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 fib_multipath_use_neigh - BOOLEAN
67 Use status of existing neighbor entry when determining nexthop for
68 multipath routes. If disabled, neighbor information is not used and
69 packets could be directed to a failed nexthop. Only valid for kernels
70 built with CONFIG_IP_ROUTE_MULTIPATH enabled.
76 fib_multipath_hash_policy - INTEGER
77 Controls which hash policy to use for multipath routes. Only valid
78 for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
84 route/max_size - INTEGER
85 Maximum number of routes allowed in the kernel. Increase
86 this when using large numbers of interfaces and/or routes.
87 From linux kernel 3.6 onwards, this is deprecated for ipv4
88 as route cache is no longer used.
90 neigh/default/gc_thresh1 - INTEGER
91 Minimum number of entries to keep. Garbage collector will not
92 purge entries if there are fewer than this number.
95 neigh/default/gc_thresh2 - INTEGER
96 Threshold when garbage collector becomes more aggressive about
97 purging entries. Entries older than 5 seconds will be cleared
98 when over this number.
101 neigh/default/gc_thresh3 - INTEGER
102 Maximum number of neighbor entries allowed. Increase this
103 when using large numbers of interfaces and when communicating
104 with large numbers of directly-connected peers.
107 neigh/default/unres_qlen_bytes - INTEGER
108 The maximum number of bytes which may be used by packets
109 queued for each unresolved address by other network layers.
111 Setting negative value is meaningless and will return error.
112 Default: SK_WMEM_MAX, (same as net.core.wmem_default).
113 Exact value depends on architecture and kernel options,
114 but should be enough to allow queuing 256 packets
117 neigh/default/unres_qlen - INTEGER
118 The maximum number of packets which may be queued for each
119 unresolved address by other network layers.
120 (deprecated in linux 3.3) : use unres_qlen_bytes instead.
121 Prior to linux 3.3, the default value is 3 which may cause
122 unexpected packet loss. The current default value is calculated
123 according to default value of unres_qlen_bytes and true size of
127 mtu_expires - INTEGER
128 Time, in seconds, that cached PMTU information is kept.
130 min_adv_mss - INTEGER
131 The advertised MSS depends on the first hop route MTU, but will
132 never be lower than this setting.
136 ipfrag_high_thresh - LONG INTEGER
137 Maximum memory used to reassemble IP fragments.
139 ipfrag_low_thresh - LONG INTEGER
140 (Obsolete since linux-4.17)
141 Maximum memory used to reassemble IP fragments before the kernel
142 begins to remove incomplete fragment queues to free up resources.
143 The kernel still accepts new fragments for defragmentation.
145 ipfrag_time - INTEGER
146 Time in seconds to keep an IP fragment in memory.
148 ipfrag_max_dist - INTEGER
149 ipfrag_max_dist is a non-negative integer value which defines the
150 maximum "disorder" which is allowed among fragments which share a
151 common IP source address. Note that reordering of packets is
152 not unusual, but if a large number of fragments arrive from a source
153 IP address while a particular fragment queue remains incomplete, it
154 probably indicates that one or more fragments belonging to that queue
155 have been lost. When ipfrag_max_dist is positive, an additional check
156 is done on fragments before they are added to a reassembly queue - if
157 ipfrag_max_dist (or more) fragments have arrived from a particular IP
158 address between additions to any IP fragment queue using that source
159 address, it's presumed that one or more fragments in the queue are
160 lost. The existing fragment queue will be dropped, and a new one
161 started. An ipfrag_max_dist value of zero disables this check.
163 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
164 result in unnecessarily dropping fragment queues when normal
165 reordering of packets occurs, which could lead to poor application
166 performance. Using a very large value, e.g. 50000, increases the
167 likelihood of incorrectly reassembling IP fragments that originate
168 from different IP datagrams, which could result in data corruption.
173 inet_peer_threshold - INTEGER
174 The approximate size of the storage. Starting from this threshold
175 entries will be thrown aggressively. This threshold also determines
176 entries' time-to-live and time intervals between garbage collection
177 passes. More entries, less time-to-live, less GC interval.
179 inet_peer_minttl - INTEGER
180 Minimum time-to-live of entries. Should be enough to cover fragment
181 time-to-live on the reassembling side. This minimum time-to-live is
182 guaranteed if the pool size is less than inet_peer_threshold.
185 inet_peer_maxttl - INTEGER
186 Maximum time-to-live of entries. Unused entries will expire after
187 this period of time if there is no memory pressure on the pool (i.e.
188 when the number of entries in the pool is very small).
194 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
195 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
198 tcp_abort_on_overflow - BOOLEAN
199 If listening service is too slow to accept new connections,
200 reset them. Default state is FALSE. It means that if overflow
201 occurred due to a burst, connection will recover. Enable this
202 option _only_ if you are really sure that listening daemon
203 cannot be tuned to accept connections faster. Enabling this
204 option can harm clients of your server.
206 tcp_adv_win_scale - INTEGER
207 Count buffering overhead as bytes/2^tcp_adv_win_scale
208 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
210 Possible values are [-31, 31], inclusive.
213 tcp_allowed_congestion_control - STRING
214 Show/set the congestion control choices available to non-privileged
215 processes. The list is a subset of those listed in
216 tcp_available_congestion_control.
217 Default is "reno" and the default setting (tcp_congestion_control).
219 tcp_app_win - INTEGER
220 Reserve max(window/2^tcp_app_win, mss) of window for application
221 buffer. Value 0 is special, it means that nothing is reserved.
224 tcp_autocorking - BOOLEAN
225 Enable TCP auto corking :
226 When applications do consecutive small write()/sendmsg() system calls,
227 we try to coalesce these small writes as much as possible, to lower
228 total amount of sent packets. This is done if at least one prior
229 packet for the flow is waiting in Qdisc queues or device transmit
230 queue. Applications can still use TCP_CORK for optimal behavior
231 when they know how/when to uncork their sockets.
234 tcp_available_congestion_control - STRING
235 Shows the available congestion control choices that are registered.
236 More congestion control algorithms may be available as modules,
239 tcp_base_mss - INTEGER
240 The initial value of search_low to be used by the packetization layer
241 Path MTU discovery (MTU probing). If MTU probing is enabled,
242 this is the initial MSS used by the connection.
244 tcp_congestion_control - STRING
245 Set the congestion control algorithm to be used for new
246 connections. The algorithm "reno" is always available, but
247 additional choices may be available based on kernel configuration.
248 Default is set as part of kernel configuration.
249 For passive connections, the listener congestion control choice
251 [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
254 Allows TCP to send "duplicate" SACKs.
256 tcp_early_retrans - INTEGER
257 Tail loss probe (TLP) converts RTOs occurring due to tail
258 losses into fast recovery (draft-ietf-tcpm-rack). Note that
259 TLP requires RACK to function properly (see tcp_recovery below)
266 Control use of Explicit Congestion Notification (ECN) by TCP.
267 ECN is used only when both ends of the TCP connection indicate
268 support for it. This feature is useful in avoiding losses due
269 to congestion by allowing supporting routers to signal
270 congestion before having to drop packets.
272 0 Disable ECN. Neither initiate nor accept ECN.
273 1 Enable ECN when requested by incoming connections and
274 also request ECN on outgoing connection attempts.
275 2 Enable ECN when requested by incoming connections
276 but do not request ECN on outgoing connections.
279 tcp_ecn_fallback - BOOLEAN
280 If the kernel detects that ECN connection misbehaves, enable fall
281 back to non-ECN. Currently, this knob implements the fallback
282 from RFC3168, section 6.1.1.1., but we reserve that in future,
283 additional detection mechanisms could be implemented under this
284 knob. The value is not used, if tcp_ecn or per route (or congestion
285 control) ECN settings are disabled.
286 Default: 1 (fallback enabled)
289 This is a legacy option, it has no effect anymore.
291 tcp_fin_timeout - INTEGER
292 The length of time an orphaned (no longer referenced by any
293 application) connection will remain in the FIN_WAIT_2 state
294 before it is aborted at the local end. While a perfectly
295 valid "receive only" state for an un-orphaned connection, an
296 orphaned connection in FIN_WAIT_2 state could otherwise wait
297 forever for the remote to close its end of the connection.
302 Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
303 F-RTO is an enhanced recovery algorithm for TCP retransmission
304 timeouts. It is particularly beneficial in networks where the
305 RTT fluctuates (e.g., wireless). F-RTO is sender-side only
306 modification. It does not require any support from the peer.
