| blob | b121c5db707fcf8708533acd7b73a51cbd286296 |
1 /proc/sys/net/ipv4/* Variables:
3 ip_forward - BOOLEAN
4 0 - disabled (default)
5 not 0 - enabled
7 Forward Packets between interfaces.
9 This variable is special, its change resets all configuration
10 parameters to their default state (RFC1122 for hosts, RFC1812
11 for routers)
13 ip_default_ttl - INTEGER
14 default 64
16 ip_no_pmtu_disc - BOOLEAN
17 Disable Path MTU Discovery.
18 default FALSE
20 min_pmtu - INTEGER
21 default 562 - minimum discovered Path MTU
23 mtu_expires - INTEGER
24 Time, in seconds, that cached PMTU information is kept.
26 min_adv_mss - INTEGER
27 The advertised MSS depends on the first hop route MTU, but will
28 never be lower than this setting.
30 rt_cache_rebuild_count - INTEGER
31 The per net-namespace route cache emergency rebuild threshold.
32 Any net-namespace having its route cache rebuilt due to
33 a hash bucket chain being too long more than this many times
34 will have its route caching disabled
36 IP Fragmentation:
38 ipfrag_high_thresh - INTEGER
39 Maximum memory used to reassemble IP fragments. When
40 ipfrag_high_thresh bytes of memory is allocated for this purpose,
41 the fragment handler will toss packets until ipfrag_low_thresh
42 is reached.
44 ipfrag_low_thresh - INTEGER
45 See ipfrag_high_thresh
47 ipfrag_time - INTEGER
48 Time in seconds to keep an IP fragment in memory.
50 ipfrag_secret_interval - INTEGER
51 Regeneration interval (in seconds) of the hash secret (or lifetime
52 for the hash secret) for IP fragments.
53 Default: 600
55 ipfrag_max_dist - INTEGER
56 ipfrag_max_dist is a non-negative integer value which defines the
57 maximum "disorder" which is allowed among fragments which share a
58 common IP source address. Note that reordering of packets is
59 not unusual, but if a large number of fragments arrive from a source
60 IP address while a particular fragment queue remains incomplete, it
61 probably indicates that one or more fragments belonging to that queue
62 have been lost. When ipfrag_max_dist is positive, an additional check
63 is done on fragments before they are added to a reassembly queue - if
64 ipfrag_max_dist (or more) fragments have arrived from a particular IP
65 address between additions to any IP fragment queue using that source
66 address, it's presumed that one or more fragments in the queue are
67 lost. The existing fragment queue will be dropped, and a new one
68 started. An ipfrag_max_dist value of zero disables this check.
70 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
71 result in unnecessarily dropping fragment queues when normal
72 reordering of packets occurs, which could lead to poor application
73 performance. Using a very large value, e.g. 50000, increases the
74 likelihood of incorrectly reassembling IP fragments that originate
75 from different IP datagrams, which could result in data corruption.
76 Default: 64
78 INET peer storage:
80 inet_peer_threshold - INTEGER
81 The approximate size of the storage. Starting from this threshold
82 entries will be thrown aggressively. This threshold also determines
83 entries' time-to-live and time intervals between garbage collection
84 passes. More entries, less time-to-live, less GC interval.
86 inet_peer_minttl - INTEGER
87 Minimum time-to-live of entries. Should be enough to cover fragment
88 time-to-live on the reassembling side. This minimum time-to-live is
89 guaranteed if the pool size is less than inet_peer_threshold.
90 Measured in seconds.
92 inet_peer_maxttl - INTEGER
93 Maximum time-to-live of entries. Unused entries will expire after
94 this period of time if there is no memory pressure on the pool (i.e.
95 when the number of entries in the pool is very small).
96 Measured in seconds.
98 inet_peer_gc_mintime - INTEGER
99 Minimum interval between garbage collection passes. This interval is
100 in effect under high memory pressure on the pool.
101 Measured in seconds.
103 inet_peer_gc_maxtime - INTEGER
104 Minimum interval between garbage collection passes. This interval is
105 in effect under low (or absent) memory pressure on the pool.
106 Measured in seconds.
108 TCP variables:
110 somaxconn - INTEGER
111 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
112 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
113 for TCP sockets.
115 tcp_abc - INTEGER
116 Controls Appropriate Byte Count (ABC) defined in RFC3465.
117 ABC is a way of increasing congestion window (cwnd) more slowly
118 in response to partial acknowledgments.
119 Possible values are:
120 0 increase cwnd once per acknowledgment (no ABC)
121 1 increase cwnd once per acknowledgment of full sized segment
122 2 allow increase cwnd by two if acknowledgment is
123 of two segments to compensate for delayed acknowledgments.
124 Default: 0 (off)
126 tcp_abort_on_overflow - BOOLEAN
127 If listening service is too slow to accept new connections,
128 reset them. Default state is FALSE. It means that if overflow
129 occurred due to a burst, connection will recover. Enable this
130 option _only_ if you are really sure that listening daemon
131 cannot be tuned to accept connections faster. Enabling this
132 option can harm clients of your server.
134 tcp_adv_win_scale - INTEGER
135 Count buffering overhead as bytes/2^tcp_adv_win_scale
136 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
137 if it is <= 0.
138 Default: 2
140 tcp_allowed_congestion_control - STRING
141 Show/set the congestion control choices available to non-privileged
142 processes. The list is a subset of those listed in
143 tcp_available_congestion_control.
144 Default is "reno" and the default setting (tcp_congestion_control).
146 tcp_app_win - INTEGER
147 Reserve max(window/2^tcp_app_win, mss) of window for application
148 buffer. Value 0 is special, it means that nothing is reserved.
149 Default: 31
151 tcp_available_congestion_control - STRING
152 Shows the available congestion control choices that are registered.
153 More congestion control algorithms may be available as modules,
154 but not loaded.
