blob: d71afa8bd82810d46e1e91f6275d8e2de3260176 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/proc/sys/net/ipv4/* Variables:
2
3ip_forward - BOOLEAN
4 0 - disabled (default)
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00005 not 0 - enabled
Linus Torvalds1da177e2005-04-16 15:20:36 -07006
7 Forward Packets between interfaces.
8
9 This variable is special, its change resets all configuration
10 parameters to their default state (RFC1122 for hosts, RFC1812
11 for routers)
12
13ip_default_ttl - INTEGER
Eric Dumazetcc6f02d2010-12-13 12:50:49 -080014 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)
Linus Torvalds1da177e2005-04-16 15:20:36 -070017
Hannes Frederic Sowacd174e62013-12-14 05:13:45 +010018ip_no_pmtu_disc - INTEGER
19 Disable Path MTU Discovery. If enabled in mode 1 and a
Hannes Frederic Sowa188b04d2013-12-14 04:42:13 +010020 fragmentation-required ICMP is received, the PMTU to this
21 destination will be set to min_pmtu (see below). You will need
22 to raise min_pmtu to the smallest interface MTU on your system
23 manually if you want to avoid locally generated fragments.
Hannes Frederic Sowacd174e62013-12-14 05:13:45 +010024
25 In mode 2 incoming Path MTU Discovery messages will be
26 discarded. Outgoing frames are handled the same as in mode 1,
27 implicitly setting IP_PMTUDISC_DONT on every created socket.
28
29 Possible values: 0-2
Hannes Frederic Sowa188b04d2013-12-14 04:42:13 +010030 Default: FALSE
Linus Torvalds1da177e2005-04-16 15:20:36 -070031
32min_pmtu - INTEGER
Eric Dumazet20db93c2011-11-08 14:21:44 -050033 default 552 - minimum discovered Path MTU
Linus Torvalds1da177e2005-04-16 15:20:36 -070034
Ben Greearcbaf0872010-11-08 09:13:48 +000035route/max_size - INTEGER
36 Maximum number of routes allowed in the kernel. Increase
37 this when using large numbers of interfaces and/or routes.
38
YOSHIFUJI Hideaki / 吉藤英明27246802013-01-22 05:20:05 +000039neigh/default/gc_thresh1 - INTEGER
40 Minimum number of entries to keep. Garbage collector will not
41 purge entries if there are fewer than this number.
Li RongQingb66c66d2013-03-14 22:49:47 +000042 Default: 128
YOSHIFUJI Hideaki / 吉藤英明27246802013-01-22 05:20:05 +000043
Ben Greearcbaf0872010-11-08 09:13:48 +000044neigh/default/gc_thresh3 - INTEGER
45 Maximum number of neighbor entries allowed. Increase this
46 when using large numbers of interfaces and when communicating
47 with large numbers of directly-connected peers.
Shan Weicc868022012-12-04 18:50:35 +000048 Default: 1024
Ben Greearcbaf0872010-11-08 09:13:48 +000049
Eric Dumazet8b5c1712011-11-09 12:07:14 +000050neigh/default/unres_qlen_bytes - INTEGER
51 The maximum number of bytes which may be used by packets
52 queued for each unresolved address by other network layers.
53 (added in linux 3.3)
stephen hemminger3b09adc2013-01-03 07:50:29 +000054 Setting negative value is meaningless and will return error.
Shan Weicc868022012-12-04 18:50:35 +000055 Default: 65536 Bytes(64KB)
Eric Dumazet8b5c1712011-11-09 12:07:14 +000056
57neigh/default/unres_qlen - INTEGER
58 The maximum number of packets which may be queued for each
59 unresolved address by other network layers.
60 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
Shan Weicc868022012-12-04 18:50:35 +000061 Prior to linux 3.3, the default value is 3 which may cause
Shan Wei5d248c42012-12-06 16:27:51 +000062 unexpected packet loss. The current default value is calculated
Shan Weicc868022012-12-04 18:50:35 +000063 according to default value of unres_qlen_bytes and true size of
64 packet.
65 Default: 31
Eric Dumazet8b5c1712011-11-09 12:07:14 +000066
Linus Torvalds1da177e2005-04-16 15:20:36 -070067mtu_expires - INTEGER
68 Time, in seconds, that cached PMTU information is kept.
69
70min_adv_mss - INTEGER
71 The advertised MSS depends on the first hop route MTU, but will
72 never be lower than this setting.
73
74IP Fragmentation:
75
76ipfrag_high_thresh - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +000077 Maximum memory used to reassemble IP fragments. When
Linus Torvalds1da177e2005-04-16 15:20:36 -070078 ipfrag_high_thresh bytes of memory is allocated for this purpose,
79 the fragment handler will toss packets until ipfrag_low_thresh
80 is reached.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +000081
Linus Torvalds1da177e2005-04-16 15:20:36 -070082ipfrag_low_thresh - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +000083 See ipfrag_high_thresh
Linus Torvalds1da177e2005-04-16 15:20:36 -070084
85ipfrag_time - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +000086 Time in seconds to keep an IP fragment in memory.
Linus Torvalds1da177e2005-04-16 15:20:36 -070087
88ipfrag_secret_interval - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +000089 Regeneration interval (in seconds) of the hash secret (or lifetime
Linus Torvalds1da177e2005-04-16 15:20:36 -070090 for the hash secret) for IP fragments.
91 Default: 600
92
Herbert Xu89cee8b2005-12-13 23:14:27 -080093ipfrag_max_dist - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +000094 ipfrag_max_dist is a non-negative integer value which defines the
95 maximum "disorder" which is allowed among fragments which share a
96 common IP source address. Note that reordering of packets is
97 not unusual, but if a large number of fragments arrive from a source
98 IP address while a particular fragment queue remains incomplete, it
99 probably indicates that one or more fragments belonging to that queue
100 have been lost. When ipfrag_max_dist is positive, an additional check
101 is done on fragments before they are added to a reassembly queue - if
102 ipfrag_max_dist (or more) fragments have arrived from a particular IP
103 address between additions to any IP fragment queue using that source
104 address, it's presumed that one or more fragments in the queue are
105 lost. The existing fragment queue will be dropped, and a new one
Herbert Xu89cee8b2005-12-13 23:14:27 -0800106 started. An ipfrag_max_dist value of zero disables this check.
107
108 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
109 result in unnecessarily dropping fragment queues when normal
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000110 reordering of packets occurs, which could lead to poor application
111 performance. Using a very large value, e.g. 50000, increases the
112 likelihood of incorrectly reassembling IP fragments that originate
Herbert Xu89cee8b2005-12-13 23:14:27 -0800113 from different IP datagrams, which could result in data corruption.
114 Default: 64
115
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116INET peer storage:
117
118inet_peer_threshold - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000119 The approximate size of the storage. Starting from this threshold
Linus Torvalds1da177e2005-04-16 15:20:36 -0700120 entries will be thrown aggressively. This threshold also determines
121 entries' time-to-live and time intervals between garbage collection
122 passes. More entries, less time-to-live, less GC interval.
123
124inet_peer_minttl - INTEGER
125 Minimum time-to-live of entries. Should be enough to cover fragment
126 time-to-live on the reassembling side. This minimum time-to-live is
127 guaranteed if the pool size is less than inet_peer_threshold.
Stephen Hemminger77a538d2008-07-01 17:22:48 -0700128 Measured in seconds.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700129
130inet_peer_maxttl - INTEGER
131 Maximum time-to-live of entries. Unused entries will expire after
132 this period of time if there is no memory pressure on the pool (i.e.
133 when the number of entries in the pool is very small).
Stephen Hemminger77a538d2008-07-01 17:22:48 -0700134 Measured in seconds.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700135
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000136TCP variables:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700137
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800138somaxconn - INTEGER
139 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
140 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
141 for TCP sockets.
142
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800143tcp_abort_on_overflow - BOOLEAN
144 If listening service is too slow to accept new connections,
145 reset them. Default state is FALSE. It means that if overflow
146 occurred due to a burst, connection will recover. Enable this
147 option _only_ if you are really sure that listening daemon
148 cannot be tuned to accept connections faster. Enabling this
149 option can harm clients of your server.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700150
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800151tcp_adv_win_scale - INTEGER
152 Count buffering overhead as bytes/2^tcp_adv_win_scale
153 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
154 if it is <= 0.
Alexey Dobriyan0147fc02010-11-22 12:54:21 +0000155 Possible values are [-31, 31], inclusive.
Eric Dumazetb49960a2012-05-02 02:28:41 +0000156 Default: 1
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800157
158tcp_allowed_congestion_control - STRING
159 Show/set the congestion control choices available to non-privileged
160 processes. The list is a subset of those listed in
161 tcp_available_congestion_control.
162 Default is "reno" and the default setting (tcp_congestion_control).
163
164tcp_app_win - INTEGER
165 Reserve max(window/2^tcp_app_win, mss) of window for application
166 buffer. Value 0 is special, it means that nothing is reserved.
167 Default: 31
168
Eric Dumazetf54b3112013-12-05 22:36:05 -0800169tcp_autocorking - BOOLEAN
170 Enable TCP auto corking :
171 When applications do consecutive small write()/sendmsg() system calls,
172 we try to coalesce these small writes as much as possible, to lower
173 total amount of sent packets. This is done if at least one prior
174 packet for the flow is waiting in Qdisc queues or device transmit
175 queue. Applications can still use TCP_CORK for optimal behavior
176 when they know how/when to uncork their sockets.
177 Default : 1
178
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800179tcp_available_congestion_control - STRING
180 Shows the available congestion control choices that are registered.
181 More congestion control algorithms may be available as modules,
182 but not loaded.
