1 /proc/sys/net/ipv4/* Variables:
7 Forward Packets between interfaces.
9 This variable is special, its change resets all configuration
10 parameters to their default state (RFC1122 for hosts, RFC1812
13 ip_default_ttl - INTEGER
14 Default value of TTL field (Time To Live) for outgoing (but not
15 forwarded) IP packets. Should be between 1 and 255 inclusive.
16 Default: 64 (as recommended by RFC1700)
18 ip_no_pmtu_disc - INTEGER
19 Disable Path MTU Discovery. If enabled in mode 1 and a
20 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.
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.
29 Mode 3 is a hardend pmtu discover mode. The kernel will only
30 accept fragmentation-needed errors if the underlying protocol
31 can verify them besides a plain socket lookup. Current
32 protocols for which pmtu events will be honored are TCP, SCTP
33 and DCCP as they verify e.g. the sequence number or the
34 association. This mode should not be enabled globally but is
35 only intended to secure e.g. name servers in namespaces where
36 TCP path mtu must still work but path MTU information of other
37 protocols should be discarded. If enabled globally this mode
38 could break other protocols.
44 default 552 - minimum discovered Path MTU
46 ip_forward_use_pmtu - BOOLEAN
47 By default we don't trust protocol path MTUs while forwarding
48 because they could be easily forged and can lead to unwanted
49 fragmentation by the router.
50 You only need to enable this if you have user-space software
51 which tries to discover path mtus by itself and depends on the
52 kernel honoring this information. This is normally not the
59 fwmark_reflect - BOOLEAN
60 Controls the fwmark of kernel-generated IPv4 reply packets that are not
61 associated with a socket for example, TCP RSTs or ICMP echo replies).
62 If unset, these packets have a fwmark of zero. If set, they have the
63 fwmark of the packet they are replying to.
66 route/max_size - INTEGER
67 Maximum number of routes allowed in the kernel. Increase
68 this when using large numbers of interfaces and/or routes.
70 neigh/default/gc_thresh1 - INTEGER
71 Minimum number of entries to keep. Garbage collector will not
72 purge entries if there are fewer than this number.
75 neigh/default/gc_thresh2 - INTEGER
76 Threshold when garbage collector becomes more aggressive about
77 purging entries. Entries older than 5 seconds will be cleared
78 when over this number.
81 neigh/default/gc_thresh3 - INTEGER
82 Maximum number of neighbor entries allowed. Increase this
83 when using large numbers of interfaces and when communicating
84 with large numbers of directly-connected peers.
87 neigh/default/unres_qlen_bytes - INTEGER
88 The maximum number of bytes which may be used by packets
89 queued for each unresolved address by other network layers.
91 Setting negative value is meaningless and will return error.
92 Default: 65536 Bytes(64KB)
94 neigh/default/unres_qlen - INTEGER
95 The maximum number of packets which may be queued for each
96 unresolved address by other network layers.
97 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
98 Prior to linux 3.3, the default value is 3 which may cause
99 unexpected packet loss. The current default value is calculated
100 according to default value of unres_qlen_bytes and true size of
104 mtu_expires - INTEGER
105 Time, in seconds, that cached PMTU information is kept.
107 min_adv_mss - INTEGER
108 The advertised MSS depends on the first hop route MTU, but will
109 never be lower than this setting.
113 ipfrag_high_thresh - INTEGER
114 Maximum memory used to reassemble IP fragments. When
115 ipfrag_high_thresh bytes of memory is allocated for this purpose,
116 the fragment handler will toss packets until ipfrag_low_thresh
117 is reached. This also serves as a maximum limit to namespaces
118 different from the initial one.
120 ipfrag_low_thresh - INTEGER
121 Maximum memory used to reassemble IP fragments before the kernel
122 begins to remove incomplete fragment queues to free up resources.
123 The kernel still accepts new fragments for defragmentation.
125 ipfrag_time - INTEGER
126 Time in seconds to keep an IP fragment in memory.
128 ipfrag_max_dist - INTEGER
129 ipfrag_max_dist is a non-negative integer value which defines the
130 maximum "disorder" which is allowed among fragments which share a
131 common IP source address. Note that reordering of packets is
132 not unusual, but if a large number of fragments arrive from a source
133 IP address while a particular fragment queue remains incomplete, it
134 probably indicates that one or more fragments belonging to that queue
135 have been lost. When ipfrag_max_dist is positive, an additional check
136 is done on fragments before they are added to a reassembly queue - if
137 ipfrag_max_dist (or more) fragments have arrived from a particular IP
138 address between additions to any IP fragment queue using that source
139 address, it's presumed that one or more fragments in the queue are
140 lost. The existing fragment queue will be dropped, and a new one
141 started. An ipfrag_max_dist value of zero disables this check.
143 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
144 result in unnecessarily dropping fragment queues when normal
145 reordering of packets occurs, which could lead to poor application
146 performance. Using a very large value, e.g. 50000, increases the
147 likelihood of incorrectly reassembling IP fragments that originate
148 from different IP datagrams, which could result in data corruption.
153 inet_peer_threshold - INTEGER
154 The approximate size of the storage. Starting from this threshold
155 entries will be thrown aggressively. This threshold also determines
156 entries' time-to-live and time intervals between garbage collection
157 passes. More entries, less time-to-live, less GC interval.
159 inet_peer_minttl - INTEGER
160 Minimum time-to-live of entries. Should be enough to cover fragment
161 time-to-live on the reassembling side. This minimum time-to-live is
162 guaranteed if the pool size is less than inet_peer_threshold.
165 inet_peer_maxttl - INTEGER
166 Maximum time-to-live of entries. Unused entries will expire after
167 this period of time if there is no memory pressure on the pool (i.e.
168 when the number of entries in the pool is very small).
174 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
175 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
178 tcp_abort_on_overflow - BOOLEAN
179 If listening service is too slow to accept new connections,
180 reset them. Default state is FALSE. It means that if overflow
181 occurred due to a burst, connection will recover. Enable this
182 option _only_ if you are really sure that listening daemon
183 cannot be tuned to accept connections faster. Enabling this
184 option can harm clients of your server.
186 tcp_adv_win_scale - INTEGER
187 Count buffering overhead as bytes/2^tcp_adv_win_scale
188 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
190 Possible values are [-31, 31], inclusive.
193 tcp_allowed_congestion_control - STRING
194 Show/set the congestion control choices available to non-privileged
195 processes. The list is a subset of those listed in
196 tcp_available_congestion_control.
197 Default is "reno" and the default setting (tcp_congestion_control).
199 tcp_app_win - INTEGER
200 Reserve max(window/2^tcp_app_win, mss) of window for application
201 buffer. Value 0 is special, it means that nothing is reserved.
204 tcp_autocorking - BOOLEAN
205 Enable TCP auto corking :
206 When applications do consecutive small write()/sendmsg() system calls,
207 we try to coalesce these small writes as much as possible, to lower
208 total amount of sent packets. This is done if at least one prior
209 packet for the flow is waiting in Qdisc queues or device transmit
210 queue. Applications can still use TCP_CORK for optimal behavior
211 when they know how/when to uncork their sockets.