308 By default it's enabled with a non-zero value. 0 disables F-RTO.
310 tcp_invalid_ratelimit - INTEGER
311 Limit the maximal rate for sending duplicate acknowledgments
312 in response to incoming TCP packets that are for an existing
313 connection but that are invalid due to any of these reasons:
315 (a) out-of-window sequence number,
316 (b) out-of-window acknowledgment number, or
317 (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
319 This can help mitigate simple "ack loop" DoS attacks, wherein
320 a buggy or malicious middlebox or man-in-the-middle can
321 rewrite TCP header fields in manner that causes each endpoint
322 to think that the other is sending invalid TCP segments, thus
323 causing each side to send an unterminating stream of duplicate
324 acknowledgments for invalid segments.
326 Using 0 disables rate-limiting of dupacks in response to
327 invalid segments; otherwise this value specifies the minimal
328 space between sending such dupacks, in milliseconds.
330 Default: 500 (milliseconds).
332 tcp_keepalive_time - INTEGER
333 How often TCP sends out keepalive messages when keepalive is enabled.
336 tcp_keepalive_probes - INTEGER
337 How many keepalive probes TCP sends out, until it decides that the
338 connection is broken. Default value: 9.
340 tcp_keepalive_intvl - INTEGER
341 How frequently the probes are send out. Multiplied by
342 tcp_keepalive_probes it is time to kill not responding connection,
343 after probes started. Default value: 75sec i.e. connection
344 will be aborted after ~11 minutes of retries.
346 tcp_l3mdev_accept - BOOLEAN
347 Enables child sockets to inherit the L3 master device index.
348 Enabling this option allows a "global" listen socket to work
349 across L3 master domains (e.g., VRFs) with connected sockets
350 derived from the listen socket to be bound to the L3 domain in
351 which the packets originated. Only valid when the kernel was
352 compiled with CONFIG_NET_L3_MASTER_DEV.
354 tcp_low_latency - BOOLEAN
355 This is a legacy option, it has no effect anymore.
357 tcp_max_orphans - INTEGER
358 Maximal number of TCP sockets not attached to any user file handle,
359 held by system. If this number is exceeded orphaned connections are
360 reset immediately and warning is printed. This limit exists
361 only to prevent simple DoS attacks, you _must_ not rely on this
362 or lower the limit artificially, but rather increase it
363 (probably, after increasing installed memory),
364 if network conditions require more than default value,
365 and tune network services to linger and kill such states
366 more aggressively. Let me to remind again: each orphan eats
367 up to ~64K of unswappable memory.
369 tcp_max_syn_backlog - INTEGER
370 Maximal number of remembered connection requests, which have not
371 received an acknowledgment from connecting client.
372 The minimal value is 128 for low memory machines, and it will
373 increase in proportion to the memory of machine.
374 If server suffers from overload, try increasing this number.
376 tcp_max_tw_buckets - INTEGER
377 Maximal number of timewait sockets held by system simultaneously.
378 If this number is exceeded time-wait socket is immediately destroyed
379 and warning is printed. This limit exists only to prevent
380 simple DoS attacks, you _must_ not lower the limit artificially,
381 but rather increase it (probably, after increasing installed memory),
382 if network conditions require more than default value.
384 tcp_mem - vector of 3 INTEGERs: min, pressure, max
385 min: below this number of pages TCP is not bothered about its
388 pressure: when amount of memory allocated by TCP exceeds this number
389 of pages, TCP moderates its memory consumption and enters memory
390 pressure mode, which is exited when memory consumption falls
393 max: number of pages allowed for queueing by all TCP sockets.
395 Defaults are calculated at boot time from amount of available
398 tcp_min_rtt_wlen - INTEGER
399 The window length of the windowed min filter to track the minimum RTT.
400 A shorter window lets a flow more quickly pick up new (higher)
401 minimum RTT when it is moved to a longer path (e.g., due to traffic
402 engineering). A longer window makes the filter more resistant to RTT
403 inflations such as transient congestion. The unit is seconds.
406 tcp_moderate_rcvbuf - BOOLEAN
407 If set, TCP performs receive buffer auto-tuning, attempting to
408 automatically size the buffer (no greater than tcp_rmem[2]) to
409 match the size required by the path for full throughput. Enabled by
412 tcp_mtu_probing - INTEGER
413 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
416 1 - Disabled by default, enabled when an ICMP black hole detected
417 2 - Always enabled, use initial MSS of tcp_base_mss.
419 tcp_probe_interval - INTEGER
420 Controls how often to start TCP Packetization-Layer Path MTU
421 Discovery reprobe. The default is reprobing every 10 minutes as
424 tcp_probe_threshold - INTEGER
425 Controls when TCP Packetization-Layer Path MTU Discovery probing
426 will stop in respect to the width of search range in bytes. Default
429 tcp_no_metrics_save - BOOLEAN
430 By default, TCP saves various connection metrics in the route cache
431 when the connection closes, so that connections established in the
432 near future can use these to set initial conditions. Usually, this
433 increases overall performance, but may sometimes cause performance
434 degradation. If set, TCP will not cache metrics on closing
437 tcp_orphan_retries - INTEGER
438 This value influences the timeout of a locally closed TCP connection,
439 when RTO retransmissions remain unacknowledged.
440 See tcp_retries2 for more details.
442 The default value is 8.
443 If your machine is a loaded WEB server,
444 you should think about lowering this value, such sockets
445 may consume significant resources. Cf. tcp_max_orphans.
447 tcp_recovery - INTEGER
448 This value is a bitmap to enable various experimental loss recovery
451 RACK: 0x1 enables the RACK loss detection for fast detection of lost
452 retransmissions and tail drops.
453 RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
457 tcp_reordering - INTEGER
458 Initial reordering level of packets in a TCP stream.
459 TCP stack can then dynamically adjust flow reordering level
460 between this initial value and tcp_max_reordering
463 tcp_max_reordering - INTEGER
464 Maximal reordering level of packets in a TCP stream.
465 300 is a fairly conservative value, but you might increase it
466 if paths are using per packet load balancing (like bonding rr mode)
469 tcp_retrans_collapse - BOOLEAN
470 Bug-to-bug compatibility with some broken printers.
471 On retransmit try to send bigger packets to work around bugs in
474 tcp_retries1 - INTEGER
475 This value influences the time, after which TCP decides, that
476 something is wrong due to unacknowledged RTO retransmissions,
477 and reports this suspicion to the network layer.
478 See tcp_retries2 for more details.
480 RFC 1122 recommends at least 3 retransmissions, which is the
483 tcp_retries2 - INTEGER
484 This value influences the timeout of an alive TCP connection,
485 when RTO retransmissions remain unacknowledged.
486 Given a value of N, a hypothetical TCP connection following
487 exponential backoff with an initial RTO of TCP_RTO_MIN would
488 retransmit N times before killing the connection at the (N+1)th RTO.
490 The default value of 15 yields a hypothetical timeout of 924.6
491 seconds and is a lower bound for the effective timeout.
492 TCP will effectively time out at the first RTO which exceeds the
493 hypothetical timeout.
495 RFC 1122 recommends at least 100 seconds for the timeout,
496 which corresponds to a value of at least 8.
498 tcp_rfc1337 - BOOLEAN
499 If set, the TCP stack behaves conforming to RFC1337. If unset,
500 we are not conforming to RFC, but prevent TCP TIME_WAIT
504 tcp_rmem - vector of 3 INTEGERs: min, default, max
505 min: Minimal size of receive buffer used by TCP sockets.
506 It is guaranteed to each TCP socket, even under moderate memory
510 default: initial size of receive buffer used by TCP sockets.
511 This value overrides net.core.rmem_default used by other protocols.
512 Default: 87380 bytes. This value results in window of 65535 with
513 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
514 less for default tcp_app_win. See below about these variables.
516 max: maximal size of receive buffer allowed for automatically
517 selected receiver buffers for TCP socket. This value does not override
518 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
519 automatic tuning of that socket's receive buffer size, in which
520 case this value is ignored.
521 Default: between 87380B and 6MB, depending on RAM size.
524 Enable select acknowledgments (SACKS).
526 tcp_slow_start_after_idle - BOOLEAN
527 If set, provide RFC2861 behavior and time out the congestion
528 window after an idle period. An idle period is defined at
529 the current RTO. If unset, the congestion window will not
530 be timed out after an idle period.
534 Use the Host requirements interpretation of the TCP urgent pointer field.
535 Most hosts use the older BSD interpretation, so if you turn this on
536 Linux might not communicate correctly with them.
539 tcp_synack_retries - INTEGER
540 Number of times SYNACKs for a passive TCP connection attempt will
541 be retransmitted. Should not be higher than 255. Default value
542 is 5, which corresponds to 31seconds till the last retransmission
543 with the current initial RTO of 1second. With this the final timeout
544 for a passive TCP connection will happen after 63seconds.
