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 - BOOLEAN
19 Disable Path MTU Discovery.
23 default 562 - minimum discovered Path MTU
25 route/max_size - INTEGER
26 Maximum number of routes allowed in the kernel. Increase
27 this when using large numbers of interfaces and/or routes.
29 neigh/default/gc_thresh3 - INTEGER
30 Maximum number of neighbor entries allowed. Increase this
31 when using large numbers of interfaces and when communicating
32 with large numbers of directly-connected peers.
35 Time, in seconds, that cached PMTU information is kept.
38 The advertised MSS depends on the first hop route MTU, but will
39 never be lower than this setting.
41 rt_cache_rebuild_count - INTEGER
42 The per net-namespace route cache emergency rebuild threshold.
43 Any net-namespace having its route cache rebuilt due to
44 a hash bucket chain being too long more than this many times
45 will have its route caching disabled
49 ipfrag_high_thresh - INTEGER
50 Maximum memory used to reassemble IP fragments. When
51 ipfrag_high_thresh bytes of memory is allocated for this purpose,
52 the fragment handler will toss packets until ipfrag_low_thresh
55 ipfrag_low_thresh - INTEGER
56 See ipfrag_high_thresh
59 Time in seconds to keep an IP fragment in memory.
61 ipfrag_secret_interval - INTEGER
62 Regeneration interval (in seconds) of the hash secret (or lifetime
63 for the hash secret) for IP fragments.
66 ipfrag_max_dist - INTEGER
67 ipfrag_max_dist is a non-negative integer value which defines the
68 maximum "disorder" which is allowed among fragments which share a
69 common IP source address. Note that reordering of packets is
70 not unusual, but if a large number of fragments arrive from a source
71 IP address while a particular fragment queue remains incomplete, it
72 probably indicates that one or more fragments belonging to that queue
73 have been lost. When ipfrag_max_dist is positive, an additional check
74 is done on fragments before they are added to a reassembly queue - if
75 ipfrag_max_dist (or more) fragments have arrived from a particular IP
76 address between additions to any IP fragment queue using that source
77 address, it's presumed that one or more fragments in the queue are
78 lost. The existing fragment queue will be dropped, and a new one
79 started. An ipfrag_max_dist value of zero disables this check.
81 Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
82 result in unnecessarily dropping fragment queues when normal
83 reordering of packets occurs, which could lead to poor application
84 performance. Using a very large value, e.g. 50000, increases the
85 likelihood of incorrectly reassembling IP fragments that originate
86 from different IP datagrams, which could result in data corruption.
91 inet_peer_threshold - INTEGER
92 The approximate size of the storage. Starting from this threshold
93 entries will be thrown aggressively. This threshold also determines
94 entries' time-to-live and time intervals between garbage collection
95 passes. More entries, less time-to-live, less GC interval.
97 inet_peer_minttl - INTEGER
98 Minimum time-to-live of entries. Should be enough to cover fragment
99 time-to-live on the reassembling side. This minimum time-to-live is
100 guaranteed if the pool size is less than inet_peer_threshold.
103 inet_peer_maxttl - INTEGER
104 Maximum time-to-live of entries. Unused entries will expire after
105 this period of time if there is no memory pressure on the pool (i.e.
106 when the number of entries in the pool is very small).
112 Limit of socket listen() backlog, known in userspace as SOMAXCONN.
113 Defaults to 128. See also tcp_max_syn_backlog for additional tuning
117 Controls Appropriate Byte Count (ABC) defined in RFC3465.
118 ABC is a way of increasing congestion window (cwnd) more slowly
119 in response to partial acknowledgments.
121 0 increase cwnd once per acknowledgment (no ABC)
122 1 increase cwnd once per acknowledgment of full sized segment
123 2 allow increase cwnd by two if acknowledgment is
124 of two segments to compensate for delayed acknowledgments.
127 tcp_abort_on_overflow - BOOLEAN
128 If listening service is too slow to accept new connections,
129 reset them. Default state is FALSE. It means that if overflow
130 occurred due to a burst, connection will recover. Enable this
131 option _only_ if you are really sure that listening daemon
132 cannot be tuned to accept connections faster. Enabling this
133 option can harm clients of your server.
135 tcp_adv_win_scale - INTEGER
136 Count buffering overhead as bytes/2^tcp_adv_win_scale
137 (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
139 Possible values are [-31, 31], inclusive.
142 tcp_allowed_congestion_control - STRING
143 Show/set the congestion control choices available to non-privileged
144 processes. The list is a subset of those listed in
145 tcp_available_congestion_control.
146 Default is "reno" and the default setting (tcp_congestion_control).
148 tcp_app_win - INTEGER
149 Reserve max(window/2^tcp_app_win, mss) of window for application
150 buffer. Value 0 is special, it means that nothing is reserved.
153 tcp_available_congestion_control - STRING
154 Shows the available congestion control choices that are registered.
155 More congestion control algorithms may be available as modules,
158 tcp_base_mss - INTEGER
159 The initial value of search_low to be used by the packetization layer
160 Path MTU discovery (MTU probing). If MTU probing is enabled,
161 this is the initial MSS used by the connection.
163 tcp_congestion_control - STRING
164 Set the congestion control algorithm to be used for new
165 connections. The algorithm "reno" is always available, but
166 additional choices may be available based on kernel configuration.
167 Default is set as part of kernel configuration.
169 tcp_cookie_size - INTEGER
170 Default size of TCP Cookie Transactions (TCPCT) option, that may be
171 overridden on a per socket basis by the TCPCT socket option.
172 Values greater than the maximum (16) are interpreted as the maximum.
173 Values greater than zero and less than the minimum (8) are interpreted
174 as the minimum. Odd values are interpreted as the next even value.
178 Allows TCP to send "duplicate" SACKs.
181 Enable Explicit Congestion Notification (ECN) in TCP. ECN is only
182 used when both ends of the TCP flow support it. It is useful to
183 avoid losses due to congestion (when the bottleneck router supports
188 2 Only server-side ECN enabled. If the other end does
189 not support ECN, behavior is like with ECN disabled.