214 tcp_available_congestion_control - STRING
215 Shows the available congestion control choices that are registered.
216 More congestion control algorithms may be available as modules,
219 tcp_base_mss - INTEGER
220 The initial value of search_low to be used by the packetization layer
221 Path MTU discovery (MTU probing). If MTU probing is enabled,
222 this is the initial MSS used by the connection.
224 tcp_congestion_control - STRING
225 Set the congestion control algorithm to be used for new
226 connections. The algorithm "reno" is always available, but
227 additional choices may be available based on kernel configuration.
228 Default is set as part of kernel configuration.
229 For passive connections, the listener congestion control choice
231 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
234 Allows TCP to send "duplicate" SACKs.
236 tcp_early_retrans - INTEGER
237 Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold
238 for triggering fast retransmit when the amount of outstanding data is
239 small and when no previously unsent data can be transmitted (such
240 that limited transmit could be used). Also controls the use of
241 Tail loss probe (TLP) that converts RTOs occurring due to tail
242 losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01).
246 2 enables ER but delays fast recovery and fast retransmit
247 by a fourth of RTT. This mitigates connection falsely
248 recovers when network has a small degree of reordering
249 (less than 3 packets).
250 3 enables delayed ER and TLP.
255 Control use of Explicit Congestion Notification (ECN) by TCP.
256 ECN is used only when both ends of the TCP connection indicate
257 support for it. This feature is useful in avoiding losses due
258 to congestion by allowing supporting routers to signal
259 congestion before having to drop packets.
261 0 Disable ECN. Neither initiate nor accept ECN.
262 1 Enable ECN when requested by incoming connections and
263 also request ECN on outgoing connection attempts.
264 2 Enable ECN when requested by incoming connections
265 but do not request ECN on outgoing connections.
269 Enable FACK congestion avoidance and fast retransmission.
270 The value is not used, if tcp_sack is not enabled.
272 tcp_fin_timeout - INTEGER
273 The length of time an orphaned (no longer referenced by any
274 application) connection will remain in the FIN_WAIT_2 state
275 before it is aborted at the local end. While a perfectly
276 valid "receive only" state for an un-orphaned connection, an
277 orphaned connection in FIN_WAIT_2 state could otherwise wait
278 forever for the remote to close its end of the connection.
283 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
284 F-RTO is an enhanced recovery algorithm for TCP retransmission
285 timeouts. It is particularly beneficial in networks where the
286 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
287 modification. It does not require any support from the peer.
289 By default it's enabled with a non-zero value. 0 disables F-RTO.
291 tcp_keepalive_time - INTEGER
292 How often TCP sends out keepalive messages when keepalive is enabled.
295 tcp_keepalive_probes - INTEGER
296 How many keepalive probes TCP sends out, until it decides that the
297 connection is broken. Default value: 9.
299 tcp_keepalive_intvl - INTEGER
300 How frequently the probes are send out. Multiplied by
301 tcp_keepalive_probes it is time to kill not responding connection,
302 after probes started. Default value: 75sec i.e. connection
303 will be aborted after ~11 minutes of retries.
305 tcp_low_latency - BOOLEAN
306 If set, the TCP stack makes decisions that prefer lower
307 latency as opposed to higher throughput. By default, this
308 option is not set meaning that higher throughput is preferred.
309 An example of an application where this default should be
310 changed would be a Beowulf compute cluster.
313 tcp_max_orphans - INTEGER
314 Maximal number of TCP sockets not attached to any user file handle,
315 held by system. If this number is exceeded orphaned connections are
316 reset immediately and warning is printed. This limit exists
317 only to prevent simple DoS attacks, you _must_ not rely on this
318 or lower the limit artificially, but rather increase it
319 (probably, after increasing installed memory),
320 if network conditions require more than default value,
321 and tune network services to linger and kill such states
322 more aggressively. Let me to remind again: each orphan eats
323 up to ~64K of unswappable memory.
325 tcp_max_syn_backlog - INTEGER
326 Maximal number of remembered connection requests, which have not
327 received an acknowledgment from connecting client.
328 The minimal value is 128 for low memory machines, and it will
329 increase in proportion to the memory of machine.
330 If server suffers from overload, try increasing this number.
332 tcp_max_tw_buckets - INTEGER
333 Maximal number of timewait sockets held by system simultaneously.
334 If this number is exceeded time-wait socket is immediately destroyed
335 and warning is printed. This limit exists only to prevent
336 simple DoS attacks, you _must_ not lower the limit artificially,
337 but rather increase it (probably, after increasing installed memory),
338 if network conditions require more than default value.
340 tcp_mem - vector of 3 INTEGERs: min, pressure, max
341 min: below this number of pages TCP is not bothered about its
344 pressure: when amount of memory allocated by TCP exceeds this number
345 of pages, TCP moderates its memory consumption and enters memory
346 pressure mode, which is exited when memory consumption falls
349 max: number of pages allowed for queueing by all TCP sockets.
351 Defaults are calculated at boot time from amount of available
354 tcp_moderate_rcvbuf - BOOLEAN
355 If set, TCP performs receive buffer auto-tuning, attempting to
356 automatically size the buffer (no greater than tcp_rmem[2]) to
357 match the size required by the path for full throughput. Enabled by
360 tcp_mtu_probing - INTEGER
361 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
364 1 - Disabled by default, enabled when an ICMP black hole detected
365 2 - Always enabled, use initial MSS of tcp_base_mss.
367 tcp_no_metrics_save - BOOLEAN
368 By default, TCP saves various connection metrics in the route cache
369 when the connection closes, so that connections established in the
370 near future can use these to set initial conditions. Usually, this
371 increases overall performance, but may sometimes cause performance
372 degradation. If set, TCP will not cache metrics on closing
375 tcp_orphan_retries - INTEGER
376 This value influences the timeout of a locally closed TCP connection,
377 when RTO retransmissions remain unacknowledged.
378 See tcp_retries2 for more details.
380 The default value is 8.
381 If your machine is a loaded WEB server,
382 you should think about lowering this value, such sockets
383 may consume significant resources. Cf. tcp_max_orphans.
385 tcp_reordering - INTEGER
386 Initial reordering level of packets in a TCP stream.
387 TCP stack can then dynamically adjust flow reordering level
388 between this initial value and tcp_max_reordering
391 tcp_max_reordering - INTEGER
392 Maximal reordering level of packets in a TCP stream.
393 300 is a fairly conservative value, but you might increase it
394 if paths are using per packet load balancing (like bonding rr mode)
397 tcp_retrans_collapse - BOOLEAN
398 Bug-to-bug compatibility with some broken printers.
399 On retransmit try to send bigger packets to work around bugs in
402 tcp_retries1 - INTEGER
403 This value influences the time, after which TCP decides, that
404 something is wrong due to unacknowledged RTO retransmissions,
405 and reports this suspicion to the network layer.
406 See tcp_retries2 for more details.
408 RFC 1122 recommends at least 3 retransmissions, which is the
411 tcp_retries2 - INTEGER
412 This value influences the timeout of an alive TCP connection,
413 when RTO retransmissions remain unacknowledged.
414 Given a value of N, a hypothetical TCP connection following
415 exponential backoff with an initial RTO of TCP_RTO_MIN would
416 retransmit N times before killing the connection at the (N+1)th RTO.