183
John Heffner71599cd2007-02-27 10:03:56 -0800184tcp_base_mss - INTEGER
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700185 The initial value of search_low to be used by the packetization layer
186 Path MTU discovery (MTU probing). If MTU probing is enabled,
187 this is the initial MSS used by the connection.
John Heffner71599cd2007-02-27 10:03:56 -0800188
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800189tcp_congestion_control - STRING
190 Set the congestion control algorithm to be used for new
191 connections. The algorithm "reno" is always available, but
192 additional choices may be available based on kernel configuration.
193 Default is set as part of kernel configuration.
Eric Dumazetd8a6e652011-11-30 01:02:41 +0000194 For passive connections, the listener congestion control choice
195 is inherited.
196 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800197
198tcp_dsack - BOOLEAN
199 Allows TCP to send "duplicate" SACKs.
200
Yuchung Chengeed530b2012-05-02 13:30:03 +0000201tcp_early_retrans - INTEGER
202 Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold
203 for triggering fast retransmit when the amount of outstanding data is
204 small and when no previously unsent data can be transmitted (such
Nandita Dukkipati6ba8a3b2013-03-11 10:00:43 +0000205 that limited transmit could be used). Also controls the use of
Masanari Iida3dd17ed2013-05-24 07:05:59 +0000206 Tail loss probe (TLP) that converts RTOs occurring due to tail
Nandita Dukkipati6ba8a3b2013-03-11 10:00:43 +0000207 losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01).
Yuchung Chengeed530b2012-05-02 13:30:03 +0000208 Possible values:
209 0 disables ER
210 1 enables ER
211 2 enables ER but delays fast recovery and fast retransmit
212 by a fourth of RTT. This mitigates connection falsely
213 recovers when network has a small degree of reordering
214 (less than 3 packets).
Nandita Dukkipati6ba8a3b2013-03-11 10:00:43 +0000215 3 enables delayed ER and TLP.
216 4 enables TLP only.
217 Default: 3
Yuchung Chengeed530b2012-05-02 13:30:03 +0000218
Peter Chubb34a6ef32011-02-02 15:39:58 -0800219tcp_ecn - INTEGER
Rick Jones7e3a2dc2012-11-28 09:53:10 +0000220 Control use of Explicit Congestion Notification (ECN) by TCP.
221 ECN is used only when both ends of the TCP connection indicate
222 support for it. This feature is useful in avoiding losses due
223 to congestion by allowing supporting routers to signal
224 congestion before having to drop packets.
Ilpo Järvinen255cac92009-05-04 11:07:36 -0700225 Possible values are:
Rick Jones7e3a2dc2012-11-28 09:53:10 +0000226 0 Disable ECN. Neither initiate nor accept ECN.
Vijay Subramanian3d55b322013-01-09 12:21:30 +0000227 1 Enable ECN when requested by incoming connections and
228 also request ECN on outgoing connection attempts.
229 2 Enable ECN when requested by incoming connections
Rick Jones7e3a2dc2012-11-28 09:53:10 +0000230 but do not request ECN on outgoing connections.
Ilpo Järvinen255cac92009-05-04 11:07:36 -0700231 Default: 2
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800232
233tcp_fack - BOOLEAN
234 Enable FACK congestion avoidance and fast retransmission.
235 The value is not used, if tcp_sack is not enabled.
236
237tcp_fin_timeout - INTEGER
Rick Jonesd825da22012-12-10 11:33:00 +0000238 The length of time an orphaned (no longer referenced by any
239 application) connection will remain in the FIN_WAIT_2 state
240 before it is aborted at the local end. While a perfectly
241 valid "receive only" state for an un-orphaned connection, an
242 orphaned connection in FIN_WAIT_2 state could otherwise wait
243 forever for the remote to close its end of the connection.
244 Cf. tcp_max_orphans
245 Default: 60 seconds
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800246
Ilpo Järvinen89808062007-02-27 10:10:55 -0800247tcp_frto - INTEGER
Yuchung Chenge33099f2013-03-20 13:33:00 +0000248 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
Ilpo Järvinencd998892007-09-20 11:35:26 -0700249 F-RTO is an enhanced recovery algorithm for TCP retransmission
Yuchung Chenge33099f2013-03-20 13:33:00 +0000250 timeouts. It is particularly beneficial in networks where the
251 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
252 modification. It does not require any support from the peer.
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700253
Yuchung Chenge33099f2013-03-20 13:33:00 +0000254 By default it's enabled with a non-zero value. 0 disables F-RTO.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700255
256tcp_keepalive_time - INTEGER
257 How often TCP sends out keepalive messages when keepalive is enabled.
258 Default: 2hours.
259
260tcp_keepalive_probes - INTEGER
261 How many keepalive probes TCP sends out, until it decides that the
262 connection is broken. Default value: 9.
263
264tcp_keepalive_intvl - INTEGER
265 How frequently the probes are send out. Multiplied by
266 tcp_keepalive_probes it is time to kill not responding connection,
267 after probes started. Default value: 75sec i.e. connection
268 will be aborted after ~11 minutes of retries.
269
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800270tcp_low_latency - BOOLEAN
271 If set, the TCP stack makes decisions that prefer lower
272 latency as opposed to higher throughput. By default, this
273 option is not set meaning that higher throughput is preferred.
274 An example of an application where this default should be
275 changed would be a Beowulf compute cluster.
276 Default: 0
Linus Torvalds1da177e2005-04-16 15:20:36 -0700277
278tcp_max_orphans - INTEGER
279 Maximal number of TCP sockets not attached to any user file handle,
280 held by system. If this number is exceeded orphaned connections are
281 reset immediately and warning is printed. This limit exists
282 only to prevent simple DoS attacks, you _must_ not rely on this
283 or lower the limit artificially, but rather increase it
284 (probably, after increasing installed memory),
285 if network conditions require more than default value,
286 and tune network services to linger and kill such states
287 more aggressively. Let me to remind again: each orphan eats
288 up to ~64K of unswappable memory.
289
Linus Torvalds1da177e2005-04-16 15:20:36 -0700290tcp_max_syn_backlog - INTEGER
Peter Pan(潘卫平)99b53bd2011-12-05 21:39:41 +0000291 Maximal number of remembered connection requests, which have not
292 received an acknowledgment from connecting client.
293 The minimal value is 128 for low memory machines, and it will
294 increase in proportion to the memory of machine.
295 If server suffers from overload, try increasing this number.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700296
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800297tcp_max_tw_buckets - INTEGER
298 Maximal number of timewait sockets held by system simultaneously.
299 If this number is exceeded time-wait socket is immediately destroyed
300 and warning is printed. This limit exists only to prevent
301 simple DoS attacks, you _must_ not lower the limit artificially,
302 but rather increase it (probably, after increasing installed memory),
303 if network conditions require more than default value.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700304
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800305tcp_mem - vector of 3 INTEGERs: min, pressure, max
306 min: below this number of pages TCP is not bothered about its
307 memory appetite.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700308
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800309 pressure: when amount of memory allocated by TCP exceeds this number
310 of pages, TCP moderates its memory consumption and enters memory
311 pressure mode, which is exited when memory consumption falls
312 under "min".
Linus Torvalds1da177e2005-04-16 15:20:36 -0700313
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800314 max: number of pages allowed for queueing by all TCP sockets.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700315
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800316 Defaults are calculated at boot time from amount of available
317 memory.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700318
John Heffner71599cd2007-02-27 10:03:56 -0800319tcp_moderate_rcvbuf - BOOLEAN
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700320 If set, TCP performs receive buffer auto-tuning, attempting to
John Heffner71599cd2007-02-27 10:03:56 -0800321 automatically size the buffer (no greater than tcp_rmem[2]) to
322 match the size required by the path for full throughput. Enabled by
323 default.
324
325tcp_mtu_probing - INTEGER
326 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
327 values:
328 0 - Disabled
329 1 - Disabled by default, enabled when an ICMP black hole detected
330 2 - Always enabled, use initial MSS of tcp_base_mss.
331
332tcp_no_metrics_save - BOOLEAN
333 By default, TCP saves various connection metrics in the route cache
334 when the connection closes, so that connections established in the
335 near future can use these to set initial conditions. Usually, this
336 increases overall performance, but may sometimes cause performance
Simon Arlott0f035b82007-10-20 01:30:25 +0200337 degradation. If set, TCP will not cache metrics on closing
John Heffner71599cd2007-02-27 10:03:56 -0800338 connections.
339
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800340tcp_orphan_retries - INTEGER
Damian Lukowski5d789222009-09-01 10:24:04 +0000341 This value influences the timeout of a locally closed TCP connection,
342 when RTO retransmissions remain unacknowledged.
343 See tcp_retries2 for more details.
344
David S. Miller06b8fc52011-07-08 09:31:31 -0700345 The default value is 8.
Damian Lukowski5d789222009-09-01 10:24:04 +0000346 If your machine is a loaded WEB server,
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800347 you should think about lowering this value, such sockets
348 may consume significant resources. Cf. tcp_max_orphans.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700349
350tcp_reordering - INTEGER
351 Maximal reordering of packets in a TCP stream.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000352 Default: 3
Linus Torvalds1da177e2005-04-16 15:20:36 -0700353
354tcp_retrans_collapse - BOOLEAN
355 Bug-to-bug compatibility with some broken printers.
356 On retransmit try to send bigger packets to work around bugs in
357 certain TCP stacks.
358
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800359tcp_retries1 - INTEGER
Damian Lukowski5d789222009-09-01 10:24:04 +0000360 This value influences the time, after which TCP decides, that
361 something is wrong due to unacknowledged RTO retransmissions,
362 and reports this suspicion to the network layer.