546 tcp_syncookies - BOOLEAN
547 Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
548 Send out syncookies when the syn backlog queue of a socket
549 overflows. This is to prevent against the common 'SYN flood attack'
552 Note, that syncookies is fallback facility.
553 It MUST NOT be used to help highly loaded servers to stand
554 against legal connection rate. If you see SYN flood warnings
555 in your logs, but investigation shows that they occur
556 because of overload with legal connections, you should tune
557 another parameters until this warning disappear.
558 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
560 syncookies seriously violate TCP protocol, do not allow
561 to use TCP extensions, can result in serious degradation
562 of some services (f.e. SMTP relaying), visible not by you,
563 but your clients and relays, contacting you. While you see
564 SYN flood warnings in logs not being really flooded, your server
565 is seriously misconfigured.
567 If you want to test which effects syncookies have to your
568 network connections you can set this knob to 2 to enable
569 unconditionally generation of syncookies.
571 tcp_fastopen - INTEGER
572 Enable TCP Fast Open (RFC7413) to send and accept data in the opening
575 The client support is enabled by flag 0x1 (on by default). The client
576 then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
577 rather than connect() to send data in SYN.
579 The server support is enabled by flag 0x2 (off by default). Then
580 either enable for all listeners with another flag (0x400) or
581 enable individual listeners via TCP_FASTOPEN socket option with
582 the option value being the length of the syn-data backlog.
584 The values (bitmap) are
585 0x1: (client) enables sending data in the opening SYN on the client.
586 0x2: (server) enables the server support, i.e., allowing data in
587 a SYN packet to be accepted and passed to the
588 application before 3-way handshake finishes.
589 0x4: (client) send data in the opening SYN regardless of cookie
590 availability and without a cookie option.
591 0x200: (server) accept data-in-SYN w/o any cookie option present.
592 0x400: (server) enable all listeners to support Fast Open by
593 default without explicit TCP_FASTOPEN socket option.
597 Note that that additional client or server features are only
598 effective if the basic support (0x1 and 0x2) are enabled respectively.
600 tcp_fastopen_blackhole_timeout_sec - INTEGER
601 Initial time period in second to disable Fastopen on active TCP sockets
602 when a TFO firewall blackhole issue happens.
603 This time period will grow exponentially when more blackhole issues
604 get detected right after Fastopen is re-enabled and will reset to
605 initial value when the blackhole issue goes away.
606 0 to disable the blackhole detection.
607 By default, it is set to 1hr.
609 tcp_syn_retries - INTEGER
610 Number of times initial SYNs for an active TCP connection attempt
611 will be retransmitted. Should not be higher than 127. Default value
612 is 6, which corresponds to 63seconds till the last retransmission
613 with the current initial RTO of 1second. With this the final timeout
614 for an active TCP connection attempt will happen after 127seconds.
616 tcp_timestamps - INTEGER
617 Enable timestamps as defined in RFC1323.
619 1: Enable timestamps as defined in RFC1323 and use random offset for
620 each connection rather than only using the current time.
621 2: Like 1, but without random offsets.
624 tcp_min_tso_segs - INTEGER
625 Minimal number of segments per TSO frame.
626 Since linux-3.12, TCP does an automatic sizing of TSO frames,
627 depending on flow rate, instead of filling 64Kbytes packets.
628 For specific usages, it's possible to force TCP to build big
629 TSO frames. Note that TCP stack might split too big TSO packets
630 if available window is too small.
633 tcp_pacing_ss_ratio - INTEGER
634 sk->sk_pacing_rate is set by TCP stack using a ratio applied
635 to current rate. (current_rate = cwnd * mss / srtt)
636 If TCP is in slow start, tcp_pacing_ss_ratio is applied
637 to let TCP probe for bigger speeds, assuming cwnd can be
638 doubled every other RTT.
641 tcp_pacing_ca_ratio - INTEGER
642 sk->sk_pacing_rate is set by TCP stack using a ratio applied
643 to current rate. (current_rate = cwnd * mss / srtt)
644 If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
645 is applied to conservatively probe for bigger throughput.
648 tcp_tso_win_divisor - INTEGER
649 This allows control over what percentage of the congestion window
650 can be consumed by a single TSO frame.
651 The setting of this parameter is a choice between burstiness and
652 building larger TSO frames.
655 tcp_tw_reuse - BOOLEAN
656 Allow to reuse TIME-WAIT sockets for new connections when it is
657 safe from protocol viewpoint. Default value is 0.
658 It should not be changed without advice/request of technical
661 tcp_window_scaling - BOOLEAN
662 Enable window scaling as defined in RFC1323.
664 tcp_wmem - vector of 3 INTEGERs: min, default, max
665 min: Amount of memory reserved for send buffers for TCP sockets.
666 Each TCP socket has rights to use it due to fact of its birth.
669 default: initial size of send buffer used by TCP sockets. This
670 value overrides net.core.wmem_default used by other protocols.
671 It is usually lower than net.core.wmem_default.
674 max: Maximal amount of memory allowed for automatically tuned
675 send buffers for TCP sockets. This value does not override
676 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
677 automatic tuning of that socket's send buffer size, in which case
678 this value is ignored.
679 Default: between 64K and 4MB, depending on RAM size.
681 tcp_notsent_lowat - UNSIGNED INTEGER
682 A TCP socket can control the amount of unsent bytes in its write queue,
683 thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
684 reports POLLOUT events if the amount of unsent bytes is below a per
685 socket value, and if the write queue is not full. sendmsg() will
686 also not add new buffers if the limit is hit.
688 This global variable controls the amount of unsent data for
689 sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
690 to the global variable has immediate effect.
692 Default: UINT_MAX (0xFFFFFFFF)
694 tcp_workaround_signed_windows - BOOLEAN
695 If set, assume no receipt of a window scaling option means the
696 remote TCP is broken and treats the window as a signed quantity.
697 If unset, assume the remote TCP is not broken even if we do
698 not receive a window scaling option from them.
701 tcp_thin_linear_timeouts - BOOLEAN
702 Enable dynamic triggering of linear timeouts for thin streams.
703 If set, a check is performed upon retransmission by timeout to
704 determine if the stream is thin (less than 4 packets in flight).
705 As long as the stream is found to be thin, up to 6 linear
706 timeouts may be performed before exponential backoff mode is
707 initiated. This improves retransmission latency for
708 non-aggressive thin streams, often found to be time-dependent.
709 For more information on thin streams, see
710 Documentation/networking/tcp-thin.txt
713 tcp_limit_output_bytes - INTEGER
714 Controls TCP Small Queue limit per tcp socket.
715 TCP bulk sender tends to increase packets in flight until it
716 gets losses notifications. With SNDBUF autotuning, this can
717 result in a large amount of packets queued in qdisc/device
718 on the local machine, hurting latency of other flows, for
719 typical pfifo_fast qdiscs.
720 tcp_limit_output_bytes limits the number of bytes on qdisc
721 or device to reduce artificial RTT/cwnd and reduce bufferbloat.
724 tcp_challenge_ack_limit - INTEGER
725 Limits number of Challenge ACK sent per second, as recommended
726 in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
731 udp_l3mdev_accept - BOOLEAN
732 Enabling this option allows a "global" bound socket to work
733 across L3 master domains (e.g., VRFs) with packets capable of
734 being received regardless of the L3 domain in which they
735 originated. Only valid when the kernel was compiled with
736 CONFIG_NET_L3_MASTER_DEV.
738 udp_mem - vector of 3 INTEGERs: min, pressure, max
739 Number of pages allowed for queueing by all UDP sockets.
741 min: Below this number of pages UDP is not bothered about its
742 memory appetite. When amount of memory allocated by UDP exceeds
743 this number, UDP starts to moderate memory usage.
745 pressure: This value was introduced to follow format of tcp_mem.
747 max: Number of pages allowed for queueing by all UDP sockets.
749 Default is calculated at boot time from amount of available memory.
751 udp_rmem_min - INTEGER
752 Minimal size of receive buffer used by UDP sockets in moderation.
753 Each UDP socket is able to use the size for receiving data, even if
754 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
757 udp_wmem_min - INTEGER
758 Minimal size of send buffer used by UDP sockets in moderation.
759 Each UDP socket is able to use the size for sending data, even if
760 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
765 cipso_cache_enable - BOOLEAN
766 If set, enable additions to and lookups from the CIPSO label mapping
767 cache. If unset, additions are ignored and lookups always result in a
768 miss. However, regardless of the setting the cache is still
769 invalidated when required when means you can safely toggle this on and
770 off and the cache will always be "safe".