418 The default value of 15 yields a hypothetical timeout of 924.6
419 seconds and is a lower bound for the effective timeout.
420 TCP will effectively time out at the first RTO which exceeds the
421 hypothetical timeout.
423 RFC 1122 recommends at least 100 seconds for the timeout,
424 which corresponds to a value of at least 8.
426 tcp_rfc1337 - BOOLEAN
427 If set, the TCP stack behaves conforming to RFC1337. If unset,
428 we are not conforming to RFC, but prevent TCP TIME_WAIT
432 tcp_rmem - vector of 3 INTEGERs: min, default, max
433 min: Minimal size of receive buffer used by TCP sockets.
434 It is guaranteed to each TCP socket, even under moderate memory
438 default: initial size of receive buffer used by TCP sockets.
439 This value overrides net.core.rmem_default used by other protocols.
440 Default: 87380 bytes. This value results in window of 65535 with
441 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
442 less for default tcp_app_win. See below about these variables.
444 max: maximal size of receive buffer allowed for automatically
445 selected receiver buffers for TCP socket. This value does not override
446 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
447 automatic tuning of that socket's receive buffer size, in which
448 case this value is ignored.
449 Default: between 87380B and 6MB, depending on RAM size.
452 Enable select acknowledgments (SACKS).
454 tcp_slow_start_after_idle - BOOLEAN
455 If set, provide RFC2861 behavior and time out the congestion
456 window after an idle period. An idle period is defined at
457 the current RTO. If unset, the congestion window will not
458 be timed out after an idle period.
462 Use the Host requirements interpretation of the TCP urgent pointer field.
463 Most hosts use the older BSD interpretation, so if you turn this on
464 Linux might not communicate correctly with them.
467 tcp_synack_retries - INTEGER
468 Number of times SYNACKs for a passive TCP connection attempt will
469 be retransmitted. Should not be higher than 255. Default value
470 is 5, which corresponds to 31seconds till the last retransmission
471 with the current initial RTO of 1second. With this the final timeout
472 for a passive TCP connection will happen after 63seconds.
474 tcp_syncookies - BOOLEAN
475 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
476 Send out syncookies when the syn backlog queue of a socket
477 overflows. This is to prevent against the common 'SYN flood attack'
480 Note, that syncookies is fallback facility.
481 It MUST NOT be used to help highly loaded servers to stand
482 against legal connection rate. If you see SYN flood warnings
483 in your logs, but investigation shows that they occur
484 because of overload with legal connections, you should tune
485 another parameters until this warning disappear.
486 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
488 syncookies seriously violate TCP protocol, do not allow
489 to use TCP extensions, can result in serious degradation
490 of some services (f.e. SMTP relaying), visible not by you,
491 but your clients and relays, contacting you. While you see
492 SYN flood warnings in logs not being really flooded, your server
493 is seriously misconfigured.
495 If you want to test which effects syncookies have to your
496 network connections you can set this knob to 2 to enable
497 unconditionally generation of syncookies.
499 tcp_fastopen - INTEGER
500 Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data
501 in the opening SYN packet. To use this feature, the client application
502 must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than
503 connect() to perform a TCP handshake automatically.
505 The values (bitmap) are
506 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN.
507 2: Enables TCP Fast Open on the server side, i.e., allowing data in
508 a SYN packet to be accepted and passed to the application before
509 3-way hand shake finishes.
510 4: Send data in the opening SYN regardless of cookie availability and
511 without a cookie option.
512 0x100: Accept SYN data w/o validating the cookie.
513 0x200: Accept data-in-SYN w/o any cookie option present.
514 0x400/0x800: Enable Fast Open on all listeners regardless of the
515 TCP_FASTOPEN socket option. The two different flags designate two
516 different ways of setting max_qlen without the TCP_FASTOPEN socket
521 Note that the client & server side Fast Open flags (1 and 2
522 respectively) must be also enabled before the rest of flags can take
525 See include/net/tcp.h and the code for more details.
527 tcp_syn_retries - INTEGER
528 Number of times initial SYNs for an active TCP connection attempt
529 will be retransmitted. Should not be higher than 255. Default value
530 is 6, which corresponds to 63seconds till the last retransmission
531 with the current initial RTO of 1second. With this the final timeout
532 for an active TCP connection attempt will happen after 127seconds.
534 tcp_timestamps - BOOLEAN
535 Enable timestamps as defined in RFC1323.
537 tcp_min_tso_segs - INTEGER
538 Minimal number of segments per TSO frame.
539 Since linux-3.12, TCP does an automatic sizing of TSO frames,
540 depending on flow rate, instead of filling 64Kbytes packets.
541 For specific usages, it's possible to force TCP to build big
542 TSO frames. Note that TCP stack might split too big TSO packets
543 if available window is too small.
546 tcp_tso_win_divisor - INTEGER
547 This allows control over what percentage of the congestion window
548 can be consumed by a single TSO frame.
549 The setting of this parameter is a choice between burstiness and
550 building larger TSO frames.
553 tcp_tw_recycle - BOOLEAN
554 Enable fast recycling TIME-WAIT sockets. Default value is 0.
555 It should not be changed without advice/request of technical
558 tcp_tw_reuse - BOOLEAN
559 Allow to reuse TIME-WAIT sockets for new connections when it is
560 safe from protocol viewpoint. Default value is 0.
561 It should not be changed without advice/request of technical
564 tcp_window_scaling - BOOLEAN
565 Enable window scaling as defined in RFC1323.
567 tcp_wmem - vector of 3 INTEGERs: min, default, max
568 min: Amount of memory reserved for send buffers for TCP sockets.
569 Each TCP socket has rights to use it due to fact of its birth.
572 default: initial size of send buffer used by TCP sockets. This
573 value overrides net.core.wmem_default used by other protocols.
574 It is usually lower than net.core.wmem_default.
577 max: Maximal amount of memory allowed for automatically tuned
578 send buffers for TCP sockets. This value does not override
579 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
580 automatic tuning of that socket's send buffer size, in which case
581 this value is ignored.
582 Default: between 64K and 4MB, depending on RAM size.
584 tcp_notsent_lowat - UNSIGNED INTEGER
585 A TCP socket can control the amount of unsent bytes in its write queue,
586 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
587 reports POLLOUT events if the amount of unsent bytes is below a per
588 socket value, and if the write queue is not full. sendmsg() will
589 also not add new buffers if the limit is hit.
591 This global variable controls the amount of unsent data for
592 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
593 to the global variable has immediate effect.
595 Default: UINT_MAX (0xFFFFFFFF)
597 tcp_workaround_signed_windows - BOOLEAN
598 If set, assume no receipt of a window scaling option means the
599 remote TCP is broken and treats the window as a signed quantity.
600 If unset, assume the remote TCP is not broken even if we do
601 not receive a window scaling option from them.
604 tcp_thin_linear_timeouts - BOOLEAN
605 Enable dynamic triggering of linear timeouts for thin streams.
606 If set, a check is performed upon retransmission by timeout to
607 determine if the stream is thin (less than 4 packets in flight).
608 As long as the stream is found to be thin, up to 6 linear
609 timeouts may be performed before exponential backoff mode is
610 initiated. This improves retransmission latency for
611 non-aggressive thin streams, often found to be time-dependent.
