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_autocorking - BOOLEAN
160 Enable TCP auto corking :
161 When applications do consecutive small write()/sendmsg() system calls,
162 we try to coalesce these small writes as much as possible, to lower
163 total amount of sent packets. This is done if at least one prior
164 packet for the flow is waiting in Qdisc queues or device transmit
165 queue. Applications can still use TCP_CORK for optimal behavior
166 when they know how/when to uncork their sockets.
169 tcp_available_congestion_control - STRING
170 Shows the available congestion control choices that are registered.
171 More congestion control algorithms may be available as modules,
174 tcp_base_mss - INTEGER
175 The initial value of search_low to be used by the packetization layer
176 Path MTU discovery (MTU probing). If MTU probing is enabled,
177 this is the initial MSS used by the connection.
179 tcp_congestion_control - STRING
180 Set the congestion control algorithm to be used for new
181 connections. The algorithm "reno" is always available, but
182 additional choices may be available based on kernel configuration.
183 Default is set as part of kernel configuration.
184 For passive connections, the listener congestion control choice
186 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
189 Allows TCP to send "duplicate" SACKs.
191 tcp_early_retrans - INTEGER
192 Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold
193 for triggering fast retransmit when the amount of outstanding data is
194 small and when no previously unsent data can be transmitted (such
195 that limited transmit could be used). Also controls the use of
196 Tail loss probe (TLP) that converts RTOs occurring due to tail
197 losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01).
201 2 enables ER but delays fast recovery and fast retransmit
202 by a fourth of RTT. This mitigates connection falsely
203 recovers when network has a small degree of reordering
204 (less than 3 packets).
205 3 enables delayed ER and TLP.
210 Control use of Explicit Congestion Notification (ECN) by TCP.
211 ECN is used only when both ends of the TCP connection indicate
212 support for it. This feature is useful in avoiding losses due
213 to congestion by allowing supporting routers to signal
214 congestion before having to drop packets.
216 0 Disable ECN. Neither initiate nor accept ECN.
217 1 Enable ECN when requested by incoming connections and
218 also request ECN on outgoing connection attempts.
219 2 Enable ECN when requested by incoming connections
220 but do not request ECN on outgoing connections.
224 Enable FACK congestion avoidance and fast retransmission.
225 The value is not used, if tcp_sack is not enabled.
227 tcp_fin_timeout - INTEGER
228 The length of time an orphaned (no longer referenced by any
229 application) connection will remain in the FIN_WAIT_2 state
230 before it is aborted at the local end. While a perfectly
231 valid "receive only" state for an un-orphaned connection, an
232 orphaned connection in FIN_WAIT_2 state could otherwise wait
233 forever for the remote to close its end of the connection.
238 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
239 F-RTO is an enhanced recovery algorithm for TCP retransmission
240 timeouts. It is particularly beneficial in networks where the
241 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
242 modification. It does not require any support from the peer.
244 By default it's enabled with a non-zero value. 0 disables F-RTO.
246 tcp_keepalive_time - INTEGER
247 How often TCP sends out keepalive messages when keepalive is enabled.
250 tcp_keepalive_probes - INTEGER
251 How many keepalive probes TCP sends out, until it decides that the
252 connection is broken. Default value: 9.
254 tcp_keepalive_intvl - INTEGER
255 How frequently the probes are send out. Multiplied by
256 tcp_keepalive_probes it is time to kill not responding connection,
257 after probes started. Default value: 75sec i.e. connection
258 will be aborted after ~11 minutes of retries.
260 tcp_low_latency - BOOLEAN
261 If set, the TCP stack makes decisions that prefer lower
262 latency as opposed to higher throughput. By default, this
263 option is not set meaning that higher throughput is preferred.
264 An example of an application where this default should be
265 changed would be a Beowulf compute cluster.
268 tcp_max_orphans - INTEGER
269 Maximal number of TCP sockets not attached to any user file handle,
270 held by system. If this number is exceeded orphaned connections are
271 reset immediately and warning is printed. This limit exists
272 only to prevent simple DoS attacks, you _must_ not rely on this
273 or lower the limit artificially, but rather increase it
274 (probably, after increasing installed memory),
275 if network conditions require more than default value,
276 and tune network services to linger and kill such states
277 more aggressively. Let me to remind again: each orphan eats
278 up to ~64K of unswappable memory.
280 tcp_max_syn_backlog - INTEGER
281 Maximal number of remembered connection requests, which have not
282 received an acknowledgment from connecting client.
283 The minimal value is 128 for low memory machines, and it will
284 increase in proportion to the memory of machine.
285 If server suffers from overload, try increasing this number.
287 tcp_max_tw_buckets - INTEGER
288 Maximal number of timewait sockets held by system simultaneously.
289 If this number is exceeded time-wait socket is immediately destroyed
290 and warning is printed. This limit exists only to prevent
291 simple DoS attacks, you _must_ not lower the limit artificially,
292 but rather increase it (probably, after increasing installed memory),
293 if network conditions require more than default value.
295 tcp_mem - vector of 3 INTEGERs: min, pressure, max
296 min: below this number of pages TCP is not bothered about its
299 pressure: when amount of memory allocated by TCP exceeds this number
300 of pages, TCP moderates its memory consumption and enters memory
301 pressure mode, which is exited when memory consumption falls
304 max: number of pages allowed for queueing by all TCP sockets.
306 Defaults are calculated at boot time from amount of available
309 tcp_moderate_rcvbuf - BOOLEAN
310 If set, TCP performs receive buffer auto-tuning, attempting to
311 automatically size the buffer (no greater than tcp_rmem[2]) to
312 match the size required by the path for full throughput. Enabled by
315 tcp_mtu_probing - INTEGER
316 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
319 1 - Disabled by default, enabled when an ICMP black hole detected
320 2 - Always enabled, use initial MSS of tcp_base_mss.