363 See tcp_retries2 for more details.
364
365 RFC 1122 recommends at least 3 retransmissions, which is the
366 default.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700367
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800368tcp_retries2 - INTEGER
Damian Lukowski5d789222009-09-01 10:24:04 +0000369 This value influences the timeout of an alive TCP connection,
370 when RTO retransmissions remain unacknowledged.
371 Given a value of N, a hypothetical TCP connection following
372 exponential backoff with an initial RTO of TCP_RTO_MIN would
373 retransmit N times before killing the connection at the (N+1)th RTO.
374
375 The default value of 15 yields a hypothetical timeout of 924.6
376 seconds and is a lower bound for the effective timeout.
377 TCP will effectively time out at the first RTO which exceeds the
378 hypothetical timeout.
379
380 RFC 1122 recommends at least 100 seconds for the timeout,
381 which corresponds to a value of at least 8.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700382
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800383tcp_rfc1337 - BOOLEAN
384 If set, the TCP stack behaves conforming to RFC1337. If unset,
385 we are not conforming to RFC, but prevent TCP TIME_WAIT
386 assassination.
387 Default: 0
Linus Torvalds1da177e2005-04-16 15:20:36 -0700388
389tcp_rmem - vector of 3 INTEGERs: min, default, max
390 min: Minimal size of receive buffer used by TCP sockets.
391 It is guaranteed to each TCP socket, even under moderate memory
392 pressure.
Max Matveev6539fef2011-06-21 21:18:13 +0000393 Default: 1 page
Linus Torvalds1da177e2005-04-16 15:20:36 -0700394
J. Bruce Fields53025f52008-07-10 16:47:41 -0700395 default: initial size of receive buffer used by TCP sockets.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700396 This value overrides net.core.rmem_default used by other protocols.
397 Default: 87380 bytes. This value results in window of 65535 with
398 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
399 less for default tcp_app_win. See below about these variables.
400
401 max: maximal size of receive buffer allowed for automatically
402 selected receiver buffers for TCP socket. This value does not override
J. Bruce Fields53025f52008-07-10 16:47:41 -0700403 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
404 automatic tuning of that socket's receive buffer size, in which
405 case this value is ignored.
Eric Dumazetb49960a2012-05-02 02:28:41 +0000406 Default: between 87380B and 6MB, depending on RAM size.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700407
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800408tcp_sack - BOOLEAN
409 Enable select acknowledgments (SACKS).
Rick Jones15d99e02006-03-20 22:40:29 -0800410
David S. Miller35089bb2006-06-13 22:33:04 -0700411tcp_slow_start_after_idle - BOOLEAN
412 If set, provide RFC2861 behavior and time out the congestion
413 window after an idle period. An idle period is defined at
414 the current RTO. If unset, the congestion window will not
415 be timed out after an idle period.
416 Default: 1
417
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800418tcp_stdurg - BOOLEAN
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700419 Use the Host requirements interpretation of the TCP urgent pointer field.
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800420 Most hosts use the older BSD interpretation, so if you turn this on
421 Linux might not communicate correctly with them.
422 Default: FALSE
423
424tcp_synack_retries - INTEGER
425 Number of times SYNACKs for a passive TCP connection attempt will
426 be retransmitted. Should not be higher than 255. Default value
Alex Bergmann6c9ff972012-08-31 02:48:31 +0000427 is 5, which corresponds to 31seconds till the last retransmission
428 with the current initial RTO of 1second. With this the final timeout
429 for a passive TCP connection will happen after 63seconds.
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800430
431tcp_syncookies - BOOLEAN
Shan Weia3c910d2013-06-21 15:18:32 +0800432 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800433 Send out syncookies when the syn backlog queue of a socket
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700434 overflows. This is to prevent against the common 'SYN flood attack'
Shan Weia3c910d2013-06-21 15:18:32 +0800435 Default: 1
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800436
437 Note, that syncookies is fallback facility.
438 It MUST NOT be used to help highly loaded servers to stand
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700439 against legal connection rate. If you see SYN flood warnings
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800440 in your logs, but investigation shows that they occur
441 because of overload with legal connections, you should tune
442 another parameters until this warning disappear.
443 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
444
445 syncookies seriously violate TCP protocol, do not allow
446 to use TCP extensions, can result in serious degradation
447 of some services (f.e. SMTP relaying), visible not by you,
448 but your clients and relays, contacting you. While you see
Stephen Hemminger4edc2f32008-07-10 16:50:26 -0700449 SYN flood warnings in logs not being really flooded, your server
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800450 is seriously misconfigured.
451
Hannes Frederic Sowa5ad37d52013-07-26 17:43:23 +0200452 If you want to test which effects syncookies have to your
453 network connections you can set this knob to 2 to enable
454 unconditionally generation of syncookies.
455
Yuchung Chengcf60af02012-07-19 06:43:09 +0000456tcp_fastopen - INTEGER
457 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
458 in the opening SYN packet. To use this feature, the client application
Jerry Chu10467162012-08-31 12:29:11 +0000459 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
460 connect() to perform a TCP handshake automatically.
Yuchung Chengcf60af02012-07-19 06:43:09 +0000461
Jerry Chu10467162012-08-31 12:29:11 +0000462 The values (bitmap) are
Yuchung Cheng0d41cca2013-10-31 09:19:32 -0700463 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN.
Jerry Chu10467162012-08-31 12:29:11 +0000464 2: Enables TCP Fast Open on the server side, i.e., allowing data in
465 a SYN packet to be accepted and passed to the application before
466 3-way hand shake finishes.
467 4: Send data in the opening SYN regardless of cookie availability and
468 without a cookie option.
469 0x100: Accept SYN data w/o validating the cookie.
470 0x200: Accept data-in-SYN w/o any cookie option present.
471 0x400/0x800: Enable Fast Open on all listeners regardless of the
472 TCP_FASTOPEN socket option. The two different flags designate two
473 different ways of setting max_qlen without the TCP_FASTOPEN socket
474 option.
Yuchung Chengcf60af02012-07-19 06:43:09 +0000475
Yuchung Cheng0d41cca2013-10-31 09:19:32 -0700476 Default: 1
Yuchung Chengcf60af02012-07-19 06:43:09 +0000477
Jerry Chu10467162012-08-31 12:29:11 +0000478 Note that the client & server side Fast Open flags (1 and 2
479 respectively) must be also enabled before the rest of flags can take
480 effect.
481
482 See include/net/tcp.h and the code for more details.
483
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800484tcp_syn_retries - INTEGER
485 Number of times initial SYNs for an active TCP connection attempt
486 will be retransmitted. Should not be higher than 255. Default value
stephen hemminger3b09adc2013-01-03 07:50:29 +0000487 is 6, which corresponds to 63seconds till the last retransmission
Alex Bergmann6c9ff972012-08-31 02:48:31 +0000488 with the current initial RTO of 1second. With this the final timeout
489 for an active TCP connection attempt will happen after 127seconds.
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800490
491tcp_timestamps - BOOLEAN
492 Enable timestamps as defined in RFC1323.
493
Eric Dumazet95bd09e2013-08-27 05:46:32 -0700494tcp_min_tso_segs - INTEGER
495 Minimal number of segments per TSO frame.
496 Since linux-3.12, TCP does an automatic sizing of TSO frames,
497 depending on flow rate, instead of filling 64Kbytes packets.
498 For specific usages, it's possible to force TCP to build big
499 TSO frames. Note that TCP stack might split too big TSO packets
500 if available window is too small.
501 Default: 2
502
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800503tcp_tso_win_divisor - INTEGER
504 This allows control over what percentage of the congestion window
505 can be consumed by a single TSO frame.
506 The setting of this parameter is a choice between burstiness and
507 building larger TSO frames.
508 Default: 3
509
510tcp_tw_recycle - BOOLEAN
511 Enable fast recycling TIME-WAIT sockets. Default value is 0.
512 It should not be changed without advice/request of technical
513 experts.
514
515tcp_tw_reuse - BOOLEAN
516 Allow to reuse TIME-WAIT sockets for new connections when it is
517 safe from protocol viewpoint. Default value is 0.
518 It should not be changed without advice/request of technical
519 experts.
520
521tcp_window_scaling - BOOLEAN
522 Enable window scaling as defined in RFC1323.
523
524tcp_wmem - vector of 3 INTEGERs: min, default, max
J. Bruce Fields53025f52008-07-10 16:47:41 -0700525 min: Amount of memory reserved for send buffers for TCP sockets.
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800526 Each TCP socket has rights to use it due to fact of its birth.
Max Matveev6539fef2011-06-21 21:18:13 +0000527 Default: 1 page
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800528
J. Bruce Fields53025f52008-07-10 16:47:41 -0700529 default: initial size of send buffer used by TCP sockets. This
530 value overrides net.core.wmem_default used by other protocols.
531 It is usually lower than net.core.wmem_default.
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800532 Default: 16K
533
J. Bruce Fields53025f52008-07-10 16:47:41 -0700534 max: Maximal amount of memory allowed for automatically tuned
535 send buffers for TCP sockets. This value does not override
536 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
537 automatic tuning of that socket's send buffer size, in which case
538 this value is ignored.
539 Default: between 64K and 4MB, depending on RAM size.
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800540
Eric Dumazetc9bee3b72013-07-22 20:27:07 -0700541tcp_notsent_lowat - UNSIGNED INTEGER
542 A TCP socket can control the amount of unsent bytes in its write queue,
543 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
544 reports POLLOUT events if the amount of unsent bytes is below a per
545 socket value, and if the write queue is not full. sendmsg() will
546 also not add new buffers if the limit is hit.