612 For more information on thin streams, see
613 Documentation/networking/tcp-thin.txt
616 tcp_thin_dupack - BOOLEAN
617 Enable dynamic triggering of retransmissions after one dupACK
618 for thin streams. If set, a check is performed upon reception
619 of a dupACK to determine if the stream is thin (less than 4
620 packets in flight). As long as the stream is found to be thin,
621 data is retransmitted on the first received dupACK. This
622 improves retransmission latency for non-aggressive thin
623 streams, often found to be time-dependent.
624 For more information on thin streams, see
625 Documentation/networking/tcp-thin.txt
628 tcp_limit_output_bytes - INTEGER
629 Controls TCP Small Queue limit per tcp socket.
630 TCP bulk sender tends to increase packets in flight until it
631 gets losses notifications. With SNDBUF autotuning, this can
632 result in a large amount of packets queued in qdisc/device
633 on the local machine, hurting latency of other flows, for
634 typical pfifo_fast qdiscs.
635 tcp_limit_output_bytes limits the number of bytes on qdisc
636 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
639 tcp_challenge_ack_limit - INTEGER
640 Limits number of Challenge ACK sent per second, as recommended
641 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
646 udp_mem - vector of 3 INTEGERs: min, pressure, max
647 Number of pages allowed for queueing by all UDP sockets.
649 min: Below this number of pages UDP is not bothered about its
650 memory appetite. When amount of memory allocated by UDP exceeds
651 this number, UDP starts to moderate memory usage.
653 pressure: This value was introduced to follow format of tcp_mem.
655 max: Number of pages allowed for queueing by all UDP sockets.
657 Default is calculated at boot time from amount of available memory.
659 udp_rmem_min - INTEGER
660 Minimal size of receive buffer used by UDP sockets in moderation.
661 Each UDP socket is able to use the size for receiving data, even if
662 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
665 udp_wmem_min - INTEGER
666 Minimal size of send buffer used by UDP sockets in moderation.
667 Each UDP socket is able to use the size for sending data, even if
668 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
673 cipso_cache_enable - BOOLEAN
674 If set, enable additions to and lookups from the CIPSO label mapping
675 cache. If unset, additions are ignored and lookups always result in a
676 miss. However, regardless of the setting the cache is still
677 invalidated when required when means you can safely toggle this on and
678 off and the cache will always be "safe".
681 cipso_cache_bucket_size - INTEGER
682 The CIPSO label cache consists of a fixed size hash table with each
683 hash bucket containing a number of cache entries. This variable limits
684 the number of entries in each hash bucket; the larger the value the
685 more CIPSO label mappings that can be cached. When the number of
686 entries in a given hash bucket reaches this limit adding new entries
687 causes the oldest entry in the bucket to be removed to make room.
690 cipso_rbm_optfmt - BOOLEAN
691 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
692 the CIPSO draft specification (see Documentation/netlabel for details).
693 This means that when set the CIPSO tag will be padded with empty
694 categories in order to make the packet data 32-bit aligned.
697 cipso_rbm_structvalid - BOOLEAN
698 If set, do a very strict check of the CIPSO option when
699 ip_options_compile() is called. If unset, relax the checks done during
700 ip_options_compile(). Either way is "safe" as errors are caught else
701 where in the CIPSO processing code but setting this to 0 (False) should
702 result in less work (i.e. it should be faster) but could cause problems
703 with other implementations that require strict checking.
708 ip_local_port_range - 2 INTEGERS
709 Defines the local port range that is used by TCP and UDP to
710 choose the local port. The first number is the first, the
711 second the last local port number. The default values are
712 32768 and 61000 respectively.
714 ip_local_reserved_ports - list of comma separated ranges
715 Specify the ports which are reserved for known third-party
716 applications. These ports will not be used by automatic port
717 assignments (e.g. when calling connect() or bind() with port
718 number 0). Explicit port allocation behavior is unchanged.
720 The format used for both input and output is a comma separated
721 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
722 10). Writing to the file will clear all previously reserved
723 ports and update the current list with the one given in the
726 Note that ip_local_port_range and ip_local_reserved_ports
727 settings are independent and both are considered by the kernel
728 when determining which ports are available for automatic port
731 You can reserve ports which are not in the current
732 ip_local_port_range, e.g.:
734 $ cat /proc/sys/net/ipv4/ip_local_port_range
736 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
739 although this is redundant. However such a setting is useful
740 if later the port range is changed to a value that will
741 include the reserved ports.
745 ip_nonlocal_bind - BOOLEAN
746 If set, allows processes to bind() to non-local IP addresses,
747 which can be quite useful - but may break some applications.
751 If set non-zero, enables support for dynamic addresses.
752 If set to a non-zero value larger than 1, a kernel log
753 message will be printed when dynamic address rewriting
757 ip_early_demux - BOOLEAN
758 Optimize input packet processing down to one demux for
759 certain kinds of local sockets. Currently we only do this
760 for established TCP sockets.
762 It may add an additional cost for pure routing workloads that
763 reduces overall throughput, in such case you should disable it.
766 icmp_echo_ignore_all - BOOLEAN
767 If set non-zero, then the kernel will ignore all ICMP ECHO
771 icmp_echo_ignore_broadcasts - BOOLEAN
772 If set non-zero, then the kernel will ignore all ICMP ECHO and
773 TIMESTAMP requests sent to it via broadcast/multicast.
776 icmp_ratelimit - INTEGER
777 Limit the maximal rates for sending ICMP packets whose type matches
778 icmp_ratemask (see below) to specific targets.
779 0 to disable any limiting,
780 otherwise the minimal space between responses in milliseconds.
781 Note that another sysctl, icmp_msgs_per_sec limits the number
782 of ICMP packets sent on all targets.
785 icmp_msgs_per_sec - INTEGER
786 Limit maximal number of ICMP packets sent per second from this host.
787 Only messages whose type matches icmp_ratemask (see below) are
788 controlled by this limit.
791 icmp_msgs_burst - INTEGER
792 icmp_msgs_per_sec controls number of ICMP packets sent per second,
793 while icmp_msgs_burst controls the burst size of these packets.
796 icmp_ratemask - INTEGER
797 Mask made of ICMP types for which rates are being limited.
798 Significant bits: IHGFEDCBA9876543210
799 Default mask: 0000001100000011000 (6168)
801 Bit definitions (see include/linux/icmp.h):
803 3 Destination Unreachable *
808 C Parameter Problem *
813 H Address Mask Request
816 * These are rate limited by default (see default mask above)
818 icmp_ignore_bogus_error_responses - BOOLEAN
819 Some routers violate RFC1122 by sending bogus responses to broadcast
820 frames. Such violations are normally logged via a kernel warning.
821 If this is set to TRUE, the kernel will not give such warnings, which
822 will avoid log file clutter.
825 icmp_errors_use_inbound_ifaddr - BOOLEAN
827 If zero, icmp error messages are sent with the primary address of
828 the exiting interface.
830 If non-zero, the message will be sent with the primary address of
831 the interface that received the packet that caused the icmp error.
832 This is the behaviour network many administrators will expect from
833 a router. And it can make debugging complicated network layouts
836 Note that if no primary address exists for the interface selected,
837 then the primary address of the first non-loopback interface that
838 has one will be used regardless of this setting.
