xref: /openbmc/linux/Documentation/networking/ip-sysctl.rst (revision e71383fb9cd15a28d6c01d2c165a96f1c0bcf418)
1.. SPDX-License-Identifier: GPL-2.0
2
3=========
4IP Sysctl
5=========
6
7/proc/sys/net/ipv4/* Variables
8==============================
9
10ip_forward - BOOLEAN
11	- 0 - disabled (default)
12	- not 0 - enabled
13
14	Forward Packets between interfaces.
15
16	This variable is special, its change resets all configuration
17	parameters to their default state (RFC1122 for hosts, RFC1812
18	for routers)
19
20ip_default_ttl - INTEGER
21	Default value of TTL field (Time To Live) for outgoing (but not
22	forwarded) IP packets. Should be between 1 and 255 inclusive.
23	Default: 64 (as recommended by RFC1700)
24
25ip_no_pmtu_disc - INTEGER
26	Disable Path MTU Discovery. If enabled in mode 1 and a
27	fragmentation-required ICMP is received, the PMTU to this
28	destination will be set to the smallest of the old MTU to
29	this destination and min_pmtu (see below). You will need
30	to raise min_pmtu to the smallest interface MTU on your system
31	manually if you want to avoid locally generated fragments.
32
33	In mode 2 incoming Path MTU Discovery messages will be
34	discarded. Outgoing frames are handled the same as in mode 1,
35	implicitly setting IP_PMTUDISC_DONT on every created socket.
36
37	Mode 3 is a hardened pmtu discover mode. The kernel will only
38	accept fragmentation-needed errors if the underlying protocol
39	can verify them besides a plain socket lookup. Current
40	protocols for which pmtu events will be honored are TCP, SCTP
41	and DCCP as they verify e.g. the sequence number or the
42	association. This mode should not be enabled globally but is
43	only intended to secure e.g. name servers in namespaces where
44	TCP path mtu must still work but path MTU information of other
45	protocols should be discarded. If enabled globally this mode
46	could break other protocols.
47
48	Possible values: 0-3
49
50	Default: FALSE
51
52min_pmtu - INTEGER
53	default 552 - minimum Path MTU. Unless this is changed manually,
54	each cached pmtu will never be lower than this setting.
55
56ip_forward_use_pmtu - BOOLEAN
57	By default we don't trust protocol path MTUs while forwarding
58	because they could be easily forged and can lead to unwanted
59	fragmentation by the router.
60	You only need to enable this if you have user-space software
61	which tries to discover path mtus by itself and depends on the
62	kernel honoring this information. This is normally not the
63	case.
64
65	Default: 0 (disabled)
66
67	Possible values:
68
69	- 0 - disabled
70	- 1 - enabled
71
72fwmark_reflect - BOOLEAN
73	Controls the fwmark of kernel-generated IPv4 reply packets that are not
74	associated with a socket for example, TCP RSTs or ICMP echo replies).
75	If unset, these packets have a fwmark of zero. If set, they have the
76	fwmark of the packet they are replying to.
77
78	Default: 0
79
80fib_multipath_use_neigh - BOOLEAN
81	Use status of existing neighbor entry when determining nexthop for
82	multipath routes. If disabled, neighbor information is not used and
83	packets could be directed to a failed nexthop. Only valid for kernels
84	built with CONFIG_IP_ROUTE_MULTIPATH enabled.
85
86	Default: 0 (disabled)
87
88	Possible values:
89
90	- 0 - disabled
91	- 1 - enabled
92
93fib_multipath_hash_policy - INTEGER
94	Controls which hash policy to use for multipath routes. Only valid
95	for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
96
97	Default: 0 (Layer 3)
98
99	Possible values:
100
101	- 0 - Layer 3
102	- 1 - Layer 4
103	- 2 - Layer 3 or inner Layer 3 if present
104	- 3 - Custom multipath hash. Fields used for multipath hash calculation
105	  are determined by fib_multipath_hash_fields sysctl
106
107fib_multipath_hash_fields - UNSIGNED INTEGER
108	When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
109	fields used for multipath hash calculation are determined by this
110	sysctl.
111
112	This value is a bitmask which enables various fields for multipath hash
113	calculation.
114
115	Possible fields are:
116
117	====== ============================
118	0x0001 Source IP address
119	0x0002 Destination IP address
120	0x0004 IP protocol
121	0x0008 Unused (Flow Label)
122	0x0010 Source port
123	0x0020 Destination port
124	0x0040 Inner source IP address
125	0x0080 Inner destination IP address
126	0x0100 Inner IP protocol
127	0x0200 Inner Flow Label
128	0x0400 Inner source port
129	0x0800 Inner destination port
130	====== ============================
131
132	Default: 0x0007 (source IP, destination IP and IP protocol)
133
134fib_sync_mem - UNSIGNED INTEGER
135	Amount of dirty memory from fib entries that can be backlogged before
136	synchronize_rcu is forced.
137
138	Default: 512kB   Minimum: 64kB   Maximum: 64MB
139
140ip_forward_update_priority - INTEGER
141	Whether to update SKB priority from "TOS" field in IPv4 header after it
142	is forwarded. The new SKB priority is mapped from TOS field value
143	according to an rt_tos2priority table (see e.g. man tc-prio).
144
145	Default: 1 (Update priority.)
146
147	Possible values:
148
149	- 0 - Do not update priority.
150	- 1 - Update priority.
151
152route/max_size - INTEGER
153	Maximum number of routes allowed in the kernel.  Increase
154	this when using large numbers of interfaces and/or routes.
155
156	From linux kernel 3.6 onwards, this is deprecated for ipv4
157	as route cache is no longer used.
158
159	From linux kernel 6.3 onwards, this is deprecated for ipv6
160	as garbage collection manages cached route entries.
161
162neigh/default/gc_thresh1 - INTEGER
163	Minimum number of entries to keep.  Garbage collector will not
164	purge entries if there are fewer than this number.
165
166	Default: 128
167
168neigh/default/gc_thresh2 - INTEGER
169	Threshold when garbage collector becomes more aggressive about
170	purging entries. Entries older than 5 seconds will be cleared
171	when over this number.
172
173	Default: 512
174
175neigh/default/gc_thresh3 - INTEGER
176	Maximum number of non-PERMANENT neighbor entries allowed.  Increase
177	this when using large numbers of interfaces and when communicating
178	with large numbers of directly-connected peers.
179
180	Default: 1024
181
182neigh/default/unres_qlen_bytes - INTEGER
183	The maximum number of bytes which may be used by packets
184	queued for each	unresolved address by other network layers.
185	(added in linux 3.3)
186
187	Setting negative value is meaningless and will return error.
188
189	Default: SK_WMEM_MAX, (same as net.core.wmem_default).
190
191		Exact value depends on architecture and kernel options,
192		but should be enough to allow queuing 256 packets
193		of medium size.
194
195neigh/default/unres_qlen - INTEGER
196	The maximum number of packets which may be queued for each
197	unresolved address by other network layers.
198
199	(deprecated in linux 3.3) : use unres_qlen_bytes instead.
200
201	Prior to linux 3.3, the default value is 3 which may cause
202	unexpected packet loss. The current default value is calculated
203	according to default value of unres_qlen_bytes and true size of
204	packet.
205
206	Default: 101
207
208neigh/default/interval_probe_time_ms - INTEGER
209	The probe interval for neighbor entries with NTF_MANAGED flag,
210	the min value is 1.
211
212	Default: 5000
213
214mtu_expires - INTEGER
215	Time, in seconds, that cached PMTU information is kept.
216
217min_adv_mss - INTEGER
218	The advertised MSS depends on the first hop route MTU, but will
219	never be lower than this setting.
220
221fib_notify_on_flag_change - INTEGER
222        Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
223        RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
224
225        After installing a route to the kernel, user space receives an
226        acknowledgment, which means the route was installed in the kernel,
227        but not necessarily in hardware.
228        It is also possible for a route already installed in hardware to change
229        its action and therefore its flags. For example, a host route that is
230        trapping packets can be "promoted" to perform decapsulation following
231        the installation of an IPinIP/VXLAN tunnel.
232        The notifications will indicate to user-space the state of the route.
233
234        Default: 0 (Do not emit notifications.)
235
236        Possible values:
237
238        - 0 - Do not emit notifications.
239        - 1 - Emit notifications.
240        - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
241
242IP Fragmentation:
243
244ipfrag_high_thresh - LONG INTEGER
245	Maximum memory used to reassemble IP fragments.
246
247ipfrag_low_thresh - LONG INTEGER
248	(Obsolete since linux-4.17)
249	Maximum memory used to reassemble IP fragments before the kernel
250	begins to remove incomplete fragment queues to free up resources.
251	The kernel still accepts new fragments for defragmentation.
252
253ipfrag_time - INTEGER
254	Time in seconds to keep an IP fragment in memory.
255
256ipfrag_max_dist - INTEGER
257	ipfrag_max_dist is a non-negative integer value which defines the
258	maximum "disorder" which is allowed among fragments which share a
259	common IP source address. Note that reordering of packets is
260	not unusual, but if a large number of fragments arrive from a source
261	IP address while a particular fragment queue remains incomplete, it
262	probably indicates that one or more fragments belonging to that queue
263	have been lost. When ipfrag_max_dist is positive, an additional check
264	is done on fragments before they are added to a reassembly queue - if
265	ipfrag_max_dist (or more) fragments have arrived from a particular IP
266	address between additions to any IP fragment queue using that source
267	address, it's presumed that one or more fragments in the queue are
268	lost. The existing fragment queue will be dropped, and a new one
269	started. An ipfrag_max_dist value of zero disables this check.
270
271	Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
272	result in unnecessarily dropping fragment queues when normal
273	reordering of packets occurs, which could lead to poor application
274	performance. Using a very large value, e.g. 50000, increases the
275	likelihood of incorrectly reassembling IP fragments that originate
276	from different IP datagrams, which could result in data corruption.
277	Default: 64
278
279bc_forwarding - INTEGER
280	bc_forwarding enables the feature described in rfc1812#section-5.3.5.2
281	and rfc2644. It allows the router to forward directed broadcast.
282	To enable this feature, the 'all' entry and the input interface entry
283	should be set to 1.
284	Default: 0
285
286INET peer storage
287=================
288
289inet_peer_threshold - INTEGER
290	The approximate size of the storage.  Starting from this threshold
291	entries will be thrown aggressively.  This threshold also determines
292	entries' time-to-live and time intervals between garbage collection
293	passes.  More entries, less time-to-live, less GC interval.
294
295inet_peer_minttl - INTEGER
296	Minimum time-to-live of entries.  Should be enough to cover fragment
297	time-to-live on the reassembling side.  This minimum time-to-live  is
298	guaranteed if the pool size is less than inet_peer_threshold.
299	Measured in seconds.
300
301inet_peer_maxttl - INTEGER
302	Maximum time-to-live of entries.  Unused entries will expire after
303	this period of time if there is no memory pressure on the pool (i.e.
304	when the number of entries in the pool is very small).
305	Measured in seconds.
306
307TCP variables
308=============
309
310somaxconn - INTEGER
311	Limit of socket listen() backlog, known in userspace as SOMAXCONN.
312	Defaults to 4096. (Was 128 before linux-5.4)
313	See also tcp_max_syn_backlog for additional tuning for TCP sockets.
314
315tcp_abort_on_overflow - BOOLEAN
316	If listening service is too slow to accept new connections,
317	reset them. Default state is FALSE. It means that if overflow
318	occurred due to a burst, connection will recover. Enable this
319	option _only_ if you are really sure that listening daemon
320	cannot be tuned to accept connections faster. Enabling this
321	option can harm clients of your server.
322
323tcp_adv_win_scale - INTEGER
324	Count buffering overhead as bytes/2^tcp_adv_win_scale
325	(if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
326	if it is <= 0.
327
328	Possible values are [-31, 31], inclusive.
329
330	Default: 1
331
332tcp_allowed_congestion_control - STRING
333	Show/set the congestion control choices available to non-privileged
334	processes. The list is a subset of those listed in
335	tcp_available_congestion_control.
336
337	Default is "reno" and the default setting (tcp_congestion_control).
338
339tcp_app_win - INTEGER
340	Reserve max(window/2^tcp_app_win, mss) of window for application
341	buffer. Value 0 is special, it means that nothing is reserved.
342
343	Possible values are [0, 31], inclusive.
344
345	Default: 31
346
347tcp_autocorking - BOOLEAN
348	Enable TCP auto corking :
349	When applications do consecutive small write()/sendmsg() system calls,
350	we try to coalesce these small writes as much as possible, to lower
351	total amount of sent packets. This is done if at least one prior
352	packet for the flow is waiting in Qdisc queues or device transmit
353	queue. Applications can still use TCP_CORK for optimal behavior
354	when they know how/when to uncork their sockets.
355
356	Default : 1
357
358tcp_available_congestion_control - STRING
359	Shows the available congestion control choices that are registered.
360	More congestion control algorithms may be available as modules,
361	but not loaded.
362
363tcp_base_mss - INTEGER
364	The initial value of search_low to be used by the packetization layer
365	Path MTU discovery (MTU probing).  If MTU probing is enabled,
366	this is the initial MSS used by the connection.
367
368tcp_mtu_probe_floor - INTEGER
369	If MTU probing is enabled this caps the minimum MSS used for search_low
370	for the connection.
371
372	Default : 48
373
374tcp_min_snd_mss - INTEGER
375	TCP SYN and SYNACK messages usually advertise an ADVMSS option,
376	as described in RFC 1122 and RFC 6691.
