN
865 Send(router) or accept(host) RFC1620 shared media redirects.
866 Overrides ip_secure_redirects.
867 shared_media for the interface will be enabled if at least one of
868 conf/{all,interface}/shared_media is set to TRUE,
869 it will be disabled otherwise
872 secure_redirects - BOOLEAN
873 Accept ICMP redirect messages only for gateways,
874 listed in default gateway list.
875 secure_redirects for the interface will be enabled if at least one of
876 conf/{all,interface}/secure_redirects is set to TRUE,
877 it will be disabled otherwise
880 send_redirects - BOOLEAN
881 Send redirects, if router.
882 send_redirects for the interface will be enabled if at least one of
883 conf/{all,interface}/send_redirects is set to TRUE,
884 it will be disabled otherwise
887 bootp_relay - BOOLEAN
888 Accept packets with source address 0.b.c.d destined
889 not to this host as local ones. It is supposed, that
890 BOOTP relay daemon will catch and forward such packets.
891 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
896 accept_source_route - BOOLEAN
897 Accept packets with SRR option.
898 conf/all/accept_source_route must also be set to TRUE to accept packets
899 with SRR option on the interface
900 default TRUE (router)
903 accept_local - BOOLEAN
904 Accept packets with local source addresses. In combination
905 with suitable routing, this can be used to direct packets
906 between two local interfaces over the wire and have them
909 rp_filter must be set to a non-zero value in order for
910 accept_local to have an effect.
914 route_localnet - BOOLEAN
915 Do not consider loopback addresses as martian source or destination
916 while routing. This enables the use of 127/8 for local routing purposes.
920 0 - No source validation.
921 1 - Strict mode as defined in RFC3704 Strict Reverse Path
922 Each incoming packet is tested against the FIB and if the interface
923 is not the best reverse path the packet check will fail.
924 By default failed packets are discarded.
925 2 - Loose mode as defined in RFC3704 Loose Reverse Path
926 Each incoming packet's source address is also tested against the FIB
927 and if the source address is not reachable via any interface
928 the packet check will fail.
930 Current recommended practice in RFC3704 is to enable strict mode
931 to prevent IP spoofing from DDos attacks. If using asymmetric routing
932 or other complicated routing, then loose mode is recommended.
934 The max value from conf/{all,interface}/rp_filter is used
935 when doing source validation on the {interface}.
937 Default value is 0. Note that some distributions enable it
941 1 - Allows you to have multiple network interfaces on the same
942 subnet, and have the ARPs for each interface be answered
943 based on whether or not the kernel would route a packet from
944 the ARP'd IP out that interface (therefore you must use source
945 based routing for this to work). In other words it allows control
946 of which cards (usually 1) will respond to an arp request.
948 0 - (default) The kernel can respond to arp requests with addresses
949 from other interfaces. This may seem wrong but it usually makes
950 sense, because it increases the chance of successful communication.
951 IP addresses are owned by the complete host on Linux, not by
952 particular interfaces. Only for more complex setups like load-
953 balancing, does this behaviour cause problems.
955 arp_filter for the interface will be enabled if at least one of
956 conf/{all,interface}/arp_filter is set to TRUE,
957 it will be disabled otherwise
959 arp_announce - INTEGER
960 Define different restriction levels for announcing the local
961 source IP address from IP packets in ARP requests sent on
963 0 - (default) Use any local address, configured on any interface
964 1 - Try to avoid local addresses that are not in the target's
965 subnet for this interface. This mode is useful when target
966 hosts reachable via this interface require the source IP
967 address in ARP requests to be part of their logical network
968 configured on the receiving interface. When we generate the
969 request we will check all our subnets that include the
970 target IP and will preserve the source address if it is from
971 such subnet. If there is no such subnet we select source
972 address according to the rules for level 2.
973 2 - Always use the best local address for this target.
974 In this mode we ignore the source address in the IP packet
975 and try to select local address that we prefer for talks with
976 the target host. Such local address is selected by looking
977 for primary IP addresses on all our subnets on the outgoing
978 interface that include the target IP address. If no suitable
979 local address is found we select the first local address
980 we have on the outgoing interface or on all other interfaces,
981 with the hope we will receive reply for our request and
982 even sometimes no matter the source IP address we announce.
984 The max value from conf/{all,interface}/arp_announce is used.
986 Increasing the restriction level gives more chance for
987 receiving answer from the resolved target while decreasing
988 the level announces more valid sender's information.