842 igmp_max_memberships - INTEGER
843 Change the maximum number of multicast groups we can subscribe to.
846 Theoretical maximum value is bounded by having to send a membership
847 report in a single datagram (i.e. the report can't span multiple
848 datagrams, or risk confusing the switch and leaving groups you don't
851 The number of supported groups 'M' is bounded by the number of group
852 report entries you can fit into a single datagram of 65535 bytes.
854 M = 65536-sizeof (ip header)/(sizeof(Group record))
856 Group records are variable length, with a minimum of 12 bytes.
857 So net.ipv4.igmp_max_memberships should not be set higher than:
859 (65536-24) / 12 = 5459
861 The value 5459 assumes no IP header options, so in practice
862 this number may be lower.
864 conf/interface/* changes special settings per interface (where
865 "interface" is the name of your network interface)
867 conf/all/* is special, changes the settings for all interfaces
870 Controls the IGMP query robustness variable (see RFC2236 8.1).
871 Default: 2 (as specified by RFC2236 8.1)
872 Minimum: 1 (as specified by RFC6636 4.5)
874 log_martians - BOOLEAN
875 Log packets with impossible addresses to kernel log.
876 log_martians for the interface will be enabled if at least one of
877 conf/{all,interface}/log_martians is set to TRUE,
878 it will be disabled otherwise
880 accept_redirects - BOOLEAN
881 Accept ICMP redirect messages.
882 accept_redirects for the interface will be enabled if:
883 - both conf/{all,interface}/accept_redirects are TRUE in the case
884 forwarding for the interface is enabled
886 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
887 case forwarding for the interface is disabled
888 accept_redirects for the interface will be disabled otherwise
893 Enable IP forwarding on this interface.
895 mc_forwarding - BOOLEAN
896 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
897 and a multicast routing daemon is required.
898 conf/all/mc_forwarding must also be set to TRUE to enable multicast
899 routing for the interface
902 Integer value used to differentiate the devices by the medium they
903 are attached to. Two devices can have different id values when
904 the broadcast packets are received only on one of them.
905 The default value 0 means that the device is the only interface
906 to its medium, value of -1 means that medium is not known.
908 Currently, it is used to change the proxy_arp behavior:
909 the proxy_arp feature is enabled for packets forwarded between
910 two devices attached to different media.
914 proxy_arp for the interface will be enabled if at least one of
915 conf/{all,interface}/proxy_arp is set to TRUE,
916 it will be disabled otherwise
918 proxy_arp_pvlan - BOOLEAN
919 Private VLAN proxy arp.
920 Basically allow proxy arp replies back to the same interface
921 (from which the ARP request/solicitation was received).
923 This is done to support (ethernet) switch features, like RFC
924 3069, where the individual ports are NOT allowed to
925 communicate with each other, but they are allowed to talk to
926 the upstream router. As described in RFC 3069, it is possible
927 to allow these hosts to communicate through the upstream
928 router by proxy_arp'ing. Don't need to be used together with
931 This technology is known by different names:
932 In RFC 3069 it is called VLAN Aggregation.
933 Cisco and Allied Telesyn call it Private VLAN.
934 Hewlett-Packard call it Source-Port filtering or port-isolation.
935 Ericsson call it MAC-Forced Forwarding (RFC Draft).
937 shared_media - BOOLEAN
938 Send(router) or accept(host) RFC1620 shared media redirects.
939 Overrides ip_secure_redirects.
940 shared_media for the interface will be enabled if at least one of
941 conf/{all,interface}/shared_media is set to TRUE,
942 it will be disabled otherwise
945 secure_redirects - BOOLEAN
946 Accept ICMP redirect messages only for gateways,
947 listed in default gateway list.
948 secure_redirects for the interface will be enabled if at least one of
949 conf/{all,interface}/secure_redirects is set to TRUE,
950 it will be disabled otherwise
953 send_redirects - BOOLEAN
954 Send redirects, if router.
955 send_redirects for the interface will be enabled if at least one of
956 conf/{all,interface}/send_redirects is set to TRUE,
957 it will be disabled otherwise
960 bootp_relay - BOOLEAN
961 Accept packets with source address 0.b.c.d destined
962 not to this host as local ones. It is supposed, that
963 BOOTP relay daemon will catch and forward such packets.
964 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
969 accept_source_route - BOOLEAN
970 Accept packets with SRR option.
971 conf/all/accept_source_route must also be set to TRUE to accept packets
972 with SRR option on the interface
973 default TRUE (router)
976 accept_local - BOOLEAN
977 Accept packets with local source addresses. In combination with
978 suitable routing, this can be used to direct packets between two
979 local interfaces over the wire and have them accepted properly.
982 route_localnet - BOOLEAN
983 Do not consider loopback addresses as martian source or destination
984 while routing. This enables the use of 127/8 for local routing purposes.
988 0 - No source validation.
989 1 - Strict mode as defined in RFC3704 Strict Reverse Path
990 Each incoming packet is tested against the FIB and if the interface
991 is not the best reverse path the packet check will fail.
992 By default failed packets are discarded.
993 2 - Loose mode as defined in RFC3704 Loose Reverse Path
994 Each incoming packet's source address is also tested against the FIB
995 and if the source address is not reachable via any interface
996 the packet check will fail.
998 Current recommended practice in RFC3704 is to enable strict mode
999 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1000 or other complicated routing, then loose mode is recommended.
1002 The max value from conf/{all,interface}/rp_filter is used
1003 when doing source validation on the {interface}.
1005 Default value is 0. Note that some distributions enable it
1008 arp_filter - BOOLEAN
1009 1 - Allows you to have multiple network interfaces on the same
1010 subnet, and have the ARPs for each interface be answered
1011 based on whether or not the kernel would route a packet from
1012 the ARP'd IP out that interface (therefore you must use source
1013 based routing for this to work). In other words it allows control
1014 of which cards (usually 1) will respond to an arp request.
1016 0 - (default) The kernel can respond to arp requests with addresses
1017 from other interfaces. This may seem wrong but it usually makes
1018 sense, because it increases the chance of successful communication.
1019 IP addresses are owned by the complete host on Linux, not by
1020 particular interfaces. Only for more complex setups like load-
1021 balancing, does this behaviour cause problems.
1023 arp_filter for the interface will be enabled if at least one of
1024 conf/{all,interface}/arp_filter is set to TRUE,
1025 it will be disabled otherwise
1027 arp_announce - INTEGER
1028 Define different restriction levels for announcing the local
1029 source IP address from IP packets in ARP requests sent on
1031 0 - (default) Use any local address, configured on any interface
1032 1 - Try to avoid local addresses that are not in the target's
1033 subnet for this interface. This mode is useful when target
1034 hosts reachable via this interface require the source IP
1035 address in ARP requests to be part of their logical network
1036 configured on the receiving interface. When we generate the
1037 request we will check all our subnets that include the
1038 target IP and will preserve the source address if it is from
1039 such subnet. If there is no such subnet we select source
1040 address according to the rules for level 2.
1041 2 - Always use the best local address for this target.