156 tcp_base_mss - INTEGER
157 The initial value of search_low to be used by the packetization layer
158 Path MTU discovery (MTU probing). If MTU probing is enabled,
159 this is the initial MSS used by the connection.
161 tcp_congestion_control - STRING
162 Set the congestion control algorithm to be used for new
163 connections. The algorithm "reno" is always available, but
164 additional choices may be available based on kernel configuration.
165 Default is set as part of kernel configuration.
167 tcp_dsack - BOOLEAN
168 Allows TCP to send "duplicate" SACKs.
170 tcp_ecn - BOOLEAN
171 Enable Explicit Congestion Notification in TCP.
173 tcp_fack - BOOLEAN
174 Enable FACK congestion avoidance and fast retransmission.
175 The value is not used, if tcp_sack is not enabled.
177 tcp_fin_timeout - INTEGER
178 Time to hold socket in state FIN-WAIT-2, if it was closed
179 by our side. Peer can be broken and never close its side,
180 or even died unexpectedly. Default value is 60sec.
181 Usual value used in 2.2 was 180 seconds, you may restore
182 it, but remember that if your machine is even underloaded WEB server,
183 you risk to overflow memory with kilotons of dead sockets,
184 FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1,
185 because they eat maximum 1.5K of memory, but they tend
186 to live longer. Cf. tcp_max_orphans.
188 tcp_frto - INTEGER
189 Enables Forward RTO-Recovery (F-RTO) defined in RFC4138.
190 F-RTO is an enhanced recovery algorithm for TCP retransmission
191 timeouts. It is particularly beneficial in wireless environments
192 where packet loss is typically due to random radio interference
193 rather than intermediate router congestion. F-RTO is sender-side
194 only modification. Therefore it does not require any support from
195 the peer.
197 If set to 1, basic version is enabled. 2 enables SACK enhanced
198 F-RTO if flow uses SACK. The basic version can be used also when
199 SACK is in use though scenario(s) with it exists where F-RTO
200 interacts badly with the packet counting of the SACK enabled TCP
201 flow.
203 tcp_frto_response - INTEGER
204 When F-RTO has detected that a TCP retransmission timeout was
205 spurious (i.e, the timeout would have been avoided had TCP set a
206 longer retransmission timeout), TCP has several options what to do
207 next. Possible values are:
208 0 Rate halving based; a smooth and conservative response,
209 results in halved cwnd and ssthresh after one RTT
210 1 Very conservative response; not recommended because even
211 though being valid, it interacts poorly with the rest of
212 Linux TCP, halves cwnd and ssthresh immediately
213 2 Aggressive response; undoes congestion control measures
214 that are now known to be unnecessary (ignoring the
215 possibility of a lost retransmission that would require
216 TCP to be more cautious), cwnd and ssthresh are restored
217 to the values prior timeout
218 Default: 0 (rate halving based)
220 tcp_keepalive_time - INTEGER
221 How often TCP sends out keepalive messages when keepalive is enabled.
222 Default: 2hours.
224 tcp_keepalive_probes - INTEGER
225 How many keepalive probes TCP sends out, until it decides that the
226 connection is broken. Default value: 9.
228 tcp_keepalive_intvl - INTEGER
229 How frequently the probes are send out. Multiplied by
230 tcp_keepalive_probes it is time to kill not responding connection,
231 after probes started. Default value: 75sec i.e. connection
232 will be aborted after ~11 minutes of retries.
234 tcp_low_latency - BOOLEAN
235 If set, the TCP stack makes decisions that prefer lower
236 latency as opposed to higher throughput. By default, this
237 option is not set meaning that higher throughput is preferred.
238 An example of an application where this default should be
239 changed would be a Beowulf compute cluster.
240 Default: 0
242 tcp_max_orphans - INTEGER
243 Maximal number of TCP sockets not attached to any user file handle,
244 held by system. If this number is exceeded orphaned connections are
245 reset immediately and warning is printed. This limit exists
246 only to prevent simple DoS attacks, you _must_ not rely on this
247 or lower the limit artificially, but rather increase it
248 (probably, after increasing installed memory),
249 if network conditions require more than default value,
250 and tune network services to linger and kill such states
251 more aggressively. Let me to remind again: each orphan eats
252 up to ~64K of unswappable memory.
254 tcp_max_syn_backlog - INTEGER
255 Maximal number of remembered connection requests, which are
256 still did not receive an acknowledgment from connecting client.
257 Default value is 1024 for systems with more than 128Mb of memory,
258 and 128 for low memory machines. If server suffers of overload,
259 try to increase this number.
261 tcp_max_tw_buckets - INTEGER
262 Maximal number of timewait sockets held by system simultaneously.
263 If this number is exceeded time-wait socket is immediately destroyed
264 and warning is printed. This limit exists only to prevent
265 simple DoS attacks, you _must_ not lower the limit artificially,
266 but rather increase it (probably, after increasing installed memory),
267 if network conditions require more than default value.
269 tcp_mem - vector of 3 INTEGERs: min, pressure, max
270 min: below this number of pages TCP is not bothered about its
271 memory appetite.
273 pressure: when amount of memory allocated by TCP exceeds this number
274 of pages, TCP moderates its memory consumption and enters memory
275 pressure mode, which is exited when memory consumption falls
276 under "min".
278 max: number of pages allowed for queueing by all TCP sockets.
280 Defaults are calculated at boot time from amount of available
281 memory.
283 tcp_moderate_rcvbuf - BOOLEAN
284 If set, TCP performs receive buffer auto-tuning, attempting to
285 automatically size the buffer (no greater than tcp_rmem[2]) to
286 match the size required by the path for full throughput. Enabled by
287 default.
