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_thresh3 - INTEGER
30 Maximum number of neighbor entries allowed. Increase this
31 when using large numbers of interfaces and when communicating
32 with large numbers of directly-connected peers.
35 neigh/default/unres_qlen_bytes - INTEGER
36 The maximum number of bytes which may be used by packets
37 queued for each unresolved address by other network layers.
39 Seting negative value is meaningless and will retrun error.
40 Default: 65536 Bytes(64KB)
42 neigh/default/unres_qlen - INTEGER
43 The maximum number of packets which may be queued for each
44 unresolved address by other network layers.
45 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
46 Prior to linux 3.3, the default value is 3 which may cause
47 unexpected packet loss. The current default value is calculated
48 according to default value of unres_qlen_bytes and true size of
53 Time, in seconds, that cached PMTU information is kept.
56 The advertised MSS depends on the first hop route MTU, but will
57 never be lower than this setting.
61 ipfrag_high_thresh - INTEGER
62 Maximum memory used to reassemble IP fragments. When
63 ipfrag_high_thresh bytes of memory is allocated for this purpose,
64 the fragment handler will toss packets until ipfrag_low_thresh
67 ipfrag_low_thresh - INTEGER
68 See ipfrag_high_thresh
71 Time in seconds to keep an IP fragment in memory.
73 ipfrag_secret_interval - INTEGER
74 Regeneration interval (in seconds) of the hash secret (or lifetime
75 for the hash secret) for IP fragments.
78 ipfrag_max_dist - INTEGER
79 ipfrag_max_dist is a non-negative integer value which defines the
80 maximum "disorder" which is allowed among fragments which share a
81 common IP source address. Note that reordering of packets is
82 not unusual, but if a large number of fragments arrive from a source
83 IP address while a particular fragment queue remains incomplete, it
84 probably indicates that one or more fragments belonging to that queue
85 have been lost. When ipfrag_max_dist is positive, an additional check
86 is done on fragments before they are added to a reassembly queue - if
87 ipfrag_max_dist (or more) fragments have arrived from a particular IP
88 address between additions to any IP fragment queue using that source
89 address, it's presumed that one or more fragments in the queue are
90 lost. The existing fragment queue will be dropped, and a new one
91 started. An ipfrag_max_dist value of zero disables this check.
93 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
94 result in unnecessarily dropping fragment queues when normal
95 reordering of packets occurs, which could lead to poor application
96 performance. Using a very large value, e.g. 50000, increases the
97 likelihood of incorrectly reassembling IP fragments that originate
98 from different IP datagrams, which could result in data corruption.
103 inet_peer_threshold - INTEGER
104 The approximate size of the storage. Starting from this threshold
105 entries will be thrown aggressively. This threshold also determines
106 entries' time-to-live and time intervals between garbage collection
107 passes. More entries, less time-to-live, less GC interval.
109 inet_peer_minttl - INTEGER
110 Minimum time-to-live of entries. Should be enough to cover fragment
111 time-to-live on the reassembling side. This minimum time-to-live is
112 guaranteed if the pool size is less than inet_peer_threshold.
115 inet_peer_maxttl - INTEGER
116 Maximum time-to-live of entries. Unused entries will expire after
117 this period of time if there is no memory pressure on the pool (i.e.
118 when the number of entries in the pool is very small).
124 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
125 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
129 Controls Appropriate Byte Count (ABC) defined in RFC3465.
130 ABC is a way of increasing congestion window (cwnd) more slowly
131 in response to partial acknowledgments.
133 0 increase cwnd once per acknowledgment (no ABC)
134 1 increase cwnd once per acknowledgment of full sized segment
135 2 allow increase cwnd by two if acknowledgment is
136 of two segments to compensate for delayed acknowledgments.
139 tcp_abort_on_overflow - BOOLEAN
140 If listening service is too slow to accept new connections,
141 reset them. Default state is FALSE. It means that if overflow
142 occurred due to a burst, connection will recover. Enable this
143 option _only_ if you are really sure that listening daemon
144 cannot be tuned to accept connections faster. Enabling this
145 option can harm clients of your server.
147 tcp_adv_win_scale - INTEGER
148 Count buffering overhead as bytes/2^tcp_adv_win_scale
149 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
151 Possible values are [-31, 31], inclusive.
154 tcp_allowed_congestion_control - STRING
155 Show/set the congestion control choices available to non-privileged
156 processes. The list is a subset of those listed in
157 tcp_available_congestion_control.
158 Default is "reno" and the default setting (tcp_congestion_control).
160 tcp_app_win - INTEGER
161 Reserve max(window/2^tcp_app_win, mss) of window for application
162 buffer. Value 0 is special, it means that nothing is reserved.
165 tcp_available_congestion_control - STRING
166 Shows the available congestion control choices that are registered.
167 More congestion control algorithms may be available as modules,
170 tcp_base_mss - INTEGER
171 The initial value of search_low to be used by the packetization layer
172 Path MTU discovery (MTU probing). If MTU probing is enabled,
173 this is the initial MSS used by the connection.
175 tcp_congestion_control - STRING
176 Set the congestion control algorithm to be used for new
177 connections. The algorithm "reno" is always available, but
178 additional choices may be available based on kernel configuration.
179 Default is set as part of kernel configuration.
180 For passive connections, the listener congestion control choice
182 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
184 tcp_cookie_size - INTEGER
185 Default size of TCP Cookie Transactions (TCPCT) option, that may be
186 overridden on a per socket basis by the TCPCT socket option.
187 Values greater than the maximum (16) are interpreted as the maximum.
188 Values greater than zero and less than the minimum (8) are interpreted
189 as the minimum. Odd values are interpreted as the next even value.
193 Allows TCP to send "duplicate" SACKs.
195 tcp_early_retrans - INTEGER
196 Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold
197 for triggering fast retransmit when the amount of outstanding data is
198 small and when no previously unsent data can be transmitted (such
199 that limited transmit could be used).
203 2 enables ER but delays fast recovery and fast retransmit
204 by a fourth of RTT. This mitigates connection falsely
205 recovers when network has a small degree of reordering
206 (less than 3 packets).
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 Always request ECN on outgoing connection attempts.
218 2 Enable ECN when requested by incomming connections
219 but do not request ECN on outgoing connections.