1042 In this mode we ignore the source address in the IP packet
1043 and try to select local address that we prefer for talks with
1044 the target host. Such local address is selected by looking
1045 for primary IP addresses on all our subnets on the outgoing
1046 interface that include the target IP address. If no suitable
1047 local address is found we select the first local address
1048 we have on the outgoing interface or on all other interfaces,
1049 with the hope we will receive reply for our request and
1050 even sometimes no matter the source IP address we announce.
1052 The max value from conf/{all,interface}/arp_announce is used.
1054 Increasing the restriction level gives more chance for
1055 receiving answer from the resolved target while decreasing
1056 the level announces more valid sender's information.
1058 arp_ignore - INTEGER
1059 Define different modes for sending replies in response to
1060 received ARP requests that resolve local target IP addresses:
1061 0 - (default): reply for any local target IP address, configured
1063 1 - reply only if the target IP address is local address
1064 configured on the incoming interface
1065 2 - reply only if the target IP address is local address
1066 configured on the incoming interface and both with the
1067 sender's IP address are part from same subnet on this interface
1068 3 - do not reply for local addresses configured with scope host,
1069 only resolutions for global and link addresses are replied
1071 8 - do not reply for all local addresses
1073 The max value from conf/{all,interface}/arp_ignore is used
1074 when ARP request is received on the {interface}
1076 arp_notify - BOOLEAN
1077 Define mode for notification of address and device changes.
1078 0 - (default): do nothing
1079 1 - Generate gratuitous arp requests when device is brought up
1080 or hardware address changes.
1082 arp_accept - BOOLEAN
1083 Define behavior for gratuitous ARP frames who's IP is not
1084 already present in the ARP table:
1085 0 - don't create new entries in the ARP table
1086 1 - create new entries in the ARP table
1088 Both replies and requests type gratuitous arp will trigger the
1089 ARP table to be updated, if this setting is on.
1091 If the ARP table already contains the IP address of the
1092 gratuitous arp frame, the arp table will be updated regardless
1093 if this setting is on or off.
1096 app_solicit - INTEGER
1097 The maximum number of probes to send to the user space ARP daemon
1098 via netlink before dropping back to multicast probes (see
1099 mcast_solicit). Defaults to 0.
1101 disable_policy - BOOLEAN
1102 Disable IPSEC policy (SPD) for this interface
1104 disable_xfrm - BOOLEAN
1105 Disable IPSEC encryption on this interface, whatever the policy
1107 igmpv2_unsolicited_report_interval - INTEGER
1108 The interval in milliseconds in which the next unsolicited
1109 IGMPv1 or IGMPv2 report retransmit will take place.
1110 Default: 10000 (10 seconds)
1112 igmpv3_unsolicited_report_interval - INTEGER
1113 The interval in milliseconds in which the next unsolicited
1114 IGMPv3 report retransmit will take place.
1115 Default: 1000 (1 seconds)
1117 promote_secondaries - BOOLEAN
1118 When a primary IP address is removed from this interface
1119 promote a corresponding secondary IP address instead of
1120 removing all the corresponding secondary IP addresses.
1124 Allows you to write a number, which can be used as required.
1128 kuznet@ms2.inr.ac.ru
1134 delon.nicolas@wanadoo.fr
1139 /proc/sys/net/ipv6/* Variables:
1141 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1142 apply to IPv6 [XXX?].
1144 bindv6only - BOOLEAN
1145 Default value for IPV6_V6ONLY socket option,
1146 which restricts use of the IPv6 socket to IPv6 communication
1148 TRUE: disable IPv4-mapped address feature
1149 FALSE: enable IPv4-mapped address feature
1151 Default: FALSE (as specified in RFC3493)
1153 flowlabel_consistency - BOOLEAN
1154 Protect the consistency (and unicity) of flow label.
1155 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1161 auto_flowlabels - BOOLEAN
1162 Automatically generate flow labels based based on a flow hash
1163 of the packet. This allows intermediate devices, such as routers,
1164 to idenfify packet flows for mechanisms like Equal Cost Multipath
1165 Routing (see RFC 6438).
1170 anycast_src_echo_reply - BOOLEAN
1171 Controls the use of anycast addresses as source addresses for ICMPv6
1178 Controls the MLD query robustness variable (see RFC3810 9.1).
1179 Default: 2 (as specified by RFC3810 9.1)
1180 Minimum: 1 (as specified by RFC6636 4.5)
1184 ip6frag_high_thresh - INTEGER
1185 Maximum memory used to reassemble IPv6 fragments. When
1186 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1187 the fragment handler will toss packets until ip6frag_low_thresh
1190 ip6frag_low_thresh - INTEGER
1191 See ip6frag_high_thresh
1193 ip6frag_time - INTEGER
1194 Time in seconds to keep an IPv6 fragment in memory.
1197 Change the interface-specific default settings.
1201 Change all the interface-specific settings.
1203 [XXX: Other special features than forwarding?]
1205 conf/all/forwarding - BOOLEAN
1206 Enable global IPv6 forwarding between all interfaces.
1208 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1209 to control which interfaces may forward packets and which not.
1211 This also sets all interfaces' Host/Router setting
1212 'forwarding' to the specified value. See below for details.
1214 This referred to as global forwarding.
1219 fwmark_reflect - BOOLEAN
1220 Controls the fwmark of kernel-generated IPv6 reply packets that are not
1221 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
1222 If unset, these packets have a fwmark of zero. If set, they have the
1223 fwmark of the packet they are replying to.
1227 Change special settings per interface.
1229 The functional behaviour for certain settings is different
1230 depending on whether local forwarding is enabled or not.
1233 Accept Router Advertisements; autoconfigure using them.
1235 It also determines whether or not to transmit Router
1236 Solicitations. If and only if the functional setting is to
1237 accept Router Advertisements, Router Solicitations will be
1240 Possible values are:
1241 0 Do not accept Router Advertisements.
1242 1 Accept Router Advertisements if forwarding is disabled.
1243 2 Overrule forwarding behaviour. Accept Router Advertisements
1244 even if forwarding is enabled.
1246 Functional default: enabled if local forwarding is disabled.
1247 disabled if local forwarding is enabled.
1249 accept_ra_defrtr - BOOLEAN
1250 Learn default router in Router Advertisement.
1252 Functional default: enabled if accept_ra is enabled.
1253 disabled if accept_ra is disabled.
1255 accept_ra_from_local - BOOLEAN
1256 Accept RA with source-address that is found on local machine
1257 if the RA is otherwise proper and able to be accepted.
1258 Default is to NOT accept these as it may be an un-intended
1262 enabled if accept_ra_from_local is enabled
1263 on a specific interface.
1264 disabled if accept_ra_from_local is disabled
1265 on a specific interface.
1267 accept_ra_pinfo - BOOLEAN
1268 Learn Prefix Information in Router Advertisement.
1270 Functional default: enabled if accept_ra is enabled.
1271 disabled if accept_ra is disabled.
1273 accept_ra_rt_info_max_plen - INTEGER
1274 Maximum prefix length of Route Information in RA.
1276 Route Information w/ prefix larger than or equal to this
1277 variable shall be ignored.
1279 Functional default: 0 if accept_ra_rtr_pref is enabled.
1280 -1 if accept_ra_rtr_pref is disabled.