547
548 This global variable controls the amount of unsent data for
549 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
550 to the global variable has immediate effect.
551
552 Default: UINT_MAX (0xFFFFFFFF)
553
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800554tcp_workaround_signed_windows - BOOLEAN
555 If set, assume no receipt of a window scaling option means the
556 remote TCP is broken and treats the window as a signed quantity.
557 If unset, assume the remote TCP is not broken even if we do
558 not receive a window scaling option from them.
559 Default: 0
560
Chris Leech72d0b7a2007-03-08 09:57:35 -0800561tcp_dma_copybreak - INTEGER
562 Lower limit, in bytes, of the size of socket reads that will be
563 offloaded to a DMA copy engine, if one is present in the system
564 and CONFIG_NET_DMA is enabled.
565 Default: 4096
566
Andreas Petlund36e31b0a2010-02-18 02:47:01 +0000567tcp_thin_linear_timeouts - BOOLEAN
568 Enable dynamic triggering of linear timeouts for thin streams.
569 If set, a check is performed upon retransmission by timeout to
570 determine if the stream is thin (less than 4 packets in flight).
571 As long as the stream is found to be thin, up to 6 linear
572 timeouts may be performed before exponential backoff mode is
573 initiated. This improves retransmission latency for
574 non-aggressive thin streams, often found to be time-dependent.
575 For more information on thin streams, see
576 Documentation/networking/tcp-thin.txt
577 Default: 0
578
Andreas Petlund7e380172010-02-18 04:48:19 +0000579tcp_thin_dupack - BOOLEAN
580 Enable dynamic triggering of retransmissions after one dupACK
581 for thin streams. If set, a check is performed upon reception
582 of a dupACK to determine if the stream is thin (less than 4
583 packets in flight). As long as the stream is found to be thin,
584 data is retransmitted on the first received dupACK. This
585 improves retransmission latency for non-aggressive thin
586 streams, often found to be time-dependent.
587 For more information on thin streams, see
588 Documentation/networking/tcp-thin.txt
589 Default: 0
590
Eric Dumazet46d3cea2012-07-11 05:50:31 +0000591tcp_limit_output_bytes - INTEGER
592 Controls TCP Small Queue limit per tcp socket.
593 TCP bulk sender tends to increase packets in flight until it
594 gets losses notifications. With SNDBUF autotuning, this can
595 result in a large amount of packets queued in qdisc/device
596 on the local machine, hurting latency of other flows, for
597 typical pfifo_fast qdiscs.
598 tcp_limit_output_bytes limits the number of bytes on qdisc
599 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
Eric Dumazet46d3cea2012-07-11 05:50:31 +0000600 Default: 131072
601
Eric Dumazet282f23c2012-07-17 10:13:05 +0200602tcp_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)
605 Default: 100
606
Hideo Aoki95766ff2007-12-31 00:29:24 -0800607UDP variables:
608
609udp_mem - vector of 3 INTEGERs: min, pressure, max
610 Number of pages allowed for queueing by all UDP sockets.
611
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.
615
616 pressure: This value was introduced to follow format of tcp_mem.
617
618 max: Number of pages allowed for queueing by all UDP sockets.
619
620 Default is calculated at boot time from amount of available memory.
621
622udp_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.
Max Matveev6539fef2011-06-21 21:18:13 +0000626 Default: 1 page
Hideo Aoki95766ff2007-12-31 00:29:24 -0800627
628udp_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.
Max Matveev6539fef2011-06-21 21:18:13 +0000632 Default: 1 page
Hideo Aoki95766ff2007-12-31 00:29:24 -0800633
Paul Moore8802f612006-08-03 16:45:49 -0700634CIPSOv4 Variables:
635
636cipso_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".
642 Default: 1
643
644cipso_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.
651 Default: 10
652
653cipso_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.
658 Default: 0
659
660cipso_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.
667 Default: 0
668
Linus Torvalds1da177e2005-04-16 15:20:36 -0700669IP Variables:
670
671ip_local_port_range - 2 INTEGERS
672 Defines the local port range that is used by TCP and UDP to
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000673 choose the local port. The first number is the first, the
Fernando Luis Vazquez Cao5d6bd862012-04-03 08:41:40 +0000674 second the last local port number. The default values are
675 32768 and 61000 respectively.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700676
Amerigo Wange3826f12010-05-05 00:27:06 +0000677ip_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.
682
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
687 input.
688
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
692 assignments.
693
694 You can reserve ports which are not in the current
695 ip_local_port_range, e.g.:
696
697 $ cat /proc/sys/net/ipv4/ip_local_port_range
698 32000 61000
699 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
700 8080,9148
701
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.
705
706 Default: Empty
707
Linus Torvalds1da177e2005-04-16 15:20:36 -0700708ip_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.
711 Default: 0
712
713ip_dynaddr - BOOLEAN
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
717 occurs.
718 Default: 0
719
Cong Wange3d73bc2013-06-11 18:54:39 +0800720ip_early_demux - BOOLEAN
721 Optimize input packet processing down to one demux for
722 certain kinds of local sockets. Currently we only do this
723 for established TCP sockets.
724
725 It may add an additional cost for pure routing workloads that
726 reduces overall throughput, in such case you should disable it.
727 Default: 1
728
Linus Torvalds1da177e2005-04-16 15:20:36 -0700729icmp_echo_ignore_all - BOOLEAN
David S. Miller7ce312462005-10-03 16:07:30 -0700730 If set non-zero, then the kernel will ignore all ICMP ECHO
731 requests sent to it.
732 Default: 0
733
Linus Torvalds1da177e2005-04-16 15:20:36 -0700734icmp_echo_ignore_broadcasts - BOOLEAN
David S. Miller7ce312462005-10-03 16:07:30 -0700735 If set non-zero, then the kernel will ignore all ICMP ECHO and
736 TIMESTAMP requests sent to it via broadcast/multicast.
737 Default: 1
Linus Torvalds1da177e2005-04-16 15:20:36 -0700738
739icmp_ratelimit - INTEGER
740 Limit the maximal rates for sending ICMP packets whose type matches
741 icmp_ratemask (see below) to specific targets.
Stephen Hemminger6dbf4bc2008-07-01 19:29:07 -0700742 0 to disable any limiting,
743 otherwise the minimal space between responses in milliseconds.
744 Default: 1000
Linus Torvalds1da177e2005-04-16 15:20:36 -0700745
746icmp_ratemask - INTEGER
747 Mask made of ICMP types for which rates are being limited.
748 Significant bits: IHGFEDCBA9876543210
749 Default mask: 0000001100000011000 (6168)
750
751 Bit definitions (see include/linux/icmp.h):
752 0 Echo Reply
753 3 Destination Unreachable *
754 4 Source Quench *
755 5 Redirect
756 8 Echo Request
757 B Time Exceeded *
758 C Parameter Problem *
759 D Timestamp Request
760 E Timestamp Reply
761 F Info Request
762 G Info Reply
763 H Address Mask Request
764 I Address Mask Reply
765
766 * These are rate limited by default (see default mask above)
767
768icmp_ignore_bogus_error_responses - BOOLEAN
769 Some routers violate RFC1122 by sending bogus responses to broadcast
770 frames. Such violations are normally logged via a kernel warning.
771 If this is set to TRUE, the kernel will not give such warnings, which
772 will avoid log file clutter.
Rami Rosene8b265e2013-06-07 20:16:19 +0000773 Default: 1
Linus Torvalds1da177e2005-04-16 15:20:36 -0700774
Horms95f7daf2006-02-02 17:02:25 -0800775icmp_errors_use_inbound_ifaddr - BOOLEAN
776
777 If zero, icmp error messages are sent with the primary address of
778 the exiting interface.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000779
Horms95f7daf2006-02-02 17:02:25 -0800780 If non-zero, the message will be sent with the primary address of
781 the interface that received the packet that caused the icmp error.
782 This is the behaviour network many administrators will expect from
783 a router. And it can make debugging complicated network layouts
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000784 much easier.
Horms95f7daf2006-02-02 17:02:25 -0800785
786 Note that if no primary address exists for the interface selected,
787 then the primary address of the first non-loopback interface that
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +0200788 has one will be used regardless of this setting.
Horms95f7daf2006-02-02 17:02:25 -0800789
790 Default: 0
791
Linus Torvalds1da177e2005-04-16 15:20:36 -0700792igmp_max_memberships - INTEGER
793 Change the maximum number of multicast groups we can subscribe to.
794 Default: 20
795
Jeremy Ederd67ef352010-11-15 05:41:31 +0000796 Theoretical maximum value is bounded by having to send a membership
797 report in a single datagram (i.e. the report can't span multiple
798 datagrams, or risk confusing the switch and leaving groups you don't
799 intend to).
Linus Torvalds1da177e2005-04-16 15:20:36 -0700800
Jeremy Ederd67ef352010-11-15 05:41:31 +0000801 The number of supported groups 'M' is bounded by the number of group
802 report entries you can fit into a single datagram of 65535 bytes.
803
804 M = 65536-sizeof (ip header)/(sizeof(Group record))
805
806 Group records are variable length, with a minimum of 12 bytes.
807 So net.ipv4.igmp_max_memberships should not be set higher than:
808
809 (65536-24) / 12 = 5459
810
811 The value 5459 assumes no IP header options, so in practice
812 this number may be lower.
