blob: af6a63ab9026a3f834a92af6746212e6b7e971f6 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001/proc/sys/net/ipv4/* Variables:
2
3ip_forward - BOOLEAN
4 0 - disabled (default)
5 not 0 - enabled
6
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
14 default 64
15
16ip_no_pmtu_disc - BOOLEAN
17 Disable Path MTU Discovery.
18 default FALSE
19
20min_pmtu - INTEGER
21 default 562 - minimum discovered Path MTU
22
23mtu_expires - INTEGER
24 Time, in seconds, that cached PMTU information is kept.
25
26min_adv_mss - INTEGER
27 The advertised MSS depends on the first hop route MTU, but will
28 never be lower than this setting.
29
30IP Fragmentation:
31
32ipfrag_high_thresh - INTEGER
33 Maximum memory used to reassemble IP fragments. When
34 ipfrag_high_thresh bytes of memory is allocated for this purpose,
35 the fragment handler will toss packets until ipfrag_low_thresh
36 is reached.
37
38ipfrag_low_thresh - INTEGER
39 See ipfrag_high_thresh
40
41ipfrag_time - INTEGER
42 Time in seconds to keep an IP fragment in memory.
43
44ipfrag_secret_interval - INTEGER
45 Regeneration interval (in seconds) of the hash secret (or lifetime
46 for the hash secret) for IP fragments.
47 Default: 600
48
Herbert Xu89cee8b2005-12-13 23:14:27 -080049ipfrag_max_dist - INTEGER
50 ipfrag_max_dist is a non-negative integer value which defines the
51 maximum "disorder" which is allowed among fragments which share a
52 common IP source address. Note that reordering of packets is
53 not unusual, but if a large number of fragments arrive from a source
54 IP address while a particular fragment queue remains incomplete, it
55 probably indicates that one or more fragments belonging to that queue
56 have been lost. When ipfrag_max_dist is positive, an additional check
57 is done on fragments before they are added to a reassembly queue - if
58 ipfrag_max_dist (or more) fragments have arrived from a particular IP
59 address between additions to any IP fragment queue using that source
60 address, it's presumed that one or more fragments in the queue are
61 lost. The existing fragment queue will be dropped, and a new one
62 started. An ipfrag_max_dist value of zero disables this check.
63
64 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
65 result in unnecessarily dropping fragment queues when normal
66 reordering of packets occurs, which could lead to poor application
67 performance. Using a very large value, e.g. 50000, increases the
68 likelihood of incorrectly reassembling IP fragments that originate
69 from different IP datagrams, which could result in data corruption.
70 Default: 64
71
Linus Torvalds1da177e2005-04-16 15:20:36 -070072INET peer storage:
73
74inet_peer_threshold - INTEGER
75 The approximate size of the storage. Starting from this threshold
76 entries will be thrown aggressively. This threshold also determines
77 entries' time-to-live and time intervals between garbage collection
78 passes. More entries, less time-to-live, less GC interval.
79
80inet_peer_minttl - INTEGER
81 Minimum time-to-live of entries. Should be enough to cover fragment
82 time-to-live on the reassembling side. This minimum time-to-live is
83 guaranteed if the pool size is less than inet_peer_threshold.
84 Measured in jiffies(1).
85
86inet_peer_maxttl - INTEGER
87 Maximum time-to-live of entries. Unused entries will expire after
88 this period of time if there is no memory pressure on the pool (i.e.
89 when the number of entries in the pool is very small).
90 Measured in jiffies(1).
91
92inet_peer_gc_mintime - INTEGER
93 Minimum interval between garbage collection passes. This interval is
94 in effect under high memory pressure on the pool.
95 Measured in jiffies(1).
96
97inet_peer_gc_maxtime - INTEGER
98 Minimum interval between garbage collection passes. This interval is
99 in effect under low (or absent) memory pressure on the pool.
100 Measured in jiffies(1).
101
102TCP variables:
103
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800104somaxconn - INTEGER
105 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
106 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
107 for TCP sockets.
108
Stephen Hemminger9772efb2005-11-10 17:09:53 -0800109tcp_abc - INTEGER
Stephen Hemmingerb3a8a402006-09-13 19:51:02 -0700110 Controls Appropriate Byte Count (ABC) defined in RFC3465.
111 ABC is a way of increasing congestion window (cwnd) more slowly
112 in response to partial acknowledgments.
113 Possible values are:
114 0 increase cwnd once per acknowledgment (no ABC)
115 1 increase cwnd once per acknowledgment of full sized segment
116 2 allow increase cwnd by two if acknowledgment is
117 of two segments to compensate for delayed acknowledgments.
118 Default: 0 (off)
Stephen Hemminger9772efb2005-11-10 17:09:53 -0800119
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800120tcp_abort_on_overflow - BOOLEAN
121 If listening service is too slow to accept new connections,
122 reset them. Default state is FALSE. It means that if overflow
123 occurred due to a burst, connection will recover. Enable this
124 option _only_ if you are really sure that listening daemon
125 cannot be tuned to accept connections faster. Enabling this
126 option can harm clients of your server.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700127
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800128tcp_adv_win_scale - INTEGER
129 Count buffering overhead as bytes/2^tcp_adv_win_scale
130 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
131 if it is <= 0.