1282 accept_ra_rtr_pref - BOOLEAN
1283 Accept Router Preference in RA.
1285 Functional default: enabled if accept_ra is enabled.
1286 disabled if accept_ra is disabled.
1288 accept_redirects - BOOLEAN
1291 Functional default: enabled if local forwarding is disabled.
1292 disabled if local forwarding is enabled.
1294 accept_source_route - INTEGER
1295 Accept source routing (routing extension header).
1297 >= 0: Accept only routing header type 2.
1298 < 0: Do not accept routing header.
1303 Autoconfigure addresses using Prefix Information in Router
1306 Functional default: enabled if accept_ra_pinfo is enabled.
1307 disabled if accept_ra_pinfo is disabled.
1309 dad_transmits - INTEGER
1310 The amount of Duplicate Address Detection probes to send.
1313 forwarding - INTEGER
1314 Configure interface-specific Host/Router behaviour.
1316 Note: It is recommended to have the same setting on all
1317 interfaces; mixed router/host scenarios are rather uncommon.
1319 Possible values are:
1320 0 Forwarding disabled
1321 1 Forwarding enabled
1325 By default, Host behaviour is assumed. This means:
1327 1. IsRouter flag is not set in Neighbour Advertisements.
1328 2. If accept_ra is TRUE (default), transmit Router
1330 3. If accept_ra is TRUE (default), accept Router
1331 Advertisements (and do autoconfiguration).
1332 4. If accept_redirects is TRUE (default), accept Redirects.
1336 If local forwarding is enabled, Router behaviour is assumed.
1337 This means exactly the reverse from the above:
1339 1. IsRouter flag is set in Neighbour Advertisements.
1340 2. Router Solicitations are not sent unless accept_ra is 2.
1341 3. Router Advertisements are ignored unless accept_ra is 2.
1342 4. Redirects are ignored.
1344 Default: 0 (disabled) if global forwarding is disabled (default),
1345 otherwise 1 (enabled).
1348 Default Hop Limit to set.
1352 Default Maximum Transfer Unit
1353 Default: 1280 (IPv6 required minimum)
1355 router_probe_interval - INTEGER
1356 Minimum interval (in seconds) between Router Probing described
1361 router_solicitation_delay - INTEGER
1362 Number of seconds to wait after interface is brought up
1363 before sending Router Solicitations.
1366 router_solicitation_interval - INTEGER
1367 Number of seconds to wait between Router Solicitations.
1370 router_solicitations - INTEGER
1371 Number of Router Solicitations to send until assuming no
1372 routers are present.
1375 use_tempaddr - INTEGER
1376 Preference for Privacy Extensions (RFC3041).
1377 <= 0 : disable Privacy Extensions
1378 == 1 : enable Privacy Extensions, but prefer public
1379 addresses over temporary addresses.
1380 > 1 : enable Privacy Extensions and prefer temporary
1381 addresses over public addresses.
1382 Default: 0 (for most devices)
1383 -1 (for point-to-point devices and loopback devices)
1385 temp_valid_lft - INTEGER
1386 valid lifetime (in seconds) for temporary addresses.
1387 Default: 604800 (7 days)
1389 temp_prefered_lft - INTEGER
1390 Preferred lifetime (in seconds) for temporary addresses.
1391 Default: 86400 (1 day)
1393 max_desync_factor - INTEGER
1394 Maximum value for DESYNC_FACTOR, which is a random value
1395 that ensures that clients don't synchronize with each
1396 other and generate new addresses at exactly the same time.
1397 value is in seconds.
1400 regen_max_retry - INTEGER
1401 Number of attempts before give up attempting to generate
1402 valid temporary addresses.
1405 max_addresses - INTEGER
1406 Maximum number of autoconfigured addresses per interface. Setting
1407 to zero disables the limitation. It is not recommended to set this
1408 value too large (or to zero) because it would be an easy way to
1409 crash the kernel by allowing too many addresses to be created.
1412 disable_ipv6 - BOOLEAN
1413 Disable IPv6 operation. If accept_dad is set to 2, this value
1414 will be dynamically set to TRUE if DAD fails for the link-local
1416 Default: FALSE (enable IPv6 operation)
1418 When this value is changed from 1 to 0 (IPv6 is being enabled),
1419 it will dynamically create a link-local address on the given
1420 interface and start Duplicate Address Detection, if necessary.
1422 When this value is changed from 0 to 1 (IPv6 is being disabled),
1423 it will dynamically delete all address on the given interface.
1425 accept_dad - INTEGER
1426 Whether to accept DAD (Duplicate Address Detection).
1428 1: Enable DAD (default)
1429 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1430 link-local address has been found.
1432 force_tllao - BOOLEAN
1433 Enable sending the target link-layer address option even when
1434 responding to a unicast neighbor solicitation.
1437 Quoting from RFC 2461, section 4.4, Target link-layer address:
1439 "The option MUST be included for multicast solicitations in order to
1440 avoid infinite Neighbor Solicitation "recursion" when the peer node
1441 does not have a cache entry to return a Neighbor Advertisements
1442 message. When responding to unicast solicitations, the option can be
1443 omitted since the sender of the solicitation has the correct link-
1444 layer address; otherwise it would not have be able to send the unicast
1445 solicitation in the first place. However, including the link-layer
1446 address in this case adds little overhead and eliminates a potential
1447 race condition where the sender deletes the cached link-layer address
1448 prior to receiving a response to a previous solicitation."
1450 ndisc_notify - BOOLEAN
1451 Define mode for notification of address and device changes.
1452 0 - (default): do nothing
1453 1 - Generate unsolicited neighbour advertisements when device is brought
1454 up or hardware address changes.
1456 mldv1_unsolicited_report_interval - INTEGER
1457 The interval in milliseconds in which the next unsolicited
1458 MLDv1 report retransmit will take place.
1459 Default: 10000 (10 seconds)
1461 mldv2_unsolicited_report_interval - INTEGER
1462 The interval in milliseconds in which the next unsolicited
1463 MLDv2 report retransmit will take place.
1464 Default: 1000 (1 second)
1466 force_mld_version - INTEGER
1467 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1468 1 - Enforce to use MLD version 1
1469 2 - Enforce to use MLD version 2
1471 suppress_frag_ndisc - INTEGER
1472 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1473 with IPv6 Neighbor Discovery) behavior:
1474 1 - (default) discard fragmented neighbor discovery packets
1475 0 - allow fragmented neighbor discovery packets
1477 optimistic_dad - BOOLEAN
1478 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
1479 0: disabled (default)
1482 use_optimistic - BOOLEAN
1483 If enabled, do not classify optimistic addresses as deprecated during
1484 source address selection. Preferred addresses will still be chosen
1485 before optimistic addresses, subject to other ranking in the source
1486 address selection algorithm.
1487 0: disabled (default)
1492 Limit the maximal rates for sending ICMPv6 packets.
1493 0 to disable any limiting,
1494 otherwise the minimal space between responses in milliseconds.
1499 Pekka Savola <pekkas@netcore.fi>
1500 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1503 /proc/sys/net/bridge/* Variables:
1505 bridge-nf-call-arptables - BOOLEAN
1506 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1510 bridge-nf-call-iptables - BOOLEAN
1511 1 : pass bridged IPv4 traffic to iptables' chains.