322 tcp_no_metrics_save - BOOLEAN
323 By default, TCP saves various connection metrics in the route cache
324 when the connection closes, so that connections established in the
325 near future can use these to set initial conditions. Usually, this
326 increases overall performance, but may sometimes cause performance
327 degradation. If set, TCP will not cache metrics on closing
330 tcp_orphan_retries - INTEGER
331 This value influences the timeout of a locally closed TCP connection,
332 when RTO retransmissions remain unacknowledged.
333 See tcp_retries2 for more details.
335 The default value is 8.
336 If your machine is a loaded WEB server,
337 you should think about lowering this value, such sockets
338 may consume significant resources. Cf. tcp_max_orphans.
340 tcp_reordering - INTEGER
341 Maximal reordering of packets in a TCP stream.
344 tcp_retrans_collapse - BOOLEAN
345 Bug-to-bug compatibility with some broken printers.
346 On retransmit try to send bigger packets to work around bugs in
349 tcp_retries1 - INTEGER
350 This value influences the time, after which TCP decides, that
351 something is wrong due to unacknowledged RTO retransmissions,
352 and reports this suspicion to the network layer.
353 See tcp_retries2 for more details.
355 RFC 1122 recommends at least 3 retransmissions, which is the
358 tcp_retries2 - INTEGER
359 This value influences the timeout of an alive TCP connection,
360 when RTO retransmissions remain unacknowledged.
361 Given a value of N, a hypothetical TCP connection following
362 exponential backoff with an initial RTO of TCP_RTO_MIN would
363 retransmit N times before killing the connection at the (N+1)th RTO.
365 The default value of 15 yields a hypothetical timeout of 924.6
366 seconds and is a lower bound for the effective timeout.
367 TCP will effectively time out at the first RTO which exceeds the
368 hypothetical timeout.
370 RFC 1122 recommends at least 100 seconds for the timeout,
371 which corresponds to a value of at least 8.
373 tcp_rfc1337 - BOOLEAN
374 If set, the TCP stack behaves conforming to RFC1337. If unset,
375 we are not conforming to RFC, but prevent TCP TIME_WAIT
379 tcp_rmem - vector of 3 INTEGERs: min, default, max
380 min: Minimal size of receive buffer used by TCP sockets.
381 It is guaranteed to each TCP socket, even under moderate memory
385 default: initial size of receive buffer used by TCP sockets.
386 This value overrides net.core.rmem_default used by other protocols.
387 Default: 87380 bytes. This value results in window of 65535 with
388 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
389 less for default tcp_app_win. See below about these variables.
391 max: maximal size of receive buffer allowed for automatically
392 selected receiver buffers for TCP socket. This value does not override
393 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
394 automatic tuning of that socket's receive buffer size, in which
395 case this value is ignored.
396 Default: between 87380B and 6MB, depending on RAM size.
399 Enable select acknowledgments (SACKS).
401 tcp_slow_start_after_idle - BOOLEAN
402 If set, provide RFC2861 behavior and time out the congestion
403 window after an idle period. An idle period is defined at
404 the current RTO. If unset, the congestion window will not
405 be timed out after an idle period.
409 Use the Host requirements interpretation of the TCP urgent pointer field.
410 Most hosts use the older BSD interpretation, so if you turn this on
411 Linux might not communicate correctly with them.
414 tcp_synack_retries - INTEGER
415 Number of times SYNACKs for a passive TCP connection attempt will
416 be retransmitted. Should not be higher than 255. Default value
417 is 5, which corresponds to 31seconds till the last retransmission
418 with the current initial RTO of 1second. With this the final timeout
419 for a passive TCP connection will happen after 63seconds.
421 tcp_syncookies - BOOLEAN
422 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
423 Send out syncookies when the syn backlog queue of a socket
424 overflows. This is to prevent against the common 'SYN flood attack'
427 Note, that syncookies is fallback facility.
428 It MUST NOT be used to help highly loaded servers to stand
429 against legal connection rate. If you see SYN flood warnings
430 in your logs, but investigation shows that they occur
431 because of overload with legal connections, you should tune
432 another parameters until this warning disappear.
433 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
435 syncookies seriously violate TCP protocol, do not allow
436 to use TCP extensions, can result in serious degradation
437 of some services (f.e. SMTP relaying), visible not by you,
438 but your clients and relays, contacting you. While you see
439 SYN flood warnings in logs not being really flooded, your server
440 is seriously misconfigured.
442 If you want to test which effects syncookies have to your
443 network connections you can set this knob to 2 to enable
444 unconditionally generation of syncookies.
446 tcp_fastopen - INTEGER
447 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
448 in the opening SYN packet. To use this feature, the client application
449 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
450 connect() to perform a TCP handshake automatically.
452 The values (bitmap) are
453 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN.
454 2: Enables TCP Fast Open on the server side, i.e., allowing data in
455 a SYN packet to be accepted and passed to the application before
456 3-way hand shake finishes.
457 4: Send data in the opening SYN regardless of cookie availability and
458 without a cookie option.
459 0x100: Accept SYN data w/o validating the cookie.
460 0x200: Accept data-in-SYN w/o any cookie option present.
461 0x400/0x800: Enable Fast Open on all listeners regardless of the
462 TCP_FASTOPEN socket option. The two different flags designate two
463 different ways of setting max_qlen without the TCP_FASTOPEN socket
468 Note that the client & server side Fast Open flags (1 and 2
469 respectively) must be also enabled before the rest of flags can take
472 See include/net/tcp.h and the code for more details.
474 tcp_syn_retries - INTEGER
475 Number of times initial SYNs for an active TCP connection attempt
476 will be retransmitted. Should not be higher than 255. Default value
477 is 6, which corresponds to 63seconds till the last retransmission
478 with the current initial RTO of 1second. With this the final timeout
479 for an active TCP connection attempt will happen after 127seconds.
481 tcp_timestamps - BOOLEAN
482 Enable timestamps as defined in RFC1323.
484 tcp_min_tso_segs - INTEGER
485 Minimal number of segments per TSO frame.
486 Since linux-3.12, TCP does an automatic sizing of TSO frames,
487 depending on flow rate, instead of filling 64Kbytes packets.
488 For specific usages, it's possible to force TCP to build big
489 TSO frames. Note that TCP stack might split too big TSO packets
490 if available window is too small.