132 Default: 2
133
134tcp_allowed_congestion_control - STRING
135 Show/set the congestion control choices available to non-privileged
136 processes. The list is a subset of those listed in
137 tcp_available_congestion_control.
138 Default is "reno" and the default setting (tcp_congestion_control).
139
140tcp_app_win - INTEGER
141 Reserve max(window/2^tcp_app_win, mss) of window for application
142 buffer. Value 0 is special, it means that nothing is reserved.
143 Default: 31
144
145tcp_available_congestion_control - STRING
146 Shows the available congestion control choices that are registered.
147 More congestion control algorithms may be available as modules,
148 but not loaded.
149
John Heffner71599cd2007-02-27 10:03:56 -0800150tcp_base_mss - INTEGER
151 The initial value of search_low to be used by Packetization Layer
152 Path MTU Discovery (MTU probing). If MTU probing is enabled,
153 this is the inital MSS used by the connection.
154
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800155tcp_congestion_control - STRING
156 Set the congestion control algorithm to be used for new
157 connections. The algorithm "reno" is always available, but
158 additional choices may be available based on kernel configuration.
159 Default is set as part of kernel configuration.
160
161tcp_dsack - BOOLEAN
162 Allows TCP to send "duplicate" SACKs.
163
164tcp_ecn - BOOLEAN
165 Enable Explicit Congestion Notification in TCP.
166
167tcp_fack - BOOLEAN
168 Enable FACK congestion avoidance and fast retransmission.
169 The value is not used, if tcp_sack is not enabled.
170
171tcp_fin_timeout - INTEGER
172 Time to hold socket in state FIN-WAIT-2, if it was closed
173 by our side. Peer can be broken and never close its side,
174 or even died unexpectedly. Default value is 60sec.
175 Usual value used in 2.2 was 180 seconds, you may restore
176 it, but remember that if your machine is even underloaded WEB server,
177 you risk to overflow memory with kilotons of dead sockets,
178 FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1,
179 because they eat maximum 1.5K of memory, but they tend
180 to live longer. Cf. tcp_max_orphans.
181
Ilpo Järvinen89808062007-02-27 10:10:55 -0800182tcp_frto - INTEGER
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800183 Enables F-RTO, an enhanced recovery algorithm for TCP retransmission
184 timeouts. It is particularly beneficial in wireless environments
185 where packet loss is typically due to random radio interference
Ilpo Järvinen127af0c2007-02-21 23:16:38 -0800186 rather than intermediate router congestion. If set to 1, basic
Ilpo Järvinen89808062007-02-27 10:10:55 -0800187 version is enabled. 2 enables SACK enhanced F-RTO, which is
Ilpo Järvinen127af0c2007-02-21 23:16:38 -0800188 EXPERIMENTAL. The basic version can be used also when SACK is
189 enabled for a flow through tcp_sack sysctl.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700190
Ilpo Järvinen89808062007-02-27 10:10:55 -0800191tcp_frto_response - INTEGER
192 When F-RTO has detected that a TCP retransmission timeout was
193 spurious (i.e, the timeout would have been avoided had TCP set a
194 longer retransmission timeout), TCP has several options what to do
195 next. Possible values are:
196 0 Rate halving based; a smooth and conservative response,
197 results in halved cwnd and ssthresh after one RTT
198 1 Very conservative response; not recommended because even
199 though being valid, it interacts poorly with the rest of
200 Linux TCP, halves cwnd and ssthresh immediately
201 2 Aggressive response; undoes congestion control measures
202 that are now known to be unnecessary (ignoring the
203 possibility of a lost retransmission that would require
204 TCP to be more cautious), cwnd and ssthresh are restored
205 to the values prior timeout
206 Default: 0 (rate halving based)
207
Linus Torvalds1da177e2005-04-16 15:20:36 -0700208tcp_keepalive_time - INTEGER
209 How often TCP sends out keepalive messages when keepalive is enabled.
210 Default: 2hours.
211
212tcp_keepalive_probes - INTEGER
213 How many keepalive probes TCP sends out, until it decides that the
214 connection is broken. Default value: 9.
215
216tcp_keepalive_intvl - INTEGER
217 How frequently the probes are send out. Multiplied by
218 tcp_keepalive_probes it is time to kill not responding connection,
219 after probes started. Default value: 75sec i.e. connection
220 will be aborted after ~11 minutes of retries.
221
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800222tcp_low_latency - BOOLEAN
223 If set, the TCP stack makes decisions that prefer lower
224 latency as opposed to higher throughput. By default, this
225 option is not set meaning that higher throughput is preferred.
226 An example of an application where this default should be
227 changed would be a Beowulf compute cluster.
228 Default: 0
Linus Torvalds1da177e2005-04-16 15:20:36 -0700229
230tcp_max_orphans - INTEGER
231 Maximal number of TCP sockets not attached to any user file handle,
232 held by system. If this number is exceeded orphaned connections are
233 reset immediately and warning is printed. This limit exists
234 only to prevent simple DoS attacks, you _must_ not rely on this
235 or lower the limit artificially, but rather increase it
236 (probably, after increasing installed memory),
237 if network conditions require more than default value,
238 and tune network services to linger and kill such states
239 more aggressively. Let me to remind again: each orphan eats
240 up to ~64K of unswappable memory.