773 cipso_cache_bucket_size - INTEGER
774 The CIPSO label cache consists of a fixed size hash table with each
775 hash bucket containing a number of cache entries. This variable limits
776 the number of entries in each hash bucket; the larger the value the
777 more CIPSO label mappings that can be cached. When the number of
778 entries in a given hash bucket reaches this limit adding new entries
779 causes the oldest entry in the bucket to be removed to make room.
782 cipso_rbm_optfmt - BOOLEAN
783 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
784 the CIPSO draft specification (see Documentation/netlabel for details).
785 This means that when set the CIPSO tag will be padded with empty
786 categories in order to make the packet data 32-bit aligned.
789 cipso_rbm_structvalid - BOOLEAN
790 If set, do a very strict check of the CIPSO option when
791 ip_options_compile() is called. If unset, relax the checks done during
792 ip_options_compile(). Either way is "safe" as errors are caught else
793 where in the CIPSO processing code but setting this to 0 (False) should
794 result in less work (i.e. it should be faster) but could cause problems
795 with other implementations that require strict checking.
800 ip_local_port_range - 2 INTEGERS
801 Defines the local port range that is used by TCP and UDP to
802 choose the local port. The first number is the first, the
803 second the last local port number.
804 If possible, it is better these numbers have different parity.
805 (one even and one odd values)
806 The default values are 32768 and 60999 respectively.
808 ip_local_reserved_ports - list of comma separated ranges
809 Specify the ports which are reserved for known third-party
810 applications. These ports will not be used by automatic port
811 assignments (e.g. when calling connect() or bind() with port
812 number 0). Explicit port allocation behavior is unchanged.
814 The format used for both input and output is a comma separated
815 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
816 10). Writing to the file will clear all previously reserved
817 ports and update the current list with the one given in the
820 Note that ip_local_port_range and ip_local_reserved_ports
821 settings are independent and both are considered by the kernel
822 when determining which ports are available for automatic port
825 You can reserve ports which are not in the current
826 ip_local_port_range, e.g.:
828 $ cat /proc/sys/net/ipv4/ip_local_port_range
830 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
833 although this is redundant. However such a setting is useful
834 if later the port range is changed to a value that will
835 include the reserved ports.
839 ip_unprivileged_port_start - INTEGER
840 This is a per-namespace sysctl. It defines the first
841 unprivileged port in the network namespace. Privileged ports
842 require root or CAP_NET_BIND_SERVICE in order to bind to them.
843 To disable all privileged ports, set this to 0. It may not
844 overlap with the ip_local_reserved_ports range.
848 ip_nonlocal_bind - BOOLEAN
849 If set, allows processes to bind() to non-local IP addresses,
850 which can be quite useful - but may break some applications.
854 If set non-zero, enables support for dynamic addresses.
855 If set to a non-zero value larger than 1, a kernel log
856 message will be printed when dynamic address rewriting
860 ip_early_demux - BOOLEAN
861 Optimize input packet processing down to one demux for
862 certain kinds of local sockets. Currently we only do this
863 for established TCP and connected UDP sockets.
865 It may add an additional cost for pure routing workloads that
866 reduces overall throughput, in such case you should disable it.
869 tcp_early_demux - BOOLEAN
870 Enable early demux for established TCP sockets.
873 udp_early_demux - BOOLEAN
874 Enable early demux for connected UDP sockets. Disable this if
875 your system could experience more unconnected load.
878 icmp_echo_ignore_all - BOOLEAN
879 If set non-zero, then the kernel will ignore all ICMP ECHO
883 icmp_echo_ignore_broadcasts - BOOLEAN
884 If set non-zero, then the kernel will ignore all ICMP ECHO and
885 TIMESTAMP requests sent to it via broadcast/multicast.
888 icmp_ratelimit - INTEGER
889 Limit the maximal rates for sending ICMP packets whose type matches
890 icmp_ratemask (see below) to specific targets.
891 0 to disable any limiting,
892 otherwise the minimal space between responses in milliseconds.
893 Note that another sysctl, icmp_msgs_per_sec limits the number
894 of ICMP packets sent on all targets.
897 icmp_msgs_per_sec - INTEGER
898 Limit maximal number of ICMP packets sent per second from this host.
899 Only messages whose type matches icmp_ratemask (see below) are
900 controlled by this limit.
903 icmp_msgs_burst - INTEGER
904 icmp_msgs_per_sec controls number of ICMP packets sent per second,
905 while icmp_msgs_burst controls the burst size of these packets.
908 icmp_ratemask - INTEGER
909 Mask made of ICMP types for which rates are being limited.
910 Significant bits: IHGFEDCBA9876543210
911 Default mask: 0000001100000011000 (6168)
913 Bit definitions (see include/linux/icmp.h):
915 3 Destination Unreachable *
920 C Parameter Problem *
925 H Address Mask Request
928 * These are rate limited by default (see default mask above)
930 icmp_ignore_bogus_error_responses - BOOLEAN
931 Some routers violate RFC1122 by sending bogus responses to broadcast
932 frames. Such violations are normally logged via a kernel warning.
933 If this is set to TRUE, the kernel will not give such warnings, which
934 will avoid log file clutter.
937 icmp_errors_use_inbound_ifaddr - BOOLEAN
939 If zero, icmp error messages are sent with the primary address of
940 the exiting interface.
942 If non-zero, the message will be sent with the primary address of
943 the interface that received the packet that caused the icmp error.
944 This is the behaviour network many administrators will expect from
945 a router. And it can make debugging complicated network layouts
948 Note that if no primary address exists for the interface selected,
949 then the primary address of the first non-loopback interface that
950 has one will be used regardless of this setting.
954 igmp_max_memberships - INTEGER
955 Change the maximum number of multicast groups we can subscribe to.
958 Theoretical maximum value is bounded by having to send a membership
959 report in a single datagram (i.e. the report can't span multiple
960 datagrams, or risk confusing the switch and leaving groups you don't
963 The number of supported groups 'M' is bounded by the number of group
964 report entries you can fit into a single datagram of 65535 bytes.
966 M = 65536-sizeof (ip header)/(sizeof(Group record))
968 Group records are variable length, with a minimum of 12 bytes.
969 So net.ipv4.igmp_max_memberships should not be set higher than:
971 (65536-24) / 12 = 5459
973 The value 5459 assumes no IP header options, so in practice
974 this number may be lower.
976 igmp_max_msf - INTEGER
977 Maximum number of addresses allowed in the source filter list for a
982 Controls the IGMP query robustness variable (see RFC2236 8.1).
983 Default: 2 (as specified by RFC2236 8.1)
984 Minimum: 1 (as specified by RFC6636 4.5)
986 force_igmp_version - INTEGER
987 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
988 allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
989 Present timer expires.
990 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
991 receive IGMPv2/v3 query.
992 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
993 IGMPv1 query message. Will reply report if receive IGMPv3 query.
994 3 - Enforce to use IGMP version 3. The same react with default 0.
996 Note: this is not the same with force_mld_version because IGMPv3 RFC3376
997 Security Considerations does not have clear description that we could
998 ignore other version messages completely as MLDv2 RFC3810. So make
999 this value as default 0 is recommended.
1001 conf/interface/* changes special settings per interface (where
1002 "interface" is the name of your network interface)
1004 conf/all/* is special, changes the settings for all interfaces
1006 log_martians - BOOLEAN
1007 Log packets with impossible addresses to kernel log.
1008 log_martians for the interface will be enabled if at least one of
1009 conf/{all,interface}/log_martians is set to TRUE,
1010 it will be disabled otherwise
1012 accept_redirects - BOOLEAN
1013 Accept ICMP redirect messages.
1014 accept_redirects for the interface will be enabled if:
1015 - both conf/{all,interface}/accept_redirects are TRUE in the case
1016 forwarding for the interface is enabled
1018 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
1019 case forwarding for the interface is disabled
1020 accept_redirects for the interface will be disabled otherwise
1024 forwarding - BOOLEAN
1025 Enable IP forwarding on this interface. This controls whether packets
1026 received _on_ this interface can be forwarded.
1028 mc_forwarding - BOOLEAN
1029 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1030 and a multicast routing daemon is required.
1031 conf/all/mc_forwarding must also be set to TRUE to enable multicast
1032 routing for the interface
1035 Integer value used to differentiate the devices by the medium they
1036 are attached to. Two devices can have different id values when
1037 the broadcast packets are received only on one of them.
1038 The default value 0 means that the device is the only interface
1039 to its medium, value of -1 means that medium is not known.
1041 Currently, it is used to change the proxy_arp behavior:
1042 the proxy_arp feature is enabled for packets forwarded between
1043 two devices attached to different media.
1047 proxy_arp for the interface will be enabled if at least one of
1048 conf/{all,interface}/proxy_arp is set to TRUE,
1049 it will be disabled otherwise
1051 proxy_arp_pvlan - BOOLEAN
1052 Private VLAN proxy arp.
1053 Basically allow proxy arp replies back to the same interface
1054 (from which the ARP request/solicitation was received).