493 tcp_tso_win_divisor - INTEGER
494 This allows control over what percentage of the congestion window
495 can be consumed by a single TSO frame.
496 The setting of this parameter is a choice between burstiness and
497 building larger TSO frames.
500 tcp_tw_recycle - BOOLEAN
501 Enable fast recycling TIME-WAIT sockets. Default value is 0.
502 It should not be changed without advice/request of technical
505 tcp_tw_reuse - BOOLEAN
506 Allow to reuse TIME-WAIT sockets for new connections when it is
507 safe from protocol viewpoint. Default value is 0.
508 It should not be changed without advice/request of technical
511 tcp_window_scaling - BOOLEAN
512 Enable window scaling as defined in RFC1323.
514 tcp_wmem - vector of 3 INTEGERs: min, default, max
515 min: Amount of memory reserved for send buffers for TCP sockets.
516 Each TCP socket has rights to use it due to fact of its birth.
519 default: initial size of send buffer used by TCP sockets. This
520 value overrides net.core.wmem_default used by other protocols.
521 It is usually lower than net.core.wmem_default.
524 max: Maximal amount of memory allowed for automatically tuned
525 send buffers for TCP sockets. This value does not override
526 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
527 automatic tuning of that socket's send buffer size, in which case
528 this value is ignored.
529 Default: between 64K and 4MB, depending on RAM size.
531 tcp_notsent_lowat - UNSIGNED INTEGER
532 A TCP socket can control the amount of unsent bytes in its write queue,
533 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
534 reports POLLOUT events if the amount of unsent bytes is below a per
535 socket value, and if the write queue is not full. sendmsg() will
536 also not add new buffers if the limit is hit.
538 This global variable controls the amount of unsent data for
539 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
540 to the global variable has immediate effect.
542 Default: UINT_MAX (0xFFFFFFFF)
544 tcp_workaround_signed_windows - BOOLEAN
545 If set, assume no receipt of a window scaling option means the
546 remote TCP is broken and treats the window as a signed quantity.
547 If unset, assume the remote TCP is not broken even if we do
548 not receive a window scaling option from them.
551 tcp_dma_copybreak - INTEGER
552 Lower limit, in bytes, of the size of socket reads that will be
553 offloaded to a DMA copy engine, if one is present in the system
554 and CONFIG_NET_DMA is enabled.
557 tcp_thin_linear_timeouts - BOOLEAN
558 Enable dynamic triggering of linear timeouts for thin streams.
559 If set, a check is performed upon retransmission by timeout to
560 determine if the stream is thin (less than 4 packets in flight).
561 As long as the stream is found to be thin, up to 6 linear
562 timeouts may be performed before exponential backoff mode is
563 initiated. This improves retransmission latency for
564 non-aggressive thin streams, often found to be time-dependent.
565 For more information on thin streams, see
566 Documentation/networking/tcp-thin.txt
569 tcp_thin_dupack - BOOLEAN
570 Enable dynamic triggering of retransmissions after one dupACK
571 for thin streams. If set, a check is performed upon reception
572 of a dupACK to determine if the stream is thin (less than 4
573 packets in flight). As long as the stream is found to be thin,
574 data is retransmitted on the first received dupACK. This
575 improves retransmission latency for non-aggressive thin
576 streams, often found to be time-dependent.
577 For more information on thin streams, see
578 Documentation/networking/tcp-thin.txt
581 tcp_limit_output_bytes - INTEGER
582 Controls TCP Small Queue limit per tcp socket.
583 TCP bulk sender tends to increase packets in flight until it
584 gets losses notifications. With SNDBUF autotuning, this can
585 result in a large amount of packets queued in qdisc/device
586 on the local machine, hurting latency of other flows, for
587 typical pfifo_fast qdiscs.
588 tcp_limit_output_bytes limits the number of bytes on qdisc
589 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
592 tcp_challenge_ack_limit - INTEGER
593 Limits number of Challenge ACK sent per second, as recommended
594 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
599 udp_mem - vector of 3 INTEGERs: min, pressure, max
600 Number of pages allowed for queueing by all UDP sockets.
602 min: Below this number of pages UDP is not bothered about its
603 memory appetite. When amount of memory allocated by UDP exceeds
604 this number, UDP starts to moderate memory usage.
606 pressure: This value was introduced to follow format of tcp_mem.
608 max: Number of pages allowed for queueing by all UDP sockets.
610 Default is calculated at boot time from amount of available memory.
612 udp_rmem_min - INTEGER
613 Minimal size of receive buffer used by UDP sockets in moderation.
614 Each UDP socket is able to use the size for receiving data, even if
615 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
618 udp_wmem_min - INTEGER
619 Minimal size of send buffer used by UDP sockets in moderation.
620 Each UDP socket is able to use the size for sending data, even if
621 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
626 cipso_cache_enable - BOOLEAN
627 If set, enable additions to and lookups from the CIPSO label mapping
628 cache. If unset, additions are ignored and lookups always result in a
629 miss. However, regardless of the setting the cache is still
630 invalidated when required when means you can safely toggle this on and
631 off and the cache will always be "safe".
634 cipso_cache_bucket_size - INTEGER
635 The CIPSO label cache consists of a fixed size hash table with each
636 hash bucket containing a number of cache entries. This variable limits
637 the number of entries in each hash bucket; the larger the value the
638 more CIPSO label mappings that can be cached. When the number of
639 entries in a given hash bucket reaches this limit adding new entries
640 causes the oldest entry in the bucket to be removed to make room.
643 cipso_rbm_optfmt - BOOLEAN
644 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
645 the CIPSO draft specification (see Documentation/netlabel for details).
646 This means that when set the CIPSO tag will be padded with empty
647 categories in order to make the packet data 32-bit aligned.
650 cipso_rbm_structvalid - BOOLEAN
651 If set, do a very strict check of the CIPSO option when
652 ip_options_compile() is called. If unset, relax the checks done during
653 ip_options_compile(). Either way is "safe" as errors are caught else
654 where in the CIPSO processing code but setting this to 0 (False) should
655 result in less work (i.e. it should be faster) but could cause problems
656 with other implementations that require strict checking.