241
Linus Torvalds1da177e2005-04-16 15:20:36 -0700242tcp_max_syn_backlog - INTEGER
243 Maximal number of remembered connection requests, which are
244 still did not receive an acknowledgment from connecting client.
245 Default value is 1024 for systems with more than 128Mb of memory,
246 and 128 for low memory machines. If server suffers of overload,
247 try to increase this number.
248
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800249tcp_max_tw_buckets - INTEGER
250 Maximal number of timewait sockets held by system simultaneously.
251 If this number is exceeded time-wait socket is immediately destroyed
252 and warning is printed. This limit exists only to prevent
253 simple DoS attacks, you _must_ not lower the limit artificially,
254 but rather increase it (probably, after increasing installed memory),
255 if network conditions require more than default value.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700256
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800257tcp_mem - vector of 3 INTEGERs: min, pressure, max
258 min: below this number of pages TCP is not bothered about its
259 memory appetite.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700260
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800261 pressure: when amount of memory allocated by TCP exceeds this number
262 of pages, TCP moderates its memory consumption and enters memory
263 pressure mode, which is exited when memory consumption falls
264 under "min".
Linus Torvalds1da177e2005-04-16 15:20:36 -0700265
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800266 max: number of pages allowed for queueing by all TCP sockets.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700267
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800268 Defaults are calculated at boot time from amount of available
269 memory.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700270
John Heffner71599cd2007-02-27 10:03:56 -0800271tcp_moderate_rcvbuf - BOOLEAN
272 If set, TCP performs receive buffer autotuning, attempting to
273 automatically size the buffer (no greater than tcp_rmem[2]) to
274 match the size required by the path for full throughput. Enabled by
275 default.
276
277tcp_mtu_probing - INTEGER
278 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
279 values:
280 0 - Disabled
281 1 - Disabled by default, enabled when an ICMP black hole detected
282 2 - Always enabled, use initial MSS of tcp_base_mss.
283
284tcp_no_metrics_save - BOOLEAN
285 By default, TCP saves various connection metrics in the route cache
286 when the connection closes, so that connections established in the
287 near future can use these to set initial conditions. Usually, this
288 increases overall performance, but may sometimes cause performance
289 degredation. If set, TCP will not cache metrics on closing
290 connections.
291
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800292tcp_orphan_retries - INTEGER
293 How may times to retry before killing TCP connection, closed
294 by our side. Default value 7 corresponds to ~50sec-16min
295 depending on RTO. If you machine is loaded WEB server,
296 you should think about lowering this value, such sockets
297 may consume significant resources. Cf. tcp_max_orphans.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700298
299tcp_reordering - INTEGER
300 Maximal reordering of packets in a TCP stream.
301 Default: 3
302
303tcp_retrans_collapse - BOOLEAN
304 Bug-to-bug compatibility with some broken printers.
305 On retransmit try to send bigger packets to work around bugs in
306 certain TCP stacks.
307
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800308tcp_retries1 - INTEGER
309 How many times to retry before deciding that something is wrong
310 and it is necessary to report this suspicion to network layer.
311 Minimal RFC value is 3, it is default, which corresponds
312 to ~3sec-8min depending on RTO.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700313
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800314tcp_retries2 - INTEGER
315 How may times to retry before killing alive TCP connection.
316 RFC1122 says that the limit should be longer than 100 sec.
317 It is too small number. Default value 15 corresponds to ~13-30min
318 depending on RTO.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700319
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800320tcp_rfc1337 - BOOLEAN
321 If set, the TCP stack behaves conforming to RFC1337. If unset,
322 we are not conforming to RFC, but prevent TCP TIME_WAIT
323 assassination.
324 Default: 0
Linus Torvalds1da177e2005-04-16 15:20:36 -0700325
326tcp_rmem - vector of 3 INTEGERs: min, default, max
327 min: Minimal size of receive buffer used by TCP sockets.
328 It is guaranteed to each TCP socket, even under moderate memory
329 pressure.
330 Default: 8K
331
332 default: default size of receive buffer used by TCP sockets.
333 This value overrides net.core.rmem_default used by other protocols.
334 Default: 87380 bytes. This value results in window of 65535 with
335 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
336 less for default tcp_app_win. See below about these variables.
337
338 max: maximal size of receive buffer allowed for automatically
339 selected receiver buffers for TCP socket. This value does not override
340 net.core.rmem_max, "static" selection via SO_RCVBUF does not use this.
341 Default: 87380*2 bytes.
342
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800343tcp_sack - BOOLEAN
344 Enable select acknowledgments (SACKS).
Rick Jones15d99e02006-03-20 22:40:29 -0800345
David S. Miller35089bb2006-06-13 22:33:04 -0700346tcp_slow_start_after_idle - BOOLEAN
347 If set, provide RFC2861 behavior and time out the congestion
348 window after an idle period. An idle period is defined at
349 the current RTO. If unset, the congestion window will not
350 be timed out after an idle period.
351 Default: 1
352
Stephen Hemmingeref56e622006-11-09 16:37:26 -0800353tcp_stdurg - BOOLEAN
354 Use the Host requirements interpretation of the TCP urg pointer field.
355 Most hosts use the older BSD interpretation, so if you turn this on
356 Linux might not communicate correctly with them.
357 Default: FALSE
358
359tcp_synack_retries - INTEGER
360 Number of times SYNACKs for a passive TCP connection attempt will
361 be retransmitted. Should not be higher than 255. Default value
362 is 5, which corresponds to ~180seconds.