193 Enable FACK congestion avoidance and fast retransmission.
194 The value is not used, if tcp_sack is not enabled.
196 tcp_fin_timeout - INTEGER
197 Time to hold socket in state FIN-WAIT-2, if it was closed
198 by our side. Peer can be broken and never close its side,
199 or even died unexpectedly. Default value is 60sec.
200 Usual value used in 2.2 was 180 seconds, you may restore
201 it, but remember that if your machine is even underloaded WEB server,
202 you risk to overflow memory with kilotons of dead sockets,
203 FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1,
204 because they eat maximum 1.5K of memory, but they tend
205 to live longer. Cf. tcp_max_orphans.
208 Enables Forward RTO-Recovery (F-RTO) defined in RFC4138.
209 F-RTO is an enhanced recovery algorithm for TCP retransmission
210 timeouts. It is particularly beneficial in wireless environments
211 where packet loss is typically due to random radio interference
212 rather than intermediate router congestion. F-RTO is sender-side
213 only modification. Therefore it does not require any support from
216 If set to 1, basic version is enabled. 2 enables SACK enhanced
217 F-RTO if flow uses SACK. The basic version can be used also when
218 SACK is in use though scenario(s) with it exists where F-RTO
219 interacts badly with the packet counting of the SACK enabled TCP
222 tcp_frto_response - INTEGER
223 When F-RTO has detected that a TCP retransmission timeout was
224 spurious (i.e, the timeout would have been avoided had TCP set a
225 longer retransmission timeout), TCP has several options what to do
226 next. Possible values are:
227 0 Rate halving based; a smooth and conservative response,
228 results in halved cwnd and ssthresh after one RTT
229 1 Very conservative response; not recommended because even
230 though being valid, it interacts poorly with the rest of
231 Linux TCP, halves cwnd and ssthresh immediately
232 2 Aggressive response; undoes congestion control measures
233 that are now known to be unnecessary (ignoring the
234 possibility of a lost retransmission that would require
235 TCP to be more cautious), cwnd and ssthresh are restored
236 to the values prior timeout
237 Default: 0 (rate halving based)
239 tcp_keepalive_time - INTEGER
240 How often TCP sends out keepalive messages when keepalive is enabled.
243 tcp_keepalive_probes - INTEGER
244 How many keepalive probes TCP sends out, until it decides that the
245 connection is broken. Default value: 9.
247 tcp_keepalive_intvl - INTEGER
248 How frequently the probes are send out. Multiplied by
249 tcp_keepalive_probes it is time to kill not responding connection,
250 after probes started. Default value: 75sec i.e. connection
251 will be aborted after ~11 minutes of retries.
253 tcp_low_latency - BOOLEAN
254 If set, the TCP stack makes decisions that prefer lower
255 latency as opposed to higher throughput. By default, this
256 option is not set meaning that higher throughput is preferred.
257 An example of an application where this default should be
258 changed would be a Beowulf compute cluster.
261 tcp_max_orphans - INTEGER
262 Maximal number of TCP sockets not attached to any user file handle,
263 held by system. If this number is exceeded orphaned connections are
264 reset immediately and warning is printed. This limit exists
265 only to prevent simple DoS attacks, you _must_ not rely on this
266 or lower the limit artificially, but rather increase it
267 (probably, after increasing installed memory),
268 if network conditions require more than default value,
269 and tune network services to linger and kill such states
270 more aggressively. Let me to remind again: each orphan eats
271 up to ~64K of unswappable memory.
273 tcp_max_ssthresh - INTEGER
274 Limited Slow-Start for TCP with large congestion windows (cwnd) defined in
275 RFC3742. Limited slow-start is a mechanism to limit growth of the cwnd
276 on the region where cwnd is larger than tcp_max_ssthresh. TCP increases cwnd
277 by at most tcp_max_ssthresh segments, and by at least tcp_max_ssthresh/2
278 segments per RTT when the cwnd is above tcp_max_ssthresh.
279 If TCP connection increased cwnd to thousands (or tens of thousands) segments,
280 and thousands of packets were being dropped during slow-start, you can set
281 tcp_max_ssthresh to improve performance for new TCP connection.
284 tcp_max_syn_backlog - INTEGER
285 Maximal number of remembered connection requests, which are
286 still did not receive an acknowledgment from connecting client.
287 Default value is 1024 for systems with more than 128Mb of memory,
288 and 128 for low memory machines. If server suffers of overload,
289 try to increase this number.
291 tcp_max_tw_buckets - INTEGER
292 Maximal number of timewait sockets held by system simultaneously.
293 If this number is exceeded time-wait socket is immediately destroyed
294 and warning is printed. This limit exists only to prevent
295 simple DoS attacks, you _must_ not lower the limit artificially,
296 but rather increase it (probably, after increasing installed memory),
297 if network conditions require more than default value.
299 tcp_mem - vector of 3 INTEGERs: min, pressure, max
300 min: below this number of pages TCP is not bothered about its
303 pressure: when amount of memory allocated by TCP exceeds this number
304 of pages, TCP moderates its memory consumption and enters memory
305 pressure mode, which is exited when memory consumption falls
308 max: number of pages allowed for queueing by all TCP sockets.
310 Defaults are calculated at boot time from amount of available
313 tcp_moderate_rcvbuf - BOOLEAN
314 If set, TCP performs receive buffer auto-tuning, attempting to
315 automatically size the buffer (no greater than tcp_rmem[2]) to
316 match the size required by the path for full throughput. Enabled by
319 tcp_mtu_probing - INTEGER
320 Controls TCP Packetization-Layer Path MTU Discovery. Takes three
323 1 - Disabled by default, enabled when an ICMP black hole detected
324 2 - Always enabled, use initial MSS of tcp_base_mss.
326 tcp_no_metrics_save - BOOLEAN
327 By default, TCP saves various connection metrics in the route cache
328 when the connection closes, so that connections established in the
329 near future can use these to set initial conditions. Usually, this
330 increases overall performance, but may sometimes cause performance
331 degradation. If set, TCP will not cache metrics on closing
334 tcp_orphan_retries - INTEGER
335 This value influences the timeout of a locally closed TCP connection,
336 when RTO retransmissions remain unacknowledged.