661 ip_local_port_range - 2 INTEGERS
662 Defines the local port range that is used by TCP and UDP to
663 choose the local port. The first number is the first, the
664 second the last local port number. The default values are
665 32768 and 61000 respectively.
667 ip_local_reserved_ports - list of comma separated ranges
668 Specify the ports which are reserved for known third-party
669 applications. These ports will not be used by automatic port
670 assignments (e.g. when calling connect() or bind() with port
671 number 0). Explicit port allocation behavior is unchanged.
673 The format used for both input and output is a comma separated
674 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
675 10). Writing to the file will clear all previously reserved
676 ports and update the current list with the one given in the
679 Note that ip_local_port_range and ip_local_reserved_ports
680 settings are independent and both are considered by the kernel
681 when determining which ports are available for automatic port
684 You can reserve ports which are not in the current
685 ip_local_port_range, e.g.:
687 $ cat /proc/sys/net/ipv4/ip_local_port_range
689 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
692 although this is redundant. However such a setting is useful
693 if later the port range is changed to a value that will
694 include the reserved ports.
698 ip_nonlocal_bind - BOOLEAN
699 If set, allows processes to bind() to non-local IP addresses,
700 which can be quite useful - but may break some applications.
704 If set non-zero, enables support for dynamic addresses.
705 If set to a non-zero value larger than 1, a kernel log
706 message will be printed when dynamic address rewriting
710 ip_early_demux - BOOLEAN
711 Optimize input packet processing down to one demux for
712 certain kinds of local sockets. Currently we only do this
713 for established TCP sockets.
715 It may add an additional cost for pure routing workloads that
716 reduces overall throughput, in such case you should disable it.
719 icmp_echo_ignore_all - BOOLEAN
720 If set non-zero, then the kernel will ignore all ICMP ECHO
724 icmp_echo_ignore_broadcasts - BOOLEAN
725 If set non-zero, then the kernel will ignore all ICMP ECHO and
726 TIMESTAMP requests sent to it via broadcast/multicast.
729 icmp_ratelimit - INTEGER
730 Limit the maximal rates for sending ICMP packets whose type matches
731 icmp_ratemask (see below) to specific targets.
732 0 to disable any limiting,
733 otherwise the minimal space between responses in milliseconds.
736 icmp_ratemask - INTEGER
737 Mask made of ICMP types for which rates are being limited.
738 Significant bits: IHGFEDCBA9876543210
739 Default mask: 0000001100000011000 (6168)
741 Bit definitions (see include/linux/icmp.h):
743 3 Destination Unreachable *
748 C Parameter Problem *
753 H Address Mask Request
756 * These are rate limited by default (see default mask above)
758 icmp_ignore_bogus_error_responses - BOOLEAN
759 Some routers violate RFC1122 by sending bogus responses to broadcast
760 frames. Such violations are normally logged via a kernel warning.
761 If this is set to TRUE, the kernel will not give such warnings, which
762 will avoid log file clutter.
765 icmp_errors_use_inbound_ifaddr - BOOLEAN
767 If zero, icmp error messages are sent with the primary address of
768 the exiting interface.
770 If non-zero, the message will be sent with the primary address of
771 the interface that received the packet that caused the icmp error.
772 This is the behaviour network many administrators will expect from
773 a router. And it can make debugging complicated network layouts
776 Note that if no primary address exists for the interface selected,
777 then the primary address of the first non-loopback interface that
778 has one will be used regardless of this setting.
782 igmp_max_memberships - INTEGER
783 Change the maximum number of multicast groups we can subscribe to.
786 Theoretical maximum value is bounded by having to send a membership
787 report in a single datagram (i.e. the report can't span multiple
788 datagrams, or risk confusing the switch and leaving groups you don't
791 The number of supported groups 'M' is bounded by the number of group
792 report entries you can fit into a single datagram of 65535 bytes.
794 M = 65536-sizeof (ip header)/(sizeof(Group record))
796 Group records are variable length, with a minimum of 12 bytes.
797 So net.ipv4.igmp_max_memberships should not be set higher than:
799 (65536-24) / 12 = 5459
801 The value 5459 assumes no IP header options, so in practice
802 this number may be lower.
804 conf/interface/* changes special settings per interface (where
805 "interface" is the name of your network interface)
807 conf/all/* is special, changes the settings for all interfaces
809 log_martians - BOOLEAN
810 Log packets with impossible addresses to kernel log.
811 log_martians for the interface will be enabled if at least one of
812 conf/{all,interface}/log_martians is set to TRUE,
813 it will be disabled otherwise
815 accept_redirects - BOOLEAN
816 Accept ICMP redirect messages.
817 accept_redirects for the interface will be enabled if:
818 - both conf/{all,interface}/accept_redirects are TRUE in the case
819 forwarding for the interface is enabled
821 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
822 case forwarding for the interface is disabled
823 accept_redirects for the interface will be disabled otherwise
828 Enable IP forwarding on this interface.
830 mc_forwarding - BOOLEAN
831 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
832 and a multicast routing daemon is required.
833 conf/all/mc_forwarding must also be set to TRUE to enable multicast
834 routing for the interface
837 Integer value used to differentiate the devices by the medium they
838 are attached to. Two devices can have different id values when
839 the broadcast packets are received only on one of them.
840 The default value 0 means that the device is the only interface
841 to its medium, value of -1 means that medium is not known.
843 Currently, it is used to change the proxy_arp behavior:
844 the proxy_arp feature is enabled for packets forwarded between
845 two devices attached to different media.
849 proxy_arp for the interface will be enabled if at least one of
850 conf/{all,interface}/proxy_arp is set to TRUE,
851 it will be disabled otherwise
853 proxy_arp_pvlan - BOOLEAN
854 Private VLAN proxy arp.
855 Basically allow proxy arp replies back to the same interface
856 (from which the ARP request/solicitation was received).