1056 This is done to support (ethernet) switch features, like RFC
1057 3069, where the individual ports are NOT allowed to
1058 communicate with each other, but they are allowed to talk to
1059 the upstream router. As described in RFC 3069, it is possible
1060 to allow these hosts to communicate through the upstream
1061 router by proxy_arp'ing. Don't need to be used together with
1064 This technology is known by different names:
1065 In RFC 3069 it is called VLAN Aggregation.
1066 Cisco and Allied Telesyn call it Private VLAN.
1067 Hewlett-Packard call it Source-Port filtering or port-isolation.
1068 Ericsson call it MAC-Forced Forwarding (RFC Draft).
1070 shared_media - BOOLEAN
1071 Send(router) or accept(host) RFC1620 shared media redirects.
1072 Overrides secure_redirects.
1073 shared_media for the interface will be enabled if at least one of
1074 conf/{all,interface}/shared_media is set to TRUE,
1075 it will be disabled otherwise
1078 secure_redirects - BOOLEAN
1079 Accept ICMP redirect messages only to gateways listed in the
1080 interface's current gateway list. Even if disabled, RFC1122 redirect
1082 Overridden by shared_media.
1083 secure_redirects for the interface will be enabled if at least one of
1084 conf/{all,interface}/secure_redirects is set to TRUE,
1085 it will be disabled otherwise
1088 send_redirects - BOOLEAN
1089 Send redirects, if router.
1090 send_redirects for the interface will be enabled if at least one of
1091 conf/{all,interface}/send_redirects is set to TRUE,
1092 it will be disabled otherwise
1095 bootp_relay - BOOLEAN
1096 Accept packets with source address 0.b.c.d destined
1097 not to this host as local ones. It is supposed, that
1098 BOOTP relay daemon will catch and forward such packets.
1099 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1102 Not Implemented Yet.
1104 accept_source_route - BOOLEAN
1105 Accept packets with SRR option.
1106 conf/all/accept_source_route must also be set to TRUE to accept packets
1107 with SRR option on the interface
1108 default TRUE (router)
1111 accept_local - BOOLEAN
1112 Accept packets with local source addresses. In combination with
1113 suitable routing, this can be used to direct packets between two
1114 local interfaces over the wire and have them accepted properly.
1117 route_localnet - BOOLEAN
1118 Do not consider loopback addresses as martian source or destination
1119 while routing. This enables the use of 127/8 for local routing purposes.
1123 0 - No source validation.
1124 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1125 Each incoming packet is tested against the FIB and if the interface
1126 is not the best reverse path the packet check will fail.
1127 By default failed packets are discarded.
1128 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1129 Each incoming packet's source address is also tested against the FIB
1130 and if the source address is not reachable via any interface
1131 the packet check will fail.
1133 Current recommended practice in RFC3704 is to enable strict mode
1134 to prevent IP spoofing from DDos attacks. If using asymmetric routing
1135 or other complicated routing, then loose mode is recommended.
1137 The max value from conf/{all,interface}/rp_filter is used
1138 when doing source validation on the {interface}.
1140 Default value is 0. Note that some distributions enable it
1143 arp_filter - BOOLEAN
1144 1 - Allows you to have multiple network interfaces on the same
1145 subnet, and have the ARPs for each interface be answered
1146 based on whether or not the kernel would route a packet from
1147 the ARP'd IP out that interface (therefore you must use source
1148 based routing for this to work). In other words it allows control
1149 of which cards (usually 1) will respond to an arp request.
1151 0 - (default) The kernel can respond to arp requests with addresses
1152 from other interfaces. This may seem wrong but it usually makes
1153 sense, because it increases the chance of successful communication.
1154 IP addresses are owned by the complete host on Linux, not by
1155 particular interfaces. Only for more complex setups like load-
1156 balancing, does this behaviour cause problems.
1158 arp_filter for the interface will be enabled if at least one of
1159 conf/{all,interface}/arp_filter is set to TRUE,
1160 it will be disabled otherwise
1162 arp_announce - INTEGER
1163 Define different restriction levels for announcing the local
1164 source IP address from IP packets in ARP requests sent on
1166 0 - (default) Use any local address, configured on any interface
1167 1 - Try to avoid local addresses that are not in the target's
1168 subnet for this interface. This mode is useful when target
1169 hosts reachable via this interface require the source IP
1170 address in ARP requests to be part of their logical network
1171 configured on the receiving interface. When we generate the
1172 request we will check all our subnets that include the
1173 target IP and will preserve the source address if it is from
1174 such subnet. If there is no such subnet we select source
1175 address according to the rules for level 2.
1176 2 - Always use the best local address for this target.
1177 In this mode we ignore the source address in the IP packet
1178 and try to select local address that we prefer for talks with
1179 the target host. Such local address is selected by looking
1180 for primary IP addresses on all our subnets on the outgoing
1181 interface that include the target IP address. If no suitable
1182 local address is found we select the first local address
1183 we have on the outgoing interface or on all other interfaces,
1184 with the hope we will receive reply for our request and
1185 even sometimes no matter the source IP address we announce.
1187 The max value from conf/{all,interface}/arp_announce is used.
1189 Increasing the restriction level gives more chance for
1190 receiving answer from the resolved target while decreasing
1191 the level announces more valid sender's information.
1193 arp_ignore - INTEGER
1194 Define different modes for sending replies in response to
1195 received ARP requests that resolve local target IP addresses:
1196 0 - (default): reply for any local target IP address, configured
1198 1 - reply only if the target IP address is local address
1199 configured on the incoming interface
1200 2 - reply only if the target IP address is local address
1201 configured on the incoming interface and both with the
1202 sender's IP address are part from same subnet on this interface
1203 3 - do not reply for local addresses configured with scope host,
1204 only resolutions for global and link addresses are replied
1206 8 - do not reply for all local addresses
1208 The max value from conf/{all,interface}/arp_ignore is used
1209 when ARP request is received on the {interface}
1211 arp_notify - BOOLEAN
1212 Define mode for notification of address and device changes.
1213 0 - (default): do nothing
1214 1 - Generate gratuitous arp requests when device is brought up
1215 or hardware address changes.
1217 arp_accept - BOOLEAN
1218 Define behavior for gratuitous ARP frames who's IP is not
1219 already present in the ARP table:
1220 0 - don't create new entries in the ARP table
1221 1 - create new entries in the ARP table
1223 Both replies and requests type gratuitous arp will trigger the
1224 ARP table to be updated, if this setting is on.
1226 If the ARP table already contains the IP address of the
1227 gratuitous arp frame, the arp table will be updated regardless
1228 if this setting is on or off.
1230 mcast_solicit - INTEGER
1231 The maximum number of multicast probes in INCOMPLETE state,
1232 when the associated hardware address is unknown. Defaults
1235 ucast_solicit - INTEGER
1236 The maximum number of unicast probes in PROBE state, when
1237 the hardware address is being reconfirmed. Defaults to 3.
1239 app_solicit - INTEGER
1240 The maximum number of probes to send to the user space ARP daemon
1241 via netlink before dropping back to multicast probes (see
1242 mcast_resolicit). Defaults to 0.
1244 mcast_resolicit - INTEGER
1245 The maximum number of multicast probes after unicast and
1246 app probes in PROBE state. Defaults to 0.
1248 disable_policy - BOOLEAN
1249 Disable IPSEC policy (SPD) for this interface
1251 disable_xfrm - BOOLEAN
1252 Disable IPSEC encryption on this interface, whatever the policy
1254 igmpv2_unsolicited_report_interval - INTEGER
1255 The interval in milliseconds in which the next unsolicited
1256 IGMPv1 or IGMPv2 report retransmit will take place.
1257 Default: 10000 (10 seconds)
1259 igmpv3_unsolicited_report_interval - INTEGER
1260 The interval in milliseconds in which the next unsolicited
1261 IGMPv3 report retransmit will take place.
1262 Default: 1000 (1 seconds)
1264 promote_secondaries - BOOLEAN
1265 When a primary IP address is removed from this interface
1266 promote a corresponding secondary IP address instead of
1267 removing all the corresponding secondary IP addresses.
1269 drop_unicast_in_l2_multicast - BOOLEAN
1270 Drop any unicast IP packets that are received in link-layer
1271 multicast (or broadcast) frames.
1272 This behavior (for multicast) is actually a SHOULD in RFC
1273 1122, but is disabled by default for compatibility reasons.
1276 drop_gratuitous_arp - BOOLEAN
1277 Drop all gratuitous ARP frames, for example if there's a known
1278 good ARP proxy on the network and such frames need not be used
1279 (or in the case of 802.11, must not be used to prevent attacks.)