813
814 conf/interface/* changes special settings per interface (where
815 "interface" is the name of your network interface)
816
817 conf/all/* is special, changes the settings for all interfaces
Linus Torvalds1da177e2005-04-16 15:20:36 -0700818
819log_martians - BOOLEAN
820 Log packets with impossible addresses to kernel log.
821 log_martians for the interface will be enabled if at least one of
822 conf/{all,interface}/log_martians is set to TRUE,
823 it will be disabled otherwise
824
825accept_redirects - BOOLEAN
826 Accept ICMP redirect messages.
827 accept_redirects for the interface will be enabled if:
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000828 - both conf/{all,interface}/accept_redirects are TRUE in the case
829 forwarding for the interface is enabled
Linus Torvalds1da177e2005-04-16 15:20:36 -0700830 or
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000831 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
832 case forwarding for the interface is disabled
Linus Torvalds1da177e2005-04-16 15:20:36 -0700833 accept_redirects for the interface will be disabled otherwise
834 default TRUE (host)
835 FALSE (router)
836
837forwarding - BOOLEAN
838 Enable IP forwarding on this interface.
839
840mc_forwarding - BOOLEAN
841 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
842 and a multicast routing daemon is required.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000843 conf/all/mc_forwarding must also be set to TRUE to enable multicast
844 routing for the interface
Linus Torvalds1da177e2005-04-16 15:20:36 -0700845
846medium_id - INTEGER
847 Integer value used to differentiate the devices by the medium they
848 are attached to. Two devices can have different id values when
849 the broadcast packets are received only on one of them.
850 The default value 0 means that the device is the only interface
851 to its medium, value of -1 means that medium is not known.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000852
Linus Torvalds1da177e2005-04-16 15:20:36 -0700853 Currently, it is used to change the proxy_arp behavior:
854 the proxy_arp feature is enabled for packets forwarded between
855 two devices attached to different media.
856
857proxy_arp - BOOLEAN
858 Do proxy arp.
859 proxy_arp for the interface will be enabled if at least one of
860 conf/{all,interface}/proxy_arp is set to TRUE,
861 it will be disabled otherwise
862
Jesper Dangaard Brouer65324142010-01-05 05:50:47 +0000863proxy_arp_pvlan - BOOLEAN
864 Private VLAN proxy arp.
865 Basically allow proxy arp replies back to the same interface
866 (from which the ARP request/solicitation was received).
867
868 This is done to support (ethernet) switch features, like RFC
869 3069, where the individual ports are NOT allowed to
870 communicate with each other, but they are allowed to talk to
871 the upstream router. As described in RFC 3069, it is possible
872 to allow these hosts to communicate through the upstream
873 router by proxy_arp'ing. Don't need to be used together with
874 proxy_arp.
875
876 This technology is known by different names:
877 In RFC 3069 it is called VLAN Aggregation.
878 Cisco and Allied Telesyn call it Private VLAN.
879 Hewlett-Packard call it Source-Port filtering or port-isolation.
880 Ericsson call it MAC-Forced Forwarding (RFC Draft).
881
Linus Torvalds1da177e2005-04-16 15:20:36 -0700882shared_media - BOOLEAN
883 Send(router) or accept(host) RFC1620 shared media redirects.
884 Overrides ip_secure_redirects.
885 shared_media for the interface will be enabled if at least one of
886 conf/{all,interface}/shared_media is set to TRUE,
887 it will be disabled otherwise
888 default TRUE
889
890secure_redirects - BOOLEAN
891 Accept ICMP redirect messages only for gateways,
892 listed in default gateway list.
893 secure_redirects for the interface will be enabled if at least one of
894 conf/{all,interface}/secure_redirects is set to TRUE,
895 it will be disabled otherwise
896 default TRUE
897
898send_redirects - BOOLEAN
899 Send redirects, if router.
900 send_redirects for the interface will be enabled if at least one of
901 conf/{all,interface}/send_redirects is set to TRUE,
902 it will be disabled otherwise
903 Default: TRUE
904
905bootp_relay - BOOLEAN
906 Accept packets with source address 0.b.c.d destined
907 not to this host as local ones. It is supposed, that
908 BOOTP relay daemon will catch and forward such packets.
909 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
910 for the interface
911 default FALSE
912 Not Implemented Yet.
913
914accept_source_route - BOOLEAN
915 Accept packets with SRR option.
916 conf/all/accept_source_route must also be set to TRUE to accept packets
917 with SRR option on the interface
918 default TRUE (router)
919 FALSE (host)
920
Patrick McHardy8153a102009-12-03 01:25:58 +0000921accept_local - BOOLEAN
David S. Millerc801e3c2012-06-30 22:39:27 -0700922 Accept packets with local source addresses. In combination
923 with suitable routing, this can be used to direct packets
924 between two local interfaces over the wire and have them
925 accepted properly.
926
927 rp_filter must be set to a non-zero value in order for
928 accept_local to have an effect.
929
Patrick McHardy8153a102009-12-03 01:25:58 +0000930 default FALSE
931
Thomas Grafd0daebc32012-06-12 00:44:01 +0000932route_localnet - BOOLEAN
933 Do not consider loopback addresses as martian source or destination
934 while routing. This enables the use of 127/8 for local routing purposes.
935 default FALSE
936
Stephen Hemmingerc1cf8422009-02-20 08:25:36 +0000937rp_filter - INTEGER
Linus Torvalds1da177e2005-04-16 15:20:36 -0700938 0 - No source validation.
Stephen Hemmingerc1cf8422009-02-20 08:25:36 +0000939 1 - Strict mode as defined in RFC3704 Strict Reverse Path
940 Each incoming packet is tested against the FIB and if the interface
941 is not the best reverse path the packet check will fail.
942 By default failed packets are discarded.
943 2 - Loose mode as defined in RFC3704 Loose Reverse Path
944 Each incoming packet's source address is also tested against the FIB
945 and if the source address is not reachable via any interface
946 the packet check will fail.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700947
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000948 Current recommended practice in RFC3704 is to enable strict mode
Jesper Dangaard Brouerbf869c32009-02-23 04:37:55 +0000949 to prevent IP spoofing from DDos attacks. If using asymmetric routing
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +0000950 or other complicated routing, then loose mode is recommended.
Stephen Hemmingerc1cf8422009-02-20 08:25:36 +0000951
Shan Wei1f5865e2009-12-02 15:39:04 -0800952 The max value from conf/{all,interface}/rp_filter is used
953 when doing source validation on the {interface}.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700954
955 Default value is 0. Note that some distributions enable it
956 in startup scripts.
957
958arp_filter - BOOLEAN
959 1 - Allows you to have multiple network interfaces on the same
960 subnet, and have the ARPs for each interface be answered
961 based on whether or not the kernel would route a packet from
962 the ARP'd IP out that interface (therefore you must use source
963 based routing for this to work). In other words it allows control
964 of which cards (usually 1) will respond to an arp request.
965
966 0 - (default) The kernel can respond to arp requests with addresses
967 from other interfaces. This may seem wrong but it usually makes
968 sense, because it increases the chance of successful communication.
969 IP addresses are owned by the complete host on Linux, not by
970 particular interfaces. Only for more complex setups like load-
971 balancing, does this behaviour cause problems.
972
973 arp_filter for the interface will be enabled if at least one of
974 conf/{all,interface}/arp_filter is set to TRUE,
975 it will be disabled otherwise
976
977arp_announce - INTEGER
978 Define different restriction levels for announcing the local
979 source IP address from IP packets in ARP requests sent on
980 interface:
981 0 - (default) Use any local address, configured on any interface
982 1 - Try to avoid local addresses that are not in the target's
983 subnet for this interface. This mode is useful when target
984 hosts reachable via this interface require the source IP
985 address in ARP requests to be part of their logical network
986 configured on the receiving interface. When we generate the
987 request we will check all our subnets that include the
988 target IP and will preserve the source address if it is from
989 such subnet. If there is no such subnet we select source
990 address according to the rules for level 2.
991 2 - Always use the best local address for this target.
992 In this mode we ignore the source address in the IP packet
993 and try to select local address that we prefer for talks with
994 the target host. Such local address is selected by looking
995 for primary IP addresses on all our subnets on the outgoing
996 interface that include the target IP address. If no suitable
997 local address is found we select the first local address
998 we have on the outgoing interface or on all other interfaces,
999 with the hope we will receive reply for our request and
1000 even sometimes no matter the source IP address we announce.
1001
1002 The max value from conf/{all,interface}/arp_announce is used.
1003
1004 Increasing the restriction level gives more chance for
1005 receiving answer from the resolved target while decreasing
1006 the level announces more valid sender's information.
1007
1008arp_ignore - INTEGER
1009 Define different modes for sending replies in response to
1010 received ARP requests that resolve local target IP addresses:
1011 0 - (default): reply for any local target IP address, configured
1012 on any interface
1013 1 - reply only if the target IP address is local address
1014 configured on the incoming interface
1015 2 - reply only if the target IP address is local address
1016 configured on the incoming interface and both with the
1017 sender's IP address are part from same subnet on this interface
1018 3 - do not reply for local addresses configured with scope host,
1019 only resolutions for global and link addresses are replied
1020 4-7 - reserved
1021 8 - do not reply for all local addresses
1022
1023 The max value from conf/{all,interface}/arp_ignore is used
1024 when ARP request is received on the {interface}
1025
Stephen Hemmingereefef1c2009-02-01 01:04:33 -08001026arp_notify - BOOLEAN
1027 Define mode for notification of address and device changes.
1028 0 - (default): do nothing
Ian Campbell3f8dc232010-05-26 00:09:41 +00001029 1 - Generate gratuitous arp requests when device is brought up
Stephen Hemmingereefef1c2009-02-01 01:04:33 -08001030 or hardware address changes.