337 See tcp_retries2 for more details.
339 The default value is 7.
340 If your machine is a loaded WEB server,
341 you should think about lowering this value, such sockets
342 may consume significant resources. Cf. tcp_max_orphans.
344 tcp_reordering - INTEGER
345 Maximal reordering of packets in a TCP stream.
348 tcp_retrans_collapse - BOOLEAN
349 Bug-to-bug compatibility with some broken printers.
350 On retransmit try to send bigger packets to work around bugs in
353 tcp_retries1 - INTEGER
354 This value influences the time, after which TCP decides, that
355 something is wrong due to unacknowledged RTO retransmissions,
356 and reports this suspicion to the network layer.
357 See tcp_retries2 for more details.
359 RFC 1122 recommends at least 3 retransmissions, which is the
362 tcp_retries2 - INTEGER
363 This value influences the timeout of an alive TCP connection,
364 when RTO retransmissions remain unacknowledged.
365 Given a value of N, a hypothetical TCP connection following
366 exponential backoff with an initial RTO of TCP_RTO_MIN would
367 retransmit N times before killing the connection at the (N+1)th RTO.
369 The default value of 15 yields a hypothetical timeout of 924.6
370 seconds and is a lower bound for the effective timeout.
371 TCP will effectively time out at the first RTO which exceeds the
372 hypothetical timeout.
374 RFC 1122 recommends at least 100 seconds for the timeout,
375 which corresponds to a value of at least 8.
377 tcp_rfc1337 - BOOLEAN
378 If set, the TCP stack behaves conforming to RFC1337. If unset,
379 we are not conforming to RFC, but prevent TCP TIME_WAIT
383 tcp_rmem - vector of 3 INTEGERs: min, default, max
384 min: Minimal size of receive buffer used by TCP sockets.
385 It is guaranteed to each TCP socket, even under moderate memory
389 default: initial size of receive buffer used by TCP sockets.
390 This value overrides net.core.rmem_default used by other protocols.
391 Default: 87380 bytes. This value results in window of 65535 with
392 default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit
393 less for default tcp_app_win. See below about these variables.
395 max: maximal size of receive buffer allowed for automatically
396 selected receiver buffers for TCP socket. This value does not override
397 net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables
398 automatic tuning of that socket's receive buffer size, in which
399 case this value is ignored.
400 Default: between 87380B and 4MB, depending on RAM size.
403 Enable select acknowledgments (SACKS).
405 tcp_slow_start_after_idle - BOOLEAN
406 If set, provide RFC2861 behavior and time out the congestion
407 window after an idle period. An idle period is defined at
408 the current RTO. If unset, the congestion window will not
409 be timed out after an idle period.
413 Use the Host requirements interpretation of the TCP urgent pointer field.
414 Most hosts use the older BSD interpretation, so if you turn this on
415 Linux might not communicate correctly with them.
418 tcp_synack_retries - INTEGER
419 Number of times SYNACKs for a passive TCP connection attempt will
420 be retransmitted. Should not be higher than 255. Default value
421 is 5, which corresponds to ~180seconds.
423 tcp_syncookies - BOOLEAN
424 Only valid when the kernel was compiled with CONFIG_SYNCOOKIES
425 Send out syncookies when the syn backlog queue of a socket
426 overflows. This is to prevent against the common 'SYN flood attack'
429 Note, that syncookies is fallback facility.
430 It MUST NOT be used to help highly loaded servers to stand
431 against legal connection rate. If you see SYN flood warnings
432 in your logs, but investigation shows that they occur
433 because of overload with legal connections, you should tune
434 another parameters until this warning disappear.
435 See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
437 syncookies seriously violate TCP protocol, do not allow
438 to use TCP extensions, can result in serious degradation
439 of some services (f.e. SMTP relaying), visible not by you,
440 but your clients and relays, contacting you. While you see
441 SYN flood warnings in logs not being really flooded, your server
442 is seriously misconfigured.
444 tcp_syn_retries - INTEGER
445 Number of times initial SYNs for an active TCP connection attempt
446 will be retransmitted. Should not be higher than 255. Default value
447 is 5, which corresponds to ~180seconds.
449 tcp_timestamps - BOOLEAN
450 Enable timestamps as defined in RFC1323.
452 tcp_tso_win_divisor - INTEGER
453 This allows control over what percentage of the congestion window
454 can be consumed by a single TSO frame.
455 The setting of this parameter is a choice between burstiness and
456 building larger TSO frames.
459 tcp_tw_recycle - BOOLEAN
460 Enable fast recycling TIME-WAIT sockets. Default value is 0.
461 It should not be changed without advice/request of technical
464 tcp_tw_reuse - BOOLEAN
465 Allow to reuse TIME-WAIT sockets for new connections when it is
466 safe from protocol viewpoint. Default value is 0.
467 It should not be changed without advice/request of technical
470 tcp_window_scaling - BOOLEAN
471 Enable window scaling as defined in RFC1323.
473 tcp_wmem - vector of 3 INTEGERs: min, default, max
474 min: Amount of memory reserved for send buffers for TCP sockets.
475 Each TCP socket has rights to use it due to fact of its birth.
478 default: initial size of send buffer used by TCP sockets. This
479 value overrides net.core.wmem_default used by other protocols.
480 It is usually lower than net.core.wmem_default.
483 max: Maximal amount of memory allowed for automatically tuned
484 send buffers for TCP sockets. This value does not override
485 net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
486 automatic tuning of that socket's send buffer size, in which case
487 this value is ignored.
488 Default: between 64K and 4MB, depending on RAM size.
490 tcp_workaround_signed_windows - BOOLEAN
491 If set, assume no receipt of a window scaling option means the
492 remote TCP is broken and treats the window as a signed quantity.
493 If unset, assume the remote TCP is not broken even if we do
494 not receive a window scaling option from them.
497 tcp_dma_copybreak - INTEGER
498 Lower limit, in bytes, of the size of socket reads that will be
499 offloaded to a DMA copy engine, if one is present in the system
500 and CONFIG_NET_DMA is enabled.