1284 Allows you to write a number, which can be used as required.
1287 xfrm4_gc_thresh - INTEGER
1288 The threshold at which we will start garbage collecting for IPv4
1289 destination cache entries. At twice this value the system will
1290 refuse new allocations.
1292 igmp_link_local_mcast_reports - BOOLEAN
1293 Enable IGMP reports for link local multicast groups in the
1298 kuznet@ms2.inr.ac.ru
1304 delon.nicolas@wanadoo.fr
1309 /proc/sys/net/ipv6/* Variables:
1311 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
1312 apply to IPv6 [XXX?].
1314 bindv6only - BOOLEAN
1315 Default value for IPV6_V6ONLY socket option,
1316 which restricts use of the IPv6 socket to IPv6 communication
1318 TRUE: disable IPv4-mapped address feature
1319 FALSE: enable IPv4-mapped address feature
1321 Default: FALSE (as specified in RFC3493)
1323 flowlabel_consistency - BOOLEAN
1324 Protect the consistency (and unicity) of flow label.
1325 You have to disable it to use IPV6_FL_F_REFLECT flag on the
1331 auto_flowlabels - INTEGER
1332 Automatically generate flow labels based on a flow hash of the
1333 packet. This allows intermediate devices, such as routers, to
1334 identify packet flows for mechanisms like Equal Cost Multipath
1335 Routing (see RFC 6438).
1336 0: automatic flow labels are completely disabled
1337 1: automatic flow labels are enabled by default, they can be
1338 disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1340 2: automatic flow labels are allowed, they may be enabled on a
1341 per socket basis using the IPV6_AUTOFLOWLABEL socket option
1342 3: automatic flow labels are enabled and enforced, they cannot
1343 be disabled by the socket option
1346 flowlabel_state_ranges - BOOLEAN
1347 Split the flow label number space into two ranges. 0-0x7FFFF is
1348 reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1349 is reserved for stateless flow labels as described in RFC6437.
1354 flowlabel_reflect - BOOLEAN
1355 Automatically reflect the flow label. Needed for Path MTU
1356 Discovery to work with Equal Cost Multipath Routing in anycast
1357 environments. See RFC 7690 and:
1358 https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
1363 fib_multipath_hash_policy - INTEGER
1364 Controls which hash policy to use for multipath routes.
1365 Default: 0 (Layer 3)
1367 0 - Layer 3 (source and destination addresses plus flow label)
1368 1 - Layer 4 (standard 5-tuple)
1370 anycast_src_echo_reply - BOOLEAN
1371 Controls the use of anycast addresses as source addresses for ICMPv6
1377 idgen_delay - INTEGER
1378 Controls the delay in seconds after which time to retry
1379 privacy stable address generation if a DAD conflict is
1381 Default: 1 (as specified in RFC7217)
1383 idgen_retries - INTEGER
1384 Controls the number of retries to generate a stable privacy
1385 address if a DAD conflict is detected.
1386 Default: 3 (as specified in RFC7217)
1389 Controls the MLD query robustness variable (see RFC3810 9.1).
1390 Default: 2 (as specified by RFC3810 9.1)
1391 Minimum: 1 (as specified by RFC6636 4.5)
1393 max_dst_opts_cnt - INTEGER
1394 Maximum number of non-padding TLVs allowed in a Destination
1395 options extension header. If this value is less than zero
1396 then unknown options are disallowed and the number of known
1397 TLVs allowed is the absolute value of this number.
1400 max_hbh_opts_cnt - INTEGER
1401 Maximum number of non-padding TLVs allowed in a Hop-by-Hop
1402 options extension header. If this value is less than zero
1403 then unknown options are disallowed and the number of known
1404 TLVs allowed is the absolute value of this number.
1407 max dst_opts_len - INTEGER
1408 Maximum length allowed for a Destination options extension
1410 Default: INT_MAX (unlimited)
1412 max hbh_opts_len - INTEGER
1413 Maximum length allowed for a Hop-by-Hop options extension
1415 Default: INT_MAX (unlimited)
1419 ip6frag_high_thresh - INTEGER
1420 Maximum memory used to reassemble IPv6 fragments. When
1421 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1422 the fragment handler will toss packets until ip6frag_low_thresh
1425 ip6frag_low_thresh - INTEGER
1426 See ip6frag_high_thresh
1428 ip6frag_time - INTEGER
1429 Time in seconds to keep an IPv6 fragment in memory.
1432 Change the interface-specific default settings.
1436 Change all the interface-specific settings.
1438 [XXX: Other special features than forwarding?]
1440 conf/all/forwarding - BOOLEAN
1441 Enable global IPv6 forwarding between all interfaces.
1443 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1444 to control which interfaces may forward packets and which not.
1446 This also sets all interfaces' Host/Router setting
1447 'forwarding' to the specified value. See below for details.
1449 This referred to as global forwarding.
1454 fwmark_reflect - BOOLEAN
1455 Controls the fwmark of kernel-generated IPv6 reply packets that are not
1456 associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
1457 If unset, these packets have a fwmark of zero. If set, they have the
1458 fwmark of the packet they are replying to.
1462 Change special settings per interface.
1464 The functional behaviour for certain settings is different
1465 depending on whether local forwarding is enabled or not.
1468 Accept Router Advertisements; autoconfigure using them.
1470 It also determines whether or not to transmit Router
1471 Solicitations. If and only if the functional setting is to
1472 accept Router Advertisements, Router Solicitations will be
1475 Possible values are:
1476 0 Do not accept Router Advertisements.
1477 1 Accept Router Advertisements if forwarding is disabled.
1478 2 Overrule forwarding behaviour. Accept Router Advertisements
1479 even if forwarding is enabled.
1481 Functional default: enabled if local forwarding is disabled.
1482 disabled if local forwarding is enabled.
1484 accept_ra_defrtr - BOOLEAN
1485 Learn default router in Router Advertisement.
1487 Functional default: enabled if accept_ra is enabled.
1488 disabled if accept_ra is disabled.
1490 accept_ra_from_local - BOOLEAN
1491 Accept RA with source-address that is found on local machine
1492 if the RA is otherwise proper and able to be accepted.
1493 Default is to NOT accept these as it may be an un-intended
1497 enabled if accept_ra_from_local is enabled
1498 on a specific interface.
1499 disabled if accept_ra_from_local is disabled
1500 on a specific interface.
1502 accept_ra_min_hop_limit - INTEGER
1503 Minimum hop limit Information in Router Advertisement.
1505 Hop limit Information in Router Advertisement less than this
1506 variable shall be ignored.
1510 accept_ra_pinfo - BOOLEAN
1511 Learn Prefix Information in Router Advertisement.
1513 Functional default: enabled if accept_ra is enabled.
1514 disabled if accept_ra is disabled.
1516 accept_ra_rt_info_min_plen - INTEGER
1517 Minimum prefix length of Route Information in RA.
1519 Route Information w/ prefix smaller than this variable shall
1522 Functional default: 0 if accept_ra_rtr_pref is enabled.
1523 -1 if accept_ra_rtr_pref is disabled.
1525 accept_ra_rt_info_max_plen - INTEGER
1526 Maximum prefix length of Route Information in RA.
1528 Route Information w/ prefix larger than this variable shall
1531 Functional default: 0 if accept_ra_rtr_pref is enabled.
1532 -1 if accept_ra_rtr_pref is disabled.
1534 accept_ra_rtr_pref - BOOLEAN
1535 Accept Router Preference in RA.
1537 Functional default: enabled if accept_ra is enabled.
1538 disabled if accept_ra is disabled.
1540 accept_ra_mtu - BOOLEAN
1541 Apply the MTU value specified in RA option 5 (RFC4861). If
1542 disabled, the MTU specified in the RA will be ignored.
1544 Functional default: enabled if accept_ra is enabled.
1545 disabled if accept_ra is disabled.
1547 accept_redirects - BOOLEAN
1550 Functional default: enabled if local forwarding is disabled.
1551 disabled if local forwarding is enabled.
1553 accept_source_route - INTEGER
1554 Accept source routing (routing extension header).
1556 >= 0: Accept only routing header type 2.
1557 < 0: Do not accept routing header.
1562 Autoconfigure addresses using Prefix Information in Router
1565 Functional default: enabled if accept_ra_pinfo is enabled.
1566 disabled if accept_ra_pinfo is disabled.
1568 dad_transmits - INTEGER
1569 The amount of Duplicate Address Detection probes to send.
1572 forwarding - INTEGER
1573 Configure interface-specific Host/Router behaviour.
1575 Note: It is recommended to have the same setting on all
1576 interfaces; mixed router/host scenarios are rather uncommon.