363
364tcp_syncookies - BOOLEAN
365 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES
366 Send out syncookies when the syn backlog queue of a socket
367 overflows. This is to prevent against the common 'syn flood attack'
368 Default: FALSE
369
370 Note, that syncookies is fallback facility.
371 It MUST NOT be used to help highly loaded servers to stand
372 against legal connection rate. If you see synflood warnings
373 in your logs, but investigation shows that they occur
374 because of overload with legal connections, you should tune
375 another parameters until this warning disappear.
376 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
377
378 syncookies seriously violate TCP protocol, do not allow
379 to use TCP extensions, can result in serious degradation
380 of some services (f.e. SMTP relaying), visible not by you,
381 but your clients and relays, contacting you. While you see
382 synflood warnings in logs not being really flooded, your server
383 is seriously misconfigured.
384
385tcp_syn_retries - INTEGER
386 Number of times initial SYNs for an active TCP connection attempt
387 will be retransmitted. Should not be higher than 255. Default value
388 is 5, which corresponds to ~180seconds.
389
390tcp_timestamps - BOOLEAN
391 Enable timestamps as defined in RFC1323.
392
393tcp_tso_win_divisor - INTEGER
394 This allows control over what percentage of the congestion window
395 can be consumed by a single TSO frame.
396 The setting of this parameter is a choice between burstiness and
397 building larger TSO frames.
398 Default: 3
399
400tcp_tw_recycle - BOOLEAN
401 Enable fast recycling TIME-WAIT sockets. Default value is 0.
402 It should not be changed without advice/request of technical
403 experts.
404
405tcp_tw_reuse - BOOLEAN
406 Allow to reuse TIME-WAIT sockets for new connections when it is
407 safe from protocol viewpoint. Default value is 0.
408 It should not be changed without advice/request of technical
409 experts.
410
411tcp_window_scaling - BOOLEAN
412 Enable window scaling as defined in RFC1323.
413
414tcp_wmem - vector of 3 INTEGERs: min, default, max
415 min: Amount of memory reserved for send buffers for TCP socket.
416 Each TCP socket has rights to use it due to fact of its birth.
417 Default: 4K
418
419 default: Amount of memory allowed for send buffers for TCP socket
420 by default. This value overrides net.core.wmem_default used
421 by other protocols, it is usually lower than net.core.wmem_default.
422 Default: 16K
423
424 max: Maximal amount of memory allowed for automatically selected
425 send buffers for TCP socket. This value does not override
426 net.core.wmem_max, "static" selection via SO_SNDBUF does not use this.
427 Default: 128K
428
429tcp_workaround_signed_windows - BOOLEAN
430 If set, assume no receipt of a window scaling option means the
431 remote TCP is broken and treats the window as a signed quantity.
432 If unset, assume the remote TCP is not broken even if we do
433 not receive a window scaling option from them.
434 Default: 0
435
Paul Moore8802f612006-08-03 16:45:49 -0700436CIPSOv4 Variables:
437
438cipso_cache_enable - BOOLEAN
439 If set, enable additions to and lookups from the CIPSO label mapping
440 cache. If unset, additions are ignored and lookups always result in a
441 miss. However, regardless of the setting the cache is still
442 invalidated when required when means you can safely toggle this on and
443 off and the cache will always be "safe".
444 Default: 1
445
446cipso_cache_bucket_size - INTEGER
447 The CIPSO label cache consists of a fixed size hash table with each
448 hash bucket containing a number of cache entries. This variable limits
449 the number of entries in each hash bucket; the larger the value the
450 more CIPSO label mappings that can be cached. When the number of
451 entries in a given hash bucket reaches this limit adding new entries
452 causes the oldest entry in the bucket to be removed to make room.
453 Default: 10
454
455cipso_rbm_optfmt - BOOLEAN
456 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
457 the CIPSO draft specification (see Documentation/netlabel for details).
458 This means that when set the CIPSO tag will be padded with empty
459 categories in order to make the packet data 32-bit aligned.
460 Default: 0
461
462cipso_rbm_structvalid - BOOLEAN
463 If set, do a very strict check of the CIPSO option when
464 ip_options_compile() is called. If unset, relax the checks done during
465 ip_options_compile(). Either way is "safe" as errors are caught else
466 where in the CIPSO processing code but setting this to 0 (False) should
467 result in less work (i.e. it should be faster) but could cause problems
468 with other implementations that require strict checking.
469 Default: 0
470
Linus Torvalds1da177e2005-04-16 15:20:36 -0700471IP Variables:
472
473ip_local_port_range - 2 INTEGERS
474 Defines the local port range that is used by TCP and UDP to
475 choose the local port. The first number is the first, the
476 second the last local port number. Default value depends on
477 amount of memory available on the system:
478 > 128Mb 32768-61000
479 < 128Mb 1024-4999 or even less.
480 This number defines number of active connections, which this
481 system can issue simultaneously to systems not supporting
482 TCP extensions (timestamps). With tcp_tw_recycle enabled
483 (i.e. by default) range 1024-4999 is enough to issue up to
484 2000 connections per second to systems supporting timestamps.