503 tcp_thin_linear_timeouts - BOOLEAN
504 Enable dynamic triggering of linear timeouts for thin streams.
505 If set, a check is performed upon retransmission by timeout to
506 determine if the stream is thin (less than 4 packets in flight).
507 As long as the stream is found to be thin, up to 6 linear
508 timeouts may be performed before exponential backoff mode is
509 initiated. This improves retransmission latency for
510 non-aggressive thin streams, often found to be time-dependent.
511 For more information on thin streams, see
512 Documentation/networking/tcp-thin.txt
515 tcp_thin_dupack - BOOLEAN
516 Enable dynamic triggering of retransmissions after one dupACK
517 for thin streams. If set, a check is performed upon reception
518 of a dupACK to determine if the stream is thin (less than 4
519 packets in flight). As long as the stream is found to be thin,
520 data is retransmitted on the first received dupACK. This
521 improves retransmission latency for non-aggressive thin
522 streams, often found to be time-dependent.
523 For more information on thin streams, see
524 Documentation/networking/tcp-thin.txt
529 udp_mem - vector of 3 INTEGERs: min, pressure, max
530 Number of pages allowed for queueing by all UDP sockets.
532 min: Below this number of pages UDP is not bothered about its
533 memory appetite. When amount of memory allocated by UDP exceeds
534 this number, UDP starts to moderate memory usage.
536 pressure: This value was introduced to follow format of tcp_mem.
538 max: Number of pages allowed for queueing by all UDP sockets.
540 Default is calculated at boot time from amount of available memory.
542 udp_rmem_min - INTEGER
543 Minimal size of receive buffer used by UDP sockets in moderation.
544 Each UDP socket is able to use the size for receiving data, even if
545 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
548 udp_wmem_min - INTEGER
549 Minimal size of send buffer used by UDP sockets in moderation.
550 Each UDP socket is able to use the size for sending data, even if
551 total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
556 cipso_cache_enable - BOOLEAN
557 If set, enable additions to and lookups from the CIPSO label mapping
558 cache. If unset, additions are ignored and lookups always result in a
559 miss. However, regardless of the setting the cache is still
560 invalidated when required when means you can safely toggle this on and
561 off and the cache will always be "safe".
564 cipso_cache_bucket_size - INTEGER
565 The CIPSO label cache consists of a fixed size hash table with each
566 hash bucket containing a number of cache entries. This variable limits
567 the number of entries in each hash bucket; the larger the value the
568 more CIPSO label mappings that can be cached. When the number of
569 entries in a given hash bucket reaches this limit adding new entries
570 causes the oldest entry in the bucket to be removed to make room.
573 cipso_rbm_optfmt - BOOLEAN
574 Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
575 the CIPSO draft specification (see Documentation/netlabel for details).
576 This means that when set the CIPSO tag will be padded with empty
577 categories in order to make the packet data 32-bit aligned.
580 cipso_rbm_structvalid - BOOLEAN
581 If set, do a very strict check of the CIPSO option when
582 ip_options_compile() is called. If unset, relax the checks done during
583 ip_options_compile(). Either way is "safe" as errors are caught else
584 where in the CIPSO processing code but setting this to 0 (False) should
585 result in less work (i.e. it should be faster) but could cause problems
586 with other implementations that require strict checking.
591 ip_local_port_range - 2 INTEGERS
592 Defines the local port range that is used by TCP and UDP to
593 choose the local port. The first number is the first, the
594 second the last local port number. Default value depends on
595 amount of memory available on the system:
597 < 128Mb 1024-4999 or even less.
598 This number defines number of active connections, which this
599 system can issue simultaneously to systems not supporting
600 TCP extensions (timestamps). With tcp_tw_recycle enabled
601 (i.e. by default) range 1024-4999 is enough to issue up to
602 2000 connections per second to systems supporting timestamps.
604 ip_local_reserved_ports - list of comma separated ranges
605 Specify the ports which are reserved for known third-party
606 applications. These ports will not be used by automatic port
607 assignments (e.g. when calling connect() or bind() with port
608 number 0). Explicit port allocation behavior is unchanged.
610 The format used for both input and output is a comma separated
611 list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
612 10). Writing to the file will clear all previously reserved
613 ports and update the current list with the one given in the
616 Note that ip_local_port_range and ip_local_reserved_ports
617 settings are independent and both are considered by the kernel
618 when determining which ports are available for automatic port
621 You can reserve ports which are not in the current
622 ip_local_port_range, e.g.:
624 $ cat /proc/sys/net/ipv4/ip_local_port_range
626 $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
629 although this is redundant. However such a setting is useful
630 if later the port range is changed to a value that will
631 include the reserved ports.
635 ip_nonlocal_bind - BOOLEAN
636 If set, allows processes to bind() to non-local IP addresses,
637 which can be quite useful - but may break some applications.
641 If set non-zero, enables support for dynamic addresses.
642 If set to a non-zero value larger than 1, a kernel log
643 message will be printed when dynamic address rewriting
647 icmp_echo_ignore_all - BOOLEAN
648 If set non-zero, then the kernel will ignore all ICMP ECHO
652 icmp_echo_ignore_broadcasts - BOOLEAN
653 If set non-zero, then the kernel will ignore all ICMP ECHO and
654 TIMESTAMP requests sent to it via broadcast/multicast.
657 icmp_ratelimit - INTEGER
658 Limit the maximal rates for sending ICMP packets whose type matches
659 icmp_ratemask (see below) to specific targets.
660 0 to disable any limiting,
661 otherwise the minimal space between responses in milliseconds.
664 icmp_ratemask - INTEGER
665 Mask made of ICMP types for which rates are being limited.
666 Significant bits: IHGFEDCBA9876543210
667 Default mask: 0000001100000011000 (6168)
669 Bit definitions (see include/linux/icmp.h):
671 3 Destination Unreachable *
676 C Parameter Problem *
681 H Address Mask Request
684 * These are rate limited by default (see default mask above)
686 icmp_ignore_bogus_error_responses - BOOLEAN
687 Some routers violate RFC1122 by sending bogus responses to broadcast
688 frames. Such violations are normally logged via a kernel warning.
689 If this is set to TRUE, the kernel will not give such warnings, which
690 will avoid log file clutter.