223 Enable FACK congestion avoidance and fast retransmission.
224 The value is not used, if tcp_sack is not enabled.
226 tcp_fin_timeout - INTEGER
227 The length of time an orphaned (no longer referenced by any
228 application) connection will remain in the FIN_WAIT_2 state
229 before it is aborted at the local end. While a perfectly
230 valid "receive only" state for an un-orphaned connection, an
231 orphaned connection in FIN_WAIT_2 state could otherwise wait
232 forever for the remote to close its end of the connection.
237 Enables Forward RTO-Recovery (F-RTO) defined in RFC4138.
238 F-RTO is an enhanced recovery algorithm for TCP retransmission
239 timeouts. It is particularly beneficial in wireless environments
240 where packet loss is typically due to random radio interference
241 rather than intermediate router congestion. F-RTO is sender-side
242 only modification. Therefore it does not require any support from
245 If set to 1, basic version is enabled. 2 enables SACK enhanced
246 F-RTO if flow uses SACK. The basic version can be used also when
247 SACK is in use though scenario(s) with it exists where F-RTO
248 interacts badly with the packet counting of the SACK enabled TCP
251 tcp_frto_response - INTEGER
252 When F-RTO has detected that a TCP retransmission timeout was
253 spurious (i.e, the timeout would have been avoided had TCP set a
254 longer retransmission timeout), TCP has several options what to do
255 next. Possible values are:
256 0 Rate halving based; a smooth and conservative response,
257 results in halved cwnd and ssthresh after one RTT
258 1 Very conservative response; not recommended because even
259 though being valid, it interacts poorly with the rest of
260 Linux TCP, halves cwnd and ssthresh immediately
261 2 Aggressive response; undoes congestion control measures
262 that are now known to be unnecessary (ignoring the
263 possibility of a lost retransmission that would require
264 TCP to be more cautious), cwnd and ssthresh are restored
265 to the values prior timeout
266 Default: 0 (rate halving based)
268 tcp_keepalive_time - INTEGER
269 How often TCP sends out keepalive messages when keepalive is enabled.
272 tcp_keepalive_probes - INTEGER
273 How many keepalive probes TCP sends out, until it decides that the
274 connection is broken. Default value: 9.
276 tcp_keepalive_intvl - INTEGER
277 How frequently the probes are send out. Multiplied by
278 tcp_keepalive_probes it is time to kill not responding connection,
279 after probes started. Default value: 75sec i.e. connection
280 will be aborted after ~11 minutes of retries.
282 tcp_low_latency - BOOLEAN
283 If set, the TCP stack makes decisions that prefer lower
284 latency as opposed to higher throughput. By default, this
285 option is not set meaning that higher throughput is preferred.
286 An example of an application where this default should be
287 changed would be a Beowulf compute cluster.
290 tcp_max_orphans - INTEGER
291 Maximal number of TCP sockets not attached to any user file handle,
292 held by system. If this number is exceeded orphaned connections are
293 reset immediately and warning is printed. This limit exists
294 only to prevent simple DoS attacks, you _must_ not rely on this
295 or lower the limit artificially, but rather increase it
296 (probably, after increasing installed memory),
297 if network conditions require more than default value,
298 and tune network services to linger and kill such states
299 more aggressively. Let me to remind again: each orphan eats
300 up to ~64K of unswappable memory.
302 tcp_max_ssthresh - INTEGER
303 Limited Slow-Start for TCP with large congestion windows (cwnd) defined in
304 RFC3742. Limited slow-start is a mechanism to limit growth of the cwnd
305 on the region where cwnd is larger than tcp_max_ssthresh. TCP increases cwnd
306 by at most tcp_max_ssthresh segments, and by at least tcp_max_ssthresh/2
307 segments per RTT when the cwnd is above tcp_max_ssthresh.
308 If TCP connection increased cwnd to thousands (or tens of thousands) segments,
309 and thousands of packets were being dropped during slow-start, you can set
310 tcp_max_ssthresh to improve performance for new TCP connection.
313 tcp_max_syn_backlog - INTEGER
314 Maximal number of remembered connection requests, which have not
315 received an acknowledgment from connecting client.
316 The minimal value is 128 for low memory machines, and it will
317 increase in proportion to the memory of machine.
318 If server suffers from overload, try increasing this number.
320 tcp_max_tw_buckets - INTEGER
321 Maximal number of timewait sockets held by system simultaneously.
322 If this number is exceeded time-wait socket is immediately destroyed
323 and warning is printed. This limit exists only to prevent
324 simple DoS attacks, you _must_ not lower the limit artificially,
325 but rather increase it (probably, after increasing installed memory),
326 if network conditions require more than default value.
328 tcp_mem - vector of 3 INTEGERs: min, pressure, max
329 min: below this number of pages TCP is not bothered about its
332 pressure: when amount of memory allocated by TCP exceeds this number
333 of pages, TCP moderates its memory consumption and enters memory
334 pressure mode, which is exited when memory consumption falls
337 max: number of pages allowed for queueing by all TCP sockets.
339 Defaults are calculated at boot time from amount of available
342 tcp_moderate_rcvbuf - BOOLEAN
343 If set, TCP performs receive buffer auto-tuning, attempting to
344 automatically size the buffer (no greater than tcp_rmem[2]) to
345 match the size required by the path for full throughput. Enabled by
348 tcp_mtu_probing - INTEGER
349 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
352 1 - Disabled by default, enabled when an ICMP black hole detected
353 2 - Always enabled, use initial MSS of tcp_base_mss.
355 tcp_no_metrics_save - BOOLEAN
356 By default, TCP saves various connection metrics in the route cache
357 when the connection closes, so that connections established in the
358 near future can use these to set initial conditions. Usually, this
359 increases overall performance, but may sometimes cause performance
360 degradation. If set, TCP will not cache metrics on closing
363 tcp_orphan_retries - INTEGER
364 This value influences the timeout of a locally closed TCP connection,
365 when RTO retransmissions remain unacknowledged.
366 See tcp_retries2 for more details.
368 The default value is 8.
369 If your machine is a loaded WEB server,
370 you should think about lowering this value, such sockets
371 may consume significant resources. Cf. tcp_max_orphans.
373 tcp_reordering - INTEGER
374 Maximal reordering of packets in a TCP stream.
377 tcp_retrans_collapse - BOOLEAN
378 Bug-to-bug compatibility with some broken printers.
379 On retransmit try to send bigger packets to work around bugs in
382 tcp_retries1 - INTEGER
383 This value influences the time, after which TCP decides, that
384 something is wrong due to unacknowledged RTO retransmissions,
385 and reports this suspicion to the network layer.