377
378	If this ADVMSS option is smaller than tcp_min_snd_mss,
379	it is silently capped to tcp_min_snd_mss.
380
381	Default : 48 (at least 8 bytes of payload per segment)
382
383tcp_congestion_control - STRING
384	Set the congestion control algorithm to be used for new
385	connections. The algorithm "reno" is always available, but
386	additional choices may be available based on kernel configuration.
387	Default is set as part of kernel configuration.
388	For passive connections, the listener congestion control choice
389	is inherited.
390
391	[see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
392
393tcp_dsack - BOOLEAN
394	Allows TCP to send "duplicate" SACKs.
395
396tcp_early_retrans - INTEGER
397	Tail loss probe (TLP) converts RTOs occurring due to tail
398	losses into fast recovery (draft-ietf-tcpm-rack). Note that
399	TLP requires RACK to function properly (see tcp_recovery below)
400
401	Possible values:
402
403		- 0 disables TLP
404		- 3 or 4 enables TLP
405
406	Default: 3
407
408tcp_ecn - INTEGER
409	Control use of Explicit Congestion Notification (ECN) by TCP.
410	ECN is used only when both ends of the TCP connection indicate
411	support for it.  This feature is useful in avoiding losses due
412	to congestion by allowing supporting routers to signal
413	congestion before having to drop packets.
414
415	Possible values are:
416
417		=  =====================================================
418		0  Disable ECN.  Neither initiate nor accept ECN.
419		1  Enable ECN when requested by incoming connections and
420		   also request ECN on outgoing connection attempts.
421		2  Enable ECN when requested by incoming connections
422		   but do not request ECN on outgoing connections.
423		=  =====================================================
424
425	Default: 2
426
427tcp_ecn_fallback - BOOLEAN
428	If the kernel detects that ECN connection misbehaves, enable fall
429	back to non-ECN. Currently, this knob implements the fallback
430	from RFC3168, section 6.1.1.1., but we reserve that in future,
431	additional detection mechanisms could be implemented under this
432	knob. The value	is not used, if tcp_ecn or per route (or congestion
433	control) ECN settings are disabled.
434
435	Default: 1 (fallback enabled)
436
437tcp_fack - BOOLEAN
438	This is a legacy option, it has no effect anymore.
439
440tcp_fin_timeout - INTEGER
441	The length of time an orphaned (no longer referenced by any
442	application) connection will remain in the FIN_WAIT_2 state
443	before it is aborted at the local end.  While a perfectly
444	valid "receive only" state for an un-orphaned connection, an
445	orphaned connection in FIN_WAIT_2 state could otherwise wait
446	forever for the remote to close its end of the connection.
447
448	Cf. tcp_max_orphans
449
450	Default: 60 seconds
451
452tcp_frto - INTEGER
453	Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
454	F-RTO is an enhanced recovery algorithm for TCP retransmission
455	timeouts.  It is particularly beneficial in networks where the
456	RTT fluctuates (e.g., wireless). F-RTO is sender-side only
457	modification. It does not require any support from the peer.
458
459	By default it's enabled with a non-zero value. 0 disables F-RTO.
460
461tcp_fwmark_accept - BOOLEAN
462	If set, incoming connections to listening sockets that do not have a
463	socket mark will set the mark of the accepting socket to the fwmark of
464	the incoming SYN packet. This will cause all packets on that connection
465	(starting from the first SYNACK) to be sent with that fwmark. The
466	listening socket's mark is unchanged. Listening sockets that already
467	have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
468	unaffected.
469
470	Default: 0
471
472tcp_invalid_ratelimit - INTEGER
473	Limit the maximal rate for sending duplicate acknowledgments
474	in response to incoming TCP packets that are for an existing
475	connection but that are invalid due to any of these reasons:
476
477	  (a) out-of-window sequence number,
478	  (b) out-of-window acknowledgment number, or
479	  (c) PAWS (Protection Against Wrapped Sequence numbers) check failure
480
481	This can help mitigate simple "ack loop" DoS attacks, wherein
482	a buggy or malicious middlebox or man-in-the-middle can
483	rewrite TCP header fields in manner that causes each endpoint
484	to think that the other is sending invalid TCP segments, thus
485	causing each side to send an unterminating stream of duplicate
486	acknowledgments for invalid segments.
487
488	Using 0 disables rate-limiting of dupacks in response to
489	invalid segments; otherwise this value specifies the minimal
490	space between sending such dupacks, in milliseconds.
491
492	Default: 500 (milliseconds).
493
494tcp_keepalive_time - INTEGER
495	How often TCP sends out keepalive messages when keepalive is enabled.
496	Default: 2hours.
497
498tcp_keepalive_probes - INTEGER
499	How many keepalive probes TCP sends out, until it decides that the
500	connection is broken. Default value: 9.
501
502tcp_keepalive_intvl - INTEGER
503	How frequently the probes are send out. Multiplied by
504	tcp_keepalive_probes it is time to kill not responding connection,
505	after probes started. Default value: 75sec i.e. connection
506	will be aborted after ~11 minutes of retries.
507
508tcp_l3mdev_accept - BOOLEAN
509	Enables child sockets to inherit the L3 master device index.
510	Enabling this option allows a "global" listen socket to work
511	across L3 master domains (e.g., VRFs) with connected sockets
512	derived from the listen socket to be bound to the L3 domain in
513	which the packets originated. Only valid when the kernel was
514	compiled with CONFIG_NET_L3_MASTER_DEV.
515
516	Default: 0 (disabled)
517
518tcp_low_latency - BOOLEAN
519	This is a legacy option, it has no effect anymore.
520
521tcp_max_orphans - INTEGER
522	Maximal number of TCP sockets not attached to any user file handle,
523	held by system.	If this number is exceeded orphaned connections are
524	reset immediately and warning is printed. This limit exists
525	only to prevent simple DoS attacks, you _must_ not rely on this
526	or lower the limit artificially, but rather increase it
527	(probably, after increasing installed memory),
528	if network conditions require more than default value,
529	and tune network services to linger and kill such states
530	more aggressively. Let me to remind again: each orphan eats
531	up to ~64K of unswappable memory.
532
533tcp_max_syn_backlog - INTEGER
534	Maximal number of remembered connection requests (SYN_RECV),
535	which have not received an acknowledgment from connecting client.
536
537	This is a per-listener limit.
538
539	The minimal value is 128 for low memory machines, and it will
540	increase in proportion to the memory of machine.
541
542	If server suffers from overload, try increasing this number.
543
544	Remember to also check /proc/sys/net/core/somaxconn
545	A SYN_RECV request socket consumes about 304 bytes of memory.
546
547tcp_max_tw_buckets - INTEGER
548	Maximal number of timewait sockets held by system simultaneously.
549	If this number is exceeded time-wait socket is immediately destroyed
550	and warning is printed. This limit exists only to prevent
551	simple DoS attacks, you _must_ not lower the limit artificially,
552	but rather increase it (probably, after increasing installed memory),
553	if network conditions require more than default value.
554
555tcp_mem - vector of 3 INTEGERs: min, pressure, max
556	min: below this number of pages TCP is not bothered about its
557	memory appetite.
558
559	pressure: when amount of memory allocated by TCP exceeds this number
560	of pages, TCP moderates its memory consumption and enters memory
561	pressure mode, which is exited when memory consumption falls
562	under "min".
563
564	max: number of pages allowed for queueing by all TCP sockets.
565
566	Defaults are calculated at boot time from amount of available
567	memory.
568
569tcp_min_rtt_wlen - INTEGER
570	The window length of the windowed min filter to track the minimum RTT.
571	A shorter window lets a flow more quickly pick up new (higher)
572	minimum RTT when it is moved to a longer path (e.g., due to traffic
573	engineering). A longer window makes the filter more resistant to RTT
574	inflations such as transient congestion. The unit is seconds.
575
576	Possible values: 0 - 86400 (1 day)
577
578	Default: 300
579
580tcp_moderate_rcvbuf - BOOLEAN
581	If set, TCP performs receive buffer auto-tuning, attempting to
582	automatically size the buffer (no greater than tcp_rmem[2]) to
583	match the size required by the path for full throughput.  Enabled by
584	default.
585
586tcp_mtu_probing - INTEGER
587	Controls TCP Packetization-Layer Path MTU Discovery.  Takes three
588	values:
589
590	- 0 - Disabled
591	- 1 - Disabled by default, enabled when an ICMP black hole detected
592	- 2 - Always enabled, use initial MSS of tcp_base_mss.
593
594tcp_probe_interval - UNSIGNED INTEGER
595	Controls how often to start TCP Packetization-Layer Path MTU
596	Discovery reprobe. The default is reprobing every 10 minutes as
597	per RFC4821.
598
599tcp_probe_threshold - INTEGER
600	Controls when TCP Packetization-Layer Path MTU Discovery probing
601	will stop in respect to the width of search range in bytes. Default
602	is 8 bytes.
603
604tcp_no_metrics_save - BOOLEAN
605	By default, TCP saves various connection metrics in the route cache
606	when the connection closes, so that connections established in the
607	near future can use these to set initial conditions.  Usually, this
608	increases overall performance, but may sometimes cause performance
609	degradation.  If set, TCP will not cache metrics on closing
610	connections.
611
612tcp_no_ssthresh_metrics_save - BOOLEAN
613	Controls whether TCP saves ssthresh metrics in the route cache.
614
615	Default is 1, which disables ssthresh metrics.
616
617tcp_orphan_retries - INTEGER
618	This value influences the timeout of a locally closed TCP connection,
619	when RTO retransmissions remain unacknowledged.
620	See tcp_retries2 for more details.
621
622	The default value is 8.
623
624	If your machine is a loaded WEB server,
625	you should think about lowering this value, such sockets
626	may consume significant resources. Cf. tcp_max_orphans.
627
628tcp_recovery - INTEGER
629	This value is a bitmap to enable various experimental loss recovery
630	features.
631
632	=========   =============================================================
633	RACK: 0x1   enables the RACK loss detection for fast detection of lost
634		    retransmissions and tail drops. It also subsumes and disables
635		    RFC6675 recovery for SACK connections.
636
637	RACK: 0x2   makes RACK's reordering window static (min_rtt/4).
638
639	RACK: 0x4   disables RACK's DUPACK threshold heuristic
640	=========   =============================================================
641
642	Default: 0x1
643
644tcp_reflect_tos - BOOLEAN
645	For listening sockets, reuse the DSCP value of the initial SYN message
646	for outgoing packets. This allows to have both directions of a TCP
647	stream to use the same DSCP value, assuming DSCP remains unchanged for
648	the lifetime of the connection.
649
650	This options affects both IPv4 and IPv6.
651
652	Default: 0 (disabled)
653
654tcp_reordering - INTEGER
655	Initial reordering level of packets in a TCP stream.
656	TCP stack can then dynamically adjust flow reordering level
657	between this initial value and tcp_max_reordering
658
659	Default: 3
660
661tcp_max_reordering - INTEGER
662	Maximal reordering level of packets in a TCP stream.
663	300 is a fairly conservative value, but you might increase it
664	if paths are using per packet load balancing (like bonding rr mode)
665
666	Default: 300
667
668tcp_retrans_collapse - BOOLEAN
669	Bug-to-bug compatibility with some broken printers.
670	On retransmit try to send bigger packets to work around bugs in
671	certain TCP stacks.
672
673tcp_retries1 - INTEGER
674	This value influences the time, after which TCP decides, that
675	something is wrong due to unacknowledged RTO retransmissions,
676	and reports this suspicion to the network layer.
677	See tcp_retries2 for more details.
678
679	RFC 1122 recommends at least 3 retransmissions, which is the
680	default.
681
682tcp_retries2 - INTEGER
683	This value influences the timeout of an alive TCP connection,
684	when RTO retransmissions remain unacknowledged.
685	Given a value of N, a hypothetical TCP connection following
686	exponential backoff with an initial RTO of TCP_RTO_MIN would
687	retransmit N times before killing the connection at the (N+1)th RTO.
688
689	The default value of 15 yields a hypothetical timeout of 924.6
690	seconds and is a lower bound for the effective timeout.
691	TCP will effectively time out at the first RTO which exceeds the
692	hypothetical timeout.
693
694	RFC 1122 recommends at least 100 seconds for the timeout,
695	which corresponds to a value of at least 8.
696
697tcp_rfc1337 - BOOLEAN
698	If set, the TCP stack behaves conforming to RFC1337. If unset,
699	we are not conforming to RFC, but prevent TCP TIME_WAIT
700	assassination.
701
702	Default: 0
703
704tcp_rmem - vector of 3 INTEGERs: min, default, max
705	min: Minimal size of receive buffer used by TCP sockets.
706	It is guaranteed to each TCP socket, even under moderate memory
707	pressure.
708
709	Default: 4K
710
711	default: initial size of receive buffer used by TCP sockets.
712	This value overrides net.core.rmem_default used by other protocols.
713	Default: 131072 bytes.
714	This value results in initial window of 65535.
715
716	max: maximal size of receive buffer allowed for automatically
717	selected receiver buffers for TCP socket. This value does not override
718	net.core.rmem_max.  Calling setsockopt() with SO_RCVBUF disables
719	automatic tuning of that socket's receive buffer size, in which
720	case this value is ignored.
721	Default: between 131072 and 6MB, depending on RAM size.
722
723tcp_sack - BOOLEAN
724	Enable select acknowledgments (SACKS).
725
726tcp_comp_sack_delay_ns - LONG INTEGER
727	TCP tries to reduce number of SACK sent, using a timer
728	based on 5% of SRTT, capped by this sysctl, in nano seconds.