289 tcp_mtu_probing - INTEGER
290 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
291 values:
292 0 - Disabled
293 1 - Disabled by default, enabled when an ICMP black hole detected
294 2 - Always enabled, use initial MSS of tcp_base_mss.
296 tcp_no_metrics_save - BOOLEAN
297 By default, TCP saves various connection metrics in the route cache
298 when the connection closes, so that connections established in the
299 near future can use these to set initial conditions. Usually, this
300 increases overall performance, but may sometimes cause performance
301 degradation. If set, TCP will not cache metrics on closing
302 connections.
304 tcp_orphan_retries - INTEGER
305 How may times to retry before killing TCP connection, closed
306 by our side. Default value 7 corresponds to ~50sec-16min
307 depending on RTO. If you machine is loaded WEB server,
308 you should think about lowering this value, such sockets
309 may consume significant resources. Cf. tcp_max_orphans.
311 tcp_reordering - INTEGER
312 Maximal reordering of packets in a TCP stream.
313 Default: 3
315 tcp_retrans_collapse - BOOLEAN
316 Bug-to-bug compatibility with some broken printers.
317 On retransmit try to send bigger packets to work around bugs in
318 certain TCP stacks.
320 tcp_retries1 - INTEGER
321 How many times to retry before deciding that something is wrong
322 and it is necessary to report this suspicion to network layer.
323 Minimal RFC value is 3, it is default, which corresponds
324 to ~3sec-8min depending on RTO.
326 tcp_retries2 - INTEGER
327 How may times to retry before killing alive TCP connection.
328 RFC1122 says that the limit should be longer than 100 sec.
329 It is too small number. Default value 15 corresponds to ~13-30min
330 depending on RTO.
332 tcp_rfc1337 - BOOLEAN
333 If set, the TCP stack behaves conforming to RFC1337. If unset,
334 we are not conforming to RFC, but prevent TCP TIME_WAIT
335 assassination.
336 Default: 0
338 tcp_rmem - vector of 3 INTEGERs: min, default, max
339 min: Minimal size of receive buffer used by TCP sockets.
340 It is guaranteed to each TCP socket, even under moderate memory
341 pressure.
342 Default: 8K
344 default: initial size of receive buffer used by TCP sockets.
345 This value overrides net.core.rmem_default used by other protocols.
346 Default: 87380 bytes. This value results in window of 65535 with
347 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
348 less for default tcp_app_win. See below about these variables.
350 max: maximal size of receive buffer allowed for automatically
351 selected receiver buffers for TCP socket. This value does not override
352 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
353 automatic tuning of that socket's receive buffer size, in which
354 case this value is ignored.
355 Default: between 87380B and 4MB, depending on RAM size.
357 tcp_sack - BOOLEAN
358 Enable select acknowledgments (SACKS).
360 tcp_slow_start_after_idle - BOOLEAN
361 If set, provide RFC2861 behavior and time out the congestion
362 window after an idle period. An idle period is defined at
363 the current RTO. If unset, the congestion window will not
364 be timed out after an idle period.
365 Default: 1
367 tcp_stdurg - BOOLEAN
368 Use the Host requirements interpretation of the TCP urgent pointer field.
369 Most hosts use the older BSD interpretation, so if you turn this on
370 Linux might not communicate correctly with them.
371 Default: FALSE
373 tcp_synack_retries - INTEGER
374 Number of times SYNACKs for a passive TCP connection attempt will
375 be retransmitted. Should not be higher than 255. Default value
376 is 5, which corresponds to ~180seconds.
378 tcp_syncookies - BOOLEAN
379 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES
380 Send out syncookies when the syn backlog queue of a socket
381 overflows. This is to prevent against the common 'SYN flood attack'
382 Default: FALSE
384 Note, that syncookies is fallback facility.
385 It MUST NOT be used to help highly loaded servers to stand
386 against legal connection rate. If you see SYN flood warnings
387 in your logs, but investigation shows that they occur
388 because of overload with legal connections, you should tune
389 another parameters until this warning disappear.
390 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
392 syncookies seriously violate TCP protocol, do not allow
393 to use TCP extensions, can result in serious degradation
394 of some services (f.e. SMTP relaying), visible not by you,
395 but your clients and relays, contacting you. While you see
396 SYN flood warnings in logs not being really flooded, your server
397 is seriously misconfigured.
399 tcp_syn_retries - INTEGER
400 Number of times initial SYNs for an active TCP connection attempt
401 will be retransmitted. Should not be higher than 255. Default value
402 is 5, which corresponds to ~180seconds.
404 tcp_timestamps - BOOLEAN
405 Enable timestamps as defined in RFC1323.
407 tcp_tso_win_divisor - INTEGER
408 This allows control over what percentage of the congestion window
409 can be consumed by a single TSO frame.
410 The setting of this parameter is a choice between burstiness and
411 building larger TSO frames.
412 Default: 3
414 tcp_tw_recycle - BOOLEAN
415 Enable fast recycling TIME-WAIT sockets. Default value is 0.
416 It should not be changed without advice/request of technical
417 experts.
419 tcp_tw_reuse - BOOLEAN
420 Allow to reuse TIME-WAIT sockets for new connections when it is
421 safe from protocol viewpoint. Default value is 0.
422 It should not be changed without advice/request of technical
423 experts.
425 tcp_window_scaling - BOOLEAN
426 Enable window scaling as defined in RFC1323.
428 tcp_wmem - vector of 3 INTEGERs: min, default, max
429 min: Amount of memory reserved for send buffers for TCP sockets.
430 Each TCP socket has rights to use it due to fact of its birth.
431 Default: 4K
433 default: initial size of send buffer used by TCP sockets. This
434 value overrides net.core.wmem_default used by other protocols.