386 See tcp_retries2 for more details.
388 RFC 1122 recommends at least 3 retransmissions, which is the
391 tcp_retries2 - INTEGER
392 This value influences the timeout of an alive TCP connection,
393 when RTO retransmissions remain unacknowledged.
394 Given a value of N, a hypothetical TCP connection following
395 exponential backoff with an initial RTO of TCP_RTO_MIN would
396 retransmit N times before killing the connection at the (N+1)th RTO.
398 The default value of 15 yields a hypothetical timeout of 924.6
399 seconds and is a lower bound for the effective timeout.
400 TCP will effectively time out at the first RTO which exceeds the
401 hypothetical timeout.
403 RFC 1122 recommends at least 100 seconds for the timeout,
404 which corresponds to a value of at least 8.
406 tcp_rfc1337 - BOOLEAN
407 If set, the TCP stack behaves conforming to RFC1337. If unset,
408 we are not conforming to RFC, but prevent TCP TIME_WAIT
412 tcp_rmem - vector of 3 INTEGERs: min, default, max
413 min: Minimal size of receive buffer used by TCP sockets.
414 It is guaranteed to each TCP socket, even under moderate memory
418 default: initial size of receive buffer used by TCP sockets.
419 This value overrides net.core.rmem_default used by other protocols.
420 Default: 87380 bytes. This value results in window of 65535 with
421 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
422 less for default tcp_app_win. See below about these variables.
424 max: maximal size of receive buffer allowed for automatically
425 selected receiver buffers for TCP socket. This value does not override
426 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
427 automatic tuning of that socket's receive buffer size, in which
428 case this value is ignored.
429 Default: between 87380B and 6MB, depending on RAM size.
432 Enable select acknowledgments (SACKS).
434 tcp_slow_start_after_idle - BOOLEAN
435 If set, provide RFC2861 behavior and time out the congestion
436 window after an idle period. An idle period is defined at
437 the current RTO. If unset, the congestion window will not
438 be timed out after an idle period.
442 Use the Host requirements interpretation of the TCP urgent pointer field.
443 Most hosts use the older BSD interpretation, so if you turn this on
444 Linux might not communicate correctly with them.
447 tcp_synack_retries - INTEGER
448 Number of times SYNACKs for a passive TCP connection attempt will
449 be retransmitted. Should not be higher than 255. Default value
450 is 5, which corresponds to 31seconds till the last retransmission
451 with the current initial RTO of 1second. With this the final timeout
452 for a passive TCP connection will happen after 63seconds.
454 tcp_syncookies - BOOLEAN
455 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES
456 Send out syncookies when the syn backlog queue of a socket
457 overflows. This is to prevent against the common 'SYN flood attack'
460 Note, that syncookies is fallback facility.
461 It MUST NOT be used to help highly loaded servers to stand
462 against legal connection rate. If you see SYN flood warnings
463 in your logs, but investigation shows that they occur
464 because of overload with legal connections, you should tune
465 another parameters until this warning disappear.
466 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
468 syncookies seriously violate TCP protocol, do not allow
469 to use TCP extensions, can result in serious degradation
470 of some services (f.e. SMTP relaying), visible not by you,
471 but your clients and relays, contacting you. While you see
472 SYN flood warnings in logs not being really flooded, your server
473 is seriously misconfigured.
475 tcp_fastopen - INTEGER
476 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
477 in the opening SYN packet. To use this feature, the client application
478 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
479 connect() to perform a TCP handshake automatically.
481 The values (bitmap) are
482 1: Enables sending data in the opening SYN on the client.
483 2: Enables TCP Fast Open on the server side, i.e., allowing data in
484 a SYN packet to be accepted and passed to the application before
485 3-way hand shake finishes.
486 4: Send data in the opening SYN regardless of cookie availability and
487 without a cookie option.
488 0x100: Accept SYN data w/o validating the cookie.
489 0x200: Accept data-in-SYN w/o any cookie option present.
490 0x400/0x800: Enable Fast Open on all listeners regardless of the
491 TCP_FASTOPEN socket option. The two different flags designate two
492 different ways of setting max_qlen without the TCP_FASTOPEN socket
497 Note that the client & server side Fast Open flags (1 and 2
498 respectively) must be also enabled before the rest of flags can take
501 See include/net/tcp.h and the code for more details.
503 tcp_syn_retries - INTEGER
504 Number of times initial SYNs for an active TCP connection attempt
505 will be retransmitted. Should not be higher than 255. Default value
506 is 6, which corresponds to 63seconds till the last restransmission
507 with the current initial RTO of 1second. With this the final timeout
508 for an active TCP connection attempt will happen after 127seconds.
510 tcp_timestamps - BOOLEAN
511 Enable timestamps as defined in RFC1323.
513 tcp_tso_win_divisor - INTEGER
514 This allows control over what percentage of the congestion window
515 can be consumed by a single TSO frame.
516 The setting of this parameter is a choice between burstiness and
517 building larger TSO frames.
520 tcp_tw_recycle - BOOLEAN
521 Enable fast recycling TIME-WAIT sockets. Default value is 0.
522 It should not be changed without advice/request of technical
525 tcp_tw_reuse - BOOLEAN
526 Allow to reuse TIME-WAIT sockets for new connections when it is
527 safe from protocol viewpoint. Default value is 0.
528 It should not be changed without advice/request of technical
531 tcp_window_scaling - BOOLEAN
532 Enable window scaling as defined in RFC1323.
534 tcp_wmem - vector of 3 INTEGERs: min, default, max
535 min: Amount of memory reserved for send buffers for TCP sockets.
536 Each TCP socket has rights to use it due to fact of its birth.
539 default: initial size of send buffer used by TCP sockets. This
540 value overrides net.core.wmem_default used by other protocols.
541 It is usually lower than net.core.wmem_default.
544 max: Maximal amount of memory allowed for automatically tuned
545 send buffers for TCP sockets. This value does not override
546 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
547 automatic tuning of that socket's send buffer size, in which case
548 this value is ignored.
549 Default: between 64K and 4MB, depending on RAM size.
551 tcp_workaround_signed_windows - BOOLEAN
552 If set, assume no receipt of a window scaling option means the
553 remote TCP is broken and treats the window as a signed quantity.
554 If unset, assume the remote TCP is not broken even if we do
555 not receive a window scaling option from them.
558 tcp_dma_copybreak - INTEGER
559 Lower limit, in bytes, of the size of socket reads that will be
560 offloaded to a DMA copy engine, if one is present in the system
561 and CONFIG_NET_DMA is enabled.