729	The default is 1ms, based on TSO autosizing period.
730
731	Default : 1,000,000 ns (1 ms)
732
733tcp_comp_sack_slack_ns - LONG INTEGER
734	This sysctl control the slack used when arming the
735	timer used by SACK compression. This gives extra time
736	for small RTT flows, and reduces system overhead by allowing
737	opportunistic reduction of timer interrupts.
738
739	Default : 100,000 ns (100 us)
740
741tcp_comp_sack_nr - INTEGER
742	Max number of SACK that can be compressed.
743	Using 0 disables SACK compression.
744
745	Default : 44
746
747tcp_slow_start_after_idle - BOOLEAN
748	If set, provide RFC2861 behavior and time out the congestion
749	window after an idle period.  An idle period is defined at
750	the current RTO.  If unset, the congestion window will not
751	be timed out after an idle period.
752
753	Default: 1
754
755tcp_stdurg - BOOLEAN
756	Use the Host requirements interpretation of the TCP urgent pointer field.
757	Most hosts use the older BSD interpretation, so if you turn this on
758	Linux might not communicate correctly with them.
759
760	Default: FALSE
761
762tcp_synack_retries - INTEGER
763	Number of times SYNACKs for a passive TCP connection attempt will
764	be retransmitted. Should not be higher than 255. Default value
765	is 5, which corresponds to 31seconds till the last retransmission
766	with the current initial RTO of 1second. With this the final timeout
767	for a passive TCP connection will happen after 63seconds.
768
769tcp_syncookies - INTEGER
770	Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
771	Send out syncookies when the syn backlog queue of a socket
772	overflows. This is to prevent against the common 'SYN flood attack'
773	Default: 1
774
775	Note, that syncookies is fallback facility.
776	It MUST NOT be used to help highly loaded servers to stand
777	against legal connection rate. If you see SYN flood warnings
778	in your logs, but investigation	shows that they occur
779	because of overload with legal connections, you should tune
780	another parameters until this warning disappear.
781	See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
782
783	syncookies seriously violate TCP protocol, do not allow
784	to use TCP extensions, can result in serious degradation
785	of some services (f.e. SMTP relaying), visible not by you,
786	but your clients and relays, contacting you. While you see
787	SYN flood warnings in logs not being really flooded, your server
788	is seriously misconfigured.
789
790	If you want to test which effects syncookies have to your
791	network connections you can set this knob to 2 to enable
792	unconditionally generation of syncookies.
793
794tcp_migrate_req - BOOLEAN
795	The incoming connection is tied to a specific listening socket when
796	the initial SYN packet is received during the three-way handshake.
797	When a listener is closed, in-flight request sockets during the
798	handshake and established sockets in the accept queue are aborted.
799
800	If the listener has SO_REUSEPORT enabled, other listeners on the
801	same port should have been able to accept such connections. This
802	option makes it possible to migrate such child sockets to another
803	listener after close() or shutdown().
804
805	The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should
806	usually be used to define the policy to pick an alive listener.
807	Otherwise, the kernel will randomly pick an alive listener only if
808	this option is enabled.
809
810	Note that migration between listeners with different settings may
811	crash applications. Let's say migration happens from listener A to
812	B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from
813	the requests migrated from A. To avoid such a situation, cancel
814	migration by returning SK_DROP in the type of eBPF program, or
815	disable this option.
816
817	Default: 0
818
819tcp_fastopen - INTEGER
820	Enable TCP Fast Open (RFC7413) to send and accept data in the opening
821	SYN packet.
822
823	The client support is enabled by flag 0x1 (on by default). The client
824	then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
825	rather than connect() to send data in SYN.
826
827	The server support is enabled by flag 0x2 (off by default). Then
828	either enable for all listeners with another flag (0x400) or
829	enable individual listeners via TCP_FASTOPEN socket option with
830	the option value being the length of the syn-data backlog.
831
832	The values (bitmap) are
833
834	=====  ======== ======================================================
835	  0x1  (client) enables sending data in the opening SYN on the client.
836	  0x2  (server) enables the server support, i.e., allowing data in
837			a SYN packet to be accepted and passed to the
838			application before 3-way handshake finishes.
839	  0x4  (client) send data in the opening SYN regardless of cookie
840			availability and without a cookie option.
841	0x200  (server) accept data-in-SYN w/o any cookie option present.
842	0x400  (server) enable all listeners to support Fast Open by
843			default without explicit TCP_FASTOPEN socket option.
844	=====  ======== ======================================================
845
846	Default: 0x1
847
848	Note that additional client or server features are only
849	effective if the basic support (0x1 and 0x2) are enabled respectively.
850
851tcp_fastopen_blackhole_timeout_sec - INTEGER
852	Initial time period in second to disable Fastopen on active TCP sockets
853	when a TFO firewall blackhole issue happens.
854	This time period will grow exponentially when more blackhole issues
855	get detected right after Fastopen is re-enabled and will reset to
856	initial value when the blackhole issue goes away.
857	0 to disable the blackhole detection.
858
859	By default, it is set to 0 (feature is disabled).
860
861tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
862	The list consists of a primary key and an optional backup key. The
863	primary key is used for both creating and validating cookies, while the
864	optional backup key is only used for validating cookies. The purpose of
865	the backup key is to maximize TFO validation when keys are rotated.
866
867	A randomly chosen primary key may be configured by the kernel if
868	the tcp_fastopen sysctl is set to 0x400 (see above), or if the
869	TCP_FASTOPEN setsockopt() optname is set and a key has not been
870	previously configured via sysctl. If keys are configured via
871	setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
872	per-socket keys will be used instead of any keys that are specified via
873	sysctl.
874
875	A key is specified as 4 8-digit hexadecimal integers which are separated
876	by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
877	omitted. A primary and a backup key may be specified by separating them
878	by a comma. If only one key is specified, it becomes the primary key and
879	any previously configured backup keys are removed.
880
881tcp_syn_retries - INTEGER
882	Number of times initial SYNs for an active TCP connection attempt
883	will be retransmitted. Should not be higher than 127. Default value
884	is 6, which corresponds to 67seconds (with tcp_syn_linear_timeouts = 4)
885	till the last retransmission with the current initial RTO of 1second.
886	With this the final timeout for an active TCP connection attempt
887	will happen after 131seconds.
888
889tcp_timestamps - INTEGER
890	Enable timestamps as defined in RFC1323.
891
892	- 0: Disabled.
893	- 1: Enable timestamps as defined in RFC1323 and use random offset for
894	  each connection rather than only using the current time.
895	- 2: Like 1, but without random offsets.
896
897	Default: 1
898
899tcp_min_tso_segs - INTEGER
900	Minimal number of segments per TSO frame.
901
902	Since linux-3.12, TCP does an automatic sizing of TSO frames,
903	depending on flow rate, instead of filling 64Kbytes packets.
904	For specific usages, it's possible to force TCP to build big
905	TSO frames. Note that TCP stack might split too big TSO packets
906	if available window is too small.
907
908	Default: 2
909
910tcp_tso_rtt_log - INTEGER
911	Adjustment of TSO packet sizes based on min_rtt
912
913	Starting from linux-5.18, TCP autosizing can be tweaked
914	for flows having small RTT.
915
916	Old autosizing was splitting the pacing budget to send 1024 TSO
917	per second.
918
919	tso_packet_size = sk->sk_pacing_rate / 1024;
920
921	With the new mechanism, we increase this TSO sizing using:
922
923	distance = min_rtt_usec / (2^tcp_tso_rtt_log)
924	tso_packet_size += gso_max_size >> distance;
925
926	This means that flows between very close hosts can use bigger
927	TSO packets, reducing their cpu costs.
928
929	If you want to use the old autosizing, set this sysctl to 0.
930
931	Default: 9  (2^9 = 512 usec)
932
933tcp_pacing_ss_ratio - INTEGER
934	sk->sk_pacing_rate is set by TCP stack using a ratio applied
935	to current rate. (current_rate = cwnd * mss / srtt)
936	If TCP is in slow start, tcp_pacing_ss_ratio is applied
937	to let TCP probe for bigger speeds, assuming cwnd can be
938	doubled every other RTT.
939
940	Default: 200
941
942tcp_pacing_ca_ratio - INTEGER
943	sk->sk_pacing_rate is set by TCP stack using a ratio applied
944	to current rate. (current_rate = cwnd * mss / srtt)
945	If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
946	is applied to conservatively probe for bigger throughput.
947
948	Default: 120
949
950tcp_syn_linear_timeouts - INTEGER
951	The number of times for an active TCP connection to retransmit SYNs with
952	a linear backoff timeout before defaulting to an exponential backoff
953	timeout. This has no effect on SYNACK at the passive TCP side.
954
955	With an initial RTO of 1 and tcp_syn_linear_timeouts = 4 we would
956	expect SYN RTOs to be: 1, 1, 1, 1, 1, 2, 4, ... (4 linear timeouts,
957	and the first exponential backoff using 2^0 * initial_RTO).
958	Default: 4
959
960tcp_tso_win_divisor - INTEGER
961	This allows control over what percentage of the congestion window
962	can be consumed by a single TSO frame.
963	The setting of this parameter is a choice between burstiness and
964	building larger TSO frames.
965
966	Default: 3
967
968tcp_tw_reuse - INTEGER
969	Enable reuse of TIME-WAIT sockets for new connections when it is
970	safe from protocol viewpoint.
971
972	- 0 - disable
973	- 1 - global enable
974	- 2 - enable for loopback traffic only
975
976	It should not be changed without advice/request of technical
977	experts.
978
979	Default: 2
980
981tcp_window_scaling - BOOLEAN
982	Enable window scaling as defined in RFC1323.
983
984tcp_wmem - vector of 3 INTEGERs: min, default, max
985	min: Amount of memory reserved for send buffers for TCP sockets.
986	Each TCP socket has rights to use it due to fact of its birth.
987
988	Default: 4K
989
990	default: initial size of send buffer used by TCP sockets.  This
991	value overrides net.core.wmem_default used by other protocols.
992
993	It is usually lower than net.core.wmem_default.
994
995	Default: 16K
996
997	max: Maximal amount of memory allowed for automatically tuned
998	send buffers for TCP sockets. This value does not override
999	net.core.wmem_max.  Calling setsockopt() with SO_SNDBUF disables
1000	automatic tuning of that socket's send buffer size, in which case
1001	this value is ignored.
1002
1003	Default: between 64K and 4MB, depending on RAM size.
1004
1005tcp_notsent_lowat - UNSIGNED INTEGER
1006	A TCP socket can control the amount of unsent bytes in its write queue,
1007	thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
1008	reports POLLOUT events if the amount of unsent bytes is below a per
1009	socket value, and if the write queue is not full. sendmsg() will
1010	also not add new buffers if the limit is hit.
1011
1012	This global variable controls the amount of unsent data for
1013	sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
1014	to the global variable has immediate effect.
1015
1016	Default: UINT_MAX (0xFFFFFFFF)
1017
1018tcp_workaround_signed_windows - BOOLEAN
1019	If set, assume no receipt of a window scaling option means the
1020	remote TCP is broken and treats the window as a signed quantity.
1021	If unset, assume the remote TCP is not broken even if we do
1022	not receive a window scaling option from them.
1023
1024	Default: 0
1025
1026tcp_thin_linear_timeouts - BOOLEAN
1027	Enable dynamic triggering of linear timeouts for thin streams.
1028	If set, a check is performed upon retransmission by timeout to
1029	determine if the stream is thin (less than 4 packets in flight).
1030	As long as the stream is found to be thin, up to 6 linear
1031	timeouts may be performed before exponential backoff mode is
1032	initiated. This improves retransmission latency for
1033	non-aggressive thin streams, often found to be time-dependent.
1034	For more information on thin streams, see
1035	Documentation/networking/tcp-thin.rst
1036
1037	Default: 0
1038
1039tcp_limit_output_bytes - INTEGER
1040	Controls TCP Small Queue limit per tcp socket.
1041	TCP bulk sender tends to increase packets in flight until it
1042	gets losses notifications. With SNDBUF autotuning, this can
1043	result in a large amount of packets queued on the local machine
1044	(e.g.: qdiscs, CPU backlog, or device) hurting latency of other
1045	flows, for typical pfifo_fast qdiscs.  tcp_limit_output_bytes
1046	limits the number of bytes on qdisc or device to reduce artificial
1047	RTT/cwnd and reduce bufferbloat.
1048
1049	Default: 1048576 (16 * 65536)
1050
1051tcp_challenge_ack_limit - INTEGER
1052	Limits number of Challenge ACK sent per second, as recommended
1053	in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
1054	Note that this per netns rate limit can allow some side channel
1055	attacks and probably should not be enabled.
1056	TCP stack implements per TCP socket limits anyway.
1057	Default: INT_MAX (unlimited)
1058
1059tcp_ehash_entries - INTEGER
1060	Show the number of hash buckets for TCP sockets in the current
1061	networking namespace.
1062
1063	A negative value means the networking namespace does not own its
1064	hash buckets and shares the initial networking namespace's one.
1065
1066tcp_child_ehash_entries - INTEGER
1067	Control the number of hash buckets for TCP sockets in the child
1068	networking namespace, which must be set before clone() or unshare().
1069
1070	If the value is not 0, the kernel uses a value rounded up to 2^n
1071	as the actual hash bucket size.  0 is a special value, meaning
1072	the child networking namespace will share the initial networking
1073	namespace's hash buckets.