858 This is done to support (ethernet) switch features, like RFC
859 3069, where the individual ports are NOT allowed to
860 communicate with each other, but they are allowed to talk to
861 the upstream router. As described in RFC 3069, it is possible
862 to allow these hosts to communicate through the upstream
863 router by proxy_arp'ing. Don't need to be used together with
866 This technology is known by different names:
867 In RFC 3069 it is called VLAN Aggregation.
868 Cisco and Allied Telesyn call it Private VLAN.
869 Hewlett-Packard call it Source-Port filtering or port-isolation.
870 Ericsson call it MAC-Forced Forwarding (RFC Draft).
872 shared_media - BOOLEAN
873 Send(router) or accept(host) RFC1620 shared media redirects.
874 Overrides ip_secure_redirects.
875 shared_media for the interface will be enabled if at least one of
876 conf/{all,interface}/shared_media is set to TRUE,
877 it will be disabled otherwise
880 secure_redirects - BOOLEAN
881 Accept ICMP redirect messages only for gateways,
882 listed in default gateway list.
883 secure_redirects for the interface will be enabled if at least one of
884 conf/{all,interface}/secure_redirects is set to TRUE,
885 it will be disabled otherwise
888 send_redirects - BOOLEAN
889 Send redirects, if router.
890 send_redirects for the interface will be enabled if at least one of
891 conf/{all,interface}/send_redirects is set to TRUE,
892 it will be disabled otherwise
895 bootp_relay - BOOLEAN
896 Accept packets with source address 0.b.c.d destined
897 not to this host as local ones. It is supposed, that
898 BOOTP relay daemon will catch and forward such packets.
899 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
904 accept_source_route - BOOLEAN
905 Accept packets with SRR option.
906 conf/all/accept_source_route must also be set to TRUE to accept packets
907 with SRR option on the interface
908 default TRUE (router)
911 accept_local - BOOLEAN
912 Accept packets with local source addresses. In combination
913 with suitable routing, this can be used to direct packets
914 between two local interfaces over the wire and have them
917 rp_filter must be set to a non-zero value in order for
918 accept_local to have an effect.
922 route_localnet - BOOLEAN
923 Do not consider loopback addresses as martian source or destination
924 while routing. This enables the use of 127/8 for local routing purposes.
928 0 - No source validation.
929 1 - Strict mode as defined in RFC3704 Strict Reverse Path
930 Each incoming packet is tested against the FIB and if the interface
931 is not the best reverse path the packet check will fail.
932 By default failed packets are discarded.
933 2 - Loose mode as defined in RFC3704 Loose Reverse Path
934 Each incoming packet's source address is also tested against the FIB
935 and if the source address is not reachable via any interface
936 the packet check will fail.
938 Current recommended practice in RFC3704 is to enable strict mode
939 to prevent IP spoofing from DDos attacks. If using asymmetric routing
940 or other complicated routing, then loose mode is recommended.
942 The max value from conf/{all,interface}/rp_filter is used
943 when doing source validation on the {interface}.
945 Default value is 0. Note that some distributions enable it
949 1 - Allows you to have multiple network interfaces on the same
950 subnet, and have the ARPs for each interface be answered
951 based on whether or not the kernel would route a packet from
952 the ARP'd IP out that interface (therefore you must use source
953 based routing for this to work). In other words it allows control
954 of which cards (usually 1) will respond to an arp request.
956 0 - (default) The kernel can respond to arp requests with addresses
957 from other interfaces. This may seem wrong but it usually makes
958 sense, because it increases the chance of successful communication.
959 IP addresses are owned by the complete host on Linux, not by
960 particular interfaces. Only for more complex setups like load-
961 balancing, does this behaviour cause problems.
963 arp_filter for the interface will be enabled if at least one of
964 conf/{all,interface}/arp_filter is set to TRUE,
965 it will be disabled otherwise
967 arp_announce - INTEGER
968 Define different restriction levels for announcing the local
969 source IP address from IP packets in ARP requests sent on
971 0 - (default) Use any local address, configured on any interface
972 1 - Try to avoid local addresses that are not in the target's
973 subnet for this interface. This mode is useful when target
974 hosts reachable via this interface require the source IP
975 address in ARP requests to be part of their logical network
976 configured on the receiving interface. When we generate the
977 request we will check all our subnets that include the
978 target IP and will preserve the source address if it is from
979 such subnet. If there is no such subnet we select source
980 address according to the rules for level 2.
981 2 - Always use the best local address for this target.
982 In this mode we ignore the source address in the IP packet
983 and try to select local address that we prefer for talks with
984 the target host. Such local address is selected by looking
985 for primary IP addresses on all our subnets on the outgoing
986 interface that include the target IP address. If no suitable
987 local address is found we select the first local address
988 we have on the outgoing interface or on all other interfaces,
989 with the hope we will receive reply for our request and
990 even sometimes no matter the source IP address we announce.
992 The max value from conf/{all,interface}/arp_announce is used.
994 Increasing the restriction level gives more chance for
995 receiving answer from the resolved target while decreasing
996 the level announces more valid sender's information.
999 Define different modes for sending replies in response to
1000 received ARP requests that resolve local target IP addresses:
1001 0 - (default): reply for any local target IP address, configured
1003 1 - reply only if the target IP address is local address
1004 configured on the incoming interface
1005 2 - reply only if the target IP address is local address
1006 configured on the incoming interface and both with the
1007 sender's IP address are part from same subnet on this interface
1008 3 - do not reply for local addresses configured with scope host,
1009 only resolutions for global and link addresses are replied
1011 8 - do not reply for all local addresses
1013 The max value from conf/{all,interface}/arp_ignore is used
1014 when ARP request is received on the {interface}
1016 arp_notify - BOOLEAN
1017 Define mode for notification of address and device changes.
1018 0 - (default): do nothing
1019 1 - Generate gratuitous arp requests when device is brought up
1020 or hardware address changes.
1022 arp_accept - BOOLEAN
1023 Define behavior for gratuitous ARP frames who's IP is not
1024 already present in the ARP table:
1025 0 - don't create new entries in the ARP table
1026 1 - create new entries in the ARP table
1028 Both replies and requests type gratuitous arp will trigger the
1029 ARP table to be updated, if this setting is on.
1031 If the ARP table already contains the IP address of the
1032 gratuitous arp frame, the arp table will be updated regardless
1033 if this setting is on or off.