564 tcp_thin_linear_timeouts - BOOLEAN
565 Enable dynamic triggering of linear timeouts for thin streams.
566 If set, a check is performed upon retransmission by timeout to
567 determine if the stream is thin (less than 4 packets in flight).
568 As long as the stream is found to be thin, up to 6 linear
569 timeouts may be performed before exponential backoff mode is
570 initiated. This improves retransmission latency for
571 non-aggressive thin streams, often found to be time-dependent.
572 For more information on thin streams, see
573 Documentation/networking/tcp-thin.txt
576 tcp_thin_dupack - BOOLEAN
577 Enable dynamic triggering of retransmissions after one dupACK
578 for thin streams. If set, a check is performed upon reception
579 of a dupACK to determine if the stream is thin (less than 4
580 packets in flight). As long as the stream is found to be thin,
581 data is retransmitted on the first received dupACK. This
582 improves retransmission latency for non-aggressive thin
583 streams, often found to be time-dependent.
584 For more information on thin streams, see
585 Documentation/networking/tcp-thin.txt
588 tcp_limit_output_bytes - INTEGER
589 Controls TCP Small Queue limit per tcp socket.
590 TCP bulk sender tends to increase packets in flight until it
591 gets losses notifications. With SNDBUF autotuning, this can
592 result in a large amount of packets queued in qdisc/device
593 on the local machine, hurting latency of other flows, for
594 typical pfifo_fast qdiscs.
595 tcp_limit_output_bytes limits the number of bytes on qdisc
596 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
597 Note: For GSO/TSO enabled flows, we try to have at least two
598 packets in flight. Reducing tcp_limit_output_bytes might also
599 reduce the size of individual GSO packet (64KB being the max)
602 tcp_challenge_ack_limit - INTEGER
603 Limits number of Challenge ACK sent per second, as recommended
604 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
609 udp_mem - vector of 3 INTEGERs: min, pressure, max
610 Number of pages allowed for queueing by all UDP sockets.
612 min: Below this number of pages UDP is not bothered about its
613 memory appetite. When amount of memory allocated by UDP exceeds
614 this number, UDP starts to moderate memory usage.
616 pressure: This value was introduced to follow format of tcp_mem.
618 max: Number of pages allowed for queueing by all UDP sockets.
620 Default is calculated at boot time from amount of available memory.
622 udp_rmem_min - INTEGER
623 Minimal size of receive buffer used by UDP sockets in moderation.
624 Each UDP socket is able to use the size for receiving data, even if
625 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
628 udp_wmem_min - INTEGER
629 Minimal size of send buffer used by UDP sockets in moderation.
630 Each UDP socket is able to use the size for sending data, even if
631 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
636 cipso_cache_enable - BOOLEAN
637 If set, enable additions to and lookups from the CIPSO label mapping
638 cache. If unset, additions are ignored and lookups always result in a
639 miss. However, regardless of the setting the cache is still
640 invalidated when required when means you can safely toggle this on and
641 off and the cache will always be "safe".
644 cipso_cache_bucket_size - INTEGER
645 The CIPSO label cache consists of a fixed size hash table with each
646 hash bucket containing a number of cache entries. This variable limits
647 the number of entries in each hash bucket; the larger the value the
648 more CIPSO label mappings that can be cached. When the number of
649 entries in a given hash bucket reaches this limit adding new entries
650 causes the oldest entry in the bucket to be removed to make room.
653 cipso_rbm_optfmt - BOOLEAN
654 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
655 the CIPSO draft specification (see Documentation/netlabel for details).
656 This means that when set the CIPSO tag will be padded with empty
657 categories in order to make the packet data 32-bit aligned.
660 cipso_rbm_structvalid - BOOLEAN
661 If set, do a very strict check of the CIPSO option when
662 ip_options_compile() is called. If unset, relax the checks done during
663 ip_options_compile(). Either way is "safe" as errors are caught else
664 where in the CIPSO processing code but setting this to 0 (False) should
665 result in less work (i.e. it should be faster) but could cause problems
666 with other implementations that require strict checking.
671 ip_local_port_range - 2 INTEGERS
672 Defines the local port range that is used by TCP and UDP to
673 choose the local port. The first number is the first, the
674 second the last local port number. The default values are
675 32768 and 61000 respectively.
677 ip_local_reserved_ports - list of comma separated ranges
678 Specify the ports which are reserved for known third-party
679 applications. These ports will not be used by automatic port
680 assignments (e.g. when calling connect() or bind() with port
681 number 0). Explicit port allocation behavior is unchanged.
683 The format used for both input and output is a comma separated
684 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
685 10). Writing to the file will clear all previously reserved
686 ports and update the current list with the one given in the
689 Note that ip_local_port_range and ip_local_reserved_ports
690 settings are independent and both are considered by the kernel
691 when determining which ports are available for automatic port
694 You can reserve ports which are not in the current
695 ip_local_port_range, e.g.:
697 $ cat /proc/sys/net/ipv4/ip_local_port_range
699 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
702 although this is redundant. However such a setting is useful
703 if later the port range is changed to a value that will
704 include the reserved ports.
708 ip_nonlocal_bind - BOOLEAN
709 If set, allows processes to bind() to non-local IP addresses,
710 which can be quite useful - but may break some applications.
714 If set non-zero, enables support for dynamic addresses.
715 If set to a non-zero value larger than 1, a kernel log
716 message will be printed when dynamic address rewriting
720 icmp_echo_ignore_all - BOOLEAN
721 If set non-zero, then the kernel will ignore all ICMP ECHO
725 icmp_echo_ignore_broadcasts - BOOLEAN
726 If set non-zero, then the kernel will ignore all ICMP ECHO and
727 TIMESTAMP requests sent to it via broadcast/multicast.
730 icmp_ratelimit - INTEGER
731 Limit the maximal rates for sending ICMP packets whose type matches
732 icmp_ratemask (see below) to specific targets.
733 0 to disable any limiting,
734 otherwise the minimal space between responses in milliseconds.
737 icmp_ratemask - INTEGER
738 Mask made of ICMP types for which rates are being limited.
739 Significant bits: IHGFEDCBA9876543210
740 Default mask: 0000001100000011000 (6168)
742 Bit definitions (see include/linux/icmp.h):
744 3 Destination Unreachable *
749 C Parameter Problem *
754 H Address Mask Request
757 * These are rate limited by default (see default mask above)
759 icmp_ignore_bogus_error_responses - BOOLEAN
760 Some routers violate RFC1122 by sending bogus responses to broadcast
761 frames. Such violations are normally logged via a kernel warning.