693 icmp_errors_use_inbound_ifaddr - BOOLEAN
695 If zero, icmp error messages are sent with the primary address of
696 the exiting interface.
698 If non-zero, the message will be sent with the primary address of
699 the interface that received the packet that caused the icmp error.
700 This is the behaviour network many administrators will expect from
701 a router. And it can make debugging complicated network layouts
704 Note that if no primary address exists for the interface selected,
705 then the primary address of the first non-loopback interface that
706 has one will be used regardless of this setting.
710 igmp_max_memberships - INTEGER
711 Change the maximum number of multicast groups we can subscribe to.
714 Theoretical maximum value is bounded by having to send a membership
715 report in a single datagram (i.e. the report can't span multiple
716 datagrams, or risk confusing the switch and leaving groups you don't
719 The number of supported groups 'M' is bounded by the number of group
720 report entries you can fit into a single datagram of 65535 bytes.
722 M = 65536-sizeof (ip header)/(sizeof(Group record))
724 Group records are variable length, with a minimum of 12 bytes.
725 So net.ipv4.igmp_max_memberships should not be set higher than:
727 (65536-24) / 12 = 5459
729 The value 5459 assumes no IP header options, so in practice
730 this number may be lower.
732 conf/interface/* changes special settings per interface (where
733 "interface" is the name of your network interface)
735 conf/all/* is special, changes the settings for all interfaces
737 log_martians - BOOLEAN
738 Log packets with impossible addresses to kernel log.
739 log_martians for the interface will be enabled if at least one of
740 conf/{all,interface}/log_martians is set to TRUE,
741 it will be disabled otherwise
743 accept_redirects - BOOLEAN
744 Accept ICMP redirect messages.
745 accept_redirects for the interface will be enabled if:
746 - both conf/{all,interface}/accept_redirects are TRUE in the case
747 forwarding for the interface is enabled
749 - at least one of conf/{all,interface}/accept_redirects is TRUE in the
750 case forwarding for the interface is disabled
751 accept_redirects for the interface will be disabled otherwise
756 Enable IP forwarding on this interface.
758 mc_forwarding - BOOLEAN
759 Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
760 and a multicast routing daemon is required.
761 conf/all/mc_forwarding must also be set to TRUE to enable multicast
762 routing for the interface
765 Integer value used to differentiate the devices by the medium they
766 are attached to. Two devices can have different id values when
767 the broadcast packets are received only on one of them.
768 The default value 0 means that the device is the only interface
769 to its medium, value of -1 means that medium is not known.
771 Currently, it is used to change the proxy_arp behavior:
772 the proxy_arp feature is enabled for packets forwarded between
773 two devices attached to different media.
777 proxy_arp for the interface will be enabled if at least one of
778 conf/{all,interface}/proxy_arp is set to TRUE,
779 it will be disabled otherwise
781 proxy_arp_pvlan - BOOLEAN
782 Private VLAN proxy arp.
783 Basically allow proxy arp replies back to the same interface
784 (from which the ARP request/solicitation was received).
786 This is done to support (ethernet) switch features, like RFC
787 3069, where the individual ports are NOT allowed to
788 communicate with each other, but they are allowed to talk to
789 the upstream router. As described in RFC 3069, it is possible
790 to allow these hosts to communicate through the upstream
791 router by proxy_arp'ing. Don't need to be used together with
794 This technology is known by different names:
795 In RFC 3069 it is called VLAN Aggregation.
796 Cisco and Allied Telesyn call it Private VLAN.
797 Hewlett-Packard call it Source-Port filtering or port-isolation.
798 Ericsson call it MAC-Forced Forwarding (RFC Draft).
800 shared_media - BOOLEAN
801 Send(router) or accept(host) RFC1620 shared media redirects.
802 Overrides ip_secure_redirects.
803 shared_media for the interface will be enabled if at least one of
804 conf/{all,interface}/shared_media is set to TRUE,
805 it will be disabled otherwise
808 secure_redirects - BOOLEAN
809 Accept ICMP redirect messages only for gateways,
810 listed in default gateway list.
811 secure_redirects for the interface will be enabled if at least one of
812 conf/{all,interface}/secure_redirects is set to TRUE,
813 it will be disabled otherwise
816 send_redirects - BOOLEAN
817 Send redirects, if router.
818 send_redirects for the interface will be enabled if at least one of
819 conf/{all,interface}/send_redirects is set to TRUE,
820 it will be disabled otherwise
823 bootp_relay - BOOLEAN
824 Accept packets with source address 0.b.c.d destined
825 not to this host as local ones. It is supposed, that
826 BOOTP relay daemon will catch and forward such packets.
827 conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
832 accept_source_route - BOOLEAN
833 Accept packets with SRR option.
834 conf/all/accept_source_route must also be set to TRUE to accept packets
835 with SRR option on the interface
836 default TRUE (router)
839 accept_local - BOOLEAN
840 Accept packets with local source addresses. In combination with
841 suitable routing, this can be used to direct packets between two
842 local interfaces over the wire and have them accepted properly.
846 0 - No source validation.
847 1 - Strict mode as defined in RFC3704 Strict Reverse Path
848 Each incoming packet is tested against the FIB and if the interface
849 is not the best reverse path the packet check will fail.
850 By default failed packets are discarded.
851 2 - Loose mode as defined in RFC3704 Loose Reverse Path
852 Each incoming packet's source address is also tested against the FIB
853 and if the source address is not reachable via any interface
854 the packet check will fail.
856 Current recommended practice in RFC3704 is to enable strict mode
857 to prevent IP spoofing from DDos attacks. If using asymmetric routing
858 or other complicated routing, then loose mode is recommended.
860 The max value from conf/{all,interface}/rp_filter is used
861 when doing source validation on the {interface}.
863 Default value is 0. Note that some distributions enable it
867 1 - Allows you to have multiple network interfaces on the same
868 subnet, and have the ARPs for each interface be answered
869 based on whether or not the kernel would route a packet from
870 the ARP'd IP out that interface (therefore you must use source
871 based routing for this to work). In other words it allows control
872 of which cards (usually 1) will respond to an arp request.