1036 app_solicit - INTEGER
1037 The maximum number of probes to send to the user space ARP daemon
1038 via netlink before dropping back to multicast probes (see
1039 mcast_solicit). Defaults to 0.
1041 disable_policy - BOOLEAN
1042 Disable IPSEC policy (SPD) for this interface
1044 disable_xfrm - BOOLEAN
1045 Disable IPSEC encryption on this interface, whatever the policy
1047 igmpv2_unsolicited_report_interval - INTEGER
1048 The interval in milliseconds in which the next unsolicited
1049 IGMPv1 or IGMPv2 report retransmit will take place.
1050 Default: 10000 (10 seconds)
1052 igmpv3_unsolicited_report_interval - INTEGER
1053 The interval in milliseconds in which the next unsolicited
1054 IGMPv3 report retransmit will take place.
1055 Default: 1000 (1 seconds)
1058 Allows you to write a number, which can be used as required.
1062 kuznet@ms2.inr.ac.ru
1068 delon.nicolas@wanadoo.fr
1073 /proc/sys/net/ipv6/* Variables:
1075 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1076 apply to IPv6 [XXX?].
1078 bindv6only - BOOLEAN
1079 Default value for IPV6_V6ONLY socket option,
1080 which restricts use of the IPv6 socket to IPv6 communication
1082 TRUE: disable IPv4-mapped address feature
1083 FALSE: enable IPv4-mapped address feature
1085 Default: FALSE (as specified in RFC3493)
1089 ip6frag_high_thresh - INTEGER
1090 Maximum memory used to reassemble IPv6 fragments. When
1091 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1092 the fragment handler will toss packets until ip6frag_low_thresh
1095 ip6frag_low_thresh - INTEGER
1096 See ip6frag_high_thresh
1098 ip6frag_time - INTEGER
1099 Time in seconds to keep an IPv6 fragment in memory.
1101 ip6frag_secret_interval - INTEGER
1102 Regeneration interval (in seconds) of the hash secret (or lifetime
1103 for the hash secret) for IPv6 fragments.
1107 Change the interface-specific default settings.
1111 Change all the interface-specific settings.
1113 [XXX: Other special features than forwarding?]
1115 conf/all/forwarding - BOOLEAN
1116 Enable global IPv6 forwarding between all interfaces.
1118 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1119 to control which interfaces may forward packets and which not.
1121 This also sets all interfaces' Host/Router setting
1122 'forwarding' to the specified value. See below for details.
1124 This referred to as global forwarding.
1130 Change special settings per interface.
1132 The functional behaviour for certain settings is different
1133 depending on whether local forwarding is enabled or not.
1136 Accept Router Advertisements; autoconfigure using them.
1138 It also determines whether or not to transmit Router
1139 Solicitations. If and only if the functional setting is to
1140 accept Router Advertisements, Router Solicitations will be
1143 Possible values are:
1144 0 Do not accept Router Advertisements.
1145 1 Accept Router Advertisements if forwarding is disabled.
1146 2 Overrule forwarding behaviour. Accept Router Advertisements
1147 even if forwarding is enabled.
1149 Functional default: enabled if local forwarding is disabled.
1150 disabled if local forwarding is enabled.
1152 accept_ra_defrtr - BOOLEAN
1153 Learn default router in Router Advertisement.
1155 Functional default: enabled if accept_ra is enabled.
1156 disabled if accept_ra is disabled.
1158 accept_ra_pinfo - BOOLEAN
1159 Learn Prefix Information in Router Advertisement.
1161 Functional default: enabled if accept_ra is enabled.
1162 disabled if accept_ra is disabled.
1164 accept_ra_rt_info_max_plen - INTEGER
1165 Maximum prefix length of Route Information in RA.
1167 Route Information w/ prefix larger than or equal to this
1168 variable shall be ignored.
1170 Functional default: 0 if accept_ra_rtr_pref is enabled.
1171 -1 if accept_ra_rtr_pref is disabled.
1173 accept_ra_rtr_pref - BOOLEAN
1174 Accept Router Preference in RA.
1176 Functional default: enabled if accept_ra is enabled.
1177 disabled if accept_ra is disabled.
1179 accept_redirects - BOOLEAN
1182 Functional default: enabled if local forwarding is disabled.
1183 disabled if local forwarding is enabled.
1185 accept_source_route - INTEGER
1186 Accept source routing (routing extension header).
1188 >= 0: Accept only routing header type 2.
1189 < 0: Do not accept routing header.
1194 Autoconfigure addresses using Prefix Information in Router
1197 Functional default: enabled if accept_ra_pinfo is enabled.
1198 disabled if accept_ra_pinfo is disabled.
1200 dad_transmits - INTEGER
1201 The amount of Duplicate Address Detection probes to send.
1204 forwarding - INTEGER
1205 Configure interface-specific Host/Router behaviour.
1207 Note: It is recommended to have the same setting on all
1208 interfaces; mixed router/host scenarios are rather uncommon.
1210 Possible values are:
1211 0 Forwarding disabled
1212 1 Forwarding enabled
1216 By default, Host behaviour is assumed. This means:
1218 1. IsRouter flag is not set in Neighbour Advertisements.
1219 2. If accept_ra is TRUE (default), transmit Router
1221 3. If accept_ra is TRUE (default), accept Router
1222 Advertisements (and do autoconfiguration).
1223 4. If accept_redirects is TRUE (default), accept Redirects.
1227 If local forwarding is enabled, Router behaviour is assumed.
1228 This means exactly the reverse from the above:
1230 1. IsRouter flag is set in Neighbour Advertisements.
1231 2. Router Solicitations are not sent unless accept_ra is 2.
1232 3. Router Advertisements are ignored unless accept_ra is 2.
1233 4. Redirects are ignored.
1235 Default: 0 (disabled) if global forwarding is disabled (default),
1236 otherwise 1 (enabled).