762 If this is set to TRUE, the kernel will not give such warnings, which
763 will avoid log file clutter.
766 icmp_errors_use_inbound_ifaddr - BOOLEAN
768 If zero, icmp error messages are sent with the primary address of
769 the exiting interface.
771 If non-zero, the message will be sent with the primary address of
772 the interface that received the packet that caused the icmp error.
773 This is the behaviour network many administrators will expect from
774 a router. And it can make debugging complicated network layouts
777 Note that if no primary address exists for the interface selected,
778 then the primary address of the first non-loopback interface that
779 has one will be used regardless of this setting.
783 igmp_max_memberships - INTEGER
784 Change the maximum number of multicast groups we can subscribe to.
787 Theoretical maximum value is bounded by having to send a membership
788 report in a single datagram (i.e. the report can't span multiple
789 datagrams, or risk confusing the switch and leaving groups you don't
792 The number of supported groups 'M' is bounded by the number of group
793 report entries you can fit into a single datagram of 65535 bytes.
795 M = 65536-sizeof (ip header)/(sizeof(Group record))
797 Group records are variable length, with a minimum of 12 bytes.
798 So net.ipv4.igmp_max_memberships should not be set higher than:
800 (65536-24) / 12 = 5459
802 The value 5459 assumes no IP header options, so in practice
803 this number may be lower.
805 conf/interface/* changes special settings per interface (where
806 "interface" is the name of your network interface)
808 conf/all/* is special, changes the settings for all interfaces
810 log_martians - BOOLEAN
811 Log packets with impossible addresses to kernel log.
812 log_martians for the interface will be enabled if at least one of
813 conf/{all,interface}/log_martians is set to TRUE,
814 it will be disabled otherwise
816 accept_redirects - BOOLEAN
817 Accept ICMP redirect messages.
818 accept_redirects for the interface will be enabled if:
819 - both conf/{all,interface}/accept_redirects are TRUE in the case
820 forwarding for the interface is enabled
822 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
823 case forwarding for the interface is disabled
824 accept_redirects for the interface will be disabled otherwise
829 Enable IP forwarding on this interface.
831 mc_forwarding - BOOLEAN
832 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
833 and a multicast routing daemon is required.
834 conf/all/mc_forwarding must also be set to TRUE to enable multicast
835 routing for the interface
838 Integer value used to differentiate the devices by the medium they
839 are attached to. Two devices can have different id values when
840 the broadcast packets are received only on one of them.
841 The default value 0 means that the device is the only interface
842 to its medium, value of -1 means that medium is not known.
844 Currently, it is used to change the proxy_arp behavior:
845 the proxy_arp feature is enabled for packets forwarded between
846 two devices attached to different media.
850 proxy_arp for the interface will be enabled if at least one of
851 conf/{all,interface}/proxy_arp is set to TRUE,
852 it will be disabled otherwise
854 proxy_arp_pvlan - BOOLEAN
855 Private VLAN proxy arp.
856 Basically allow proxy arp replies back to the same interface
857 (from which the ARP request/solicitation was received).
859 This is done to support (ethernet) switch features, like RFC
860 3069, where the individual ports are NOT allowed to
861 communicate with each other, but they are allowed to talk to
862 the upstream router. As described in RFC 3069, it is possible
863 to allow these hosts to communicate through the upstream
864 router by proxy_arp'ing. Don't need to be used together with
867 This technology is known by different names:
868 In RFC 3069 it is called VLAN Aggregation.
869 Cisco and Allied Telesyn call it Private VLAN.
870 Hewlett-Packard call it Source-Port filtering or port-isolation.
871 Ericsson call it MAC-Forced Forwarding (RFC Draft).
873 shared_media - BOOLEAN
874 Send(router) or accept(host) RFC1620 shared media redirects.
875 Overrides ip_secure_redirects.
876 shared_media for the interface will be enabled if at least one of
877 conf/{all,interface}/shared_media is set to TRUE,
878 it will be disabled otherwise
881 secure_redirects - BOOLEAN
882 Accept ICMP redirect messages only for gateways,
883 listed in default gateway list.
884 secure_redirects for the interface will be enabled if at least one of
885 conf/{all,interface}/secure_redirects is set to TRUE,
886 it will be disabled otherwise
889 send_redirects - BOOLEAN
890 Send redirects, if router.
891 send_redirects for the interface will be enabled if at least one of
892 conf/{all,interface}/send_redirects is set to TRUE,
893 it will be disabled otherwise
896 bootp_relay - BOOLEAN
897 Accept packets with source address 0.b.c.d destined
898 not to this host as local ones. It is supposed, that
899 BOOTP relay daemon will catch and forward such packets.
900 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
905 accept_source_route - BOOLEAN
906 Accept packets with SRR option.
907 conf/all/accept_source_route must also be set to TRUE to accept packets
908 with SRR option on the interface
909 default TRUE (router)
912 accept_local - BOOLEAN
913 Accept packets with local source addresses. In combination
914 with suitable routing, this can be used to direct packets
915 between two local interfaces over the wire and have them
918 rp_filter must be set to a non-zero value in order for
919 accept_local to have an effect.
923 route_localnet - BOOLEAN
924 Do not consider loopback addresses as martian source or destination
925 while routing. This enables the use of 127/8 for local routing purposes.
929 0 - No source validation.
930 1 - Strict mode as defined in RFC3704 Strict Reverse Path
931 Each incoming packet is tested against the FIB and if the interface
932 is not the best reverse path the packet check will fail.
933 By default failed packets are discarded.
934 2 - Loose mode as defined in RFC3704 Loose Reverse Path
935 Each incoming packet's source address is also tested against the FIB
936 and if the source address is not reachable via any interface
937 the packet check will fail.
939 Current recommended practice in RFC3704 is to enable strict mode
940 to prevent IP spoofing from DDos attacks. If using asymmetric routing
941 or other complicated routing, then loose mode is recommended.
943 The max value from conf/{all,interface}/rp_filter is used
944 when doing source validation on the {interface}.
946 Default value is 0. Note that some distributions enable it
950 1 - Allows you to have multiple network interfaces on the same
951 subnet, and have the ARPs for each interface be answered
952 based on whether or not the kernel would route a packet from
953 the ARP'd IP out that interface (therefore you must use source
954 based routing for this to work). In other words it allows control
955 of which cards (usually 1) will respond to an arp request.