1515 bridge-nf-call-ip6tables - BOOLEAN
1516 1 : pass bridged IPv6 traffic to ip6tables' chains.
1520 bridge-nf-filter-vlan-tagged - BOOLEAN
1521 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1525 bridge-nf-filter-pppoe-tagged - BOOLEAN
1526 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1530 bridge-nf-pass-vlan-input-dev - BOOLEAN
1531 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1532 interface on the bridge and set the netfilter input device to the vlan.
1533 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1534 target work with vlan-on-top-of-bridge interfaces. When no matching
1535 vlan interface is found, or this switch is off, the input device is
1536 set to the bridge interface.
1537 0: disable bridge netfilter vlan interface lookup.
1540 proc/sys/net/sctp/* Variables:
1542 addip_enable - BOOLEAN
1543 Enable or disable extension of Dynamic Address Reconfiguration
1544 (ADD-IP) functionality specified in RFC5061. This extension provides
1545 the ability to dynamically add and remove new addresses for the SCTP
1548 1: Enable extension.
1550 0: Disable extension.
1554 addip_noauth_enable - BOOLEAN
1555 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1556 authentication to protect the operations of adding or removing new
1557 addresses. This requirement is mandated so that unauthorized hosts
1558 would not be able to hijack associations. However, older
1559 implementations may not have implemented this requirement while
1560 allowing the ADD-IP extension. For reasons of interoperability,
1561 we provide this variable to control the enforcement of the
1562 authentication requirement.
1564 1: Allow ADD-IP extension to be used without authentication. This
1565 should only be set in a closed environment for interoperability
1566 with older implementations.
1568 0: Enforce the authentication requirement
1572 auth_enable - BOOLEAN
1573 Enable or disable Authenticated Chunks extension. This extension
1574 provides the ability to send and receive authenticated chunks and is
1575 required for secure operation of Dynamic Address Reconfiguration
1578 1: Enable this extension.
1579 0: Disable this extension.
1583 prsctp_enable - BOOLEAN
1584 Enable or disable the Partial Reliability extension (RFC3758) which
1585 is used to notify peers that a given DATA should no longer be expected.
1593 The limit of the number of new packets that can be initially sent. It
1594 controls how bursty the generated traffic can be.
1598 association_max_retrans - INTEGER
1599 Set the maximum number for retransmissions that an association can
1600 attempt deciding that the remote end is unreachable. If this value
1601 is exceeded, the association is terminated.
1605 max_init_retransmits - INTEGER
1606 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1607 that an association will attempt before declaring the destination
1608 unreachable and terminating.
1612 path_max_retrans - INTEGER
1613 The maximum number of retransmissions that will be attempted on a given
1614 path. Once this threshold is exceeded, the path is considered
1615 unreachable, and new traffic will use a different path when the
1616 association is multihomed.
1620 pf_retrans - INTEGER
1621 The number of retransmissions that will be attempted on a given path
1622 before traffic is redirected to an alternate transport (should one
1623 exist). Note this is distinct from path_max_retrans, as a path that
1624 passes the pf_retrans threshold can still be used. Its only
1625 deprioritized when a transmission path is selected by the stack. This
1626 setting is primarily used to enable fast failover mechanisms without
1627 having to reduce path_max_retrans to a very low value. See:
1628 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1629 for details. Note also that a value of pf_retrans > path_max_retrans
1630 disables this feature
1634 rto_initial - INTEGER
1635 The initial round trip timeout value in milliseconds that will be used
1636 in calculating round trip times. This is the initial time interval
1637 for retransmissions.
1642 The maximum value (in milliseconds) of the round trip timeout. This
1643 is the largest time interval that can elapse between retransmissions.
1648 The minimum value (in milliseconds) of the round trip timeout. This
1649 is the smallest time interval the can elapse between retransmissions.
1653 hb_interval - INTEGER
1654 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1655 are sent at the specified interval on idle paths to probe the state of
1656 a given path between 2 associations.
1660 sack_timeout - INTEGER
1661 The amount of time (in milliseconds) that the implementation will wait
1666 valid_cookie_life - INTEGER
1667 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1668 is used during association establishment.
1672 cookie_preserve_enable - BOOLEAN
1673 Enable or disable the ability to extend the lifetime of the SCTP cookie
1674 that is used during the establishment phase of SCTP association
1676 1: Enable cookie lifetime extension.
1681 cookie_hmac_alg - STRING
1682 Select the hmac algorithm used when generating the cookie value sent by
1683 a listening sctp socket to a connecting client in the INIT-ACK chunk.
1688 Ability to assign md5 or sha1 as the selected alg is predicated on the
1689 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
1690 CONFIG_CRYPTO_SHA1).
1692 Default: Dependent on configuration. MD5 if available, else SHA1 if
1693 available, else none.
1695 rcvbuf_policy - INTEGER
1696 Determines if the receive buffer is attributed to the socket or to
1697 association. SCTP supports the capability to create multiple
1698 associations on a single socket. When using this capability, it is
1699 possible that a single stalled association that's buffering a lot
1700 of data may block other associations from delivering their data by
1701 consuming all of the receive buffer space. To work around this,
1702 the rcvbuf_policy could be set to attribute the receiver buffer space
1703 to each association instead of the socket. This prevents the described
1706 1: rcvbuf space is per association
1707 0: rcvbuf space is per socket
1711 sndbuf_policy - INTEGER
1712 Similar to rcvbuf_policy above, this applies to send buffer space.
1714 1: Send buffer is tracked per association
1715 0: Send buffer is tracked per socket.
1719 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1720 Number of pages allowed for queueing by all SCTP sockets.
1722 min: Below this number of pages SCTP is not bothered about its
1723 memory appetite. When amount of memory allocated by SCTP exceeds
1724 this number, SCTP starts to moderate memory usage.
1726 pressure: This value was introduced to follow format of tcp_mem.
1728 max: Number of pages allowed for queueing by all SCTP sockets.
1730 Default is calculated at boot time from amount of available memory.
1732 sctp_rmem - vector of 3 INTEGERs: min, default, max
1733 Only the first value ("min") is used, "default" and "max" are
1736 min: Minimal size of receive buffer used by SCTP socket.
1737 It is guaranteed to each SCTP socket (but not association) even
1738 under moderate memory pressure.
1742 sctp_wmem - vector of 3 INTEGERs: min, default, max
1743 Currently this tunable has no effect.
1745 addr_scope_policy - INTEGER
1746 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1748 0 - Disable IPv4 address scoping
1749 1 - Enable IPv4 address scoping
1750 2 - Follow draft but allow IPv4 private addresses
1751 3 - Follow draft but allow IPv4 link local addresses
1756 /proc/sys/net/core/*
1757 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
1760 /proc/sys/net/unix/*
1761 max_dgram_qlen - INTEGER
1762 The maximum length of dgram socket receive queue
1769 /proc/sys/net/irda/*
1770 fast_poll_increase FIXME
1771 warn_noreply_time FIXME
1772 discovery_slots FIXME
1775 discovery_timeout FIXME
1776 lap_keepalive_time FIXME
1777 max_noreply_time FIXME
1778 max_tx_data_size FIXME
1780 min_tx_turn_time FIXME