1074
1075	Note that the child will use the global one in case the kernel
1076	fails to allocate enough memory.  In addition, the global hash
1077	buckets are spread over available NUMA nodes, but the allocation
1078	of the child hash table depends on the current process's NUMA
1079	policy, which could result in performance differences.
1080
1081	Note also that the default value of tcp_max_tw_buckets and
1082	tcp_max_syn_backlog depend on the hash bucket size.
1083
1084	Possible values: 0, 2^n (n: 0 - 24 (16Mi))
1085
1086	Default: 0
1087
1088tcp_plb_enabled - BOOLEAN
1089	If set and the underlying congestion control (e.g. DCTCP) supports
1090	and enables PLB feature, TCP PLB (Protective Load Balancing) is
1091	enabled. PLB is described in the following paper:
1092	https://doi.org/10.1145/3544216.3544226. Based on PLB parameters,
1093	upon sensing sustained congestion, TCP triggers a change in
1094	flow label field for outgoing IPv6 packets. A change in flow label
1095	field potentially changes the path of outgoing packets for switches
1096	that use ECMP/WCMP for routing.
1097
1098	PLB changes socket txhash which results in a change in IPv6 Flow Label
1099	field, and currently no-op for IPv4 headers. It is possible
1100	to apply PLB for IPv4 with other network header fields (e.g. TCP
1101	or IPv4 options) or using encapsulation where outer header is used
1102	by switches to determine next hop. In either case, further host
1103	and switch side changes will be needed.
1104
1105	When set, PLB assumes that congestion signal (e.g. ECN) is made
1106	available and used by congestion control module to estimate a
1107	congestion measure (e.g. ce_ratio). PLB needs a congestion measure to
1108	make repathing decisions.
1109
1110	Default: FALSE
1111
1112tcp_plb_idle_rehash_rounds - INTEGER
1113	Number of consecutive congested rounds (RTT) seen after which
1114	a rehash can be performed, given there are no packets in flight.
1115	This is referred to as M in PLB paper:
1116	https://doi.org/10.1145/3544216.3544226.
1117
1118	Possible Values: 0 - 31
1119
1120	Default: 3
1121
1122tcp_plb_rehash_rounds - INTEGER
1123	Number of consecutive congested rounds (RTT) seen after which
1124	a forced rehash can be performed. Be careful when setting this
1125	parameter, as a small value increases the risk of retransmissions.
1126	This is referred to as N in PLB paper:
1127	https://doi.org/10.1145/3544216.3544226.
1128
1129	Possible Values: 0 - 31
1130
1131	Default: 12
1132
1133tcp_plb_suspend_rto_sec - INTEGER
1134	Time, in seconds, to suspend PLB in event of an RTO. In order to avoid
1135	having PLB repath onto a connectivity "black hole", after an RTO a TCP
1136	connection suspends PLB repathing for a random duration between 1x and
1137	2x of this parameter. Randomness is added to avoid concurrent rehashing
1138	of multiple TCP connections. This should be set corresponding to the
1139	amount of time it takes to repair a failed link.
1140
1141	Possible Values: 0 - 255
1142
1143	Default: 60
1144
1145tcp_plb_cong_thresh - INTEGER
1146	Fraction of packets marked with congestion over a round (RTT) to
1147	tag that round as congested. This is referred to as K in the PLB paper:
1148	https://doi.org/10.1145/3544216.3544226.
1149
1150	The 0-1 fraction range is mapped to 0-256 range to avoid floating
1151	point operations. For example, 128 means that if at least 50% of
1152	the packets in a round were marked as congested then the round
1153	will be tagged as congested.
1154
1155	Setting threshold to 0 means that PLB repaths every RTT regardless
1156	of congestion. This is not intended behavior for PLB and should be
1157	used only for experimentation purpose.
1158
1159	Possible Values: 0 - 256
1160
1161	Default: 128
1162
1163UDP variables
1164=============
1165
1166udp_l3mdev_accept - BOOLEAN
1167	Enabling this option allows a "global" bound socket to work
1168	across L3 master domains (e.g., VRFs) with packets capable of
1169	being received regardless of the L3 domain in which they
1170	originated. Only valid when the kernel was compiled with
1171	CONFIG_NET_L3_MASTER_DEV.
1172
1173	Default: 0 (disabled)
1174
1175udp_mem - vector of 3 INTEGERs: min, pressure, max
1176	Number of pages allowed for queueing by all UDP sockets.
1177
1178	min: Number of pages allowed for queueing by all UDP sockets.
1179
1180	pressure: This value was introduced to follow format of tcp_mem.
1181
1182	max: This value was introduced to follow format of tcp_mem.
1183
1184	Default is calculated at boot time from amount of available memory.
1185
1186udp_rmem_min - INTEGER
1187	Minimal size of receive buffer used by UDP sockets in moderation.
1188	Each UDP socket is able to use the size for receiving data, even if
1189	total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
1190
1191	Default: 4K
1192
1193udp_wmem_min - INTEGER
1194	UDP does not have tx memory accounting and this tunable has no effect.
1195
1196udp_hash_entries - INTEGER
1197	Show the number of hash buckets for UDP sockets in the current
1198	networking namespace.
1199
1200	A negative value means the networking namespace does not own its
1201	hash buckets and shares the initial networking namespace's one.
1202
1203udp_child_ehash_entries - INTEGER
1204	Control the number of hash buckets for UDP sockets in the child
1205	networking namespace, which must be set before clone() or unshare().
1206
1207	If the value is not 0, the kernel uses a value rounded up to 2^n
1208	as the actual hash bucket size.  0 is a special value, meaning
1209	the child networking namespace will share the initial networking
1210	namespace's hash buckets.
1211
1212	Note that the child will use the global one in case the kernel
1213	fails to allocate enough memory.  In addition, the global hash
1214	buckets are spread over available NUMA nodes, but the allocation
1215	of the child hash table depends on the current process's NUMA
1216	policy, which could result in performance differences.
1217
1218	Possible values: 0, 2^n (n: 7 (128) - 16 (64K))
1219
1220	Default: 0
1221
1222
1223RAW variables
1224=============
1225
1226raw_l3mdev_accept - BOOLEAN
1227	Enabling this option allows a "global" bound socket to work
1228	across L3 master domains (e.g., VRFs) with packets capable of
1229	being received regardless of the L3 domain in which they
1230	originated. Only valid when the kernel was compiled with
1231	CONFIG_NET_L3_MASTER_DEV.
1232
1233	Default: 1 (enabled)
1234
1235CIPSOv4 Variables
1236=================
1237
1238cipso_cache_enable - BOOLEAN
1239	If set, enable additions to and lookups from the CIPSO label mapping
1240	cache.  If unset, additions are ignored and lookups always result in a
1241	miss.  However, regardless of the setting the cache is still
1242	invalidated when required when means you can safely toggle this on and
1243	off and the cache will always be "safe".
1244
1245	Default: 1
1246
1247cipso_cache_bucket_size - INTEGER
1248	The CIPSO label cache consists of a fixed size hash table with each
1249	hash bucket containing a number of cache entries.  This variable limits
1250	the number of entries in each hash bucket; the larger the value is, the
1251	more CIPSO label mappings that can be cached.  When the number of
1252	entries in a given hash bucket reaches this limit adding new entries
1253	causes the oldest entry in the bucket to be removed to make room.
1254
1255	Default: 10
1256
1257cipso_rbm_optfmt - BOOLEAN
1258	Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
1259	the CIPSO draft specification (see Documentation/netlabel for details).
1260	This means that when set the CIPSO tag will be padded with empty
1261	categories in order to make the packet data 32-bit aligned.
1262
1263	Default: 0
1264
1265cipso_rbm_structvalid - BOOLEAN
1266	If set, do a very strict check of the CIPSO option when
1267	ip_options_compile() is called.  If unset, relax the checks done during
1268	ip_options_compile().  Either way is "safe" as errors are caught else
1269	where in the CIPSO processing code but setting this to 0 (False) should
1270	result in less work (i.e. it should be faster) but could cause problems
1271	with other implementations that require strict checking.
1272
1273	Default: 0
1274
1275IP Variables
1276============
1277
1278ip_local_port_range - 2 INTEGERS
1279	Defines the local port range that is used by TCP and UDP to
1280	choose the local port. The first number is the first, the
1281	second the last local port number.
1282	If possible, it is better these numbers have different parity
1283	(one even and one odd value).
1284	Must be greater than or equal to ip_unprivileged_port_start.
1285	The default values are 32768 and 60999 respectively.
1286
1287ip_local_reserved_ports - list of comma separated ranges
1288	Specify the ports which are reserved for known third-party
1289	applications. These ports will not be used by automatic port
1290	assignments (e.g. when calling connect() or bind() with port
1291	number 0). Explicit port allocation behavior is unchanged.
1292
1293	The format used for both input and output is a comma separated
1294	list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
1295	10). Writing to the file will clear all previously reserved
1296	ports and update the current list with the one given in the
1297	input.
1298
1299	Note that ip_local_port_range and ip_local_reserved_ports
1300	settings are independent and both are considered by the kernel
1301	when determining which ports are available for automatic port
1302	assignments.
1303
1304	You can reserve ports which are not in the current
1305	ip_local_port_range, e.g.::
1306
1307	    $ cat /proc/sys/net/ipv4/ip_local_port_range
1308	    32000	60999
1309	    $ cat /proc/sys/net/ipv4/ip_local_reserved_ports
1310	    8080,9148
1311
1312	although this is redundant. However such a setting is useful
1313	if later the port range is changed to a value that will
1314	include the reserved ports. Also keep in mind, that overlapping
1315	of these ranges may affect probability of selecting ephemeral
1316	ports which are right after block of reserved ports.
1317
1318	Default: Empty
1319
1320ip_unprivileged_port_start - INTEGER
1321	This is a per-namespace sysctl.  It defines the first
1322	unprivileged port in the network namespace.  Privileged ports
1323	require root or CAP_NET_BIND_SERVICE in order to bind to them.
1324	To disable all privileged ports, set this to 0.  They must not
1325	overlap with the ip_local_port_range.
1326
1327	Default: 1024
1328
1329ip_nonlocal_bind - BOOLEAN
1330	If set, allows processes to bind() to non-local IP addresses,
1331	which can be quite useful - but may break some applications.
1332
1333	Default: 0
1334
1335ip_autobind_reuse - BOOLEAN
1336	By default, bind() does not select the ports automatically even if
1337	the new socket and all sockets bound to the port have SO_REUSEADDR.
1338	ip_autobind_reuse allows bind() to reuse the port and this is useful
1339	when you use bind()+connect(), but may break some applications.
1340	The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
1341	option should only be set by experts.
1342	Default: 0
1343
1344ip_dynaddr - INTEGER
1345	If set non-zero, enables support for dynamic addresses.
1346	If set to a non-zero value larger than 1, a kernel log
1347	message will be printed when dynamic address rewriting
1348	occurs.
1349
1350	Default: 0
1351
1352ip_early_demux - BOOLEAN
1353	Optimize input packet processing down to one demux for
1354	certain kinds of local sockets.  Currently we only do this
1355	for established TCP and connected UDP sockets.
1356
1357	It may add an additional cost for pure routing workloads that
1358	reduces overall throughput, in such case you should disable it.
1359
1360	Default: 1
1361
1362ping_group_range - 2 INTEGERS
1363	Restrict ICMP_PROTO datagram sockets to users in the group range.
1364	The default is "1 0", meaning, that nobody (not even root) may
1365	create ping sockets.  Setting it to "100 100" would grant permissions
1366	to the single group. "0 4294967294" would enable it for the world, "100
1367	4294967294" would enable it for the users, but not daemons.
1368
1369tcp_early_demux - BOOLEAN
1370	Enable early demux for established TCP sockets.
1371
1372	Default: 1
1373
1374udp_early_demux - BOOLEAN
1375	Enable early demux for connected UDP sockets. Disable this if
1376	your system could experience more unconnected load.
1377
1378	Default: 1
1379
1380icmp_echo_ignore_all - BOOLEAN
1381	If set non-zero, then the kernel will ignore all ICMP ECHO
1382	requests sent to it.
1383
1384	Default: 0
1385
1386icmp_echo_enable_probe - BOOLEAN
1387        If set to one, then the kernel will respond to RFC 8335 PROBE
1388        requests sent to it.
1389
1390        Default: 0
1391
1392icmp_echo_ignore_broadcasts - BOOLEAN
1393	If set non-zero, then the kernel will ignore all ICMP ECHO and
1394	TIMESTAMP requests sent to it via broadcast/multicast.
1395
1396	Default: 1
1397
1398icmp_ratelimit - INTEGER
1399	Limit the maximal rates for sending ICMP packets whose type matches
1400	icmp_ratemask (see below) to specific targets.
1401	0 to disable any limiting,
1402	otherwise the minimal space between responses in milliseconds.
1403	Note that another sysctl, icmp_msgs_per_sec limits the number
1404	of ICMP packets	sent on all targets.
1405
1406	Default: 1000
1407
1408icmp_msgs_per_sec - INTEGER
1409	Limit maximal number of ICMP packets sent per second from this host.
1410	Only messages whose type matches icmp_ratemask (see below) are
1411	controlled by this limit. For security reasons, the precise count
1412	of messages per second is randomized.
1413
1414	Default: 1000
1415
1416icmp_msgs_burst - INTEGER
1417	icmp_msgs_per_sec controls number of ICMP packets sent per second,
1418	while icmp_msgs_burst controls the burst size of these packets.