435 It is usually lower than net.core.wmem_default.
436 Default: 16K
438 max: Maximal amount of memory allowed for automatically tuned
439 send buffers for TCP sockets. This value does not override
440 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
441 automatic tuning of that socket's send buffer size, in which case
442 this value is ignored.
443 Default: between 64K and 4MB, depending on RAM size.
445 tcp_workaround_signed_windows - BOOLEAN
446 If set, assume no receipt of a window scaling option means the
447 remote TCP is broken and treats the window as a signed quantity.
448 If unset, assume the remote TCP is not broken even if we do
449 not receive a window scaling option from them.
450 Default: 0
452 tcp_dma_copybreak - INTEGER
453 Lower limit, in bytes, of the size of socket reads that will be
454 offloaded to a DMA copy engine, if one is present in the system
455 and CONFIG_NET_DMA is enabled.
456 Default: 4096
458 UDP variables:
460 udp_mem - vector of 3 INTEGERs: min, pressure, max
461 Number of pages allowed for queueing by all UDP sockets.
463 min: Below this number of pages UDP is not bothered about its
464 memory appetite. When amount of memory allocated by UDP exceeds
465 this number, UDP starts to moderate memory usage.
467 pressure: This value was introduced to follow format of tcp_mem.
469 max: Number of pages allowed for queueing by all UDP sockets.
471 Default is calculated at boot time from amount of available memory.
473 udp_rmem_min - INTEGER
474 Minimal size of receive buffer used by UDP sockets in moderation.
475 Each UDP socket is able to use the size for receiving data, even if
476 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
477 Default: 4096
479 udp_wmem_min - INTEGER
480 Minimal size of send buffer used by UDP sockets in moderation.
481 Each UDP socket is able to use the size for sending data, even if
482 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
483 Default: 4096
485 CIPSOv4 Variables:
487 cipso_cache_enable - BOOLEAN
488 If set, enable additions to and lookups from the CIPSO label mapping
489 cache. If unset, additions are ignored and lookups always result in a
490 miss. However, regardless of the setting the cache is still
491 invalidated when required when means you can safely toggle this on and
492 off and the cache will always be "safe".
493 Default: 1
495 cipso_cache_bucket_size - INTEGER
496 The CIPSO label cache consists of a fixed size hash table with each
497 hash bucket containing a number of cache entries. This variable limits
498 the number of entries in each hash bucket; the larger the value the
499 more CIPSO label mappings that can be cached. When the number of
500 entries in a given hash bucket reaches this limit adding new entries
501 causes the oldest entry in the bucket to be removed to make room.
502 Default: 10
504 cipso_rbm_optfmt - BOOLEAN
505 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
506 the CIPSO draft specification (see Documentation/netlabel for details).
507 This means that when set the CIPSO tag will be padded with empty
508 categories in order to make the packet data 32-bit aligned.
509 Default: 0
511 cipso_rbm_structvalid - BOOLEAN
512 If set, do a very strict check of the CIPSO option when
513 ip_options_compile() is called. If unset, relax the checks done during
514 ip_options_compile(). Either way is "safe" as errors are caught else
515 where in the CIPSO processing code but setting this to 0 (False) should
516 result in less work (i.e. it should be faster) but could cause problems
517 with other implementations that require strict checking.
518 Default: 0
520 IP Variables:
522 ip_local_port_range - 2 INTEGERS
523 Defines the local port range that is used by TCP and UDP to
524 choose the local port. The first number is the first, the
525 second the last local port number. Default value depends on
526 amount of memory available on the system:
527 > 128Mb 32768-61000
528 < 128Mb 1024-4999 or even less.
529 This number defines number of active connections, which this
530 system can issue simultaneously to systems not supporting
531 TCP extensions (timestamps). With tcp_tw_recycle enabled
532 (i.e. by default) range 1024-4999 is enough to issue up to
533 2000 connections per second to systems supporting timestamps.
535 ip_nonlocal_bind - BOOLEAN
536 If set, allows processes to bind() to non-local IP addresses,
537 which can be quite useful - but may break some applications.
538 Default: 0
540 ip_dynaddr - BOOLEAN
541 If set non-zero, enables support for dynamic addresses.
542 If set to a non-zero value larger than 1, a kernel log
543 message will be printed when dynamic address rewriting
544 occurs.
545 Default: 0
547 icmp_echo_ignore_all - BOOLEAN
548 If set non-zero, then the kernel will ignore all ICMP ECHO
549 requests sent to it.
550 Default: 0
552 icmp_echo_ignore_broadcasts - BOOLEAN
553 If set non-zero, then the kernel will ignore all ICMP ECHO and
554 TIMESTAMP requests sent to it via broadcast/multicast.
555 Default: 1
557 icmp_ratelimit - INTEGER
558 Limit the maximal rates for sending ICMP packets whose type matches
559 icmp_ratemask (see below) to specific targets.
560 0 to disable any limiting,
561 otherwise the minimal space between responses in milliseconds.
562 Default: 1000
564 icmp_ratemask - INTEGER
565 Mask made of ICMP types for which rates are being limited.
566 Significant bits: IHGFEDCBA9876543210
567 Default mask: 0000001100000011000 (6168)
569 Bit definitions (see include/linux/icmp.h):
570 0 Echo Reply
571 3 Destination Unreachable *
572 4 Source Quench *
573 5 Redirect
574 8 Echo Request
575 B Time Exceeded *
576 C Parameter Problem *
577 D Timestamp Request
578 E Timestamp Reply
579 F Info Request
580 G Info Reply
581 H Address Mask Request
582 I Address Mask Reply
584 * These are rate limited by default (see default mask above)
586 icmp_ignore_bogus_error_responses - BOOLEAN
587 Some routers violate RFC1122 by sending bogus responses to broadcast
588 frames. Such violations are normally logged via a kernel warning.