991 Define different modes for sending replies in response to
992 received ARP requests that resolve local target IP addresses:
993 0 - (default): reply for any local target IP address, configured
995 1 - reply only if the target IP address is local address
996 configured on the incoming interface
997 2 - reply only if the target IP address is local address
998 configured on the incoming interface and both with the
999 sender's IP address are part from same subnet on this interface
1000 3 - do not reply for local addresses configured with scope host,
1001 only resolutions for global and link addresses are replied
1003 8 - do not reply for all local addresses
1005 The max value from conf/{all,interface}/arp_ignore is used
1006 when ARP request is received on the {interface}
1008 arp_notify - BOOLEAN
1009 Define mode for notification of address and device changes.
1010 0 - (default): do nothing
1011 1 - Generate gratuitous arp requests when device is brought up
1012 or hardware address changes.
1014 arp_accept - BOOLEAN
1015 Define behavior for gratuitous ARP frames who's IP is not
1016 already present in the ARP table:
1017 0 - don't create new entries in the ARP table
1018 1 - create new entries in the ARP table
1020 Both replies and requests type gratuitous arp will trigger the
1021 ARP table to be updated, if this setting is on.
1023 If the ARP table already contains the IP address of the
1024 gratuitous arp frame, the arp table will be updated regardless
1025 if this setting is on or off.
1028 app_solicit - INTEGER
1029 The maximum number of probes to send to the user space ARP daemon
1030 via netlink before dropping back to multicast probes (see
1031 mcast_solicit). Defaults to 0.
1033 disable_policy - BOOLEAN
1034 Disable IPSEC policy (SPD) for this interface
1036 disable_xfrm - BOOLEAN
1037 Disable IPSEC encryption on this interface, whatever the policy
1042 Allows you to write a number, which can be used as required.
1046 kuznet@ms2.inr.ac.ru
1052 delon.nicolas@wanadoo.fr
1057 /proc/sys/net/ipv6/* Variables:
1059 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1060 apply to IPv6 [XXX?].
1062 bindv6only - BOOLEAN
1063 Default value for IPV6_V6ONLY socket option,
1064 which restricts use of the IPv6 socket to IPv6 communication
1066 TRUE: disable IPv4-mapped address feature
1067 FALSE: enable IPv4-mapped address feature
1069 Default: FALSE (as specified in RFC3493)
1073 ip6frag_high_thresh - INTEGER
1074 Maximum memory used to reassemble IPv6 fragments. When
1075 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1076 the fragment handler will toss packets until ip6frag_low_thresh
1079 ip6frag_low_thresh - INTEGER
1080 See ip6frag_high_thresh
1082 ip6frag_time - INTEGER
1083 Time in seconds to keep an IPv6 fragment in memory.
1085 ip6frag_secret_interval - INTEGER
1086 Regeneration interval (in seconds) of the hash secret (or lifetime
1087 for the hash secret) for IPv6 fragments.
1091 Change the interface-specific default settings.
1095 Change all the interface-specific settings.
1097 [XXX: Other special features than forwarding?]
1099 conf/all/forwarding - BOOLEAN
1100 Enable global IPv6 forwarding between all interfaces.
1102 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1103 to control which interfaces may forward packets and which not.
1105 This also sets all interfaces' Host/Router setting
1106 'forwarding' to the specified value. See below for details.
1108 This referred to as global forwarding.
1114 Change special settings per interface.
1116 The functional behaviour for certain settings is different
1117 depending on whether local forwarding is enabled or not.
1120 Accept Router Advertisements; autoconfigure using them.
1122 It also determines whether or not to transmit Router
1123 Solicitations. If and only if the functional setting is to
1124 accept Router Advertisements, Router Solicitations will be
1127 Possible values are:
1128 0 Do not accept Router Advertisements.
1129 1 Accept Router Advertisements if forwarding is disabled.
1130 2 Overrule forwarding behaviour. Accept Router Advertisements
1131 even if forwarding is enabled.
1133 Functional default: enabled if local forwarding is disabled.
1134 disabled if local forwarding is enabled.
1136 accept_ra_defrtr - BOOLEAN
1137 Learn default router in Router Advertisement.
1139 Functional default: enabled if accept_ra is enabled.
1140 disabled if accept_ra is disabled.
1142 accept_ra_pinfo - BOOLEAN
1143 Learn Prefix Information in Router Advertisement.
1145 Functional default: enabled if accept_ra is enabled.
1146 disabled if accept_ra is disabled.
1148 accept_ra_rt_info_max_plen - INTEGER
1149 Maximum prefix length of Route Information in RA.
1151 Route Information w/ prefix larger than or equal to this
1152 variable shall be ignored.