1419	For security reasons, the precise burst size is randomized.
1420
1421	Default: 50
1422
1423icmp_ratemask - INTEGER
1424	Mask made of ICMP types for which rates are being limited.
1425
1426	Significant bits: IHGFEDCBA9876543210
1427
1428	Default mask:     0000001100000011000 (6168)
1429
1430	Bit definitions (see include/linux/icmp.h):
1431
1432		= =========================
1433		0 Echo Reply
1434		3 Destination Unreachable [1]_
1435		4 Source Quench [1]_
1436		5 Redirect
1437		8 Echo Request
1438		B Time Exceeded [1]_
1439		C Parameter Problem [1]_
1440		D Timestamp Request
1441		E Timestamp Reply
1442		F Info Request
1443		G Info Reply
1444		H Address Mask Request
1445		I Address Mask Reply
1446		= =========================
1447
1448	.. [1] These are rate limited by default (see default mask above)
1449
1450icmp_ignore_bogus_error_responses - BOOLEAN
1451	Some routers violate RFC1122 by sending bogus responses to broadcast
1452	frames.  Such violations are normally logged via a kernel warning.
1453	If this is set to TRUE, the kernel will not give such warnings, which
1454	will avoid log file clutter.
1455
1456	Default: 1
1457
1458icmp_errors_use_inbound_ifaddr - BOOLEAN
1459
1460	If zero, icmp error messages are sent with the primary address of
1461	the exiting interface.
1462
1463	If non-zero, the message will be sent with the primary address of
1464	the interface that received the packet that caused the icmp error.
1465	This is the behaviour many network administrators will expect from
1466	a router. And it can make debugging complicated network layouts
1467	much easier.
1468
1469	Note that if no primary address exists for the interface selected,
1470	then the primary address of the first non-loopback interface that
1471	has one will be used regardless of this setting.
1472
1473	Default: 0
1474
1475igmp_max_memberships - INTEGER
1476	Change the maximum number of multicast groups we can subscribe to.
1477	Default: 20
1478
1479	Theoretical maximum value is bounded by having to send a membership
1480	report in a single datagram (i.e. the report can't span multiple
1481	datagrams, or risk confusing the switch and leaving groups you don't
1482	intend to).
1483
1484	The number of supported groups 'M' is bounded by the number of group
1485	report entries you can fit into a single datagram of 65535 bytes.
1486
1487	M = 65536-sizeof (ip header)/(sizeof(Group record))
1488
1489	Group records are variable length, with a minimum of 12 bytes.
1490	So net.ipv4.igmp_max_memberships should not be set higher than:
1491
1492	(65536-24) / 12 = 5459
1493
1494	The value 5459 assumes no IP header options, so in practice
1495	this number may be lower.
1496
1497igmp_max_msf - INTEGER
1498	Maximum number of addresses allowed in the source filter list for a
1499	multicast group.
1500
1501	Default: 10
1502
1503igmp_qrv - INTEGER
1504	Controls the IGMP query robustness variable (see RFC2236 8.1).
1505
1506	Default: 2 (as specified by RFC2236 8.1)
1507
1508	Minimum: 1 (as specified by RFC6636 4.5)
1509
1510force_igmp_version - INTEGER
1511	- 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
1512	  allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
1513	  Present timer expires.
1514	- 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
1515	  receive IGMPv2/v3 query.
1516	- 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
1517	  IGMPv1 query message. Will reply report if receive IGMPv3 query.
1518	- 3 - Enforce to use IGMP version 3. The same react with default 0.
1519
1520	.. note::
1521
1522	   this is not the same with force_mld_version because IGMPv3 RFC3376
1523	   Security Considerations does not have clear description that we could
1524	   ignore other version messages completely as MLDv2 RFC3810. So make
1525	   this value as default 0 is recommended.
1526
1527``conf/interface/*``
1528	changes special settings per interface (where
1529	interface" is the name of your network interface)
1530
1531``conf/all/*``
1532	  is special, changes the settings for all interfaces
1533
1534log_martians - BOOLEAN
1535	Log packets with impossible addresses to kernel log.
1536	log_martians for the interface will be enabled if at least one of
1537	conf/{all,interface}/log_martians is set to TRUE,
1538	it will be disabled otherwise
1539
1540accept_redirects - BOOLEAN
1541	Accept ICMP redirect messages.
1542	accept_redirects for the interface will be enabled if:
1543
1544	- both conf/{all,interface}/accept_redirects are TRUE in the case
1545	  forwarding for the interface is enabled
1546
1547	or
1548
1549	- at least one of conf/{all,interface}/accept_redirects is TRUE in the
1550	  case forwarding for the interface is disabled
1551
1552	accept_redirects for the interface will be disabled otherwise
1553
1554	default:
1555
1556		- TRUE (host)
1557		- FALSE (router)
1558
1559forwarding - BOOLEAN
1560	Enable IP forwarding on this interface.  This controls whether packets
1561	received _on_ this interface can be forwarded.
1562
1563mc_forwarding - BOOLEAN
1564	Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1565	and a multicast routing daemon is required.
1566	conf/all/mc_forwarding must also be set to TRUE to enable multicast
1567	routing	for the interface
1568
1569medium_id - INTEGER
1570	Integer value used to differentiate the devices by the medium they
1571	are attached to. Two devices can have different id values when
1572	the broadcast packets are received only on one of them.
1573	The default value 0 means that the device is the only interface
1574	to its medium, value of -1 means that medium is not known.
1575
1576	Currently, it is used to change the proxy_arp behavior:
1577	the proxy_arp feature is enabled for packets forwarded between
1578	two devices attached to different media.
1579
1580proxy_arp - BOOLEAN
1581	Do proxy arp.
1582
1583	proxy_arp for the interface will be enabled if at least one of
1584	conf/{all,interface}/proxy_arp is set to TRUE,
1585	it will be disabled otherwise
1586
1587proxy_arp_pvlan - BOOLEAN
1588	Private VLAN proxy arp.
1589
1590	Basically allow proxy arp replies back to the same interface
1591	(from which the ARP request/solicitation was received).
1592
1593	This is done to support (ethernet) switch features, like RFC
1594	3069, where the individual ports are NOT allowed to
1595	communicate with each other, but they are allowed to talk to
1596	the upstream router.  As described in RFC 3069, it is possible
1597	to allow these hosts to communicate through the upstream
1598	router by proxy_arp'ing. Don't need to be used together with
1599	proxy_arp.
1600
1601	This technology is known by different names:
1602
1603	  In RFC 3069 it is called VLAN Aggregation.
1604	  Cisco and Allied Telesyn call it Private VLAN.
1605	  Hewlett-Packard call it Source-Port filtering or port-isolation.
1606	  Ericsson call it MAC-Forced Forwarding (RFC Draft).
1607
1608proxy_delay - INTEGER
1609	Delay proxy response.
1610
1611	Delay response to a neighbor solicitation when proxy_arp
1612	or proxy_ndp is enabled. A random value between [0, proxy_delay)
1613	will be chosen, setting to zero means reply with no delay.
1614	Value in jiffies. Defaults to 80.
1615
1616shared_media - BOOLEAN
1617	Send(router) or accept(host) RFC1620 shared media redirects.
1618	Overrides secure_redirects.
1619
1620	shared_media for the interface will be enabled if at least one of
1621	conf/{all,interface}/shared_media is set to TRUE,
1622	it will be disabled otherwise
1623
1624	default TRUE
1625
1626secure_redirects - BOOLEAN
1627	Accept ICMP redirect messages only to gateways listed in the
1628	interface's current gateway list. Even if disabled, RFC1122 redirect
1629	rules still apply.
1630
1631	Overridden by shared_media.
1632
1633	secure_redirects for the interface will be enabled if at least one of
1634	conf/{all,interface}/secure_redirects is set to TRUE,
1635	it will be disabled otherwise
1636
1637	default TRUE
1638
1639send_redirects - BOOLEAN
1640	Send redirects, if router.
1641
1642	send_redirects for the interface will be enabled if at least one of
1643	conf/{all,interface}/send_redirects is set to TRUE,
1644	it will be disabled otherwise
1645
1646	Default: TRUE
1647
1648bootp_relay - BOOLEAN
1649	Accept packets with source address 0.b.c.d destined
1650	not to this host as local ones. It is supposed, that
1651	BOOTP relay daemon will catch and forward such packets.
1652	conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1653	for the interface
1654
1655	default FALSE
1656
1657	Not Implemented Yet.
1658
1659accept_source_route - BOOLEAN
1660	Accept packets with SRR option.
1661	conf/all/accept_source_route must also be set to TRUE to accept packets
1662	with SRR option on the interface
1663
1664	default
1665
1666		- TRUE (router)
1667		- FALSE (host)
1668
1669accept_local - BOOLEAN
1670	Accept packets with local source addresses. In combination with
1671	suitable routing, this can be used to direct packets between two
1672	local interfaces over the wire and have them accepted properly.
1673	default FALSE
1674
1675route_localnet - BOOLEAN
1676	Do not consider loopback addresses as martian source or destination
1677	while routing. This enables the use of 127/8 for local routing purposes.
1678
1679	default FALSE
1680
1681rp_filter - INTEGER
1682	- 0 - No source validation.
1683	- 1 - Strict mode as defined in RFC3704 Strict Reverse Path
1684	  Each incoming packet is tested against the FIB and if the interface
1685	  is not the best reverse path the packet check will fail.
1686	  By default failed packets are discarded.
1687	- 2 - Loose mode as defined in RFC3704 Loose Reverse Path
1688	  Each incoming packet's source address is also tested against the FIB
1689	  and if the source address is not reachable via any interface
1690	  the packet check will fail.
1691
1692	Current recommended practice in RFC3704 is to enable strict mode
1693	to prevent IP spoofing from DDos attacks. If using asymmetric routing
1694	or other complicated routing, then loose mode is recommended.
1695
1696	The max value from conf/{all,interface}/rp_filter is used
1697	when doing source validation on the {interface}.
1698
1699	Default value is 0. Note that some distributions enable it
1700	in startup scripts.
1701
1702src_valid_mark - BOOLEAN
1703	- 0 - The fwmark of the packet is not included in reverse path
1704	  route lookup.  This allows for asymmetric routing configurations
1705	  utilizing the fwmark in only one direction, e.g., transparent
1706	  proxying.
1707
1708	- 1 - The fwmark of the packet is included in reverse path route
1709	  lookup.  This permits rp_filter to function when the fwmark is
1710	  used for routing traffic in both directions.
1711
1712	This setting also affects the utilization of fmwark when
1713	performing source address selection for ICMP replies, or
1714	determining addresses stored for the IPOPT_TS_TSANDADDR and
1715	IPOPT_RR IP options.
1716
1717	The max value from conf/{all,interface}/src_valid_mark is used.
1718
1719	Default value is 0.
1720
1721arp_filter - BOOLEAN
1722	- 1 - Allows you to have multiple network interfaces on the same
1723	  subnet, and have the ARPs for each interface be answered
1724	  based on whether or not the kernel would route a packet from
1725	  the ARP'd IP out that interface (therefore you must use source
1726	  based routing for this to work). In other words it allows control
1727	  of which cards (usually 1) will respond to an arp request.
1728
1729	- 0 - (default) The kernel can respond to arp requests with addresses
1730	  from other interfaces. This may seem wrong but it usually makes
1731	  sense, because it increases the chance of successful communication.
1732	  IP addresses are owned by the complete host on Linux, not by
1733	  particular interfaces. Only for more complex setups like load-
1734	  balancing, does this behaviour cause problems.
1735
1736	arp_filter for the interface will be enabled if at least one of
1737	conf/{all,interface}/arp_filter is set to TRUE,
1738	it will be disabled otherwise
1739
1740arp_announce - INTEGER
1741	Define different restriction levels for announcing the local
1742	source IP address from IP packets in ARP requests sent on
1743	interface:
1744
1745	- 0 - (default) Use any local address, configured on any interface
1746	- 1 - Try to avoid local addresses that are not in the target's
1747	  subnet for this interface. This mode is useful when target
1748	  hosts reachable via this interface require the source IP
1749	  address in ARP requests to be part of their logical network
1750	  configured on the receiving interface. When we generate the
1751	  request we will check all our subnets that include the
1752	  target IP and will preserve the source address if it is from
1753	  such subnet. If there is no such subnet we select source
1754	  address according to the rules for level 2.
1755	- 2 - Always use the best local address for this target.
1756	  In this mode we ignore the source address in the IP packet
1757	  and try to select local address that we prefer for talks with
1758	  the target host. Such local address is selected by looking
1759	  for primary IP addresses on all our subnets on the outgoing
1760	  interface that include the target IP address. If no suitable
1761	  local address is found we select the first local address
1762	  we have on the outgoing interface or on all other interfaces,
1763	  with the hope we will receive reply for our request and
1764	  even sometimes no matter the source IP address we announce.
1765
1766	The max value from conf/{all,interface}/arp_announce is used.
1767
1768	Increasing the restriction level gives more chance for
1769	receiving answer from the resolved target while decreasing
1770	the level announces more valid sender's information.
1771
1772arp_ignore - INTEGER
1773	Define different modes for sending replies in response to
1774	received ARP requests that resolve local target IP addresses:
1775
1776	- 0 - (default): reply for any local target IP address, configured
1777	  on any interface
1778	- 1 - reply only if the target IP address is local address
1779	  configured on the incoming interface
1780	- 2 - reply only if the target IP address is local address
1781	  configured on the incoming interface and both with the
1782	  sender's IP address are part from same subnet on this interface
1783	- 3 - do not reply for local addresses configured with scope host,
1784	  only resolutions for global and link addresses are replied
1785	- 4-7 - reserved
1786	- 8 - do not reply for all local addresses
1787
1788	The max value from conf/{all,interface}/arp_ignore is used
1789	when ARP request is received on the {interface}
1790
1791arp_notify - BOOLEAN
1792	Define mode for notification of address and device changes.
1793
1794	 ==  ==========================================================
1795	  0  (default): do nothing
1796	  1  Generate gratuitous arp requests when device is brought up
1797	     or hardware address changes.