589 If this is set to TRUE, the kernel will not give such warnings, which
590 will avoid log file clutter.
591 Default: FALSE
593 icmp_errors_use_inbound_ifaddr - BOOLEAN
595 If zero, icmp error messages are sent with the primary address of
596 the exiting interface.
598 If non-zero, the message will be sent with the primary address of
599 the interface that received the packet that caused the icmp error.
600 This is the behaviour network many administrators will expect from
601 a router. And it can make debugging complicated network layouts
602 much easier.
604 Note that if no primary address exists for the interface selected,
605 then the primary address of the first non-loopback interface that
606 has one will be used regardless of this setting.
608 Default: 0
610 igmp_max_memberships - INTEGER
611 Change the maximum number of multicast groups we can subscribe to.
612 Default: 20
614 conf/interface/* changes special settings per interface (where "interface" is
615 the name of your network interface)
616 conf/all/* is special, changes the settings for all interfaces
619 log_martians - BOOLEAN
620 Log packets with impossible addresses to kernel log.
621 log_martians for the interface will be enabled if at least one of
622 conf/{all,interface}/log_martians is set to TRUE,
623 it will be disabled otherwise
625 accept_redirects - BOOLEAN
626 Accept ICMP redirect messages.
627 accept_redirects for the interface will be enabled if:
628 - both conf/{all,interface}/accept_redirects are TRUE in the case
629 forwarding for the interface is enabled
630 or
631 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
632 case forwarding for the interface is disabled
633 accept_redirects for the interface will be disabled otherwise
634 default TRUE (host)
635 FALSE (router)
637 forwarding - BOOLEAN
638 Enable IP forwarding on this interface.
640 mc_forwarding - BOOLEAN
641 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
642 and a multicast routing daemon is required.
643 conf/all/mc_forwarding must also be set to TRUE to enable multicast
644 routing for the interface
646 medium_id - INTEGER
647 Integer value used to differentiate the devices by the medium they
648 are attached to. Two devices can have different id values when
649 the broadcast packets are received only on one of them.
650 The default value 0 means that the device is the only interface
651 to its medium, value of -1 means that medium is not known.
653 Currently, it is used to change the proxy_arp behavior:
654 the proxy_arp feature is enabled for packets forwarded between
655 two devices attached to different media.
657 proxy_arp - BOOLEAN
658 Do proxy arp.
659 proxy_arp for the interface will be enabled if at least one of
660 conf/{all,interface}/proxy_arp is set to TRUE,
661 it will be disabled otherwise
663 shared_media - BOOLEAN
664 Send(router) or accept(host) RFC1620 shared media redirects.
665 Overrides ip_secure_redirects.
666 shared_media for the interface will be enabled if at least one of
667 conf/{all,interface}/shared_media is set to TRUE,
668 it will be disabled otherwise
669 default TRUE
671 secure_redirects - BOOLEAN
672 Accept ICMP redirect messages only for gateways,
673 listed in default gateway list.
674 secure_redirects for the interface will be enabled if at least one of
675 conf/{all,interface}/secure_redirects is set to TRUE,
676 it will be disabled otherwise
677 default TRUE
679 send_redirects - BOOLEAN
680 Send redirects, if router.
681 send_redirects for the interface will be enabled if at least one of
682 conf/{all,interface}/send_redirects is set to TRUE,
683 it will be disabled otherwise
684 Default: TRUE
686 bootp_relay - BOOLEAN
687 Accept packets with source address 0.b.c.d destined
688 not to this host as local ones. It is supposed, that
689 BOOTP relay daemon will catch and forward such packets.
690 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
691 for the interface
692 default FALSE
693 Not Implemented Yet.
695 accept_source_route - BOOLEAN
696 Accept packets with SRR option.
697 conf/all/accept_source_route must also be set to TRUE to accept packets
698 with SRR option on the interface
699 default TRUE (router)
700 FALSE (host)
702 rp_filter - INTEGER
703 0 - No source validation.
704 1 - Strict mode as defined in RFC3704 Strict Reverse Path
705 Each incoming packet is tested against the FIB and if the interface
706 is not the best reverse path the packet check will fail.
707 By default failed packets are discarded.
708 2 - Loose mode as defined in RFC3704 Loose Reverse Path
709 Each incoming packet's source address is also tested against the FIB
710 and if the source address is not reachable via any interface
711 the packet check will fail.
713 Current recommended practice in RFC3704 is to enable strict mode
714 to prevent IP spoofing from DDos attacks. If using asymmetric routing
715 or other complicated routing, then loose mode is recommended.
717 conf/all/rp_filter must also be set to non-zero to do source validation
718 on the interface
720 Default value is 0. Note that some distributions enable it
721 in startup scripts.
723 arp_filter - BOOLEAN
724 1 - Allows you to have multiple network interfaces on the same
725 subnet, and have the ARPs for each interface be answered
726 based on whether or not the kernel would route a packet from
727 the ARP'd IP out that interface (therefore you must use source
728 based routing for this to work). In other words it allows control
729 of which cards (usually 1) will respond to an arp request.
731 0 - (default) The kernel can respond to arp requests with addresses
732 from other interfaces. This may seem wrong but it usually makes
733 sense, because it increases the chance of successful communication.
734 IP addresses are owned by the complete host on Linux, not by
735 particular interfaces. Only for more complex setups like load-
736 balancing, does this behaviour cause problems.
738 arp_filter for the interface will be enabled if at least one of
739 conf/{all,interface}/arp_filter is set to TRUE,
740 it will be disabled otherwise
742 arp_announce - INTEGER
743 Define different restriction levels for announcing the local
744 source IP address from IP packets in ARP requests sent on
745 interface:
746 0 - (default) Use any local address, configured on any interface
747 1 - Try to avoid local addresses that are not in the target's
748 subnet for this interface. This mode is useful when target
749 hosts reachable via this interface require the source IP
750 address in ARP requests to be part of their logical network
751 configured on the receiving interface. When we generate the
752 request we will check all our subnets that include the
753 target IP and will preserve the source address if it is from
754 such subnet. If there is no such subnet we select source
755 address according to the rules for level 2.