1031
Neil Hormanc1b1bce2006-03-20 22:40:03 -08001032arp_accept - BOOLEAN
Octavian Purdila6d955182010-01-18 12:58:44 +00001033 Define behavior for gratuitous ARP frames who's IP is not
1034 already present in the ARP table:
1035 0 - don't create new entries in the ARP table
1036 1 - create new entries in the ARP table
1037
1038 Both replies and requests type gratuitous arp will trigger the
1039 ARP table to be updated, if this setting is on.
1040
1041 If the ARP table already contains the IP address of the
1042 gratuitous arp frame, the arp table will be updated regardless
1043 if this setting is on or off.
1044
Neil Hormanc1b1bce2006-03-20 22:40:03 -08001045
Linus Torvalds1da177e2005-04-16 15:20:36 -07001046app_solicit - INTEGER
1047 The maximum number of probes to send to the user space ARP daemon
1048 via netlink before dropping back to multicast probes (see
1049 mcast_solicit). Defaults to 0.
1050
1051disable_policy - BOOLEAN
1052 Disable IPSEC policy (SPD) for this interface
1053
1054disable_xfrm - BOOLEAN
1055 Disable IPSEC encryption on this interface, whatever the policy
1056
Hannes Frederic Sowafc4eba52013-08-14 01:03:46 +02001057igmpv2_unsolicited_report_interval - INTEGER
1058 The interval in milliseconds in which the next unsolicited
1059 IGMPv1 or IGMPv2 report retransmit will take place.
1060 Default: 10000 (10 seconds)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001061
Hannes Frederic Sowafc4eba52013-08-14 01:03:46 +02001062igmpv3_unsolicited_report_interval - INTEGER
1063 The interval in milliseconds in which the next unsolicited
1064 IGMPv3 report retransmit will take place.
1065 Default: 1000 (1 seconds)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001066
1067tag - INTEGER
1068 Allows you to write a number, which can be used as required.
1069 Default value is 0.
1070
Linus Torvalds1da177e2005-04-16 15:20:36 -07001071Alexey Kuznetsov.
1072kuznet@ms2.inr.ac.ru
1073
1074Updated by:
1075Andi Kleen
1076ak@muc.de
1077Nicolas Delon
1078delon.nicolas@wanadoo.fr
1079
1080
1081
1082
1083/proc/sys/net/ipv6/* Variables:
1084
1085IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1086apply to IPv6 [XXX?].
1087
1088bindv6only - BOOLEAN
1089 Default value for IPV6_V6ONLY socket option,
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001090 which restricts use of the IPv6 socket to IPv6 communication
Linus Torvalds1da177e2005-04-16 15:20:36 -07001091 only.
1092 TRUE: disable IPv4-mapped address feature
1093 FALSE: enable IPv4-mapped address feature
1094
Geoffrey Thomasd5c073ca2011-08-22 11:28:57 -07001095 Default: FALSE (as specified in RFC3493)
Linus Torvalds1da177e2005-04-16 15:20:36 -07001096
1097IPv6 Fragmentation:
1098
1099ip6frag_high_thresh - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001100 Maximum memory used to reassemble IPv6 fragments. When
Linus Torvalds1da177e2005-04-16 15:20:36 -07001101 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1102 the fragment handler will toss packets until ip6frag_low_thresh
1103 is reached.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001104
Linus Torvalds1da177e2005-04-16 15:20:36 -07001105ip6frag_low_thresh - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001106 See ip6frag_high_thresh
Linus Torvalds1da177e2005-04-16 15:20:36 -07001107
1108ip6frag_time - INTEGER
1109 Time in seconds to keep an IPv6 fragment in memory.
1110
1111ip6frag_secret_interval - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001112 Regeneration interval (in seconds) of the hash secret (or lifetime
Linus Torvalds1da177e2005-04-16 15:20:36 -07001113 for the hash secret) for IPv6 fragments.
1114 Default: 600
1115
1116conf/default/*:
1117 Change the interface-specific default settings.
1118
1119
1120conf/all/*:
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001121 Change all the interface-specific settings.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001122
1123 [XXX: Other special features than forwarding?]
1124
1125conf/all/forwarding - BOOLEAN
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001126 Enable global IPv6 forwarding between all interfaces.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001127
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001128 IPv4 and IPv6 work differently here; e.g. netfilter must be used
Linus Torvalds1da177e2005-04-16 15:20:36 -07001129 to control which interfaces may forward packets and which not.
1130
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001131 This also sets all interfaces' Host/Router setting
Linus Torvalds1da177e2005-04-16 15:20:36 -07001132 'forwarding' to the specified value. See below for details.
1133
1134 This referred to as global forwarding.
1135
YOSHIFUJI Hideakifbea49e2006-09-22 14:43:49 -07001136proxy_ndp - BOOLEAN
1137 Do proxy ndp.
1138
Linus Torvalds1da177e2005-04-16 15:20:36 -07001139conf/interface/*:
1140 Change special settings per interface.
1141
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001142 The functional behaviour for certain settings is different
Linus Torvalds1da177e2005-04-16 15:20:36 -07001143 depending on whether local forwarding is enabled or not.
1144
Roy.Li605b91c2011-09-28 19:51:54 +00001145accept_ra - INTEGER
Linus Torvalds1da177e2005-04-16 15:20:36 -07001146 Accept Router Advertisements; autoconfigure using them.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001147
Tore Anderson026359b2011-08-28 23:47:33 +00001148 It also determines whether or not to transmit Router
1149 Solicitations. If and only if the functional setting is to
1150 accept Router Advertisements, Router Solicitations will be
1151 transmitted.
1152
Thomas Grafae8abfa2010-09-03 05:47:30 +00001153 Possible values are:
1154 0 Do not accept Router Advertisements.
1155 1 Accept Router Advertisements if forwarding is disabled.
1156 2 Overrule forwarding behaviour. Accept Router Advertisements
1157 even if forwarding is enabled.
1158
Linus Torvalds1da177e2005-04-16 15:20:36 -07001159 Functional default: enabled if local forwarding is disabled.
1160 disabled if local forwarding is enabled.
1161
YOSHIFUJI Hideaki65f5c7c2006-03-20 16:55:08 -08001162accept_ra_defrtr - BOOLEAN
1163 Learn default router in Router Advertisement.
1164
1165 Functional default: enabled if accept_ra is enabled.
1166 disabled if accept_ra is disabled.
1167
YOSHIFUJI Hideakic4fd30e2006-03-20 16:55:26 -08001168accept_ra_pinfo - BOOLEAN
Matt LaPlante2fe0ae72006-10-03 22:50:39 +02001169 Learn Prefix Information in Router Advertisement.
YOSHIFUJI Hideakic4fd30e2006-03-20 16:55:26 -08001170
1171 Functional default: enabled if accept_ra is enabled.
1172 disabled if accept_ra is disabled.
1173
YOSHIFUJI Hideaki09c884d2006-03-20 17:07:03 -08001174accept_ra_rt_info_max_plen - INTEGER
1175 Maximum prefix length of Route Information in RA.
1176
1177 Route Information w/ prefix larger than or equal to this
1178 variable shall be ignored.
1179
1180 Functional default: 0 if accept_ra_rtr_pref is enabled.
1181 -1 if accept_ra_rtr_pref is disabled.
1182
YOSHIFUJI Hideaki930d6ff2006-03-20 17:05:30 -08001183accept_ra_rtr_pref - BOOLEAN
1184 Accept Router Preference in RA.
1185
1186 Functional default: enabled if accept_ra is enabled.
1187 disabled if accept_ra is disabled.
1188
Linus Torvalds1da177e2005-04-16 15:20:36 -07001189accept_redirects - BOOLEAN
1190 Accept Redirects.
1191
1192 Functional default: enabled if local forwarding is disabled.
1193 disabled if local forwarding is enabled.
1194
YOSHIFUJI Hideaki0bcbc922007-04-24 14:58:30 -07001195accept_source_route - INTEGER
1196 Accept source routing (routing extension header).
1197
YOSHIFUJI Hideakibb4dbf92007-07-10 22:55:49 -07001198 >= 0: Accept only routing header type 2.
YOSHIFUJI Hideaki0bcbc922007-04-24 14:58:30 -07001199 < 0: Do not accept routing header.
1200
1201 Default: 0
1202
Linus Torvalds1da177e2005-04-16 15:20:36 -07001203autoconf - BOOLEAN
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001204 Autoconfigure addresses using Prefix Information in Router
Linus Torvalds1da177e2005-04-16 15:20:36 -07001205 Advertisements.
1206
YOSHIFUJI Hideakic4fd30e2006-03-20 16:55:26 -08001207 Functional default: enabled if accept_ra_pinfo is enabled.
1208 disabled if accept_ra_pinfo is disabled.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001209
1210dad_transmits - INTEGER
1211 The amount of Duplicate Address Detection probes to send.
1212 Default: 1
Linus Torvalds1da177e2005-04-16 15:20:36 -07001213
Roy.Li605b91c2011-09-28 19:51:54 +00001214forwarding - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001215 Configure interface-specific Host/Router behaviour.
1216
1217 Note: It is recommended to have the same setting on all
Linus Torvalds1da177e2005-04-16 15:20:36 -07001218 interfaces; mixed router/host scenarios are rather uncommon.
1219
Thomas Grafae8abfa2010-09-03 05:47:30 +00001220 Possible values are:
1221 0 Forwarding disabled
1222 1 Forwarding enabled
Thomas Grafae8abfa2010-09-03 05:47:30 +00001223
1224 FALSE (0):
Linus Torvalds1da177e2005-04-16 15:20:36 -07001225
1226 By default, Host behaviour is assumed. This means:
1227
1228 1. IsRouter flag is not set in Neighbour Advertisements.