1578 Possible values are:
1579 0 Forwarding disabled
1580 1 Forwarding enabled
1584 By default, Host behaviour is assumed. This means:
1586 1. IsRouter flag is not set in Neighbour Advertisements.
1587 2. If accept_ra is TRUE (default), transmit Router
1589 3. If accept_ra is TRUE (default), accept Router
1590 Advertisements (and do autoconfiguration).
1591 4. If accept_redirects is TRUE (default), accept Redirects.
1595 If local forwarding is enabled, Router behaviour is assumed.
1596 This means exactly the reverse from the above:
1598 1. IsRouter flag is set in Neighbour Advertisements.
1599 2. Router Solicitations are not sent unless accept_ra is 2.
1600 3. Router Advertisements are ignored unless accept_ra is 2.
1601 4. Redirects are ignored.
1603 Default: 0 (disabled) if global forwarding is disabled (default),
1604 otherwise 1 (enabled).
1607 Default Hop Limit to set.
1611 Default Maximum Transfer Unit
1612 Default: 1280 (IPv6 required minimum)
1614 ip_nonlocal_bind - BOOLEAN
1615 If set, allows processes to bind() to non-local IPv6 addresses,
1616 which can be quite useful - but may break some applications.
1619 router_probe_interval - INTEGER
1620 Minimum interval (in seconds) between Router Probing described
1625 router_solicitation_delay - INTEGER
1626 Number of seconds to wait after interface is brought up
1627 before sending Router Solicitations.
1630 router_solicitation_interval - INTEGER
1631 Number of seconds to wait between Router Solicitations.
1634 router_solicitations - INTEGER
1635 Number of Router Solicitations to send until assuming no
1636 routers are present.
1639 use_oif_addrs_only - BOOLEAN
1640 When enabled, the candidate source addresses for destinations
1641 routed via this interface are restricted to the set of addresses
1642 configured on this interface (vis. RFC 6724, section 4).
1646 use_tempaddr - INTEGER
1647 Preference for Privacy Extensions (RFC3041).
1648 <= 0 : disable Privacy Extensions
1649 == 1 : enable Privacy Extensions, but prefer public
1650 addresses over temporary addresses.
1651 > 1 : enable Privacy Extensions and prefer temporary
1652 addresses over public addresses.
1653 Default: 0 (for most devices)
1654 -1 (for point-to-point devices and loopback devices)
1656 temp_valid_lft - INTEGER
1657 valid lifetime (in seconds) for temporary addresses.
1658 Default: 604800 (7 days)
1660 temp_prefered_lft - INTEGER
1661 Preferred lifetime (in seconds) for temporary addresses.
1662 Default: 86400 (1 day)
1664 keep_addr_on_down - INTEGER
1665 Keep all IPv6 addresses on an interface down event. If set static
1666 global addresses with no expiration time are not flushed.
1671 Default: 0 (addresses are removed)
1673 max_desync_factor - INTEGER
1674 Maximum value for DESYNC_FACTOR, which is a random value
1675 that ensures that clients don't synchronize with each
1676 other and generate new addresses at exactly the same time.
1677 value is in seconds.
1680 regen_max_retry - INTEGER
1681 Number of attempts before give up attempting to generate
1682 valid temporary addresses.
1685 max_addresses - INTEGER
1686 Maximum number of autoconfigured addresses per interface. Setting
1687 to zero disables the limitation. It is not recommended to set this
1688 value too large (or to zero) because it would be an easy way to
1689 crash the kernel by allowing too many addresses to be created.
1692 disable_ipv6 - BOOLEAN
1693 Disable IPv6 operation. If accept_dad is set to 2, this value
1694 will be dynamically set to TRUE if DAD fails for the link-local
1696 Default: FALSE (enable IPv6 operation)
1698 When this value is changed from 1 to 0 (IPv6 is being enabled),
1699 it will dynamically create a link-local address on the given
1700 interface and start Duplicate Address Detection, if necessary.
1702 When this value is changed from 0 to 1 (IPv6 is being disabled),
1703 it will dynamically delete all addresses and routes on the given
1704 interface. From now on it will not possible to add addresses/routes
1705 to the selected interface.
1707 accept_dad - INTEGER
1708 Whether to accept DAD (Duplicate Address Detection).
1710 1: Enable DAD (default)
1711 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1712 link-local address has been found.
1714 DAD operation and mode on a given interface will be selected according
1715 to the maximum value of conf/{all,interface}/accept_dad.
1717 force_tllao - BOOLEAN
1718 Enable sending the target link-layer address option even when
1719 responding to a unicast neighbor solicitation.
1722 Quoting from RFC 2461, section 4.4, Target link-layer address:
1724 "The option MUST be included for multicast solicitations in order to
1725 avoid infinite Neighbor Solicitation "recursion" when the peer node
1726 does not have a cache entry to return a Neighbor Advertisements
1727 message. When responding to unicast solicitations, the option can be
1728 omitted since the sender of the solicitation has the correct link-
1729 layer address; otherwise it would not have be able to send the unicast
1730 solicitation in the first place. However, including the link-layer
1731 address in this case adds little overhead and eliminates a potential
1732 race condition where the sender deletes the cached link-layer address
1733 prior to receiving a response to a previous solicitation."
1735 ndisc_notify - BOOLEAN
1736 Define mode for notification of address and device changes.
1737 0 - (default): do nothing
1738 1 - Generate unsolicited neighbour advertisements when device is brought
1739 up or hardware address changes.
1741 ndisc_tclass - INTEGER
1742 The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
1743 Discovery (Router Solicitation, Router Advertisement, Neighbor
1744 Solicitation, Neighbor Advertisement, Redirect) messages.
1745 These 8 bits can be interpreted as 6 high order bits holding the DSCP
1746 value and 2 low order bits representing ECN (which you probably want
1750 mldv1_unsolicited_report_interval - INTEGER
1751 The interval in milliseconds in which the next unsolicited
1752 MLDv1 report retransmit will take place.
1753 Default: 10000 (10 seconds)
1755 mldv2_unsolicited_report_interval - INTEGER
1756 The interval in milliseconds in which the next unsolicited
1757 MLDv2 report retransmit will take place.
1758 Default: 1000 (1 second)
1760 force_mld_version - INTEGER
1761 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
1762 1 - Enforce to use MLD version 1
1763 2 - Enforce to use MLD version 2
1765 suppress_frag_ndisc - INTEGER
1766 Control RFC 6980 (Security Implications of IPv6 Fragmentation
1767 with IPv6 Neighbor Discovery) behavior:
1768 1 - (default) discard fragmented neighbor discovery packets
1769 0 - allow fragmented neighbor discovery packets
1771 optimistic_dad - BOOLEAN
1772 Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
1773 0: disabled (default)
1776 Optimistic Duplicate Address Detection for the interface will be enabled
1777 if at least one of conf/{all,interface}/optimistic_dad is set to 1,
1778 it will be disabled otherwise.
1780 use_optimistic - BOOLEAN
1781 If enabled, do not classify optimistic addresses as deprecated during
1782 source address selection. Preferred addresses will still be chosen
1783 before optimistic addresses, subject to other ranking in the source
1784 address selection algorithm.
1785 0: disabled (default)
1788 This will be enabled if at least one of
1789 conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
1791 stable_secret - IPv6 address
1792 This IPv6 address will be used as a secret to generate IPv6
1793 addresses for link-local addresses and autoconfigured
1794 ones. All addresses generated after setting this secret will
1795 be stable privacy ones by default. This can be changed via the
1796 addrgenmode ip-link. conf/default/stable_secret is used as the
1797 secret for the namespace, the interface specific ones can
1798 overwrite that. Writes to conf/all/stable_secret are refused.
1800 It is recommended to generate this secret during installation
1801 of a system and keep it stable after that.
1803 By default the stable secret is unset.
1805 drop_unicast_in_l2_multicast - BOOLEAN
1806 Drop any unicast IPv6 packets that are received in link-layer
1807 multicast (or broadcast) frames.
1809 By default this is turned off.
1811 drop_unsolicited_na - BOOLEAN
1812 Drop all unsolicited neighbor advertisements, for example if there's
1813 a known good NA proxy on the network and such frames need not be used
1814 (or in the case of 802.11, must not be used to prevent attacks.)
1816 By default this is turned off.
1818 enhanced_dad - BOOLEAN
1819 Include a nonce option in the IPv6 neighbor solicitation messages used for
1820 duplicate address detection per RFC7527. A received DAD NS will only signal
1821 a duplicate address if the nonce is different. This avoids any false
1822 detection of duplicates due to loopback of the NS messages that we send.
1823 The nonce option will be sent on an interface unless both of
1824 conf/{all,interface}/enhanced_dad are set to FALSE.