1798	 ==  ==========================================================
1799
1800arp_accept - INTEGER
1801	Define behavior for accepting gratuitous ARP (garp) frames from devices
1802	that are not already present in the ARP table:
1803
1804	- 0 - don't create new entries in the ARP table
1805	- 1 - create new entries in the ARP table
1806	- 2 - create new entries only if the source IP address is in the same
1807	  subnet as an address configured on the interface that received the
1808	  garp message.
1809
1810	Both replies and requests type gratuitous arp will trigger the
1811	ARP table to be updated, if this setting is on.
1812
1813	If the ARP table already contains the IP address of the
1814	gratuitous arp frame, the arp table will be updated regardless
1815	if this setting is on or off.
1816
1817arp_evict_nocarrier - BOOLEAN
1818	Clears the ARP cache on NOCARRIER events. This option is important for
1819	wireless devices where the ARP cache should not be cleared when roaming
1820	between access points on the same network. In most cases this should
1821	remain as the default (1).
1822
1823	- 1 - (default): Clear the ARP cache on NOCARRIER events
1824	- 0 - Do not clear ARP cache on NOCARRIER events
1825
1826mcast_solicit - INTEGER
1827	The maximum number of multicast probes in INCOMPLETE state,
1828	when the associated hardware address is unknown.  Defaults
1829	to 3.
1830
1831ucast_solicit - INTEGER
1832	The maximum number of unicast probes in PROBE state, when
1833	the hardware address is being reconfirmed.  Defaults to 3.
1834
1835app_solicit - INTEGER
1836	The maximum number of probes to send to the user space ARP daemon
1837	via netlink before dropping back to multicast probes (see
1838	mcast_resolicit).  Defaults to 0.
1839
1840mcast_resolicit - INTEGER
1841	The maximum number of multicast probes after unicast and
1842	app probes in PROBE state.  Defaults to 0.
1843
1844disable_policy - BOOLEAN
1845	Disable IPSEC policy (SPD) for this interface
1846
1847disable_xfrm - BOOLEAN
1848	Disable IPSEC encryption on this interface, whatever the policy
1849
1850igmpv2_unsolicited_report_interval - INTEGER
1851	The interval in milliseconds in which the next unsolicited
1852	IGMPv1 or IGMPv2 report retransmit will take place.
1853
1854	Default: 10000 (10 seconds)
1855
1856igmpv3_unsolicited_report_interval - INTEGER
1857	The interval in milliseconds in which the next unsolicited
1858	IGMPv3 report retransmit will take place.
1859
1860	Default: 1000 (1 seconds)
1861
1862ignore_routes_with_linkdown - BOOLEAN
1863        Ignore routes whose link is down when performing a FIB lookup.
1864
1865promote_secondaries - BOOLEAN
1866	When a primary IP address is removed from this interface
1867	promote a corresponding secondary IP address instead of
1868	removing all the corresponding secondary IP addresses.
1869
1870drop_unicast_in_l2_multicast - BOOLEAN
1871	Drop any unicast IP packets that are received in link-layer
1872	multicast (or broadcast) frames.
1873
1874	This behavior (for multicast) is actually a SHOULD in RFC
1875	1122, but is disabled by default for compatibility reasons.
1876
1877	Default: off (0)
1878
1879drop_gratuitous_arp - BOOLEAN
1880	Drop all gratuitous ARP frames, for example if there's a known
1881	good ARP proxy on the network and such frames need not be used
1882	(or in the case of 802.11, must not be used to prevent attacks.)
1883
1884	Default: off (0)
1885
1886
1887tag - INTEGER
1888	Allows you to write a number, which can be used as required.
1889
1890	Default value is 0.
1891
1892xfrm4_gc_thresh - INTEGER
1893	(Obsolete since linux-4.14)
1894	The threshold at which we will start garbage collecting for IPv4
1895	destination cache entries.  At twice this value the system will
1896	refuse new allocations.
1897
1898igmp_link_local_mcast_reports - BOOLEAN
1899	Enable IGMP reports for link local multicast groups in the
1900	224.0.0.X range.
1901
1902	Default TRUE
1903
1904Alexey Kuznetsov.
1905kuznet@ms2.inr.ac.ru
1906
1907Updated by:
1908
1909- Andi Kleen
1910  ak@muc.de
1911- Nicolas Delon
1912  delon.nicolas@wanadoo.fr
1913
1914
1915
1916
1917/proc/sys/net/ipv6/* Variables
1918==============================
1919
1920IPv6 has no global variables such as tcp_*.  tcp_* settings under ipv4/ also
1921apply to IPv6 [XXX?].
1922
1923bindv6only - BOOLEAN
1924	Default value for IPV6_V6ONLY socket option,
1925	which restricts use of the IPv6 socket to IPv6 communication
1926	only.
1927
1928		- TRUE: disable IPv4-mapped address feature
1929		- FALSE: enable IPv4-mapped address feature
1930
1931	Default: FALSE (as specified in RFC3493)
1932
1933flowlabel_consistency - BOOLEAN
1934	Protect the consistency (and unicity) of flow label.
1935	You have to disable it to use IPV6_FL_F_REFLECT flag on the
1936	flow label manager.
1937
1938	- TRUE: enabled
1939	- FALSE: disabled
1940
1941	Default: TRUE
1942
1943auto_flowlabels - INTEGER
1944	Automatically generate flow labels based on a flow hash of the
1945	packet. This allows intermediate devices, such as routers, to
1946	identify packet flows for mechanisms like Equal Cost Multipath
1947	Routing (see RFC 6438).
1948
1949	=  ===========================================================
1950	0  automatic flow labels are completely disabled
1951	1  automatic flow labels are enabled by default, they can be
1952	   disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
1953	   socket option
1954	2  automatic flow labels are allowed, they may be enabled on a
1955	   per socket basis using the IPV6_AUTOFLOWLABEL socket option
1956	3  automatic flow labels are enabled and enforced, they cannot
1957	   be disabled by the socket option
1958	=  ===========================================================
1959
1960	Default: 1
1961
1962flowlabel_state_ranges - BOOLEAN
1963	Split the flow label number space into two ranges. 0-0x7FFFF is
1964	reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
1965	is reserved for stateless flow labels as described in RFC6437.
1966
1967	- TRUE: enabled
1968	- FALSE: disabled
1969
1970	Default: true
1971
1972flowlabel_reflect - INTEGER
1973	Control flow label reflection. Needed for Path MTU
1974	Discovery to work with Equal Cost Multipath Routing in anycast
1975	environments. See RFC 7690 and:
1976	https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
1977
1978	This is a bitmask.
1979
1980	- 1: enabled for established flows
1981
1982	  Note that this prevents automatic flowlabel changes, as done
1983	  in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
1984	  and "tcp: Change txhash on every SYN and RTO retransmit"
1985
1986	- 2: enabled for TCP RESET packets (no active listener)
1987	  If set, a RST packet sent in response to a SYN packet on a closed
1988	  port will reflect the incoming flow label.
1989
1990	- 4: enabled for ICMPv6 echo reply messages.
1991
1992	Default: 0
1993
1994fib_multipath_hash_policy - INTEGER
1995	Controls which hash policy to use for multipath routes.
1996
1997	Default: 0 (Layer 3)
1998
1999	Possible values:
2000
2001	- 0 - Layer 3 (source and destination addresses plus flow label)
2002	- 1 - Layer 4 (standard 5-tuple)
2003	- 2 - Layer 3 or inner Layer 3 if present
2004	- 3 - Custom multipath hash. Fields used for multipath hash calculation
2005	  are determined by fib_multipath_hash_fields sysctl
2006
2007fib_multipath_hash_fields - UNSIGNED INTEGER
2008	When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
2009	fields used for multipath hash calculation are determined by this
2010	sysctl.
2011
2012	This value is a bitmask which enables various fields for multipath hash
2013	calculation.
2014
2015	Possible fields are:
2016
2017	====== ============================
2018	0x0001 Source IP address
2019	0x0002 Destination IP address
2020	0x0004 IP protocol
2021	0x0008 Flow Label
2022	0x0010 Source port
2023	0x0020 Destination port
2024	0x0040 Inner source IP address
2025	0x0080 Inner destination IP address
2026	0x0100 Inner IP protocol
2027	0x0200 Inner Flow Label
2028	0x0400 Inner source port
2029	0x0800 Inner destination port
2030	====== ============================
2031
2032	Default: 0x0007 (source IP, destination IP and IP protocol)
2033
2034anycast_src_echo_reply - BOOLEAN
2035	Controls the use of anycast addresses as source addresses for ICMPv6
2036	echo reply
2037
2038	- TRUE:  enabled
2039	- FALSE: disabled
2040
2041	Default: FALSE
2042
2043idgen_delay - INTEGER
2044	Controls the delay in seconds after which time to retry
2045	privacy stable address generation if a DAD conflict is
2046	detected.
2047
2048	Default: 1 (as specified in RFC7217)
2049
2050idgen_retries - INTEGER
2051	Controls the number of retries to generate a stable privacy
2052	address if a DAD conflict is detected.
2053
2054	Default: 3 (as specified in RFC7217)
2055
2056mld_qrv - INTEGER
2057	Controls the MLD query robustness variable (see RFC3810 9.1).
2058
2059	Default: 2 (as specified by RFC3810 9.1)
2060
2061	Minimum: 1 (as specified by RFC6636 4.5)
2062
2063max_dst_opts_number - INTEGER
2064	Maximum number of non-padding TLVs allowed in a Destination
2065	options extension header. If this value is less than zero
2066	then unknown options are disallowed and the number of known
2067	TLVs allowed is the absolute value of this number.
2068
2069	Default: 8
2070
2071max_hbh_opts_number - INTEGER
2072	Maximum number of non-padding TLVs allowed in a Hop-by-Hop
2073	options extension header. If this value is less than zero
2074	then unknown options are disallowed and the number of known
2075	TLVs allowed is the absolute value of this number.
2076
2077	Default: 8
2078
2079max_dst_opts_length - INTEGER
2080	Maximum length allowed for a Destination options extension
2081	header.
2082
2083	Default: INT_MAX (unlimited)
2084
2085max_hbh_length - INTEGER
2086	Maximum length allowed for a Hop-by-Hop options extension
2087	header.
2088
2089	Default: INT_MAX (unlimited)
2090
2091skip_notify_on_dev_down - BOOLEAN
2092	Controls whether an RTM_DELROUTE message is generated for routes
2093	removed when a device is taken down or deleted. IPv4 does not
2094	generate this message; IPv6 does by default. Setting this sysctl
2095	to true skips the message, making IPv4 and IPv6 on par in relying
2096	on userspace caches to track link events and evict routes.
2097
2098	Default: false (generate message)
2099
2100nexthop_compat_mode - BOOLEAN
2101	New nexthop API provides a means for managing nexthops independent of
2102	prefixes. Backwards compatibility with old route format is enabled by
2103	default which means route dumps and notifications contain the new
2104	nexthop attribute but also the full, expanded nexthop definition.
2105	Further, updates or deletes of a nexthop configuration generate route
2106	notifications for each fib entry using the nexthop. Once a system
2107	understands the new API, this sysctl can be disabled to achieve full
2108	performance benefits of the new API by disabling the nexthop expansion
2109	and extraneous notifications.
2110	Default: true (backward compat mode)
2111
2112fib_notify_on_flag_change - INTEGER
2113        Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
2114        RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
2115
2116        After installing a route to the kernel, user space receives an
2117        acknowledgment, which means the route was installed in the kernel,
2118        but not necessarily in hardware.
2119        It is also possible for a route already installed in hardware to change
2120        its action and therefore its flags. For example, a host route that is
2121        trapping packets can be "promoted" to perform decapsulation following
2122        the installation of an IPinIP/VXLAN tunnel.
2123        The notifications will indicate to user-space the state of the route.
2124
2125        Default: 0 (Do not emit notifications.)
2126
2127        Possible values:
2128
2129        - 0 - Do not emit notifications.
2130        - 1 - Emit notifications.
2131        - 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
2132
2133ioam6_id - INTEGER
2134        Define the IOAM id of this node. Uses only 24 bits out of 32 in total.
2135
2136        Min: 0
2137        Max: 0xFFFFFF
2138
2139        Default: 0xFFFFFF
2140
2141ioam6_id_wide - LONG INTEGER
2142        Define the wide IOAM id of this node. Uses only 56 bits out of 64 in
2143        total. Can be different from ioam6_id.
2144
2145        Min: 0
2146        Max: 0xFFFFFFFFFFFFFF
2147
2148        Default: 0xFFFFFFFFFFFFFF
2149
2150IPv6 Fragmentation:
2151
2152ip6frag_high_thresh - INTEGER
2153	Maximum memory used to reassemble IPv6 fragments. When
2154	ip6frag_high_thresh bytes of memory is allocated for this purpose,
2155	the fragment handler will toss packets until ip6frag_low_thresh
2156	is reached.
2157
2158ip6frag_low_thresh - INTEGER
2159	See ip6frag_high_thresh
2160
2161ip6frag_time - INTEGER
2162	Time in seconds to keep an IPv6 fragment in memory.
2163
2164``conf/default/*``:
2165	Change the interface-specific default settings.
2166
2167	These settings would be used during creating new interfaces.
2168
2169
2170``conf/all/*``:
2171	Change all the interface-specific settings.
2172
2173	[XXX:  Other special features than forwarding?]