874 0 - (default) The kernel can respond to arp requests with addresses
875 from other interfaces. This may seem wrong but it usually makes
876 sense, because it increases the chance of successful communication.
877 IP addresses are owned by the complete host on Linux, not by
878 particular interfaces. Only for more complex setups like load-
879 balancing, does this behaviour cause problems.
881 arp_filter for the interface will be enabled if at least one of
882 conf/{all,interface}/arp_filter is set to TRUE,
883 it will be disabled otherwise
885 arp_announce - INTEGER
886 Define different restriction levels for announcing the local
887 source IP address from IP packets in ARP requests sent on
889 0 - (default) Use any local address, configured on any interface
890 1 - Try to avoid local addresses that are not in the target's
891 subnet for this interface. This mode is useful when target
892 hosts reachable via this interface require the source IP
893 address in ARP requests to be part of their logical network
894 configured on the receiving interface. When we generate the
895 request we will check all our subnets that include the
896 target IP and will preserve the source address if it is from
897 such subnet. If there is no such subnet we select source
898 address according to the rules for level 2.
899 2 - Always use the best local address for this target.
900 In this mode we ignore the source address in the IP packet
901 and try to select local address that we prefer for talks with
902 the target host. Such local address is selected by looking
903 for primary IP addresses on all our subnets on the outgoing
904 interface that include the target IP address. If no suitable
905 local address is found we select the first local address
906 we have on the outgoing interface or on all other interfaces,
907 with the hope we will receive reply for our request and
908 even sometimes no matter the source IP address we announce.
910 The max value from conf/{all,interface}/arp_announce is used.
912 Increasing the restriction level gives more chance for
913 receiving answer from the resolved target while decreasing
914 the level announces more valid sender's information.
917 Define different modes for sending replies in response to
918 received ARP requests that resolve local target IP addresses:
919 0 - (default): reply for any local target IP address, configured
921 1 - reply only if the target IP address is local address
922 configured on the incoming interface
923 2 - reply only if the target IP address is local address
924 configured on the incoming interface and both with the
925 sender's IP address are part from same subnet on this interface
926 3 - do not reply for local addresses configured with scope host,
927 only resolutions for global and link addresses are replied
929 8 - do not reply for all local addresses
931 The max value from conf/{all,interface}/arp_ignore is used
932 when ARP request is received on the {interface}
935 Define mode for notification of address and device changes.
936 0 - (default): do nothing
937 1 - Generate gratuitous arp requests when device is brought up
938 or hardware address changes.
941 Define behavior for gratuitous ARP frames who's IP is not
942 already present in the ARP table:
943 0 - don't create new entries in the ARP table
944 1 - create new entries in the ARP table
946 Both replies and requests type gratuitous arp will trigger the
947 ARP table to be updated, if this setting is on.
949 If the ARP table already contains the IP address of the
950 gratuitous arp frame, the arp table will be updated regardless
951 if this setting is on or off.
954 app_solicit - INTEGER
955 The maximum number of probes to send to the user space ARP daemon
956 via netlink before dropping back to multicast probes (see
957 mcast_solicit). Defaults to 0.
959 disable_policy - BOOLEAN
960 Disable IPSEC policy (SPD) for this interface
962 disable_xfrm - BOOLEAN
963 Disable IPSEC encryption on this interface, whatever the policy
968 Allows you to write a number, which can be used as required.
978 delon.nicolas@wanadoo.fr
983 /proc/sys/net/ipv6/* Variables:
985 IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
986 apply to IPv6 [XXX?].
989 Default value for IPV6_V6ONLY socket option,
990 which restricts use of the IPv6 socket to IPv6 communication
992 TRUE: disable IPv4-mapped address feature
993 FALSE: enable IPv4-mapped address feature
995 Default: FALSE (as specified in RFC2553bis)
999 ip6frag_high_thresh - INTEGER
1000 Maximum memory used to reassemble IPv6 fragments. When
1001 ip6frag_high_thresh bytes of memory is allocated for this purpose,
1002 the fragment handler will toss packets until ip6frag_low_thresh
1005 ip6frag_low_thresh - INTEGER
1006 See ip6frag_high_thresh
1008 ip6frag_time - INTEGER
1009 Time in seconds to keep an IPv6 fragment in memory.
1011 ip6frag_secret_interval - INTEGER
1012 Regeneration interval (in seconds) of the hash secret (or lifetime
1013 for the hash secret) for IPv6 fragments.
1017 Change the interface-specific default settings.
1021 Change all the interface-specific settings.
1023 [XXX: Other special features than forwarding?]
1025 conf/all/forwarding - BOOLEAN
1026 Enable global IPv6 forwarding between all interfaces.
1028 IPv4 and IPv6 work differently here; e.g. netfilter must be used
1029 to control which interfaces may forward packets and which not.
1031 This also sets all interfaces' Host/Router setting
1032 'forwarding' to the specified value. See below for details.
1034 This referred to as global forwarding.
1040 Change special settings per interface.
1042 The functional behaviour for certain settings is different
1043 depending on whether local forwarding is enabled or not.
1046 Accept Router Advertisements; autoconfigure using them.
1048 Possible values are:
1049 0 Do not accept Router Advertisements.
1050 1 Accept Router Advertisements if forwarding is disabled.
1051 2 Overrule forwarding behaviour. Accept Router Advertisements
1052 even if forwarding is enabled.
1054 Functional default: enabled if local forwarding is disabled.
1055 disabled if local forwarding is enabled.
1057 accept_ra_defrtr - BOOLEAN
1058 Learn default router in Router Advertisement.
1060 Functional default: enabled if accept_ra is enabled.
1061 disabled if accept_ra is disabled.
1063 accept_ra_pinfo - BOOLEAN
1064 Learn Prefix Information in Router Advertisement.
1066 Functional default: enabled if accept_ra is enabled.
1067 disabled if accept_ra is disabled.
1069 accept_ra_rt_info_max_plen - INTEGER
1070 Maximum prefix length of Route Information in RA.
1072 Route Information w/ prefix larger than or equal to this
1073 variable shall be ignored.
1075 Functional default: 0 if accept_ra_rtr_pref is enabled.
1076 -1 if accept_ra_rtr_pref is disabled.