1154 Functional default: 0 if accept_ra_rtr_pref is enabled.
1155 -1 if accept_ra_rtr_pref is disabled.
1157 accept_ra_rtr_pref - BOOLEAN
1158 Accept Router Preference in RA.
1160 Functional default: enabled if accept_ra is enabled.
1161 disabled if accept_ra is disabled.
1163 accept_redirects - BOOLEAN
1166 Functional default: enabled if local forwarding is disabled.
1167 disabled if local forwarding is enabled.
1169 accept_source_route - INTEGER
1170 Accept source routing (routing extension header).
1172 >= 0: Accept only routing header type 2.
1173 < 0: Do not accept routing header.
1178 Autoconfigure addresses using Prefix Information in Router
1181 Functional default: enabled if accept_ra_pinfo is enabled.
1182 disabled if accept_ra_pinfo is disabled.
1184 dad_transmits - INTEGER
1185 The amount of Duplicate Address Detection probes to send.
1188 forwarding - INTEGER
1189 Configure interface-specific Host/Router behaviour.
1191 Note: It is recommended to have the same setting on all
1192 interfaces; mixed router/host scenarios are rather uncommon.
1194 Possible values are:
1195 0 Forwarding disabled
1196 1 Forwarding enabled
1200 By default, Host behaviour is assumed. This means:
1202 1. IsRouter flag is not set in Neighbour Advertisements.
1203 2. If accept_ra is TRUE (default), transmit Router
1205 3. If accept_ra is TRUE (default), accept Router
1206 Advertisements (and do autoconfiguration).
1207 4. If accept_redirects is TRUE (default), accept Redirects.
1211 If local forwarding is enabled, Router behaviour is assumed.
1212 This means exactly the reverse from the above:
1214 1. IsRouter flag is set in Neighbour Advertisements.
1215 2. Router Solicitations are not sent unless accept_ra is 2.
1216 3. Router Advertisements are ignored unless accept_ra is 2.
1217 4. Redirects are ignored.
1219 Default: 0 (disabled) if global forwarding is disabled (default),
1220 otherwise 1 (enabled).
1223 Default Hop Limit to set.
1227 Default Maximum Transfer Unit
1228 Default: 1280 (IPv6 required minimum)
1230 router_probe_interval - INTEGER
1231 Minimum interval (in seconds) between Router Probing described
1236 router_solicitation_delay - INTEGER
1237 Number of seconds to wait after interface is brought up
1238 before sending Router Solicitations.
1241 router_solicitation_interval - INTEGER
1242 Number of seconds to wait between Router Solicitations.
1245 router_solicitations - INTEGER
1246 Number of Router Solicitations to send until assuming no
1247 routers are present.
1250 use_tempaddr - INTEGER
1251 Preference for Privacy Extensions (RFC3041).
1252 <= 0 : disable Privacy Extensions
1253 == 1 : enable Privacy Extensions, but prefer public
1254 addresses over temporary addresses.
1255 > 1 : enable Privacy Extensions and prefer temporary
1256 addresses over public addresses.
1257 Default: 0 (for most devices)
1258 -1 (for point-to-point devices and loopback devices)
1260 temp_valid_lft - INTEGER
1261 valid lifetime (in seconds) for temporary addresses.
1262 Default: 604800 (7 days)
1264 temp_prefered_lft - INTEGER
1265 Preferred lifetime (in seconds) for temporary addresses.
1266 Default: 86400 (1 day)
1268 max_desync_factor - INTEGER
1269 Maximum value for DESYNC_FACTOR, which is a random value
1270 that ensures that clients don't synchronize with each
1271 other and generate new addresses at exactly the same time.
1272 value is in seconds.
1275 regen_max_retry - INTEGER
1276 Number of attempts before give up attempting to generate
1277 valid temporary addresses.
1280 max_addresses - INTEGER
1281 Maximum number of autoconfigured addresses per interface. Setting
1282 to zero disables the limitation. It is not recommended to set this
1283 value too large (or to zero) because it would be an easy way to
1284 crash the kernel by allowing too many addresses to be created.
1287 disable_ipv6 - BOOLEAN
1288 Disable IPv6 operation. If accept_dad is set to 2, this value
1289 will be dynamically set to TRUE if DAD fails for the link-local
1291 Default: FALSE (enable IPv6 operation)
1293 When this value is changed from 1 to 0 (IPv6 is being enabled),
1294 it will dynamically create a link-local address on the given
1295 interface and start Duplicate Address Detection, if necessary.