957 0 - (default) The kernel can respond to arp requests with addresses
958 from other interfaces. This may seem wrong but it usually makes
959 sense, because it increases the chance of successful communication.
960 IP addresses are owned by the complete host on Linux, not by
961 particular interfaces. Only for more complex setups like load-
962 balancing, does this behaviour cause problems.
964 arp_filter for the interface will be enabled if at least one of
965 conf/{all,interface}/arp_filter is set to TRUE,
966 it will be disabled otherwise
968 arp_announce - INTEGER
969 Define different restriction levels for announcing the local
970 source IP address from IP packets in ARP requests sent on
972 0 - (default) Use any local address, configured on any interface
973 1 - Try to avoid local addresses that are not in the target's
974 subnet for this interface. This mode is useful when target
975 hosts reachable via this interface require the source IP
976 address in ARP requests to be part of their logical network
977 configured on the receiving interface. When we generate the
978 request we will check all our subnets that include the
979 target IP and will preserve the source address if it is from
980 such subnet. If there is no such subnet we select source
981 address according to the rules for level 2.
982 2 - Always use the best local address for this target.
983 In this mode we ignore the source address in the IP packet
984 and try to select local address that we prefer for talks with
985 the target host. Such local address is selected by looking
986 for primary IP addresses on all our subnets on the outgoing
987 interface that include the target IP address. If no suitable
988 local address is found we select the first local address
989 we have on the outgoing interface or on all other interfaces,
990 with the hope we will receive reply for our request and
991 even sometimes no matter the source IP address we announce.
993 The max value from conf/{all,interface}/arp_announce is used.
995 Increasing the restriction level gives more chance for
996 receiving answer from the resolved target while decreasing
997 the level announces more valid sender's information.
1000 Define different modes for sending replies in response to
1001 received ARP requests that resolve local target IP addresses:
1002 0 - (default): reply for any local target IP address, configured
1004 1 - reply only if the target IP address is local address
1005 configured on the incoming interface
1006 2 - reply only if the target IP address is local address
1007 configured on the incoming interface and both with the
1008 sender's IP address are part from same subnet on this interface
1009 3 - do not reply for local addresses configured with scope host,
1010 only resolutions for global and link addresses are replied
1012 8 - do not reply for all local addresses
1014 The max value from conf/{all,interface}/arp_ignore is used
1015 when ARP request is received on the {interface}
1017 arp_notify - BOOLEAN
1018 Define mode for notification of address and device changes.
1019 0 - (default): do nothing
1020 1 - Generate gratuitous arp requests when device is brought up
1021 or hardware address changes.
1023 arp_accept - BOOLEAN
1024 Define behavior for gratuitous ARP frames who's IP is not
1025 already present in the ARP table:
1026 0 - don't create new entries in the ARP table
1027 1 - create new entries in the ARP table
1029 Both replies and requests type gratuitous arp will trigger the
1030 ARP table to be updated, if this setting is on.
1032 If the ARP table already contains the IP address of the
1033 gratuitous arp frame, the arp table will be updated regardless
1034 if this setting is on or off.
1037 app_solicit - INTEGER
1038 The maximum number of probes to send to the user space ARP daemon
1039 via netlink before dropping back to multicast probes (see
1040 mcast_solicit). Defaults to 0.
1042 disable_policy - BOOLEAN
1043 Disable IPSEC policy (SPD) for this interface
1045 disable_xfrm - BOOLEAN
1046 Disable IPSEC encryption on this interface, whatever the policy
1051 Allows you to write a number, which can be used as required.
1055 kuznet@ms2.inr.ac.ru
1061 delon.nicolas@wanadoo.fr
1066 /proc/sys/net/ipv6/* Variables:
1068 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1069 apply to IPv6 [XXX?].
1071 bindv6only - BOOLEAN
1072 Default value for IPV6_V6ONLY socket option,
1073 which restricts use of the IPv6 socket to IPv6 communication
1075 TRUE: disable IPv4-mapped address feature
1076 FALSE: enable IPv4-mapped address feature
1078 Default: FALSE (as specified in RFC3493)
1082 ip6frag_high_thresh - INTEGER
1083 Maximum memory used to reassemble IPv6 fragments. When
1084 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1085 the fragment handler will toss packets until ip6frag_low_thresh
1088 ip6frag_low_thresh - INTEGER
1089 See ip6frag_high_thresh
1091 ip6frag_time - INTEGER
1092 Time in seconds to keep an IPv6 fragment in memory.
1094 ip6frag_secret_interval - INTEGER
1095 Regeneration interval (in seconds) of the hash secret (or lifetime
1096 for the hash secret) for IPv6 fragments.
1100 Change the interface-specific default settings.
1104 Change all the interface-specific settings.
1106 [XXX: Other special features than forwarding?]
1108 conf/all/forwarding - BOOLEAN
1109 Enable global IPv6 forwarding between all interfaces.
1111 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1112 to control which interfaces may forward packets and which not.
1114 This also sets all interfaces' Host/Router setting
1115 'forwarding' to the specified value. See below for details.
1117 This referred to as global forwarding.
1123 Change special settings per interface.
1125 The functional behaviour for certain settings is different
1126 depending on whether local forwarding is enabled or not.
1129 Accept Router Advertisements; autoconfigure using them.
1131 It also determines whether or not to transmit Router
1132 Solicitations. If and only if the functional setting is to
1133 accept Router Advertisements, Router Solicitations will be
1136 Possible values are:
1137 0 Do not accept Router Advertisements.
1138 1 Accept Router Advertisements if forwarding is disabled.
1139 2 Overrule forwarding behaviour. Accept Router Advertisements
1140 even if forwarding is enabled.
1142 Functional default: enabled if local forwarding is disabled.
1143 disabled if local forwarding is enabled.
1145 accept_ra_defrtr - BOOLEAN
1146 Learn default router in Router Advertisement.
1148 Functional default: enabled if accept_ra is enabled.
1149 disabled if accept_ra is disabled.
1151 accept_ra_pinfo - BOOLEAN
1152 Learn Prefix Information in Router Advertisement.
1154 Functional default: enabled if accept_ra is enabled.
1155 disabled if accept_ra is disabled.
1157 accept_ra_rt_info_max_plen - INTEGER
1158 Maximum prefix length of Route Information in RA.
1160 Route Information w/ prefix larger than or equal to this
1161 variable shall be ignored.
1163 Functional default: 0 if accept_ra_rtr_pref is enabled.
1164 -1 if accept_ra_rtr_pref is disabled.