756 2 - Always use the best local address for this target.
757 In this mode we ignore the source address in the IP packet
758 and try to select local address that we prefer for talks with
759 the target host. Such local address is selected by looking
760 for primary IP addresses on all our subnets on the outgoing
761 interface that include the target IP address. If no suitable
762 local address is found we select the first local address
763 we have on the outgoing interface or on all other interfaces,
764 with the hope we will receive reply for our request and
765 even sometimes no matter the source IP address we announce.
767 The max value from conf/{all,interface}/arp_announce is used.
769 Increasing the restriction level gives more chance for
770 receiving answer from the resolved target while decreasing
771 the level announces more valid sender's information.
773 arp_ignore - INTEGER
774 Define different modes for sending replies in response to
775 received ARP requests that resolve local target IP addresses:
776 0 - (default): reply for any local target IP address, configured
777 on any interface
778 1 - reply only if the target IP address is local address
779 configured on the incoming interface
780 2 - reply only if the target IP address is local address
781 configured on the incoming interface and both with the
782 sender's IP address are part from same subnet on this interface
783 3 - do not reply for local addresses configured with scope host,
784 only resolutions for global and link addresses are replied
785 4-7 - reserved
786 8 - do not reply for all local addresses
788 The max value from conf/{all,interface}/arp_ignore is used
789 when ARP request is received on the {interface}
791 arp_notify - BOOLEAN
792 Define mode for notification of address and device changes.
793 0 - (default): do nothing
794 1 - Generate gratuitous arp replies when device is brought up
795 or hardware address changes.
797 arp_accept - BOOLEAN
798 Define behavior when gratuitous arp replies are received:
799 0 - drop gratuitous arp frames
800 1 - accept gratuitous arp frames
802 app_solicit - INTEGER
803 The maximum number of probes to send to the user space ARP daemon
804 via netlink before dropping back to multicast probes (see
805 mcast_solicit). Defaults to 0.
807 disable_policy - BOOLEAN
808 Disable IPSEC policy (SPD) for this interface
810 disable_xfrm - BOOLEAN
811 Disable IPSEC encryption on this interface, whatever the policy
815 tag - INTEGER
816 Allows you to write a number, which can be used as required.
817 Default value is 0.
819 Alexey Kuznetsov.
820 kuznet@ms2.inr.ac.ru
822 Updated by:
823 Andi Kleen
824 ak@muc.de
825 Nicolas Delon
826 delon.nicolas@wanadoo.fr
831 /proc/sys/net/ipv6/* Variables:
833 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
834 apply to IPv6 [XXX?].
836 bindv6only - BOOLEAN
837 Default value for IPV6_V6ONLY socket option,
838 which restricts use of the IPv6 socket to IPv6 communication
839 only.
840 TRUE: disable IPv4-mapped address feature
841 FALSE: enable IPv4-mapped address feature
843 Default: FALSE (as specified in RFC2553bis)
845 IPv6 Fragmentation:
847 ip6frag_high_thresh - INTEGER
848 Maximum memory used to reassemble IPv6 fragments. When
849 ip6frag_high_thresh bytes of memory is allocated for this purpose,
850 the fragment handler will toss packets until ip6frag_low_thresh
851 is reached.
853 ip6frag_low_thresh - INTEGER
854 See ip6frag_high_thresh
856 ip6frag_time - INTEGER
857 Time in seconds to keep an IPv6 fragment in memory.
859 ip6frag_secret_interval - INTEGER
860 Regeneration interval (in seconds) of the hash secret (or lifetime
861 for the hash secret) for IPv6 fragments.
862 Default: 600
864 conf/default/*:
865 Change the interface-specific default settings.
868 conf/all/*:
869 Change all the interface-specific settings.
871 [XXX: Other special features than forwarding?]
873 conf/all/forwarding - BOOLEAN
874 Enable global IPv6 forwarding between all interfaces.
876 IPv4 and IPv6 work differently here; e.g. netfilter must be used
877 to control which interfaces may forward packets and which not.
879 This also sets all interfaces' Host/Router setting
880 'forwarding' to the specified value. See below for details.
882 This referred to as global forwarding.
884 proxy_ndp - BOOLEAN
885 Do proxy ndp.
887 conf/interface/*:
888 Change special settings per interface.
890 The functional behaviour for certain settings is different
891 depending on whether local forwarding is enabled or not.
893 accept_ra - BOOLEAN
894 Accept Router Advertisements; autoconfigure using them.
896 Functional default: enabled if local forwarding is disabled.
897 disabled if local forwarding is enabled.
899 accept_ra_defrtr - BOOLEAN
900 Learn default router in Router Advertisement.
902 Functional default: enabled if accept_ra is enabled.
903 disabled if accept_ra is disabled.
905 accept_ra_pinfo - BOOLEAN
906 Learn Prefix Information in Router Advertisement.
908 Functional default: enabled if accept_ra is enabled.
909 disabled if accept_ra is disabled.
911 accept_ra_rt_info_max_plen - INTEGER
912 Maximum prefix length of Route Information in RA.
914 Route Information w/ prefix larger than or equal to this
915 variable shall be ignored.
917 Functional default: 0 if accept_ra_rtr_pref is enabled.
918 -1 if accept_ra_rtr_pref is disabled.
920 accept_ra_rtr_pref - BOOLEAN
921 Accept Router Preference in RA.
923 Functional default: enabled if accept_ra is enabled.
924 disabled if accept_ra is disabled.