1297 When this value is changed from 0 to 1 (IPv6 is being disabled),
1298 it will dynamically delete all address on the given interface.
1300 accept_dad - INTEGER
1301 Whether to accept DAD (Duplicate Address Detection).
1303 1: Enable DAD (default)
1304 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1305 link-local address has been found.
1307 force_tllao - BOOLEAN
1308 Enable sending the target link-layer address option even when
1309 responding to a unicast neighbor solicitation.
1312 Quoting from RFC 2461, section 4.4, Target link-layer address:
1314 "The option MUST be included for multicast solicitations in order to
1315 avoid infinite Neighbor Solicitation "recursion" when the peer node
1316 does not have a cache entry to return a Neighbor Advertisements
1317 message. When responding to unicast solicitations, the option can be
1318 omitted since the sender of the solicitation has the correct link-
1319 layer address; otherwise it would not have be able to send the unicast
1320 solicitation in the first place. However, including the link-layer
1321 address in this case adds little overhead and eliminates a potential
1322 race condition where the sender deletes the cached link-layer address
1323 prior to receiving a response to a previous solicitation."
1325 ndisc_notify - BOOLEAN
1326 Define mode for notification of address and device changes.
1327 0 - (default): do nothing
1328 1 - Generate unsolicited neighbour advertisements when device is brought
1329 up or hardware address changes.
1333 Limit the maximal rates for sending ICMPv6 packets.
1334 0 to disable any limiting,
1335 otherwise the minimal space between responses in milliseconds.
1340 Pekka Savola <pekkas@netcore.fi>
1341 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1344 /proc/sys/net/bridge/* Variables:
1346 bridge-nf-call-arptables - BOOLEAN
1347 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1351 bridge-nf-call-iptables - BOOLEAN
1352 1 : pass bridged IPv4 traffic to iptables' chains.
1356 bridge-nf-call-ip6tables - BOOLEAN
1357 1 : pass bridged IPv6 traffic to ip6tables' chains.
1361 bridge-nf-filter-vlan-tagged - BOOLEAN
1362 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1366 bridge-nf-filter-pppoe-tagged - BOOLEAN
1367 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1371 bridge-nf-pass-vlan-input-dev - BOOLEAN
1372 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1373 interface on the bridge and set the netfilter input device to the vlan.
1374 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1375 target work with vlan-on-top-of-bridge interfaces. When no matching
1376 vlan interface is found, or this switch is off, the input device is
1377 set to the bridge interface.
1378 0: disable bridge netfilter vlan interface lookup.
1381 proc/sys/net/sctp/* Variables:
1383 addip_enable - BOOLEAN
1384 Enable or disable extension of Dynamic Address Reconfiguration
1385 (ADD-IP) functionality specified in RFC5061. This extension provides
1386 the ability to dynamically add and remove new addresses for the SCTP
1389 1: Enable extension.
1391 0: Disable extension.
1395 addip_noauth_enable - BOOLEAN
1396 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1397 authentication to protect the operations of adding or removing new
1398 addresses. This requirement is mandated so that unauthorized hosts
1399 would not be able to hijack associations. However, older
1400 implementations may not have implemented this requirement while
1401 allowing the ADD-IP extension. For reasons of interoperability,
1402 we provide this variable to control the enforcement of the
1403 authentication requirement.
1405 1: Allow ADD-IP extension to be used without authentication. This
1406 should only be set in a closed environment for interoperability
1407 with older implementations.
1409 0: Enforce the authentication requirement
1413 auth_enable - BOOLEAN
1414 Enable or disable Authenticated Chunks extension. This extension
1415 provides the ability to send and receive authenticated chunks and is
1416 required for secure operation of Dynamic Address Reconfiguration
1419 1: Enable this extension.
1420 0: Disable this extension.
1424 prsctp_enable - BOOLEAN
1425 Enable or disable the Partial Reliability extension (RFC3758) which
1426 is used to notify peers that a given DATA should no longer be expected.
1434 The limit of the number of new packets that can be initially sent. It
1435 controls how bursty the generated traffic can be.
1439 association_max_retrans - INTEGER
1440 Set the maximum number for retransmissions that an association can
1441 attempt deciding that the remote end is unreachable. If this value
1442 is exceeded, the association is terminated.
1446 max_init_retransmits - INTEGER
1447 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1448 that an association will attempt before declaring the destination
1449 unreachable and terminating.