1166 accept_ra_rtr_pref - BOOLEAN
1167 Accept Router Preference in RA.
1169 Functional default: enabled if accept_ra is enabled.
1170 disabled if accept_ra is disabled.
1172 accept_redirects - BOOLEAN
1175 Functional default: enabled if local forwarding is disabled.
1176 disabled if local forwarding is enabled.
1178 accept_source_route - INTEGER
1179 Accept source routing (routing extension header).
1181 >= 0: Accept only routing header type 2.
1182 < 0: Do not accept routing header.
1187 Autoconfigure addresses using Prefix Information in Router
1190 Functional default: enabled if accept_ra_pinfo is enabled.
1191 disabled if accept_ra_pinfo is disabled.
1193 dad_transmits - INTEGER
1194 The amount of Duplicate Address Detection probes to send.
1197 forwarding - INTEGER
1198 Configure interface-specific Host/Router behaviour.
1200 Note: It is recommended to have the same setting on all
1201 interfaces; mixed router/host scenarios are rather uncommon.
1203 Possible values are:
1204 0 Forwarding disabled
1205 1 Forwarding enabled
1209 By default, Host behaviour is assumed. This means:
1211 1. IsRouter flag is not set in Neighbour Advertisements.
1212 2. If accept_ra is TRUE (default), transmit Router
1214 3. If accept_ra is TRUE (default), accept Router
1215 Advertisements (and do autoconfiguration).
1216 4. If accept_redirects is TRUE (default), accept Redirects.
1220 If local forwarding is enabled, Router behaviour is assumed.
1221 This means exactly the reverse from the above:
1223 1. IsRouter flag is set in Neighbour Advertisements.
1224 2. Router Solicitations are not sent unless accept_ra is 2.
1225 3. Router Advertisements are ignored unless accept_ra is 2.
1226 4. Redirects are ignored.
1228 Default: 0 (disabled) if global forwarding is disabled (default),
1229 otherwise 1 (enabled).
1232 Default Hop Limit to set.
1236 Default Maximum Transfer Unit
1237 Default: 1280 (IPv6 required minimum)
1239 router_probe_interval - INTEGER
1240 Minimum interval (in seconds) between Router Probing described
1245 router_solicitation_delay - INTEGER
1246 Number of seconds to wait after interface is brought up
1247 before sending Router Solicitations.
1250 router_solicitation_interval - INTEGER
1251 Number of seconds to wait between Router Solicitations.
1254 router_solicitations - INTEGER
1255 Number of Router Solicitations to send until assuming no
1256 routers are present.
1259 use_tempaddr - INTEGER
1260 Preference for Privacy Extensions (RFC3041).
1261 <= 0 : disable Privacy Extensions
1262 == 1 : enable Privacy Extensions, but prefer public
1263 addresses over temporary addresses.
1264 > 1 : enable Privacy Extensions and prefer temporary
1265 addresses over public addresses.
1266 Default: 0 (for most devices)
1267 -1 (for point-to-point devices and loopback devices)
1269 temp_valid_lft - INTEGER
1270 valid lifetime (in seconds) for temporary addresses.
1271 Default: 604800 (7 days)
1273 temp_prefered_lft - INTEGER
1274 Preferred lifetime (in seconds) for temporary addresses.
1275 Default: 86400 (1 day)
1277 max_desync_factor - INTEGER
1278 Maximum value for DESYNC_FACTOR, which is a random value
1279 that ensures that clients don't synchronize with each
1280 other and generate new addresses at exactly the same time.
1281 value is in seconds.
1284 regen_max_retry - INTEGER
1285 Number of attempts before give up attempting to generate
1286 valid temporary addresses.
1289 max_addresses - INTEGER
1290 Maximum number of autoconfigured addresses per interface. Setting
1291 to zero disables the limitation. It is not recommended to set this
1292 value too large (or to zero) because it would be an easy way to
1293 crash the kernel by allowing too many addresses to be created.
1296 disable_ipv6 - BOOLEAN
1297 Disable IPv6 operation. If accept_dad is set to 2, this value
1298 will be dynamically set to TRUE if DAD fails for the link-local
1300 Default: FALSE (enable IPv6 operation)
1302 When this value is changed from 1 to 0 (IPv6 is being enabled),
1303 it will dynamically create a link-local address on the given
1304 interface and start Duplicate Address Detection, if necessary.
1306 When this value is changed from 0 to 1 (IPv6 is being disabled),
1307 it will dynamically delete all address on the given interface.
1309 accept_dad - INTEGER
1310 Whether to accept DAD (Duplicate Address Detection).
1312 1: Enable DAD (default)
1313 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1314 link-local address has been found.
1316 force_tllao - BOOLEAN
1317 Enable sending the target link-layer address option even when
1318 responding to a unicast neighbor solicitation.
1321 Quoting from RFC 2461, section 4.4, Target link-layer address:
1323 "The option MUST be included for multicast solicitations in order to
1324 avoid infinite Neighbor Solicitation "recursion" when the peer node
1325 does not have a cache entry to return a Neighbor Advertisements
1326 message. When responding to unicast solicitations, the option can be
1327 omitted since the sender of the solicitation has the correct link-
1328 layer address; otherwise it would not have be able to send the unicast
1329 solicitation in the first place. However, including the link-layer
1330 address in this case adds little overhead and eliminates a potential
1331 race condition where the sender deletes the cached link-layer address
1332 prior to receiving a response to a previous solicitation."
1336 Limit the maximal rates for sending ICMPv6 packets.
1337 0 to disable any limiting,
1338 otherwise the minimal space between responses in milliseconds.
1343 Pekka Savola <pekkas@netcore.fi>
1344 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1347 /proc/sys/net/bridge/* Variables:
1349 bridge-nf-call-arptables - BOOLEAN
1350 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1354 bridge-nf-call-iptables - BOOLEAN
1355 1 : pass bridged IPv4 traffic to iptables' chains.
1359 bridge-nf-call-ip6tables - BOOLEAN
1360 1 : pass bridged IPv6 traffic to ip6tables' chains.
1364 bridge-nf-filter-vlan-tagged - BOOLEAN
1365 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1369 bridge-nf-filter-pppoe-tagged - BOOLEAN
1370 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1374 bridge-nf-pass-vlan-input-dev - BOOLEAN
1375 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1376 interface on the bridge and set the netfilter input device to the vlan.
1377 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1378 target work with vlan-on-top-of-bridge interfaces. When no matching
1379 vlan interface is found, or this switch is off, the input device is
1380 set to the bridge interface.