Tore Anderson026359b2011-08-28 23:47:33 +00001229 2. If accept_ra is TRUE (default), transmit Router
1230 Solicitations.
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001231 3. If accept_ra is TRUE (default), accept Router
Linus Torvalds1da177e2005-04-16 15:20:36 -07001232 Advertisements (and do autoconfiguration).
1233 4. If accept_redirects is TRUE (default), accept Redirects.
1234
Thomas Grafae8abfa2010-09-03 05:47:30 +00001235 TRUE (1):
Linus Torvalds1da177e2005-04-16 15:20:36 -07001236
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001237 If local forwarding is enabled, Router behaviour is assumed.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001238 This means exactly the reverse from the above:
1239
1240 1. IsRouter flag is set in Neighbour Advertisements.
Tore Anderson026359b2011-08-28 23:47:33 +00001241 2. Router Solicitations are not sent unless accept_ra is 2.
Thomas Grafae8abfa2010-09-03 05:47:30 +00001242 3. Router Advertisements are ignored unless accept_ra is 2.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001243 4. Redirects are ignored.
1244
Thomas Grafae8abfa2010-09-03 05:47:30 +00001245 Default: 0 (disabled) if global forwarding is disabled (default),
1246 otherwise 1 (enabled).
Linus Torvalds1da177e2005-04-16 15:20:36 -07001247
1248hop_limit - INTEGER
1249 Default Hop Limit to set.
1250 Default: 64
1251
1252mtu - INTEGER
1253 Default Maximum Transfer Unit
1254 Default: 1280 (IPv6 required minimum)
1255
YOSHIFUJI Hideaki52e16352006-03-20 17:05:47 -08001256router_probe_interval - INTEGER
1257 Minimum interval (in seconds) between Router Probing described
1258 in RFC4191.
1259
1260 Default: 60
1261
Linus Torvalds1da177e2005-04-16 15:20:36 -07001262router_solicitation_delay - INTEGER
1263 Number of seconds to wait after interface is brought up
1264 before sending Router Solicitations.
1265 Default: 1
1266
1267router_solicitation_interval - INTEGER
1268 Number of seconds to wait between Router Solicitations.
1269 Default: 4
1270
1271router_solicitations - INTEGER
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001272 Number of Router Solicitations to send until assuming no
Linus Torvalds1da177e2005-04-16 15:20:36 -07001273 routers are present.
1274 Default: 3
1275
1276use_tempaddr - INTEGER
1277 Preference for Privacy Extensions (RFC3041).
1278 <= 0 : disable Privacy Extensions
1279 == 1 : enable Privacy Extensions, but prefer public
1280 addresses over temporary addresses.
1281 > 1 : enable Privacy Extensions and prefer temporary
1282 addresses over public addresses.
1283 Default: 0 (for most devices)
1284 -1 (for point-to-point devices and loopback devices)
1285
1286temp_valid_lft - INTEGER
1287 valid lifetime (in seconds) for temporary addresses.
1288 Default: 604800 (7 days)
1289
1290temp_prefered_lft - INTEGER
1291 Preferred lifetime (in seconds) for temporary addresses.
1292 Default: 86400 (1 day)
1293
1294max_desync_factor - INTEGER
1295 Maximum value for DESYNC_FACTOR, which is a random value
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001296 that ensures that clients don't synchronize with each
Linus Torvalds1da177e2005-04-16 15:20:36 -07001297 other and generate new addresses at exactly the same time.
1298 value is in seconds.
1299 Default: 600
Jesper Dangaard Brouere18f5fe2009-02-23 04:39:04 +00001300
Linus Torvalds1da177e2005-04-16 15:20:36 -07001301regen_max_retry - INTEGER
1302 Number of attempts before give up attempting to generate
1303 valid temporary addresses.
1304 Default: 5
1305
1306max_addresses - INTEGER
Brian Haleye79dc482010-02-22 12:27:21 +00001307 Maximum number of autoconfigured addresses per interface. Setting
1308 to zero disables the limitation. It is not recommended to set this
1309 value too large (or to zero) because it would be an easy way to
1310 crash the kernel by allowing too many addresses to be created.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001311 Default: 16
1312
YOSHIFUJI Hideaki778d80b2008-06-28 14:17:11 +09001313disable_ipv6 - BOOLEAN
Brian Haley9bdd8d42009-03-18 18:22:48 -07001314 Disable IPv6 operation. If accept_dad is set to 2, this value
1315 will be dynamically set to TRUE if DAD fails for the link-local
1316 address.
YOSHIFUJI Hideaki778d80b2008-06-28 14:17:11 +09001317 Default: FALSE (enable IPv6 operation)
1318
Brian Haley56d417b2009-06-01 03:07:33 -07001319 When this value is changed from 1 to 0 (IPv6 is being enabled),
1320 it will dynamically create a link-local address on the given
1321 interface and start Duplicate Address Detection, if necessary.
1322
1323 When this value is changed from 0 to 1 (IPv6 is being disabled),
1324 it will dynamically delete all address on the given interface.
1325
YOSHIFUJI Hideaki1b34be72008-06-28 14:18:38 +09001326accept_dad - INTEGER
1327 Whether to accept DAD (Duplicate Address Detection).
1328 0: Disable DAD
1329 1: Enable DAD (default)
1330 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1331 link-local address has been found.
1332
Octavian Purdilaf7734fd2009-10-02 11:39:15 +00001333force_tllao - BOOLEAN
1334 Enable sending the target link-layer address option even when
1335 responding to a unicast neighbor solicitation.
1336 Default: FALSE
1337
1338 Quoting from RFC 2461, section 4.4, Target link-layer address:
1339
1340 "The option MUST be included for multicast solicitations in order to
1341 avoid infinite Neighbor Solicitation "recursion" when the peer node
1342 does not have a cache entry to return a Neighbor Advertisements
1343 message. When responding to unicast solicitations, the option can be
1344 omitted since the sender of the solicitation has the correct link-
1345 layer address; otherwise it would not have be able to send the unicast
1346 solicitation in the first place. However, including the link-layer
1347 address in this case adds little overhead and eliminates a potential
1348 race condition where the sender deletes the cached link-layer address
1349 prior to receiving a response to a previous solicitation."
1350
Hannes Frederic Sowadb2b6202013-01-01 00:35:31 +00001351ndisc_notify - BOOLEAN
1352 Define mode for notification of address and device changes.
1353 0 - (default): do nothing
1354 1 - Generate unsolicited neighbour advertisements when device is brought
1355 up or hardware address changes.
1356
Hannes Frederic Sowafc4eba52013-08-14 01:03:46 +02001357mldv1_unsolicited_report_interval - INTEGER
1358 The interval in milliseconds in which the next unsolicited
1359 MLDv1 report retransmit will take place.
1360 Default: 10000 (10 seconds)
1361
1362mldv2_unsolicited_report_interval - INTEGER
1363 The interval in milliseconds in which the next unsolicited
1364 MLDv2 report retransmit will take place.
1365 Default: 1000 (1 second)
1366
Daniel Borkmannf2127812013-09-04 00:19:44 +02001367force_mld_version - INTEGER
1368 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1369 1 - Enforce to use MLD version 1
1370 2 - Enforce to use MLD version 2
1371
Hannes Frederic Sowab800c3b2013-08-27 01:36:51 +02001372suppress_frag_ndisc - INTEGER
1373 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1374 with IPv6 Neighbor Discovery) behavior:
1375 1 - (default) discard fragmented neighbor discovery packets
1376 0 - allow fragmented neighbor discovery packets
1377
Linus Torvalds1da177e2005-04-16 15:20:36 -07001378icmp/*:
1379ratelimit - INTEGER
1380 Limit the maximal rates for sending ICMPv6 packets.
Stephen Hemminger6dbf4bc2008-07-01 19:29:07 -07001381 0 to disable any limiting,
1382 otherwise the minimal space between responses in milliseconds.
1383 Default: 1000
Linus Torvalds1da177e2005-04-16 15:20:36 -07001384
1385
1386IPv6 Update by:
1387Pekka Savola <pekkas@netcore.fi>
1388YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1389
1390
1391/proc/sys/net/bridge/* Variables:
1392
1393bridge-nf-call-arptables - BOOLEAN
1394 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1395 0 : disable this.
1396 Default: 1
1397
1398bridge-nf-call-iptables - BOOLEAN
1399 1 : pass bridged IPv4 traffic to iptables' chains.
1400 0 : disable this.
1401 Default: 1
1402
1403bridge-nf-call-ip6tables - BOOLEAN
1404 1 : pass bridged IPv6 traffic to ip6tables' chains.
1405 0 : disable this.
1406 Default: 1
1407
1408bridge-nf-filter-vlan-tagged - BOOLEAN
Michael Milner516299d2007-04-12 22:14:23 -07001409 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1410 0 : disable this.
Pablo Neira Ayuso49816822012-05-08 19:36:44 +02001411 Default: 0
Michael Milner516299d2007-04-12 22:14:23 -07001412
1413bridge-nf-filter-pppoe-tagged - BOOLEAN
1414 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001415 0 : disable this.
Pablo Neira Ayuso49816822012-05-08 19:36:44 +02001416 Default: 0
Linus Torvalds1da177e2005-04-16 15:20:36 -07001417
Pablo Neira Ayuso49816822012-05-08 19:36:44 +02001418bridge-nf-pass-vlan-input-dev - BOOLEAN
1419 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1420 interface on the bridge and set the netfilter input device to the vlan.