1078 accept_ra_rtr_pref - BOOLEAN
1079 Accept Router Preference in RA.
1081 Functional default: enabled if accept_ra is enabled.
1082 disabled if accept_ra is disabled.
1084 accept_redirects - BOOLEAN
1087 Functional default: enabled if local forwarding is disabled.
1088 disabled if local forwarding is enabled.
1090 accept_source_route - INTEGER
1091 Accept source routing (routing extension header).
1093 >= 0: Accept only routing header type 2.
1094 < 0: Do not accept routing header.
1099 Autoconfigure addresses using Prefix Information in Router
1102 Functional default: enabled if accept_ra_pinfo is enabled.
1103 disabled if accept_ra_pinfo is disabled.
1105 dad_transmits - INTEGER
1106 The amount of Duplicate Address Detection probes to send.
1109 forwarding - BOOLEAN
1110 Configure interface-specific Host/Router behaviour.
1112 Note: It is recommended to have the same setting on all
1113 interfaces; mixed router/host scenarios are rather uncommon.
1115 Possible values are:
1116 0 Forwarding disabled
1117 1 Forwarding enabled
1118 2 Forwarding enabled (Hybrid Mode)
1122 By default, Host behaviour is assumed. This means:
1124 1. IsRouter flag is not set in Neighbour Advertisements.
1125 2. Router Solicitations are being sent when necessary.
1126 3. If accept_ra is TRUE (default), accept Router
1127 Advertisements (and do autoconfiguration).
1128 4. If accept_redirects is TRUE (default), accept Redirects.
1132 If local forwarding is enabled, Router behaviour is assumed.
1133 This means exactly the reverse from the above:
1135 1. IsRouter flag is set in Neighbour Advertisements.
1136 2. Router Solicitations are not sent.
1137 3. Router Advertisements are ignored unless accept_ra is 2.
1138 4. Redirects are ignored.
1142 Hybrid mode. Same behaviour as TRUE, except for:
1144 2. Router Solicitations are being sent when necessary.
1146 Default: 0 (disabled) if global forwarding is disabled (default),
1147 otherwise 1 (enabled).
1150 Default Hop Limit to set.
1154 Default Maximum Transfer Unit
1155 Default: 1280 (IPv6 required minimum)
1157 router_probe_interval - INTEGER
1158 Minimum interval (in seconds) between Router Probing described
1163 router_solicitation_delay - INTEGER
1164 Number of seconds to wait after interface is brought up
1165 before sending Router Solicitations.
1168 router_solicitation_interval - INTEGER
1169 Number of seconds to wait between Router Solicitations.
1172 router_solicitations - INTEGER
1173 Number of Router Solicitations to send until assuming no
1174 routers are present.
1177 use_tempaddr - INTEGER
1178 Preference for Privacy Extensions (RFC3041).
1179 <= 0 : disable Privacy Extensions
1180 == 1 : enable Privacy Extensions, but prefer public
1181 addresses over temporary addresses.
1182 > 1 : enable Privacy Extensions and prefer temporary
1183 addresses over public addresses.
1184 Default: 0 (for most devices)
1185 -1 (for point-to-point devices and loopback devices)
1187 temp_valid_lft - INTEGER
1188 valid lifetime (in seconds) for temporary addresses.
1189 Default: 604800 (7 days)
1191 temp_prefered_lft - INTEGER
1192 Preferred lifetime (in seconds) for temporary addresses.
1193 Default: 86400 (1 day)
1195 max_desync_factor - INTEGER
1196 Maximum value for DESYNC_FACTOR, which is a random value
1197 that ensures that clients don't synchronize with each
1198 other and generate new addresses at exactly the same time.
1199 value is in seconds.
1202 regen_max_retry - INTEGER
1203 Number of attempts before give up attempting to generate
1204 valid temporary addresses.
1207 max_addresses - INTEGER
1208 Maximum number of autoconfigured addresses per interface. Setting
1209 to zero disables the limitation. It is not recommended to set this
1210 value too large (or to zero) because it would be an easy way to
1211 crash the kernel by allowing too many addresses to be created.
1214 disable_ipv6 - BOOLEAN
1215 Disable IPv6 operation. If accept_dad is set to 2, this value
1216 will be dynamically set to TRUE if DAD fails for the link-local
1218 Default: FALSE (enable IPv6 operation)
1220 When this value is changed from 1 to 0 (IPv6 is being enabled),
1221 it will dynamically create a link-local address on the given
1222 interface and start Duplicate Address Detection, if necessary.
1224 When this value is changed from 0 to 1 (IPv6 is being disabled),
1225 it will dynamically delete all address on the given interface.
1227 accept_dad - INTEGER
1228 Whether to accept DAD (Duplicate Address Detection).
1230 1: Enable DAD (default)
1231 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate
1232 link-local address has been found.
1234 force_tllao - BOOLEAN
1235 Enable sending the target link-layer address option even when
1236 responding to a unicast neighbor solicitation.
1239 Quoting from RFC 2461, section 4.4, Target link-layer address:
1241 "The option MUST be included for multicast solicitations in order to
1242 avoid infinite Neighbor Solicitation "recursion" when the peer node
1243 does not have a cache entry to return a Neighbor Advertisements
1244 message. When responding to unicast solicitations, the option can be
1245 omitted since the sender of the solicitation has the correct link-
1246 layer address; otherwise it would not have be able to send the unicast
1247 solicitation in the first place. However, including the link-layer
1248 address in this case adds little overhead and eliminates a potential
1249 race condition where the sender deletes the cached link-layer address
1250 prior to receiving a response to a previous solicitation."
1254 Limit the maximal rates for sending ICMPv6 packets.
1255 0 to disable any limiting,
1256 otherwise the minimal space between responses in milliseconds.
1261 Pekka Savola <pekkas@netcore.fi>
1262 YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
1265 /proc/sys/net/bridge/* Variables:
1267 bridge-nf-call-arptables - BOOLEAN
1268 1 : pass bridged ARP traffic to arptables' FORWARD chain.
1272 bridge-nf-call-iptables - BOOLEAN
1273 1 : pass bridged IPv4 traffic to iptables' chains.