1381 0: disable bridge netfilter vlan interface lookup.
1384 proc/sys/net/sctp/* Variables:
1386 addip_enable - BOOLEAN
1387 Enable or disable extension of Dynamic Address Reconfiguration
1388 (ADD-IP) functionality specified in RFC5061. This extension provides
1389 the ability to dynamically add and remove new addresses for the SCTP
1392 1: Enable extension.
1394 0: Disable extension.
1398 addip_noauth_enable - BOOLEAN
1399 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1400 authentication to protect the operations of adding or removing new
1401 addresses. This requirement is mandated so that unauthorized hosts
1402 would not be able to hijack associations. However, older
1403 implementations may not have implemented this requirement while
1404 allowing the ADD-IP extension. For reasons of interoperability,
1405 we provide this variable to control the enforcement of the
1406 authentication requirement.
1408 1: Allow ADD-IP extension to be used without authentication. This
1409 should only be set in a closed environment for interoperability
1410 with older implementations.
1412 0: Enforce the authentication requirement
1416 auth_enable - BOOLEAN
1417 Enable or disable Authenticated Chunks extension. This extension
1418 provides the ability to send and receive authenticated chunks and is
1419 required for secure operation of Dynamic Address Reconfiguration
1422 1: Enable this extension.
1423 0: Disable this extension.
1427 prsctp_enable - BOOLEAN
1428 Enable or disable the Partial Reliability extension (RFC3758) which
1429 is used to notify peers that a given DATA should no longer be expected.
1437 The limit of the number of new packets that can be initially sent. It
1438 controls how bursty the generated traffic can be.
1442 association_max_retrans - INTEGER
1443 Set the maximum number for retransmissions that an association can
1444 attempt deciding that the remote end is unreachable. If this value
1445 is exceeded, the association is terminated.
1449 max_init_retransmits - INTEGER
1450 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1451 that an association will attempt before declaring the destination
1452 unreachable and terminating.
1456 path_max_retrans - INTEGER
1457 The maximum number of retransmissions that will be attempted on a given
1458 path. Once this threshold is exceeded, the path is considered
1459 unreachable, and new traffic will use a different path when the
1460 association is multihomed.
1464 pf_retrans - INTEGER
1465 The number of retransmissions that will be attempted on a given path
1466 before traffic is redirected to an alternate transport (should one
1467 exist). Note this is distinct from path_max_retrans, as a path that
1468 passes the pf_retrans threshold can still be used. Its only
1469 deprioritized when a transmission path is selected by the stack. This
1470 setting is primarily used to enable fast failover mechanisms without
1471 having to reduce path_max_retrans to a very low value. See:
1472 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1473 for details. Note also that a value of pf_retrans > path_max_retrans
1474 disables this feature
1478 rto_initial - INTEGER
1479 The initial round trip timeout value in milliseconds that will be used
1480 in calculating round trip times. This is the initial time interval
1481 for retransmissions.
1486 The maximum value (in milliseconds) of the round trip timeout. This
1487 is the largest time interval that can elapse between retransmissions.
1492 The minimum value (in milliseconds) of the round trip timeout. This
1493 is the smallest time interval the can elapse between retransmissions.
1497 hb_interval - INTEGER
1498 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1499 are sent at the specified interval on idle paths to probe the state of
1500 a given path between 2 associations.
1504 sack_timeout - INTEGER
1505 The amount of time (in milliseconds) that the implementation will wait
1510 valid_cookie_life - INTEGER
1511 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1512 is used during association establishment.
1516 cookie_preserve_enable - BOOLEAN
1517 Enable or disable the ability to extend the lifetime of the SCTP cookie
1518 that is used during the establishment phase of SCTP association
1520 1: Enable cookie lifetime extension.
1525 cookie_hmac_alg - STRING
1526 Select the hmac algorithm used when generating the cookie value sent by
1527 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1532 Ability to assign md5 or sha1 as the selected alg is predicated on the
1533 configuarion of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1534 CONFIG_CRYPTO_SHA1).
1536 Default: Dependent on configuration. MD5 if available, else SHA1 if
1537 available, else none.
1539 rcvbuf_policy - INTEGER
1540 Determines if the receive buffer is attributed to the socket or to
1541 association. SCTP supports the capability to create multiple
1542 associations on a single socket. When using this capability, it is
1543 possible that a single stalled association that's buffering a lot
1544 of data may block other associations from delivering their data by
1545 consuming all of the receive buffer space. To work around this,
1546 the rcvbuf_policy could be set to attribute the receiver buffer space
1547 to each association instead of the socket. This prevents the described
1550 1: rcvbuf space is per association
1551 0: recbuf space is per socket
1555 sndbuf_policy - INTEGER
1556 Similar to rcvbuf_policy above, this applies to send buffer space.
1558 1: Send buffer is tracked per association
1559 0: Send buffer is tracked per socket.
1563 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1564 Number of pages allowed for queueing by all SCTP sockets.
1566 min: Below this number of pages SCTP is not bothered about its
1567 memory appetite. When amount of memory allocated by SCTP exceeds
1568 this number, SCTP starts to moderate memory usage.
1570 pressure: This value was introduced to follow format of tcp_mem.
1572 max: Number of pages allowed for queueing by all SCTP sockets.
1574 Default is calculated at boot time from amount of available memory.
1576 sctp_rmem - vector of 3 INTEGERs: min, default, max
1577 Only the first value ("min") is used, "default" and "max" are
1580 min: Minimal size of receive buffer used by SCTP socket.
1581 It is guaranteed to each SCTP socket (but not association) even
1582 under moderate memory pressure.
1586 sctp_wmem - vector of 3 INTEGERs: min, default, max
1587 Currently this tunable has no effect.
1589 addr_scope_policy - INTEGER
1590 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1592 0 - Disable IPv4 address scoping
1593 1 - Enable IPv4 address scoping
1594 2 - Follow draft but allow IPv4 private addresses
1595 3 - Follow draft but allow IPv4 link local addresses
1600 /proc/sys/net/core/*
1601 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1604 /proc/sys/net/unix/*
1605 max_dgram_qlen - INTEGER
1606 The maximum length of dgram socket receive queue
1613 /proc/sys/net/irda/*
1614 fast_poll_increase FIXME
1615 warn_noreply_time FIXME
1616 discovery_slots FIXME
1619 discovery_timeout FIXME
1620 lap_keepalive_time FIXME
1621 max_noreply_time FIXME
1622 max_tx_data_size FIXME
1624 min_tx_turn_time FIXME