485
486ip_nonlocal_bind - BOOLEAN
487 If set, allows processes to bind() to non-local IP addresses,
488 which can be quite useful - but may break some applications.
489 Default: 0
490
491ip_dynaddr - BOOLEAN
492 If set non-zero, enables support for dynamic addresses.
493 If set to a non-zero value larger than 1, a kernel log
494 message will be printed when dynamic address rewriting
495 occurs.
496 Default: 0
497
498icmp_echo_ignore_all - BOOLEAN
David S. Miller7ce312462005-10-03 16:07:30 -0700499 If set non-zero, then the kernel will ignore all ICMP ECHO
500 requests sent to it.
501 Default: 0
502
Linus Torvalds1da177e2005-04-16 15:20:36 -0700503icmp_echo_ignore_broadcasts - BOOLEAN
David S. Miller7ce312462005-10-03 16:07:30 -0700504 If set non-zero, then the kernel will ignore all ICMP ECHO and
505 TIMESTAMP requests sent to it via broadcast/multicast.
506 Default: 1
Linus Torvalds1da177e2005-04-16 15:20:36 -0700507
508icmp_ratelimit - INTEGER
509 Limit the maximal rates for sending ICMP packets whose type matches
510 icmp_ratemask (see below) to specific targets.
511 0 to disable any limiting, otherwise the maximal rate in jiffies(1)
512 Default: 100
513
514icmp_ratemask - INTEGER
515 Mask made of ICMP types for which rates are being limited.
516 Significant bits: IHGFEDCBA9876543210
517 Default mask: 0000001100000011000 (6168)
518
519 Bit definitions (see include/linux/icmp.h):
520 0 Echo Reply
521 3 Destination Unreachable *
522 4 Source Quench *
523 5 Redirect
524 8 Echo Request
525 B Time Exceeded *
526 C Parameter Problem *
527 D Timestamp Request
528 E Timestamp Reply
529 F Info Request
530 G Info Reply
531 H Address Mask Request
532 I Address Mask Reply
533
534 * These are rate limited by default (see default mask above)
535
536icmp_ignore_bogus_error_responses - BOOLEAN
537 Some routers violate RFC1122 by sending bogus responses to broadcast
538 frames. Such violations are normally logged via a kernel warning.
539 If this is set to TRUE, the kernel will not give such warnings, which
540 will avoid log file clutter.
541 Default: FALSE
542
Horms95f7daf2006-02-02 17:02:25 -0800543icmp_errors_use_inbound_ifaddr - BOOLEAN
544
545 If zero, icmp error messages are sent with the primary address of
546 the exiting interface.
547
548 If non-zero, the message will be sent with the primary address of
549 the interface that received the packet that caused the icmp error.
550 This is the behaviour network many administrators will expect from
551 a router. And it can make debugging complicated network layouts
552 much easier.
553
554 Note that if no primary address exists for the interface selected,
555 then the primary address of the first non-loopback interface that
Matt LaPlanted6bc8ac2006-10-03 22:54:15 +0200556 has one will be used regardless of this setting.
Horms95f7daf2006-02-02 17:02:25 -0800557
558 Default: 0
559
Linus Torvalds1da177e2005-04-16 15:20:36 -0700560igmp_max_memberships - INTEGER
561 Change the maximum number of multicast groups we can subscribe to.
562 Default: 20
563
564conf/interface/* changes special settings per interface (where "interface" is
565 the name of your network interface)
566conf/all/* is special, changes the settings for all interfaces
567
568
569log_martians - BOOLEAN
570 Log packets with impossible addresses to kernel log.
571 log_martians for the interface will be enabled if at least one of
572 conf/{all,interface}/log_martians is set to TRUE,
573 it will be disabled otherwise
574
575accept_redirects - BOOLEAN
576 Accept ICMP redirect messages.
577 accept_redirects for the interface will be enabled if:
578 - both conf/{all,interface}/accept_redirects are TRUE in the case forwarding
579 for the interface is enabled
580 or
581 - at least one of conf/{all,interface}/accept_redirects is TRUE in the case
582 forwarding for the interface is disabled
583 accept_redirects for the interface will be disabled otherwise
584 default TRUE (host)
585 FALSE (router)
586
587forwarding - BOOLEAN
588 Enable IP forwarding on this interface.
589
590mc_forwarding - BOOLEAN
591 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
592 and a multicast routing daemon is required.
593 conf/all/mc_forwarding must also be set to TRUE to enable multicast routing
594 for the interface
595
596medium_id - INTEGER
597 Integer value used to differentiate the devices by the medium they
598 are attached to. Two devices can have different id values when
599 the broadcast packets are received only on one of them.
600 The default value 0 means that the device is the only interface
601 to its medium, value of -1 means that medium is not known.
602
603 Currently, it is used to change the proxy_arp behavior:
604 the proxy_arp feature is enabled for packets forwarded between
605 two devices attached to different media.
606
607proxy_arp - BOOLEAN
608 Do proxy arp.
609 proxy_arp for the interface will be enabled if at least one of
610 conf/{all,interface}/proxy_arp is set to TRUE,
611 it will be disabled otherwise
612
613shared_media - BOOLEAN
614 Send(router) or accept(host) RFC1620 shared media redirects.
615 Overrides ip_secure_redirects.
616 shared_media for the interface will be enabled if at least one of
617 conf/{all,interface}/shared_media is set to TRUE,
618 it will be disabled otherwise
619 default TRUE
620
621secure_redirects - BOOLEAN
622 Accept ICMP redirect messages only for gateways,
623 listed in default gateway list.