1239 Default Hop Limit to set.
1243 Default Maximum Transfer Unit
1244 Default: 1280 (IPv6 required minimum)
1246 router_probe_interval - INTEGER
1247 Minimum interval (in seconds) between Router Probing described
1252 router_solicitation_delay - INTEGER
1253 Number of seconds to wait after interface is brought up
1254 before sending Router Solicitations.
1257 router_solicitation_interval - INTEGER
1258 Number of seconds to wait between Router Solicitations.
1261 router_solicitations - INTEGER
1262 Number of Router Solicitations to send until assuming no
1263 routers are present.
1266 use_tempaddr - INTEGER
1267 Preference for Privacy Extensions (RFC3041).
1268 <= 0 : disable Privacy Extensions
1269 == 1 : enable Privacy Extensions, but prefer public
1270 addresses over temporary addresses.
1271 > 1 : enable Privacy Extensions and prefer temporary
1272 addresses over public addresses.
1273 Default: 0 (for most devices)
1274 -1 (for point-to-point devices and loopback devices)
1276 temp_valid_lft - INTEGER
1277 valid lifetime (in seconds) for temporary addresses.
1278 Default: 604800 (7 days)
1280 temp_prefered_lft - INTEGER
1281 Preferred lifetime (in seconds) for temporary addresses.
1282 Default: 86400 (1 day)
1284 max_desync_factor - INTEGER
1285 Maximum value for DESYNC_FACTOR, which is a random value
1286 that ensures that clients don't synchronize with each
1287 other and generate new addresses at exactly the same time.
1288 value is in seconds.
1291 regen_max_retry - INTEGER
1292 Number of attempts before give up attempting to generate
1293 valid temporary addresses.
1296 max_addresses - INTEGER
1297 Maximum number of autoconfigured addresses per interface. Setting
1298 to zero disables the limitation. It is not recommended to set this
1299 value too large (or to zero) because it would be an easy way to
1300 crash the kernel by allowing too many addresses to be created.
1303 disable_ipv6 - BOOLEAN
1304 Disable IPv6 operation. If accept_dad is set to 2, this value
1305 will be dynamically set to TRUE if DAD fails for the link-local
1307 Default: FALSE (enable IPv6 operation)
1309 When this value is changed from 1 to 0 (IPv6 is being enabled),
1310 it will dynamically create a link-local address on the given
1311 interface and start Duplicate Address Detection, if necessary.
1313 When this value is changed from 0 to 1 (IPv6 is being disabled),
1314 it will dynamically delete all address on the given interface.
1316 accept_dad - INTEGER
1317 Whether to accept DAD (Duplicate Address Detection).
1319 1: Enable DAD (default)
1320 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1321 link-local address has been found.
1323 force_tllao - BOOLEAN
1324 Enable sending the target link-layer address option even when
1325 responding to a unicast neighbor solicitation.
1328 Quoting from RFC 2461, section 4.4, Target link-layer address:
1330 "The option MUST be included for multicast solicitations in order to
1331 avoid infinite Neighbor Solicitation "recursion" when the peer node
1332 does not have a cache entry to return a Neighbor Advertisements
1333 message. When responding to unicast solicitations, the option can be
1334 omitted since the sender of the solicitation has the correct link-
1335 layer address; otherwise it would not have be able to send the unicast
1336 solicitation in the first place. However, including the link-layer
1337 address in this case adds little overhead and eliminates a potential
1338 race condition where the sender deletes the cached link-layer address
1339 prior to receiving a response to a previous solicitation."
1341 ndisc_notify - BOOLEAN
1342 Define mode for notification of address and device changes.
1343 0 - (default): do nothing
1344 1 - Generate unsolicited neighbour advertisements when device is brought
1345 up or hardware address changes.
1347 mldv1_unsolicited_report_interval - INTEGER
1348 The interval in milliseconds in which the next unsolicited
1349 MLDv1 report retransmit will take place.
1350 Default: 10000 (10 seconds)
1352 mldv2_unsolicited_report_interval - INTEGER
1353 The interval in milliseconds in which the next unsolicited
1354 MLDv2 report retransmit will take place.
1355 Default: 1000 (1 second)
1357 force_mld_version - INTEGER
1358 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1359 1 - Enforce to use MLD version 1
1360 2 - Enforce to use MLD version 2
1362 suppress_frag_ndisc - INTEGER
1363 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1364 with IPv6 Neighbor Discovery) behavior:
1365 1 - (default) discard fragmented neighbor discovery packets
1366 0 - allow fragmented neighbor discovery packets
1370 Limit the maximal rates for sending ICMPv6 packets.
1371 0 to disable any limiting,
1372 otherwise the minimal space between responses in milliseconds.
1377 Pekka Savola <pekkas@netcore.fi>
1378 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1381 /proc/sys/net/bridge/* Variables:
1383 bridge-nf-call-arptables - BOOLEAN
1384 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1388 bridge-nf-call-iptables - BOOLEAN
1389 1 : pass bridged IPv4 traffic to iptables' chains.
1393 bridge-nf-call-ip6tables - BOOLEAN
1394 1 : pass bridged IPv6 traffic to ip6tables' chains.
1398 bridge-nf-filter-vlan-tagged - BOOLEAN
1399 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1403 bridge-nf-filter-pppoe-tagged - BOOLEAN
1404 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1408 bridge-nf-pass-vlan-input-dev - BOOLEAN
1409 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1410 interface on the bridge and set the netfilter input device to the vlan.
1411 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1412 target work with vlan-on-top-of-bridge interfaces. When no matching
1413 vlan interface is found, or this switch is off, the input device is
1414 set to the bridge interface.
1415 0: disable bridge netfilter vlan interface lookup.
1418 proc/sys/net/sctp/* Variables:
1420 addip_enable - BOOLEAN
1421 Enable or disable extension of Dynamic Address Reconfiguration
1422 (ADD-IP) functionality specified in RFC5061. This extension provides
1423 the ability to dynamically add and remove new addresses for the SCTP
1426 1: Enable extension.
1428 0: Disable extension.