2174
2175conf/all/disable_ipv6 - BOOLEAN
2176	Changing this value is same as changing ``conf/default/disable_ipv6``
2177	setting and also all per-interface ``disable_ipv6`` settings to the same
2178	value.
2179
2180	Reading this value does not have any particular meaning. It does not say
2181	whether IPv6 support is enabled or disabled. Returned value can be 1
2182	also in the case when some interface has ``disable_ipv6`` set to 0 and
2183	has configured IPv6 addresses.
2184
2185conf/all/forwarding - BOOLEAN
2186	Enable global IPv6 forwarding between all interfaces.
2187
2188	IPv4 and IPv6 work differently here; e.g. netfilter must be used
2189	to control which interfaces may forward packets and which not.
2190
2191	This also sets all interfaces' Host/Router setting
2192	'forwarding' to the specified value.  See below for details.
2193
2194	This referred to as global forwarding.
2195
2196proxy_ndp - BOOLEAN
2197	Do proxy ndp.
2198
2199fwmark_reflect - BOOLEAN
2200	Controls the fwmark of kernel-generated IPv6 reply packets that are not
2201	associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
2202	If unset, these packets have a fwmark of zero. If set, they have the
2203	fwmark of the packet they are replying to.
2204
2205	Default: 0
2206
2207``conf/interface/*``:
2208	Change special settings per interface.
2209
2210	The functional behaviour for certain settings is different
2211	depending on whether local forwarding is enabled or not.
2212
2213accept_ra - INTEGER
2214	Accept Router Advertisements; autoconfigure using them.
2215
2216	It also determines whether or not to transmit Router
2217	Solicitations. If and only if the functional setting is to
2218	accept Router Advertisements, Router Solicitations will be
2219	transmitted.
2220
2221	Possible values are:
2222
2223		==  ===========================================================
2224		 0  Do not accept Router Advertisements.
2225		 1  Accept Router Advertisements if forwarding is disabled.
2226		 2  Overrule forwarding behaviour. Accept Router Advertisements
2227		    even if forwarding is enabled.
2228		==  ===========================================================
2229
2230	Functional default:
2231
2232		- enabled if local forwarding is disabled.
2233		- disabled if local forwarding is enabled.
2234
2235accept_ra_defrtr - BOOLEAN
2236	Learn default router in Router Advertisement.
2237
2238	Functional default:
2239
2240		- enabled if accept_ra is enabled.
2241		- disabled if accept_ra is disabled.
2242
2243ra_defrtr_metric - UNSIGNED INTEGER
2244	Route metric for default route learned in Router Advertisement. This value
2245	will be assigned as metric for the default route learned via IPv6 Router
2246	Advertisement. Takes affect only if accept_ra_defrtr is enabled.
2247
2248	Possible values:
2249		1 to 0xFFFFFFFF
2250
2251		Default: IP6_RT_PRIO_USER i.e. 1024.
2252
2253accept_ra_from_local - BOOLEAN
2254	Accept RA with source-address that is found on local machine
2255	if the RA is otherwise proper and able to be accepted.
2256
2257	Default is to NOT accept these as it may be an un-intended
2258	network loop.
2259
2260	Functional default:
2261
2262	   - enabled if accept_ra_from_local is enabled
2263	     on a specific interface.
2264	   - disabled if accept_ra_from_local is disabled
2265	     on a specific interface.
2266
2267accept_ra_min_hop_limit - INTEGER
2268	Minimum hop limit Information in Router Advertisement.
2269
2270	Hop limit Information in Router Advertisement less than this
2271	variable shall be ignored.
2272
2273	Default: 1
2274
2275accept_ra_pinfo - BOOLEAN
2276	Learn Prefix Information in Router Advertisement.
2277
2278	Functional default:
2279
2280		- enabled if accept_ra is enabled.
2281		- disabled if accept_ra is disabled.
2282
2283accept_ra_rt_info_min_plen - INTEGER
2284	Minimum prefix length of Route Information in RA.
2285
2286	Route Information w/ prefix smaller than this variable shall
2287	be ignored.
2288
2289	Functional default:
2290
2291		* 0 if accept_ra_rtr_pref is enabled.
2292		* -1 if accept_ra_rtr_pref is disabled.
2293
2294accept_ra_rt_info_max_plen - INTEGER
2295	Maximum prefix length of Route Information in RA.
2296
2297	Route Information w/ prefix larger than this variable shall
2298	be ignored.
2299
2300	Functional default:
2301
2302		* 0 if accept_ra_rtr_pref is enabled.
2303		* -1 if accept_ra_rtr_pref is disabled.
2304
2305accept_ra_rtr_pref - BOOLEAN
2306	Accept Router Preference in RA.
2307
2308	Functional default:
2309
2310		- enabled if accept_ra is enabled.
2311		- disabled if accept_ra is disabled.
2312
2313accept_ra_mtu - BOOLEAN
2314	Apply the MTU value specified in RA option 5 (RFC4861). If
2315	disabled, the MTU specified in the RA will be ignored.
2316
2317	Functional default:
2318
2319		- enabled if accept_ra is enabled.
2320		- disabled if accept_ra is disabled.
2321
2322accept_redirects - BOOLEAN
2323	Accept Redirects.
2324
2325	Functional default:
2326
2327		- enabled if local forwarding is disabled.
2328		- disabled if local forwarding is enabled.
2329
2330accept_source_route - INTEGER
2331	Accept source routing (routing extension header).
2332
2333	- >= 0: Accept only routing header type 2.
2334	- < 0: Do not accept routing header.
2335
2336	Default: 0
2337
2338autoconf - BOOLEAN
2339	Autoconfigure addresses using Prefix Information in Router
2340	Advertisements.
2341
2342	Functional default:
2343
2344		- enabled if accept_ra_pinfo is enabled.
2345		- disabled if accept_ra_pinfo is disabled.
2346
2347dad_transmits - INTEGER
2348	The amount of Duplicate Address Detection probes to send.
2349
2350	Default: 1
2351
2352forwarding - INTEGER
2353	Configure interface-specific Host/Router behaviour.
2354
2355	.. note::
2356
2357	   It is recommended to have the same setting on all
2358	   interfaces; mixed router/host scenarios are rather uncommon.
2359
2360	Possible values are:
2361
2362		- 0 Forwarding disabled
2363		- 1 Forwarding enabled
2364
2365	**FALSE (0)**:
2366
2367	By default, Host behaviour is assumed.  This means:
2368
2369	1. IsRouter flag is not set in Neighbour Advertisements.
2370	2. If accept_ra is TRUE (default), transmit Router
2371	   Solicitations.
2372	3. If accept_ra is TRUE (default), accept Router
2373	   Advertisements (and do autoconfiguration).
2374	4. If accept_redirects is TRUE (default), accept Redirects.
2375
2376	**TRUE (1)**:
2377
2378	If local forwarding is enabled, Router behaviour is assumed.
2379	This means exactly the reverse from the above:
2380
2381	1. IsRouter flag is set in Neighbour Advertisements.
2382	2. Router Solicitations are not sent unless accept_ra is 2.
2383	3. Router Advertisements are ignored unless accept_ra is 2.
2384	4. Redirects are ignored.
2385
2386	Default: 0 (disabled) if global forwarding is disabled (default),
2387	otherwise 1 (enabled).
2388
2389hop_limit - INTEGER
2390	Default Hop Limit to set.
2391
2392	Default: 64
2393
2394mtu - INTEGER
2395	Default Maximum Transfer Unit
2396
2397	Default: 1280 (IPv6 required minimum)
2398
2399ip_nonlocal_bind - BOOLEAN
2400	If set, allows processes to bind() to non-local IPv6 addresses,
2401	which can be quite useful - but may break some applications.
2402
2403	Default: 0
2404
2405router_probe_interval - INTEGER
2406	Minimum interval (in seconds) between Router Probing described
2407	in RFC4191.
2408
2409	Default: 60
2410
2411router_solicitation_delay - INTEGER
2412	Number of seconds to wait after interface is brought up
2413	before sending Router Solicitations.
2414
2415	Default: 1
2416
2417router_solicitation_interval - INTEGER
2418	Number of seconds to wait between Router Solicitations.
2419
2420	Default: 4
2421
2422router_solicitations - INTEGER
2423	Number of Router Solicitations to send until assuming no
2424	routers are present.
2425
2426	Default: 3
2427
2428use_oif_addrs_only - BOOLEAN
2429	When enabled, the candidate source addresses for destinations
2430	routed via this interface are restricted to the set of addresses
2431	configured on this interface (vis. RFC 6724, section 4).
2432
2433	Default: false
2434
2435use_tempaddr - INTEGER
2436	Preference for Privacy Extensions (RFC3041).
2437
2438	  * <= 0 : disable Privacy Extensions
2439	  * == 1 : enable Privacy Extensions, but prefer public
2440	    addresses over temporary addresses.
2441	  * >  1 : enable Privacy Extensions and prefer temporary
2442	    addresses over public addresses.
2443
2444	Default:
2445
2446		* 0 (for most devices)
2447		* -1 (for point-to-point devices and loopback devices)
2448
2449temp_valid_lft - INTEGER
2450	valid lifetime (in seconds) for temporary addresses.
2451
2452	Default: 172800 (2 days)
2453
2454temp_prefered_lft - INTEGER
2455	Preferred lifetime (in seconds) for temporary addresses.
2456
2457	Default: 86400 (1 day)
2458
2459keep_addr_on_down - INTEGER
2460	Keep all IPv6 addresses on an interface down event. If set static
2461	global addresses with no expiration time are not flushed.
2462
2463	*   >0 : enabled
2464	*    0 : system default
2465	*   <0 : disabled
2466
2467	Default: 0 (addresses are removed)
2468
2469max_desync_factor - INTEGER
2470	Maximum value for DESYNC_FACTOR, which is a random value
2471	that ensures that clients don't synchronize with each
2472	other and generate new addresses at exactly the same time.
2473	value is in seconds.
2474
2475	Default: 600
2476
2477regen_max_retry - INTEGER
2478	Number of attempts before give up attempting to generate
2479	valid temporary addresses.
2480
2481	Default: 5
2482
2483max_addresses - INTEGER
2484	Maximum number of autoconfigured addresses per interface.  Setting
2485	to zero disables the limitation.  It is not recommended to set this
2486	value too large (or to zero) because it would be an easy way to
2487	crash the kernel by allowing too many addresses to be created.
2488
2489	Default: 16
2490
2491disable_ipv6 - BOOLEAN
2492	Disable IPv6 operation.  If accept_dad is set to 2, this value
2493	will be dynamically set to TRUE if DAD fails for the link-local
2494	address.
2495
2496	Default: FALSE (enable IPv6 operation)
2497
2498	When this value is changed from 1 to 0 (IPv6 is being enabled),
2499	it will dynamically create a link-local address on the given
2500	interface and start Duplicate Address Detection, if necessary.
2501
2502	When this value is changed from 0 to 1 (IPv6 is being disabled),
2503	it will dynamically delete all addresses and routes on the given
2504	interface. From now on it will not possible to add addresses/routes
2505	to the selected interface.
2506
2507accept_dad - INTEGER
2508	Whether to accept DAD (Duplicate Address Detection).
2509
2510	 == ==============================================================
2511	  0  Disable DAD
2512	  1  Enable DAD (default)
2513	  2  Enable DAD, and disable IPv6 operation if MAC-based duplicate
2514	     link-local address has been found.
2515	 == ==============================================================
2516
2517	DAD operation and mode on a given interface will be selected according
2518	to the maximum value of conf/{all,interface}/accept_dad.
2519
2520force_tllao - BOOLEAN
2521	Enable sending the target link-layer address option even when
2522	responding to a unicast neighbor solicitation.
2523
2524	Default: FALSE
2525
2526	Quoting from RFC 2461, section 4.4, Target link-layer address:
2527
2528	"The option MUST be included for multicast solicitations in order to
2529	avoid infinite Neighbor Solicitation "recursion" when the peer node
2530	does not have a cache entry to return a Neighbor Advertisements
2531	message.  When responding to unicast solicitations, the option can be
2532	omitted since the sender of the solicitation has the correct link-
2533	layer address; otherwise it would not have be able to send the unicast
2534	solicitation in the first place. However, including the link-layer
2535	address in this case adds little overhead and eliminates a potential
2536	race condition where the sender deletes the cached link-layer address
2537	prior to receiving a response to a previous solicitation."
2538
2539ndisc_notify - BOOLEAN
2540	Define mode for notification of address and device changes.
2541
2542	* 0 - (default): do nothing
2543	* 1 - Generate unsolicited neighbour advertisements when device is brought
2544	  up or hardware address changes.
2545
2546ndisc_tclass - INTEGER
2547	The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
2548	Discovery (Router Solicitation, Router Advertisement, Neighbor
2549	Solicitation, Neighbor Advertisement, Redirect) messages.
2550	These 8 bits can be interpreted as 6 high order bits holding the DSCP
2551	value and 2 low order bits representing ECN (which you probably want
2552	to leave cleared).
2553
2554	* 0 - (default)
2555
2556ndisc_evict_nocarrier - BOOLEAN
2557	Clears the neighbor discovery table on NOCARRIER events. This option is
2558	important for wireless devices where the neighbor discovery cache should
2559	not be cleared when roaming between access points on the same network.