624 secure_redirects for the interface will be enabled if at least one of
625 conf/{all,interface}/secure_redirects is set to TRUE,
626 it will be disabled otherwise
627 default TRUE
628
629send_redirects - BOOLEAN
630 Send redirects, if router.
631 send_redirects for the interface will be enabled if at least one of
632 conf/{all,interface}/send_redirects is set to TRUE,
633 it will be disabled otherwise
634 Default: TRUE
635
636bootp_relay - BOOLEAN
637 Accept packets with source address 0.b.c.d destined
638 not to this host as local ones. It is supposed, that
639 BOOTP relay daemon will catch and forward such packets.
640 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
641 for the interface
642 default FALSE
643 Not Implemented Yet.
644
645accept_source_route - BOOLEAN
646 Accept packets with SRR option.
647 conf/all/accept_source_route must also be set to TRUE to accept packets
648 with SRR option on the interface
649 default TRUE (router)
650 FALSE (host)
651
652rp_filter - BOOLEAN
653 1 - do source validation by reversed path, as specified in RFC1812
654 Recommended option for single homed hosts and stub network
655 routers. Could cause troubles for complicated (not loop free)
656 networks running a slow unreliable protocol (sort of RIP),
657 or using static routes.
658
659 0 - No source validation.
660
661 conf/all/rp_filter must also be set to TRUE to do source validation
662 on the interface
663
664 Default value is 0. Note that some distributions enable it
665 in startup scripts.
666
667arp_filter - BOOLEAN
668 1 - Allows you to have multiple network interfaces on the same
669 subnet, and have the ARPs for each interface be answered
670 based on whether or not the kernel would route a packet from
671 the ARP'd IP out that interface (therefore you must use source
672 based routing for this to work). In other words it allows control
673 of which cards (usually 1) will respond to an arp request.
674
675 0 - (default) The kernel can respond to arp requests with addresses
676 from other interfaces. This may seem wrong but it usually makes
677 sense, because it increases the chance of successful communication.
678 IP addresses are owned by the complete host on Linux, not by
679 particular interfaces. Only for more complex setups like load-
680 balancing, does this behaviour cause problems.
681
682 arp_filter for the interface will be enabled if at least one of
683 conf/{all,interface}/arp_filter is set to TRUE,
684 it will be disabled otherwise
685
686arp_announce - INTEGER
687 Define different restriction levels for announcing the local
688 source IP address from IP packets in ARP requests sent on
689 interface:
690 0 - (default) Use any local address, configured on any interface
691 1 - Try to avoid local addresses that are not in the target's
692 subnet for this interface. This mode is useful when target
693 hosts reachable via this interface require the source IP
694 address in ARP requests to be part of their logical network
695 configured on the receiving interface. When we generate the
696 request we will check all our subnets that include the
697 target IP and will preserve the source address if it is from
698 such subnet. If there is no such subnet we select source
699 address according to the rules for level 2.
700 2 - Always use the best local address for this target.
701 In this mode we ignore the source address in the IP packet
702 and try to select local address that we prefer for talks with
703 the target host. Such local address is selected by looking
704 for primary IP addresses on all our subnets on the outgoing
705 interface that include the target IP address. If no suitable
706 local address is found we select the first local address
707 we have on the outgoing interface or on all other interfaces,
708 with the hope we will receive reply for our request and
709 even sometimes no matter the source IP address we announce.
710
711 The max value from conf/{all,interface}/arp_announce is used.
712
713 Increasing the restriction level gives more chance for
714 receiving answer from the resolved target while decreasing
715 the level announces more valid sender's information.
716
717arp_ignore - INTEGER
718 Define different modes for sending replies in response to
719 received ARP requests that resolve local target IP addresses:
720 0 - (default): reply for any local target IP address, configured
721 on any interface
722 1 - reply only if the target IP address is local address
723 configured on the incoming interface
724 2 - reply only if the target IP address is local address
725 configured on the incoming interface and both with the
726 sender's IP address are part from same subnet on this interface
727 3 - do not reply for local addresses configured with scope host,
728 only resolutions for global and link addresses are replied
729 4-7 - reserved
730 8 - do not reply for all local addresses
731
732 The max value from conf/{all,interface}/arp_ignore is used
733 when ARP request is received on the {interface}
734
Neil Hormanc1b1bce2006-03-20 22:40:03 -0800735arp_accept - BOOLEAN
736 Define behavior when gratuitous arp replies are received:
737 0 - drop gratuitous arp frames
738 1 - accept gratuitous arp frames
739
Linus Torvalds1da177e2005-04-16 15:20:36 -0700740app_solicit - INTEGER
741 The maximum number of probes to send to the user space ARP daemon
742 via netlink before dropping back to multicast probes (see
743 mcast_solicit). Defaults to 0.
744
745disable_policy - BOOLEAN
746 Disable IPSEC policy (SPD) for this interface
747
748disable_xfrm - BOOLEAN
749 Disable IPSEC encryption on this interface, whatever the policy
750
751
752
753tag - INTEGER
754 Allows you to write a number, which can be used as required.
755 Default value is 0.
756
757(1) Jiffie: internal timeunit for the kernel. On the i386 1/100s, on the
758Alpha 1/1024s. See the HZ define in /usr/include/asm/param.h for the exact
759value on your system.