1453 path_max_retrans - INTEGER
1454 The maximum number of retransmissions that will be attempted on a given
1455 path. Once this threshold is exceeded, the path is considered
1456 unreachable, and new traffic will use a different path when the
1457 association is multihomed.
1461 pf_retrans - INTEGER
1462 The number of retransmissions that will be attempted on a given path
1463 before traffic is redirected to an alternate transport (should one
1464 exist). Note this is distinct from path_max_retrans, as a path that
1465 passes the pf_retrans threshold can still be used. Its only
1466 deprioritized when a transmission path is selected by the stack. This
1467 setting is primarily used to enable fast failover mechanisms without
1468 having to reduce path_max_retrans to a very low value. See:
1469 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1470 for details. Note also that a value of pf_retrans > path_max_retrans
1471 disables this feature
1475 rto_initial - INTEGER
1476 The initial round trip timeout value in milliseconds that will be used
1477 in calculating round trip times. This is the initial time interval
1478 for retransmissions.
1483 The maximum value (in milliseconds) of the round trip timeout. This
1484 is the largest time interval that can elapse between retransmissions.
1489 The minimum value (in milliseconds) of the round trip timeout. This
1490 is the smallest time interval the can elapse between retransmissions.
1494 hb_interval - INTEGER
1495 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1496 are sent at the specified interval on idle paths to probe the state of
1497 a given path between 2 associations.
1501 sack_timeout - INTEGER
1502 The amount of time (in milliseconds) that the implementation will wait
1507 valid_cookie_life - INTEGER
1508 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1509 is used during association establishment.
1513 cookie_preserve_enable - BOOLEAN
1514 Enable or disable the ability to extend the lifetime of the SCTP cookie
1515 that is used during the establishment phase of SCTP association
1517 1: Enable cookie lifetime extension.
1522 cookie_hmac_alg - STRING
1523 Select the hmac algorithm used when generating the cookie value sent by
1524 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1529 Ability to assign md5 or sha1 as the selected alg is predicated on the
1530 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1531 CONFIG_CRYPTO_SHA1).
1533 Default: Dependent on configuration. MD5 if available, else SHA1 if
1534 available, else none.
1536 rcvbuf_policy - INTEGER
1537 Determines if the receive buffer is attributed to the socket or to
1538 association. SCTP supports the capability to create multiple
1539 associations on a single socket. When using this capability, it is
1540 possible that a single stalled association that's buffering a lot
1541 of data may block other associations from delivering their data by
1542 consuming all of the receive buffer space. To work around this,
1543 the rcvbuf_policy could be set to attribute the receiver buffer space
1544 to each association instead of the socket. This prevents the described
1547 1: rcvbuf space is per association
1548 0: rcvbuf space is per socket
1552 sndbuf_policy - INTEGER
1553 Similar to rcvbuf_policy above, this applies to send buffer space.
1555 1: Send buffer is tracked per association
1556 0: Send buffer is tracked per socket.
1560 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1561 Number of pages allowed for queueing by all SCTP sockets.
1563 min: Below this number of pages SCTP is not bothered about its
1564 memory appetite. When amount of memory allocated by SCTP exceeds
1565 this number, SCTP starts to moderate memory usage.
1567 pressure: This value was introduced to follow format of tcp_mem.
1569 max: Number of pages allowed for queueing by all SCTP sockets.
1571 Default is calculated at boot time from amount of available memory.
1573 sctp_rmem - vector of 3 INTEGERs: min, default, max
1574 Only the first value ("min") is used, "default" and "max" are
1577 min: Minimal size of receive buffer used by SCTP socket.
1578 It is guaranteed to each SCTP socket (but not association) even
1579 under moderate memory pressure.
1583 sctp_wmem - vector of 3 INTEGERs: min, default, max
1584 Currently this tunable has no effect.
1586 addr_scope_policy - INTEGER
1587 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1589 0 - Disable IPv4 address scoping
1590 1 - Enable IPv4 address scoping
1591 2 - Follow draft but allow IPv4 private addresses
1592 3 - Follow draft but allow IPv4 link local addresses
1597 /proc/sys/net/core/*
1598 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1601 /proc/sys/net/unix/*
1602 max_dgram_qlen - INTEGER
1603 The maximum length of dgram socket receive queue
1610 /proc/sys/net/irda/*
1611 fast_poll_increase FIXME
1612 warn_noreply_time FIXME
1613 discovery_slots FIXME
1616 discovery_timeout FIXME
1617 lap_keepalive_time FIXME
1618 max_noreply_time FIXME
1619 max_tx_data_size FIXME
1621 min_tx_turn_time FIXME