926 accept_redirects - BOOLEAN
927 Accept Redirects.
929 Functional default: enabled if local forwarding is disabled.
930 disabled if local forwarding is enabled.
932 accept_source_route - INTEGER
933 Accept source routing (routing extension header).
935 >= 0: Accept only routing header type 2.
936 < 0: Do not accept routing header.
938 Default: 0
940 autoconf - BOOLEAN
941 Autoconfigure addresses using Prefix Information in Router
942 Advertisements.
944 Functional default: enabled if accept_ra_pinfo is enabled.
945 disabled if accept_ra_pinfo is disabled.
947 dad_transmits - INTEGER
948 The amount of Duplicate Address Detection probes to send.
949 Default: 1
951 forwarding - BOOLEAN
952 Configure interface-specific Host/Router behaviour.
954 Note: It is recommended to have the same setting on all
955 interfaces; mixed router/host scenarios are rather uncommon.
957 FALSE:
959 By default, Host behaviour is assumed. This means:
961 1. IsRouter flag is not set in Neighbour Advertisements.
962 2. Router Solicitations are being sent when necessary.
963 3. If accept_ra is TRUE (default), accept Router
964 Advertisements (and do autoconfiguration).
965 4. If accept_redirects is TRUE (default), accept Redirects.
967 TRUE:
969 If local forwarding is enabled, Router behaviour is assumed.
970 This means exactly the reverse from the above:
972 1. IsRouter flag is set in Neighbour Advertisements.
973 2. Router Solicitations are not sent.
974 3. Router Advertisements are ignored.
975 4. Redirects are ignored.
977 Default: FALSE if global forwarding is disabled (default),
978 otherwise TRUE.
980 hop_limit - INTEGER
981 Default Hop Limit to set.
982 Default: 64
984 mtu - INTEGER
985 Default Maximum Transfer Unit
986 Default: 1280 (IPv6 required minimum)
988 router_probe_interval - INTEGER
989 Minimum interval (in seconds) between Router Probing described
990 in RFC4191.
992 Default: 60
994 router_solicitation_delay - INTEGER
995 Number of seconds to wait after interface is brought up
996 before sending Router Solicitations.
997 Default: 1
999 router_solicitation_interval - INTEGER
1000 Number of seconds to wait between Router Solicitations.
1001 Default: 4
1003 router_solicitations - INTEGER
1004 Number of Router Solicitations to send until assuming no
1005 routers are present.
1006 Default: 3
1008 use_tempaddr - INTEGER
1009 Preference for Privacy Extensions (RFC3041).
1010 <= 0 : disable Privacy Extensions
1011 == 1 : enable Privacy Extensions, but prefer public
1012 addresses over temporary addresses.
1013 > 1 : enable Privacy Extensions and prefer temporary
1014 addresses over public addresses.
1015 Default: 0 (for most devices)
1016 -1 (for point-to-point devices and loopback devices)
1018 temp_valid_lft - INTEGER
1019 valid lifetime (in seconds) for temporary addresses.
1020 Default: 604800 (7 days)
1022 temp_prefered_lft - INTEGER
1023 Preferred lifetime (in seconds) for temporary addresses.
1024 Default: 86400 (1 day)
1026 max_desync_factor - INTEGER
1027 Maximum value for DESYNC_FACTOR, which is a random value
1028 that ensures that clients don't synchronize with each
1029 other and generate new addresses at exactly the same time.
1030 value is in seconds.
1031 Default: 600
1033 regen_max_retry - INTEGER
1034 Number of attempts before give up attempting to generate
1035 valid temporary addresses.
1036 Default: 5
1038 max_addresses - INTEGER
1039 Number of maximum addresses per interface. 0 disables limitation.
1040 It is recommended not set too large value (or 0) because it would
1041 be too easy way to crash kernel to allow to create too much of
1042 autoconfigured addresses.
1043 Default: 16
1045 disable_ipv6 - BOOLEAN
1046 Disable IPv6 operation. If accept_dad is set to 2, this value
1047 will be dynamically set to TRUE if DAD fails for the link-local
1048 address.
1049 Default: FALSE (enable IPv6 operation)
1051 accept_dad - INTEGER
1052 Whether to accept DAD (Duplicate Address Detection).
1053 0: Disable DAD
1054 1: Enable DAD (default)
1055 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1056 link-local address has been found.
1058 icmp/*:
1059 ratelimit - INTEGER
1060 Limit the maximal rates for sending ICMPv6 packets.
1061 0 to disable any limiting,
1062 otherwise the minimal space between responses in milliseconds.
1063 Default: 1000
1066 IPv6 Update by:
1067 Pekka Savola <pekkas@netcore.fi>
1068 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1071 /proc/sys/net/bridge/* Variables:
1073 bridge-nf-call-arptables - BOOLEAN
1074 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1075 0 : disable this.
1076 Default: 1
1078 bridge-nf-call-iptables - BOOLEAN
1079 1 : pass bridged IPv4 traffic to iptables' chains.
1080 0 : disable this.
1081 Default: 1
1083 bridge-nf-call-ip6tables - BOOLEAN
1084 1 : pass bridged IPv6 traffic to ip6tables' chains.
1085 0 : disable this.
1086 Default: 1
1088 bridge-nf-filter-vlan-tagged - BOOLEAN
1089 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1090 0 : disable this.
1091 Default: 1
1093 bridge-nf-filter-pppoe-tagged - BOOLEAN
1094 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1095 0 : disable this.
1096 Default: 1
1099 proc/sys/net/sctp/* Variables:
1101 addip_enable - BOOLEAN
1102 Enable or disable extension of Dynamic Address Reconfiguration
1103 (ADD-IP) functionality specified in RFC5061. This extension provides
1104 the ability to dynamically add and remove new addresses for the SCTP
1105 associations.