760
761Alexey Kuznetsov.
762kuznet@ms2.inr.ac.ru
763
764Updated by:
765Andi Kleen
766ak@muc.de
767Nicolas Delon
768delon.nicolas@wanadoo.fr
769
770
771
772
773/proc/sys/net/ipv6/* Variables:
774
775IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
776apply to IPv6 [XXX?].
777
778bindv6only - BOOLEAN
779 Default value for IPV6_V6ONLY socket option,
780 which restricts use of the IPv6 socket to IPv6 communication
781 only.
782 TRUE: disable IPv4-mapped address feature
783 FALSE: enable IPv4-mapped address feature
784
785 Default: FALSE (as specified in RFC2553bis)
786
787IPv6 Fragmentation:
788
789ip6frag_high_thresh - INTEGER
790 Maximum memory used to reassemble IPv6 fragments. When
791 ip6frag_high_thresh bytes of memory is allocated for this purpose,
792 the fragment handler will toss packets until ip6frag_low_thresh
793 is reached.
794
795ip6frag_low_thresh - INTEGER
796 See ip6frag_high_thresh
797
798ip6frag_time - INTEGER
799 Time in seconds to keep an IPv6 fragment in memory.
800
801ip6frag_secret_interval - INTEGER
802 Regeneration interval (in seconds) of the hash secret (or lifetime
803 for the hash secret) for IPv6 fragments.
804 Default: 600
805
806conf/default/*:
807 Change the interface-specific default settings.
808
809
810conf/all/*:
811 Change all the interface-specific settings.
812
813 [XXX: Other special features than forwarding?]
814
815conf/all/forwarding - BOOLEAN
816 Enable global IPv6 forwarding between all interfaces.
817
818 IPv4 and IPv6 work differently here; e.g. netfilter must be used
819 to control which interfaces may forward packets and which not.
820
821 This also sets all interfaces' Host/Router setting
822 'forwarding' to the specified value. See below for details.
823
824 This referred to as global forwarding.
825
YOSHIFUJI Hideakifbea49e2006-09-22 14:43:49 -0700826proxy_ndp - BOOLEAN
827 Do proxy ndp.
828
Linus Torvalds1da177e2005-04-16 15:20:36 -0700829conf/interface/*:
830 Change special settings per interface.
831
832 The functional behaviour for certain settings is different
833 depending on whether local forwarding is enabled or not.
834
835accept_ra - BOOLEAN
836 Accept Router Advertisements; autoconfigure using them.
837
838 Functional default: enabled if local forwarding is disabled.
839 disabled if local forwarding is enabled.
840
YOSHIFUJI Hideaki65f5c7c2006-03-20 16:55:08 -0800841accept_ra_defrtr - BOOLEAN
842 Learn default router in Router Advertisement.
843
844 Functional default: enabled if accept_ra is enabled.
845 disabled if accept_ra is disabled.
846
YOSHIFUJI Hideakic4fd30e2006-03-20 16:55:26 -0800847accept_ra_pinfo - BOOLEAN
Matt LaPlante2fe0ae72006-10-03 22:50:39 +0200848 Learn Prefix Information in Router Advertisement.
YOSHIFUJI Hideakic4fd30e2006-03-20 16:55:26 -0800849
850 Functional default: enabled if accept_ra is enabled.
851 disabled if accept_ra is disabled.
852
YOSHIFUJI Hideaki09c884d2006-03-20 17:07:03 -0800853accept_ra_rt_info_max_plen - INTEGER
854 Maximum prefix length of Route Information in RA.
855
856 Route Information w/ prefix larger than or equal to this
857 variable shall be ignored.
858
859 Functional default: 0 if accept_ra_rtr_pref is enabled.
860 -1 if accept_ra_rtr_pref is disabled.
861
YOSHIFUJI Hideaki930d6ff2006-03-20 17:05:30 -0800862accept_ra_rtr_pref - BOOLEAN
863 Accept Router Preference in RA.
864
865 Functional default: enabled if accept_ra is enabled.
866 disabled if accept_ra is disabled.
867
Linus Torvalds1da177e2005-04-16 15:20:36 -0700868accept_redirects - BOOLEAN
869 Accept Redirects.
870
871 Functional default: enabled if local forwarding is disabled.
872 disabled if local forwarding is enabled.
873
YOSHIFUJI Hideaki0bcbc922007-04-24 14:58:30 -0700874accept_source_route - INTEGER
875 Accept source routing (routing extension header).
876
877 > 0: Accept routing header.
878 = 0: Accept only routing header type 2.
879 < 0: Do not accept routing header.
880
881 Default: 0
882
Linus Torvalds1da177e2005-04-16 15:20:36 -0700883autoconf - BOOLEAN
884 Autoconfigure addresses using Prefix Information in Router
885 Advertisements.
886
YOSHIFUJI Hideakic4fd30e2006-03-20 16:55:26 -0800887 Functional default: enabled if accept_ra_pinfo is enabled.