1432 addip_noauth_enable - BOOLEAN
1433 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1434 authentication to protect the operations of adding or removing new
1435 addresses. This requirement is mandated so that unauthorized hosts
1436 would not be able to hijack associations. However, older
1437 implementations may not have implemented this requirement while
1438 allowing the ADD-IP extension. For reasons of interoperability,
1439 we provide this variable to control the enforcement of the
1440 authentication requirement.
1442 1: Allow ADD-IP extension to be used without authentication. This
1443 should only be set in a closed environment for interoperability
1444 with older implementations.
1446 0: Enforce the authentication requirement
1450 auth_enable - BOOLEAN
1451 Enable or disable Authenticated Chunks extension. This extension
1452 provides the ability to send and receive authenticated chunks and is
1453 required for secure operation of Dynamic Address Reconfiguration
1456 1: Enable this extension.
1457 0: Disable this extension.
1461 prsctp_enable - BOOLEAN
1462 Enable or disable the Partial Reliability extension (RFC3758) which
1463 is used to notify peers that a given DATA should no longer be expected.
1471 The limit of the number of new packets that can be initially sent. It
1472 controls how bursty the generated traffic can be.
1476 association_max_retrans - INTEGER
1477 Set the maximum number for retransmissions that an association can
1478 attempt deciding that the remote end is unreachable. If this value
1479 is exceeded, the association is terminated.
1483 max_init_retransmits - INTEGER
1484 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1485 that an association will attempt before declaring the destination
1486 unreachable and terminating.
1490 path_max_retrans - INTEGER
1491 The maximum number of retransmissions that will be attempted on a given
1492 path. Once this threshold is exceeded, the path is considered
1493 unreachable, and new traffic will use a different path when the
1494 association is multihomed.
1498 pf_retrans - INTEGER
1499 The number of retransmissions that will be attempted on a given path
1500 before traffic is redirected to an alternate transport (should one
1501 exist). Note this is distinct from path_max_retrans, as a path that
1502 passes the pf_retrans threshold can still be used. Its only
1503 deprioritized when a transmission path is selected by the stack. This
1504 setting is primarily used to enable fast failover mechanisms without
1505 having to reduce path_max_retrans to a very low value. See:
1506 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1507 for details. Note also that a value of pf_retrans > path_max_retrans
1508 disables this feature
1512 rto_initial - INTEGER
1513 The initial round trip timeout value in milliseconds that will be used
1514 in calculating round trip times. This is the initial time interval
1515 for retransmissions.
1520 The maximum value (in milliseconds) of the round trip timeout. This
1521 is the largest time interval that can elapse between retransmissions.
1526 The minimum value (in milliseconds) of the round trip timeout. This
1527 is the smallest time interval the can elapse between retransmissions.
1531 hb_interval - INTEGER
1532 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1533 are sent at the specified interval on idle paths to probe the state of
1534 a given path between 2 associations.
1538 sack_timeout - INTEGER
1539 The amount of time (in milliseconds) that the implementation will wait
1544 valid_cookie_life - INTEGER
1545 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1546 is used during association establishment.
1550 cookie_preserve_enable - BOOLEAN
1551 Enable or disable the ability to extend the lifetime of the SCTP cookie
1552 that is used during the establishment phase of SCTP association
1554 1: Enable cookie lifetime extension.
1559 cookie_hmac_alg - STRING
1560 Select the hmac algorithm used when generating the cookie value sent by
1561 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1566 Ability to assign md5 or sha1 as the selected alg is predicated on the
1567 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1568 CONFIG_CRYPTO_SHA1).
1570 Default: Dependent on configuration. MD5 if available, else SHA1 if
1571 available, else none.
1573 rcvbuf_policy - INTEGER
1574 Determines if the receive buffer is attributed to the socket or to
1575 association. SCTP supports the capability to create multiple
1576 associations on a single socket. When using this capability, it is
1577 possible that a single stalled association that's buffering a lot
1578 of data may block other associations from delivering their data by
1579 consuming all of the receive buffer space. To work around this,
1580 the rcvbuf_policy could be set to attribute the receiver buffer space
1581 to each association instead of the socket. This prevents the described
1584 1: rcvbuf space is per association
1585 0: rcvbuf space is per socket
1589 sndbuf_policy - INTEGER
1590 Similar to rcvbuf_policy above, this applies to send buffer space.
1592 1: Send buffer is tracked per association
1593 0: Send buffer is tracked per socket.
1597 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1598 Number of pages allowed for queueing by all SCTP sockets.
1600 min: Below this number of pages SCTP is not bothered about its
1601 memory appetite. When amount of memory allocated by SCTP exceeds
1602 this number, SCTP starts to moderate memory usage.
1604 pressure: This value was introduced to follow format of tcp_mem.
1606 max: Number of pages allowed for queueing by all SCTP sockets.
1608 Default is calculated at boot time from amount of available memory.
1610 sctp_rmem - vector of 3 INTEGERs: min, default, max
1611 Only the first value ("min") is used, "default" and "max" are
1614 min: Minimal size of receive buffer used by SCTP socket.
1615 It is guaranteed to each SCTP socket (but not association) even
1616 under moderate memory pressure.
1620 sctp_wmem - vector of 3 INTEGERs: min, default, max
1621 Currently this tunable has no effect.
1623 addr_scope_policy - INTEGER
1624 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1626 0 - Disable IPv4 address scoping
1627 1 - Enable IPv4 address scoping
1628 2 - Follow draft but allow IPv4 private addresses
1629 3 - Follow draft but allow IPv4 link local addresses
1634 /proc/sys/net/core/*
1635 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1638 /proc/sys/net/unix/*
1639 max_dgram_qlen - INTEGER
1640 The maximum length of dgram socket receive queue
1647 /proc/sys/net/irda/*
1648 fast_poll_increase FIXME
1649 warn_noreply_time FIXME
1650 discovery_slots FIXME
1653 discovery_timeout FIXME
1654 lap_keepalive_time FIXME
1655 max_noreply_time FIXME
1656 max_tx_data_size FIXME
1658 min_tx_turn_time FIXME