1421 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1422 target work with vlan-on-top-of-bridge interfaces. When no matching
1423 vlan interface is found, or this switch is off, the input device is
1424 set to the bridge interface.
1425 0: disable bridge netfilter vlan interface lookup.
1426 Default: 0
Linus Torvalds1da177e2005-04-16 15:20:36 -07001427
Vlad Yasevich32e8d492008-07-08 16:43:29 -07001428proc/sys/net/sctp/* Variables:
1429
1430addip_enable - BOOLEAN
1431 Enable or disable extension of Dynamic Address Reconfiguration
1432 (ADD-IP) functionality specified in RFC5061. This extension provides
1433 the ability to dynamically add and remove new addresses for the SCTP
1434 associations.
1435
1436 1: Enable extension.
1437
1438 0: Disable extension.
1439
1440 Default: 0
1441
1442addip_noauth_enable - BOOLEAN
1443 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1444 authentication to protect the operations of adding or removing new
1445 addresses. This requirement is mandated so that unauthorized hosts
1446 would not be able to hijack associations. However, older
1447 implementations may not have implemented this requirement while
1448 allowing the ADD-IP extension. For reasons of interoperability,
1449 we provide this variable to control the enforcement of the
1450 authentication requirement.
1451
1452 1: Allow ADD-IP extension to be used without authentication. This
1453 should only be set in a closed environment for interoperability
1454 with older implementations.
1455
1456 0: Enforce the authentication requirement
1457
1458 Default: 0
1459
1460auth_enable - BOOLEAN
1461 Enable or disable Authenticated Chunks extension. This extension
1462 provides the ability to send and receive authenticated chunks and is
1463 required for secure operation of Dynamic Address Reconfiguration
1464 (ADD-IP) extension.
1465
1466 1: Enable this extension.
1467 0: Disable this extension.
1468
1469 Default: 0
1470
1471prsctp_enable - BOOLEAN
1472 Enable or disable the Partial Reliability extension (RFC3758) which
1473 is used to notify peers that a given DATA should no longer be expected.
1474
1475 1: Enable extension
1476 0: Disable
1477
1478 Default: 1
1479
1480max_burst - INTEGER
1481 The limit of the number of new packets that can be initially sent. It
1482 controls how bursty the generated traffic can be.
1483
1484 Default: 4
1485
1486association_max_retrans - INTEGER
1487 Set the maximum number for retransmissions that an association can
1488 attempt deciding that the remote end is unreachable. If this value
1489 is exceeded, the association is terminated.
1490
1491 Default: 10
1492
1493max_init_retransmits - INTEGER
1494 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1495 that an association will attempt before declaring the destination
1496 unreachable and terminating.
1497
1498 Default: 8
1499
1500path_max_retrans - INTEGER
1501 The maximum number of retransmissions that will be attempted on a given
1502 path. Once this threshold is exceeded, the path is considered
1503 unreachable, and new traffic will use a different path when the
1504 association is multihomed.
1505
1506 Default: 5
1507
Neil Horman5aa93bc2012-07-21 07:56:07 +00001508pf_retrans - INTEGER
1509 The number of retransmissions that will be attempted on a given path
1510 before traffic is redirected to an alternate transport (should one
1511 exist). Note this is distinct from path_max_retrans, as a path that
1512 passes the pf_retrans threshold can still be used. Its only
1513 deprioritized when a transmission path is selected by the stack. This
1514 setting is primarily used to enable fast failover mechanisms without
1515 having to reduce path_max_retrans to a very low value. See:
1516 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1517 for details. Note also that a value of pf_retrans > path_max_retrans
1518 disables this feature
1519
1520 Default: 0
1521
Vlad Yasevich32e8d492008-07-08 16:43:29 -07001522rto_initial - INTEGER
1523 The initial round trip timeout value in milliseconds that will be used
1524 in calculating round trip times. This is the initial time interval
1525 for retransmissions.
1526
1527 Default: 3000
1528
1529rto_max - INTEGER
1530 The maximum value (in milliseconds) of the round trip timeout. This
1531 is the largest time interval that can elapse between retransmissions.
1532
1533 Default: 60000
1534
1535rto_min - INTEGER
1536 The minimum value (in milliseconds) of the round trip timeout. This
1537 is the smallest time interval the can elapse between retransmissions.
1538
1539 Default: 1000
1540
1541hb_interval - INTEGER
1542 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1543 are sent at the specified interval on idle paths to probe the state of
1544 a given path between 2 associations.
1545
1546 Default: 30000
1547
1548sack_timeout - INTEGER
1549 The amount of time (in milliseconds) that the implementation will wait
1550 to send a SACK.
1551
1552 Default: 200
1553
1554valid_cookie_life - INTEGER
1555 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1556 is used during association establishment.
1557
1558 Default: 60000
1559
1560cookie_preserve_enable - BOOLEAN
1561 Enable or disable the ability to extend the lifetime of the SCTP cookie
1562 that is used during the establishment phase of SCTP association
1563
1564 1: Enable cookie lifetime extension.
1565 0: Disable
1566
1567 Default: 1
1568
Neil Horman3c681982012-10-24 09:20:03 +00001569cookie_hmac_alg - STRING
1570 Select the hmac algorithm used when generating the cookie value sent by
1571 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1572 Valid values are:
1573 * md5
1574 * sha1
1575 * none
1576 Ability to assign md5 or sha1 as the selected alg is predicated on the
stephen hemminger3b09adc2013-01-03 07:50:29 +00001577 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
Neil Horman3c681982012-10-24 09:20:03 +00001578 CONFIG_CRYPTO_SHA1).
1579
1580 Default: Dependent on configuration. MD5 if available, else SHA1 if
1581 available, else none.
1582
Vlad Yasevich32e8d492008-07-08 16:43:29 -07001583rcvbuf_policy - INTEGER
1584 Determines if the receive buffer is attributed to the socket or to
1585 association. SCTP supports the capability to create multiple
1586 associations on a single socket. When using this capability, it is
1587 possible that a single stalled association that's buffering a lot
1588 of data may block other associations from delivering their data by
1589 consuming all of the receive buffer space. To work around this,
1590 the rcvbuf_policy could be set to attribute the receiver buffer space
1591 to each association instead of the socket. This prevents the described
1592 blocking.
1593
1594 1: rcvbuf space is per association
stephen hemminger3b09adc2013-01-03 07:50:29 +00001595 0: rcvbuf space is per socket
Vlad Yasevich32e8d492008-07-08 16:43:29 -07001596
1597 Default: 0
1598
1599sndbuf_policy - INTEGER
1600 Similar to rcvbuf_policy above, this applies to send buffer space.
1601
1602 1: Send buffer is tracked per association
1603 0: Send buffer is tracked per socket.
1604
1605 Default: 0
1606
1607sctp_mem - vector of 3 INTEGERs: min, pressure, max
1608 Number of pages allowed for queueing by all SCTP sockets.
1609
1610 min: Below this number of pages SCTP is not bothered about its
1611 memory appetite. When amount of memory allocated by SCTP exceeds
1612 this number, SCTP starts to moderate memory usage.
1613
1614 pressure: This value was introduced to follow format of tcp_mem.
1615
1616 max: Number of pages allowed for queueing by all SCTP sockets.
1617
1618 Default is calculated at boot time from amount of available memory.
1619
1620sctp_rmem - vector of 3 INTEGERs: min, default, max
Max Matveeva6e12042011-06-19 22:08:10 +00001621 Only the first value ("min") is used, "default" and "max" are
1622 ignored.
1623
1624 min: Minimal size of receive buffer used by SCTP socket.
1625 It is guaranteed to each SCTP socket (but not association) even
1626 under moderate memory pressure.
1627
1628 Default: 1 page
Vlad Yasevich32e8d492008-07-08 16:43:29 -07001629
1630sctp_wmem - vector of 3 INTEGERs: min, default, max
Max Matveeva6e12042011-06-19 22:08:10 +00001631 Currently this tunable has no effect.
Vlad Yasevich32e8d492008-07-08 16:43:29 -07001632
Bhaskar Dutta72388432009-09-03 17:25:47 +05301633addr_scope_policy - INTEGER
1634 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1635
1636 0 - Disable IPv4 address scoping
1637 1 - Enable IPv4 address scoping
1638 2 - Follow draft but allow IPv4 private addresses
1639 3 - Follow draft but allow IPv4 link local addresses
1640
1641 Default: 1
1642
Linus Torvalds1da177e2005-04-16 15:20:36 -07001643
Stephen Hemminger4edc2f32008-07-10 16:50:26 -07001644/proc/sys/net/core/*
Shan Weic60f6aa2012-04-26 16:52:52 +00001645 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
Wang Tinggong705efc32009-05-14 22:49:36 +00001646
Linus Torvalds1da177e2005-04-16 15:20:36 -07001647
Stephen Hemminger4edc2f32008-07-10 16:50:26 -07001648/proc/sys/net/unix/*
Wang Tinggong705efc32009-05-14 22:49:36 +00001649max_dgram_qlen - INTEGER
1650 The maximum length of dgram socket receive queue
1651
1652 Default: 10
1653
1654
1655UNDOCUMENTED:
Stephen Hemminger4edc2f32008-07-10 16:50:26 -07001656
1657/proc/sys/net/irda/*
1658 fast_poll_increase FIXME
1659 warn_noreply_time FIXME
1660 discovery_slots FIXME
1661 slot_timeout FIXME
1662 max_baud_rate FIXME
1663 discovery_timeout FIXME
1664 lap_keepalive_time FIXME
1665 max_noreply_time FIXME
1666 max_tx_data_size FIXME
1667 max_tx_window FIXME
1668 min_tx_turn_time FIXME