888 disabled if accept_ra_pinfo is disabled.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700889
890dad_transmits - INTEGER
891 The amount of Duplicate Address Detection probes to send.
892 Default: 1
893
894forwarding - BOOLEAN
895 Configure interface-specific Host/Router behaviour.
896
897 Note: It is recommended to have the same setting on all
898 interfaces; mixed router/host scenarios are rather uncommon.
899
900 FALSE:
901
902 By default, Host behaviour is assumed. This means:
903
904 1. IsRouter flag is not set in Neighbour Advertisements.
905 2. Router Solicitations are being sent when necessary.
906 3. If accept_ra is TRUE (default), accept Router
907 Advertisements (and do autoconfiguration).
908 4. If accept_redirects is TRUE (default), accept Redirects.
909
910 TRUE:
911
912 If local forwarding is enabled, Router behaviour is assumed.
913 This means exactly the reverse from the above:
914
915 1. IsRouter flag is set in Neighbour Advertisements.
916 2. Router Solicitations are not sent.
917 3. Router Advertisements are ignored.
918 4. Redirects are ignored.
919
920 Default: FALSE if global forwarding is disabled (default),
921 otherwise TRUE.
922
923hop_limit - INTEGER
924 Default Hop Limit to set.
925 Default: 64
926
927mtu - INTEGER
928 Default Maximum Transfer Unit
929 Default: 1280 (IPv6 required minimum)
930
YOSHIFUJI Hideaki52e16352006-03-20 17:05:47 -0800931router_probe_interval - INTEGER
932 Minimum interval (in seconds) between Router Probing described
933 in RFC4191.
934
935 Default: 60
936
Linus Torvalds1da177e2005-04-16 15:20:36 -0700937router_solicitation_delay - INTEGER
938 Number of seconds to wait after interface is brought up
939 before sending Router Solicitations.
940 Default: 1
941
942router_solicitation_interval - INTEGER
943 Number of seconds to wait between Router Solicitations.
944 Default: 4
945
946router_solicitations - INTEGER
947 Number of Router Solicitations to send until assuming no
948 routers are present.
949 Default: 3
950
951use_tempaddr - INTEGER
952 Preference for Privacy Extensions (RFC3041).
953 <= 0 : disable Privacy Extensions
954 == 1 : enable Privacy Extensions, but prefer public
955 addresses over temporary addresses.
956 > 1 : enable Privacy Extensions and prefer temporary
957 addresses over public addresses.
958 Default: 0 (for most devices)
959 -1 (for point-to-point devices and loopback devices)
960
961temp_valid_lft - INTEGER
962 valid lifetime (in seconds) for temporary addresses.
963 Default: 604800 (7 days)
964
965temp_prefered_lft - INTEGER
966 Preferred lifetime (in seconds) for temporary addresses.
967 Default: 86400 (1 day)
968
969max_desync_factor - INTEGER
970 Maximum value for DESYNC_FACTOR, which is a random value
971 that ensures that clients don't synchronize with each
972 other and generate new addresses at exactly the same time.
973 value is in seconds.
974 Default: 600
975
976regen_max_retry - INTEGER
977 Number of attempts before give up attempting to generate
978 valid temporary addresses.
979 Default: 5
980
981max_addresses - INTEGER
982 Number of maximum addresses per interface. 0 disables limitation.
983 It is recommended not set too large value (or 0) because it would
984 be too easy way to crash kernel to allow to create too much of
985 autoconfigured addresses.
986 Default: 16
987
988icmp/*:
989ratelimit - INTEGER
990 Limit the maximal rates for sending ICMPv6 packets.
991 0 to disable any limiting, otherwise the maximal rate in jiffies(1)
992 Default: 100
993
994
995IPv6 Update by:
996Pekka Savola <pekkas@netcore.fi>
997YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
998
999
1000/proc/sys/net/bridge/* Variables:
1001
1002bridge-nf-call-arptables - BOOLEAN
1003 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1004 0 : disable this.
1005 Default: 1
1006
1007bridge-nf-call-iptables - BOOLEAN
1008 1 : pass bridged IPv4 traffic to iptables' chains.
1009 0 : disable this.
1010 Default: 1
1011
1012bridge-nf-call-ip6tables - BOOLEAN
1013 1 : pass bridged IPv6 traffic to ip6tables' chains.
1014 0 : disable this.
1015 Default: 1
1016
1017bridge-nf-filter-vlan-tagged - BOOLEAN
Michael Milner516299d2007-04-12 22:14:23 -07001018 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1019 0 : disable this.
1020 Default: 1
1021
1022bridge-nf-filter-pppoe-tagged - BOOLEAN
1023 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001024 0 : disable this.
1025 Default: 1
1026
1027
1028UNDOCUMENTED:
1029
1030dev_weight FIXME
1031discovery_slots FIXME
1032discovery_timeout FIXME
1033fast_poll_increase FIXME
1034ip6_queue_maxlen FIXME
1035lap_keepalive_time FIXME
1036lo_cong FIXME
1037max_baud_rate FIXME
1038max_dgram_qlen FIXME
1039max_noreply_time FIXME
1040max_tx_data_size FIXME
1041max_tx_window FIXME
1042min_tx_turn_time FIXME
1043mod_cong FIXME
1044no_cong FIXME
1045no_cong_thresh FIXME
1046slot_timeout FIXME
1047warn_noreply_time FIXME
1048