1829 Limit the maximal rates for sending ICMPv6 packets.
1830 0 to disable any limiting,
1831 otherwise the minimal space between responses in milliseconds.
1834 xfrm6_gc_thresh - INTEGER
1835 The threshold at which we will start garbage collecting for IPv6
1836 destination cache entries. At twice this value the system will
1837 refuse new allocations.
1841 Pekka Savola <pekkas@netcore.fi>
1842 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1845 /proc/sys/net/bridge/* Variables:
1847 bridge-nf-call-arptables - BOOLEAN
1848 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1852 bridge-nf-call-iptables - BOOLEAN
1853 1 : pass bridged IPv4 traffic to iptables' chains.
1857 bridge-nf-call-ip6tables - BOOLEAN
1858 1 : pass bridged IPv6 traffic to ip6tables' chains.
1862 bridge-nf-filter-vlan-tagged - BOOLEAN
1863 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1867 bridge-nf-filter-pppoe-tagged - BOOLEAN
1868 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1872 bridge-nf-pass-vlan-input-dev - BOOLEAN
1873 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
1874 interface on the bridge and set the netfilter input device to the vlan.
1875 This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT
1876 target work with vlan-on-top-of-bridge interfaces. When no matching
1877 vlan interface is found, or this switch is off, the input device is
1878 set to the bridge interface.
1879 0: disable bridge netfilter vlan interface lookup.
1882 proc/sys/net/sctp/* Variables:
1884 addip_enable - BOOLEAN
1885 Enable or disable extension of Dynamic Address Reconfiguration
1886 (ADD-IP) functionality specified in RFC5061. This extension provides
1887 the ability to dynamically add and remove new addresses for the SCTP
1890 1: Enable extension.
1892 0: Disable extension.
1897 Enable or disable pf (pf is short for potentially failed) state. A value
1898 of pf_retrans > path_max_retrans also disables pf state. That is, one of
1899 both pf_enable and pf_retrans > path_max_retrans can disable pf state.
1900 Since pf_retrans and path_max_retrans can be changed by userspace
1901 application, sometimes user expects to disable pf state by the value of
1902 pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
1903 or path_max_retrans is changed by the user application, this pf state is
1904 enabled. As such, it is necessary to add this to dynamically enable
1905 and disable pf state. See:
1906 https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
1915 addip_noauth_enable - BOOLEAN
1916 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1917 authentication to protect the operations of adding or removing new
1918 addresses. This requirement is mandated so that unauthorized hosts
1919 would not be able to hijack associations. However, older
1920 implementations may not have implemented this requirement while
1921 allowing the ADD-IP extension. For reasons of interoperability,
1922 we provide this variable to control the enforcement of the
1923 authentication requirement.
1925 1: Allow ADD-IP extension to be used without authentication. This
1926 should only be set in a closed environment for interoperability
1927 with older implementations.
1929 0: Enforce the authentication requirement
1933 auth_enable - BOOLEAN
1934 Enable or disable Authenticated Chunks extension. This extension
1935 provides the ability to send and receive authenticated chunks and is
1936 required for secure operation of Dynamic Address Reconfiguration
1939 1: Enable this extension.
1940 0: Disable this extension.
1944 prsctp_enable - BOOLEAN
1945 Enable or disable the Partial Reliability extension (RFC3758) which
1946 is used to notify peers that a given DATA should no longer be expected.
1954 The limit of the number of new packets that can be initially sent. It
1955 controls how bursty the generated traffic can be.
1959 association_max_retrans - INTEGER
1960 Set the maximum number for retransmissions that an association can
1961 attempt deciding that the remote end is unreachable. If this value
1962 is exceeded, the association is terminated.
1966 max_init_retransmits - INTEGER
1967 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1968 that an association will attempt before declaring the destination
1969 unreachable and terminating.
1973 path_max_retrans - INTEGER
1974 The maximum number of retransmissions that will be attempted on a given
1975 path. Once this threshold is exceeded, the path is considered
1976 unreachable, and new traffic will use a different path when the
1977 association is multihomed.
1981 pf_retrans - INTEGER
1982 The number of retransmissions that will be attempted on a given path
1983 before traffic is redirected to an alternate transport (should one
1984 exist). Note this is distinct from path_max_retrans, as a path that
1985 passes the pf_retrans threshold can still be used. Its only
1986 deprioritized when a transmission path is selected by the stack. This
1987 setting is primarily used to enable fast failover mechanisms without
1988 having to reduce path_max_retrans to a very low value. See:
1989 http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
1990 for details. Note also that a value of pf_retrans > path_max_retrans
1991 disables this feature. Since both pf_retrans and path_max_retrans can
1992 be changed by userspace application, a variable pf_enable is used to
1997 rto_initial - INTEGER
1998 The initial round trip timeout value in milliseconds that will be used
1999 in calculating round trip times. This is the initial time interval
2000 for retransmissions.
2005 The maximum value (in milliseconds) of the round trip timeout. This
2006 is the largest time interval that can elapse between retransmissions.
2011 The minimum value (in milliseconds) of the round trip timeout. This
2012 is the smallest time interval the can elapse between retransmissions.
2016 hb_interval - INTEGER
2017 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
2018 are sent at the specified interval on idle paths to probe the state of
2019 a given path between 2 associations.
2023 sack_timeout - INTEGER
2024 The amount of time (in milliseconds) that the implementation will wait
2029 valid_cookie_life - INTEGER
2030 The default lifetime of the SCTP cookie (in milliseconds). The cookie
2031 is used during association establishment.
2035 cookie_preserve_enable - BOOLEAN
2036 Enable or disable the ability to extend the lifetime of the SCTP cookie
2037 that is used during the establishment phase of SCTP association
2039 1: Enable cookie lifetime extension.
2044 cookie_hmac_alg - STRING
2045 Select the hmac algorithm used when generating the cookie value sent by
2046 a listening sctp socket to a connecting client in the INIT-ACK chunk.
2051 Ability to assign md5 or sha1 as the selected alg is predicated on the
2052 configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
2053 CONFIG_CRYPTO_SHA1).
2055 Default: Dependent on configuration. MD5 if available, else SHA1 if
2056 available, else none.
2058 rcvbuf_policy - INTEGER
2059 Determines if the receive buffer is attributed to the socket or to
2060 association. SCTP supports the capability to create multiple
2061 associations on a single socket. When using this capability, it is
2062 possible that a single stalled association that's buffering a lot
2063 of data may block other associations from delivering their data by
2064 consuming all of the receive buffer space. To work around this,
2065 the rcvbuf_policy could be set to attribute the receiver buffer space
2066 to each association instead of the socket. This prevents the described
2069 1: rcvbuf space is per association
2070 0: rcvbuf space is per socket
2074 sndbuf_policy - INTEGER
2075 Similar to rcvbuf_policy above, this applies to send buffer space.
2077 1: Send buffer is tracked per association
2078 0: Send buffer is tracked per socket.
2082 sctp_mem - vector of 3 INTEGERs: min, pressure, max
2083 Number of pages allowed for queueing by all SCTP sockets.
2085 min: Below this number of pages SCTP is not bothered about its
2086 memory appetite. When amount of memory allocated by SCTP exceeds
2087 this number, SCTP starts to moderate memory usage.
2089 pressure: This value was introduced to follow format of tcp_mem.
2091 max: Number of pages allowed for queueing by all SCTP sockets.
2093 Default is calculated at boot time from amount of available memory.
2095 sctp_rmem - vector of 3 INTEGERs: min, default, max
2096 Only the first value ("min") is used, "default" and "max" are
2099 min: Minimal size of receive buffer used by SCTP socket.
2100 It is guaranteed to each SCTP socket (but not association) even
2101 under moderate memory pressure.
2105 sctp_wmem - vector of 3 INTEGERs: min, default, max
2106 Currently this tunable has no effect.
2108 addr_scope_policy - INTEGER
2109 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
2111 0 - Disable IPv4 address scoping
2112 1 - Enable IPv4 address scoping
2113 2 - Follow draft but allow IPv4 private addresses
2114 3 - Follow draft but allow IPv4 link local addresses
2119 /proc/sys/net/core/*
2120 Please see: Documentation/sysctl/net.txt for descriptions of these entries.
2123 /proc/sys/net/unix/*
2124 max_dgram_qlen - INTEGER
2125 The maximum length of dgram socket receive queue
2132 /proc/sys/net/irda/*
2133 fast_poll_increase FIXME
2134 warn_noreply_time FIXME
2135 discovery_slots FIXME
2138 discovery_timeout FIXME
2139 lap_keepalive_time FIXME
2140 max_noreply_time FIXME
2141 max_tx_data_size FIXME
2143 min_tx_turn_time FIXME