1277 bridge-nf-call-ip6tables - BOOLEAN
1278 1 : pass bridged IPv6 traffic to ip6tables' chains.
1282 bridge-nf-filter-vlan-tagged - BOOLEAN
1283 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
1287 bridge-nf-filter-pppoe-tagged - BOOLEAN
1288 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
1293 proc/sys/net/sctp/* Variables:
1295 addip_enable - BOOLEAN
1296 Enable or disable extension of Dynamic Address Reconfiguration
1297 (ADD-IP) functionality specified in RFC5061. This extension provides
1298 the ability to dynamically add and remove new addresses for the SCTP
1301 1: Enable extension.
1303 0: Disable extension.
1307 addip_noauth_enable - BOOLEAN
1308 Dynamic Address Reconfiguration (ADD-IP) requires the use of
1309 authentication to protect the operations of adding or removing new
1310 addresses. This requirement is mandated so that unauthorized hosts
1311 would not be able to hijack associations. However, older
1312 implementations may not have implemented this requirement while
1313 allowing the ADD-IP extension. For reasons of interoperability,
1314 we provide this variable to control the enforcement of the
1315 authentication requirement.
1317 1: Allow ADD-IP extension to be used without authentication. This
1318 should only be set in a closed environment for interoperability
1319 with older implementations.
1321 0: Enforce the authentication requirement
1325 auth_enable - BOOLEAN
1326 Enable or disable Authenticated Chunks extension. This extension
1327 provides the ability to send and receive authenticated chunks and is
1328 required for secure operation of Dynamic Address Reconfiguration
1331 1: Enable this extension.
1332 0: Disable this extension.
1336 prsctp_enable - BOOLEAN
1337 Enable or disable the Partial Reliability extension (RFC3758) which
1338 is used to notify peers that a given DATA should no longer be expected.
1346 The limit of the number of new packets that can be initially sent. It
1347 controls how bursty the generated traffic can be.
1351 association_max_retrans - INTEGER
1352 Set the maximum number for retransmissions that an association can
1353 attempt deciding that the remote end is unreachable. If this value
1354 is exceeded, the association is terminated.
1358 max_init_retransmits - INTEGER
1359 The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
1360 that an association will attempt before declaring the destination
1361 unreachable and terminating.
1365 path_max_retrans - INTEGER
1366 The maximum number of retransmissions that will be attempted on a given
1367 path. Once this threshold is exceeded, the path is considered
1368 unreachable, and new traffic will use a different path when the
1369 association is multihomed.
1373 rto_initial - INTEGER
1374 The initial round trip timeout value in milliseconds that will be used
1375 in calculating round trip times. This is the initial time interval
1376 for retransmissions.
1381 The maximum value (in milliseconds) of the round trip timeout. This
1382 is the largest time interval that can elapse between retransmissions.
1387 The minimum value (in milliseconds) of the round trip timeout. This
1388 is the smallest time interval the can elapse between retransmissions.
1392 hb_interval - INTEGER
1393 The interval (in milliseconds) between HEARTBEAT chunks. These chunks
1394 are sent at the specified interval on idle paths to probe the state of
1395 a given path between 2 associations.
1399 sack_timeout - INTEGER
1400 The amount of time (in milliseconds) that the implementation will wait
1405 valid_cookie_life - INTEGER
1406 The default lifetime of the SCTP cookie (in milliseconds). The cookie
1407 is used during association establishment.
1411 cookie_preserve_enable - BOOLEAN
1412 Enable or disable the ability to extend the lifetime of the SCTP cookie
1413 that is used during the establishment phase of SCTP association
1415 1: Enable cookie lifetime extension.
1420 rcvbuf_policy - INTEGER
1421 Determines if the receive buffer is attributed to the socket or to
1422 association. SCTP supports the capability to create multiple
1423 associations on a single socket. When using this capability, it is
1424 possible that a single stalled association that's buffering a lot
1425 of data may block other associations from delivering their data by
1426 consuming all of the receive buffer space. To work around this,
1427 the rcvbuf_policy could be set to attribute the receiver buffer space
1428 to each association instead of the socket. This prevents the described
1431 1: rcvbuf space is per association
1432 0: recbuf space is per socket
1436 sndbuf_policy - INTEGER
1437 Similar to rcvbuf_policy above, this applies to send buffer space.
1439 1: Send buffer is tracked per association
1440 0: Send buffer is tracked per socket.
1444 sctp_mem - vector of 3 INTEGERs: min, pressure, max
1445 Number of pages allowed for queueing by all SCTP sockets.
1447 min: Below this number of pages SCTP is not bothered about its
1448 memory appetite. When amount of memory allocated by SCTP exceeds
1449 this number, SCTP starts to moderate memory usage.
1451 pressure: This value was introduced to follow format of tcp_mem.
1453 max: Number of pages allowed for queueing by all SCTP sockets.
1455 Default is calculated at boot time from amount of available memory.
1457 sctp_rmem - vector of 3 INTEGERs: min, default, max
1458 See tcp_rmem for a description.
1460 sctp_wmem - vector of 3 INTEGERs: min, default, max
1461 See tcp_wmem for a description.
1463 addr_scope_policy - INTEGER
1464 Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
1466 0 - Disable IPv4 address scoping
1467 1 - Enable IPv4 address scoping
1468 2 - Follow draft but allow IPv4 private addresses
1469 3 - Follow draft but allow IPv4 link local addresses
1474 /proc/sys/net/core/*
1475 dev_weight - INTEGER
1476 The maximum number of packets that kernel can handle on a NAPI
1477 interrupt, it's a Per-CPU variable.
1481 /proc/sys/net/unix/*
1482 max_dgram_qlen - INTEGER
1483 The maximum length of dgram socket receive queue
1490 /proc/sys/net/irda/*
1491 fast_poll_increase FIXME
1492 warn_noreply_time FIXME
1493 discovery_slots FIXME
1496 discovery_timeout FIXME
1497 lap_keepalive_time FIXME
1498 max_noreply_time FIXME
1499 max_tx_data_size FIXME
1501 min_tx_turn_time FIXME