1107 1: Enable extension.
1109 0: Disable extension.
1111 Default: 0
1113 addip_noauth_enable - BOOLEAN
1114 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1115 authentication to protect the operations of adding or removing new
1116 addresses. This requirement is mandated so that unauthorized hosts
1117 would not be able to hijack associations. However, older
1118 implementations may not have implemented this requirement while
1119 allowing the ADD-IP extension. For reasons of interoperability,
1120 we provide this variable to control the enforcement of the
1121 authentication requirement.
1123 1: Allow ADD-IP extension to be used without authentication. This
1124 should only be set in a closed environment for interoperability
1125 with older implementations.
1127 0: Enforce the authentication requirement
1129 Default: 0
1131 auth_enable - BOOLEAN
1132 Enable or disable Authenticated Chunks extension. This extension
1133 provides the ability to send and receive authenticated chunks and is
1134 required for secure operation of Dynamic Address Reconfiguration
1135 (ADD-IP) extension.
1137 1: Enable this extension.
1138 0: Disable this extension.
1140 Default: 0
1142 prsctp_enable - BOOLEAN
1143 Enable or disable the Partial Reliability extension (RFC3758) which
1144 is used to notify peers that a given DATA should no longer be expected.
1146 1: Enable extension
1147 0: Disable
1149 Default: 1
1151 max_burst - INTEGER
1152 The limit of the number of new packets that can be initially sent. It
1153 controls how bursty the generated traffic can be.
1155 Default: 4
1157 association_max_retrans - INTEGER
1158 Set the maximum number for retransmissions that an association can
1159 attempt deciding that the remote end is unreachable. If this value
1160 is exceeded, the association is terminated.
1162 Default: 10
1164 max_init_retransmits - INTEGER
1165 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1166 that an association will attempt before declaring the destination
1167 unreachable and terminating.
1169 Default: 8
1171 path_max_retrans - INTEGER
1172 The maximum number of retransmissions that will be attempted on a given
1173 path. Once this threshold is exceeded, the path is considered
1174 unreachable, and new traffic will use a different path when the
1175 association is multihomed.
1177 Default: 5
1179 rto_initial - INTEGER
1180 The initial round trip timeout value in milliseconds that will be used
1181 in calculating round trip times. This is the initial time interval
1182 for retransmissions.
1184 Default: 3000
1186 rto_max - INTEGER
1187 The maximum value (in milliseconds) of the round trip timeout. This
1188 is the largest time interval that can elapse between retransmissions.
1190 Default: 60000
1192 rto_min - INTEGER
1193 The minimum value (in milliseconds) of the round trip timeout. This
1194 is the smallest time interval the can elapse between retransmissions.
1196 Default: 1000
1198 hb_interval - INTEGER
1199 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1200 are sent at the specified interval on idle paths to probe the state of
1201 a given path between 2 associations.
1203 Default: 30000
1205 sack_timeout - INTEGER
1206 The amount of time (in milliseconds) that the implementation will wait
1207 to send a SACK.
1209 Default: 200
1211 valid_cookie_life - INTEGER
1212 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1213 is used during association establishment.
1215 Default: 60000
1217 cookie_preserve_enable - BOOLEAN
1218 Enable or disable the ability to extend the lifetime of the SCTP cookie
1219 that is used during the establishment phase of SCTP association
1221 1: Enable cookie lifetime extension.
1222 0: Disable
1224 Default: 1
1226 rcvbuf_policy - INTEGER
1227 Determines if the receive buffer is attributed to the socket or to
1228 association. SCTP supports the capability to create multiple
1229 associations on a single socket. When using this capability, it is
1230 possible that a single stalled association that's buffering a lot
1231 of data may block other associations from delivering their data by
1232 consuming all of the receive buffer space. To work around this,
1233 the rcvbuf_policy could be set to attribute the receiver buffer space
1234 to each association instead of the socket. This prevents the described
1235 blocking.
1237 1: rcvbuf space is per association
1238 0: recbuf space is per socket
1240 Default: 0
1242 sndbuf_policy - INTEGER
1243 Similar to rcvbuf_policy above, this applies to send buffer space.
1245 1: Send buffer is tracked per association
1246 0: Send buffer is tracked per socket.
1248 Default: 0
1250 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1251 Number of pages allowed for queueing by all SCTP sockets.
1253 min: Below this number of pages SCTP is not bothered about its
1254 memory appetite. When amount of memory allocated by SCTP exceeds
1255 this number, SCTP starts to moderate memory usage.
1257 pressure: This value was introduced to follow format of tcp_mem.
1259 max: Number of pages allowed for queueing by all SCTP sockets.
1261 Default is calculated at boot time from amount of available memory.
1263 sctp_rmem - vector of 3 INTEGERs: min, default, max
1264 See tcp_rmem for a description.
1266 sctp_wmem - vector of 3 INTEGERs: min, default, max
1267 See tcp_wmem for a description.
1270 /proc/sys/net/core/*
1271 dev_weight - INTEGER
1272 The maximum number of packets that kernel can handle on a NAPI
1273 interrupt, it's a Per-CPU variable.
1275 Default: 64
1277 /proc/sys/net/unix/*
1278 max_dgram_qlen - INTEGER
1279 The maximum length of dgram socket receive queue
1281 Default: 10
1284 UNDOCUMENTED:
1286 /proc/sys/net/irda/*
1287 fast_poll_increase FIXME
1288 warn_noreply_time FIXME
1289 discovery_slots FIXME
1290 slot_timeout FIXME
1291 max_baud_rate FIXME
1292 discovery_timeout FIXME
1293 lap_keepalive_time FIXME
1294 max_noreply_time FIXME
1295 max_tx_data_size FIXME
1296 max_tx_window FIXME
1297 min_tx_turn_time FIXME