2560	In most cases this should remain as the default (1).
2561
2562	- 1 - (default): Clear neighbor discover cache on NOCARRIER events.
2563	- 0 - Do not clear neighbor discovery cache on NOCARRIER events.
2564
2565mldv1_unsolicited_report_interval - INTEGER
2566	The interval in milliseconds in which the next unsolicited
2567	MLDv1 report retransmit will take place.
2568
2569	Default: 10000 (10 seconds)
2570
2571mldv2_unsolicited_report_interval - INTEGER
2572	The interval in milliseconds in which the next unsolicited
2573	MLDv2 report retransmit will take place.
2574
2575	Default: 1000 (1 second)
2576
2577force_mld_version - INTEGER
2578	* 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
2579	* 1 - Enforce to use MLD version 1
2580	* 2 - Enforce to use MLD version 2
2581
2582suppress_frag_ndisc - INTEGER
2583	Control RFC 6980 (Security Implications of IPv6 Fragmentation
2584	with IPv6 Neighbor Discovery) behavior:
2585
2586	* 1 - (default) discard fragmented neighbor discovery packets
2587	* 0 - allow fragmented neighbor discovery packets
2588
2589optimistic_dad - BOOLEAN
2590	Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
2591
2592	* 0: disabled (default)
2593	* 1: enabled
2594
2595	Optimistic Duplicate Address Detection for the interface will be enabled
2596	if at least one of conf/{all,interface}/optimistic_dad is set to 1,
2597	it will be disabled otherwise.
2598
2599use_optimistic - BOOLEAN
2600	If enabled, do not classify optimistic addresses as deprecated during
2601	source address selection.  Preferred addresses will still be chosen
2602	before optimistic addresses, subject to other ranking in the source
2603	address selection algorithm.
2604
2605	* 0: disabled (default)
2606	* 1: enabled
2607
2608	This will be enabled if at least one of
2609	conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
2610
2611stable_secret - IPv6 address
2612	This IPv6 address will be used as a secret to generate IPv6
2613	addresses for link-local addresses and autoconfigured
2614	ones. All addresses generated after setting this secret will
2615	be stable privacy ones by default. This can be changed via the
2616	addrgenmode ip-link. conf/default/stable_secret is used as the
2617	secret for the namespace, the interface specific ones can
2618	overwrite that. Writes to conf/all/stable_secret are refused.
2619
2620	It is recommended to generate this secret during installation
2621	of a system and keep it stable after that.
2622
2623	By default the stable secret is unset.
2624
2625addr_gen_mode - INTEGER
2626	Defines how link-local and autoconf addresses are generated.
2627
2628	=  =================================================================
2629	0  generate address based on EUI64 (default)
2630	1  do no generate a link-local address, use EUI64 for addresses
2631	   generated from autoconf
2632	2  generate stable privacy addresses, using the secret from
2633	   stable_secret (RFC7217)
2634	3  generate stable privacy addresses, using a random secret if unset
2635	=  =================================================================
2636
2637drop_unicast_in_l2_multicast - BOOLEAN
2638	Drop any unicast IPv6 packets that are received in link-layer
2639	multicast (or broadcast) frames.
2640
2641	By default this is turned off.
2642
2643drop_unsolicited_na - BOOLEAN
2644	Drop all unsolicited neighbor advertisements, for example if there's
2645	a known good NA proxy on the network and such frames need not be used
2646	(or in the case of 802.11, must not be used to prevent attacks.)
2647
2648	By default this is turned off.
2649
2650accept_untracked_na - INTEGER
2651	Define behavior for accepting neighbor advertisements from devices that
2652	are absent in the neighbor cache:
2653
2654	- 0 - (default) Do not accept unsolicited and untracked neighbor
2655	  advertisements.
2656
2657	- 1 - Add a new neighbor cache entry in STALE state for routers on
2658	  receiving a neighbor advertisement (either solicited or unsolicited)
2659	  with target link-layer address option specified if no neighbor entry
2660	  is already present for the advertised IPv6 address. Without this knob,
2661	  NAs received for untracked addresses (absent in neighbor cache) are
2662	  silently ignored.
2663
2664	  This is as per router-side behavior documented in RFC9131.
2665
2666	  This has lower precedence than drop_unsolicited_na.
2667
2668	  This will optimize the return path for the initial off-link
2669	  communication that is initiated by a directly connected host, by
2670	  ensuring that the first-hop router which turns on this setting doesn't
2671	  have to buffer the initial return packets to do neighbor-solicitation.
2672	  The prerequisite is that the host is configured to send unsolicited
2673	  neighbor advertisements on interface bringup. This setting should be
2674	  used in conjunction with the ndisc_notify setting on the host to
2675	  satisfy this prerequisite.
2676
2677	- 2 - Extend option (1) to add a new neighbor cache entry only if the
2678	  source IP address is in the same subnet as an address configured on
2679	  the interface that received the neighbor advertisement.
2680
2681enhanced_dad - BOOLEAN
2682	Include a nonce option in the IPv6 neighbor solicitation messages used for
2683	duplicate address detection per RFC7527. A received DAD NS will only signal
2684	a duplicate address if the nonce is different. This avoids any false
2685	detection of duplicates due to loopback of the NS messages that we send.
2686	The nonce option will be sent on an interface unless both of
2687	conf/{all,interface}/enhanced_dad are set to FALSE.
2688
2689	Default: TRUE
2690
2691``icmp/*``:
2692===========
2693
2694ratelimit - INTEGER
2695	Limit the maximal rates for sending ICMPv6 messages.
2696
2697	0 to disable any limiting,
2698	otherwise the minimal space between responses in milliseconds.
2699
2700	Default: 1000
2701
2702ratemask - list of comma separated ranges
2703	For ICMPv6 message types matching the ranges in the ratemask, limit
2704	the sending of the message according to ratelimit parameter.
2705
2706	The format used for both input and output is a comma separated
2707	list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
2708	129). Writing to the file will clear all previous ranges of ICMPv6
2709	message types and update the current list with the input.
2710
2711	Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
2712	for numerical values of ICMPv6 message types, e.g. echo request is 128
2713	and echo reply is 129.
2714
2715	Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
2716
2717echo_ignore_all - BOOLEAN
2718	If set non-zero, then the kernel will ignore all ICMP ECHO
2719	requests sent to it over the IPv6 protocol.
2720
2721	Default: 0
2722
2723echo_ignore_multicast - BOOLEAN
2724	If set non-zero, then the kernel will ignore all ICMP ECHO
2725	requests sent to it over the IPv6 protocol via multicast.
2726
2727	Default: 0
2728
2729echo_ignore_anycast - BOOLEAN
2730	If set non-zero, then the kernel will ignore all ICMP ECHO
2731	requests sent to it over the IPv6 protocol destined to anycast address.
2732
2733	Default: 0
2734
2735error_anycast_as_unicast - BOOLEAN
2736	If set to 1, then the kernel will respond with ICMP Errors
2737	resulting from requests sent to it over the IPv6 protocol destined
2738	to anycast address essentially treating anycast as unicast.
2739
2740	Default: 0
2741
2742xfrm6_gc_thresh - INTEGER
2743	(Obsolete since linux-4.14)
2744	The threshold at which we will start garbage collecting for IPv6
2745	destination cache entries.  At twice this value the system will
2746	refuse new allocations.
2747
2748
2749IPv6 Update by:
2750Pekka Savola <pekkas@netcore.fi>
2751YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org>
2752
2753
2754/proc/sys/net/bridge/* Variables:
2755=================================
2756
2757bridge-nf-call-arptables - BOOLEAN
2758	- 1 : pass bridged ARP traffic to arptables' FORWARD chain.
2759	- 0 : disable this.
2760
2761	Default: 1
2762
2763bridge-nf-call-iptables - BOOLEAN
2764	- 1 : pass bridged IPv4 traffic to iptables' chains.
2765	- 0 : disable this.
2766
2767	Default: 1
2768
2769bridge-nf-call-ip6tables - BOOLEAN
2770	- 1 : pass bridged IPv6 traffic to ip6tables' chains.
2771	- 0 : disable this.
2772
2773	Default: 1
2774
2775bridge-nf-filter-vlan-tagged - BOOLEAN
2776	- 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables.
2777	- 0 : disable this.
2778
2779	Default: 0
2780
2781bridge-nf-filter-pppoe-tagged - BOOLEAN
2782	- 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
2783	- 0 : disable this.
2784
2785	Default: 0
2786
2787bridge-nf-pass-vlan-input-dev - BOOLEAN
2788	- 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
2789	  interface on the bridge and set the netfilter input device to the
2790	  vlan. This allows use of e.g. "iptables -i br0.1" and makes the
2791	  REDIRECT target work with vlan-on-top-of-bridge interfaces.  When no
2792	  matching vlan interface is found, or this switch is off, the input
2793	  device is set to the bridge interface.
2794
2795	- 0: disable bridge netfilter vlan interface lookup.
2796
2797	Default: 0
2798
2799``proc/sys/net/sctp/*`` Variables:
2800==================================
2801
2802addip_enable - BOOLEAN
2803	Enable or disable extension of  Dynamic Address Reconfiguration
2804	(ADD-IP) functionality specified in RFC5061.  This extension provides
2805	the ability to dynamically add and remove new addresses for the SCTP
2806	associations.
2807
2808	1: Enable extension.
2809
2810	0: Disable extension.
2811
2812	Default: 0
2813
2814pf_enable - INTEGER
2815	Enable or disable pf (pf is short for potentially failed) state. A value
2816	of pf_retrans > path_max_retrans also disables pf state. That is, one of
2817	both pf_enable and pf_retrans > path_max_retrans can disable pf state.
2818	Since pf_retrans and path_max_retrans can be changed by userspace
2819	application, sometimes user expects to disable pf state by the value of
2820	pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
2821	or path_max_retrans is changed by the user application, this pf state is
2822	enabled. As such, it is necessary to add this to dynamically enable
2823	and disable pf state. See:
2824	https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
2825	details.
2826
2827	1: Enable pf.
2828
2829	0: Disable pf.
2830
2831	Default: 1
2832
2833pf_expose - INTEGER
2834	Unset or enable/disable pf (pf is short for potentially failed) state
2835	exposure.  Applications can control the exposure of the PF path state
2836	in the SCTP_PEER_ADDR_CHANGE event and the SCTP_GET_PEER_ADDR_INFO
2837	sockopt.   When it's unset, no SCTP_PEER_ADDR_CHANGE event with
2838	SCTP_ADDR_PF state will be sent and a SCTP_PF-state transport info
2839	can be got via SCTP_GET_PEER_ADDR_INFO sockopt;  When it's enabled,
2840	a SCTP_PEER_ADDR_CHANGE event will be sent for a transport becoming
2841	SCTP_PF state and a SCTP_PF-state transport info can be got via
2842	SCTP_GET_PEER_ADDR_INFO sockopt;  When it's disabled, no
2843	SCTP_PEER_ADDR_CHANGE event will be sent and it returns -EACCES when
2844	trying to get a SCTP_PF-state transport info via SCTP_GET_PEER_ADDR_INFO
2845	sockopt.
2846
2847	0: Unset pf state exposure, Compatible with old applications.
2848
2849	1: Disable pf state exposure.
2850
2851	2: Enable pf state exposure.
2852
2853	Default: 0
2854
2855addip_noauth_enable - BOOLEAN
2856	Dynamic Address Reconfiguration (ADD-IP) requires the use of
2857	authentication to protect the operations of adding or removing new
2858	addresses.  This requirement is mandated so that unauthorized hosts
2859	would not be able to hijack associations.  However, older
2860	implementations may not have implemented this requirement while
2861	allowing the ADD-IP extension.  For reasons of interoperability,
2862	we provide this variable to control the enforcement of the
2863	authentication requirement.
2864
2865	== ===============================================================
2866	1  Allow ADD-IP extension to be used without authentication.  This
2867	   should only be set in a closed environment for interoperability
2868	   with older implementations.
2869
2870	0  Enforce the authentication requirement
2871	== ===============================================================
2872
2873	Default: 0
2874
2875auth_enable - BOOLEAN
2876	Enable or disable Authenticated Chunks extension.  This extension
2877	provides the ability to send and receive authenticated chunks and is
2878	required for secure operation of Dynamic Address Reconfiguration
2879	(ADD-IP) extension.
2880
2881	- 1: Enable this extension.
2882	- 0: Disable this extension.
2883
2884	Default: 0
2885
2886prsctp_enable - BOOLEAN
2887	Enable or disable the Partial Reliability extension (RFC3758) which
2888	is used to notify peers that a given DATA should no longer be expected.
2889
2890	- 1: Enable extension
2891	- 0: Disable
2892
2893	Default: 1
2894
2895max_burst - INTEGER
2896	The limit of the number of new packets that can be initially sent.  It
2897	controls how bursty the generated traffic can be.
2898
2899	Default: 4
2900
2901association_max_retrans - INTEGER
2902	Set the maximum number for retransmissions that an association can
2903	attempt deciding that the remote end is unreachable.  If this value
2904	is exceeded, the association is terminated.
2905
2906	Default: 10
2907
2908max_init_retransmits - INTEGER
2909	The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
2910	that an association will attempt before declaring the destination
2911	unreachable and terminating.
2912
2913	Default: 8
2914
2915path_max_retrans - INTEGER
2916	The maximum number of retransmissions that will be attempted on a given
2917	path.  Once this threshold is exceeded, the path is considered
2918	unreachable, and new traffic will use a different path when the
2919	association is multihomed.
2920
2921	Default: 5
2922
2923pf_retrans - INTEGER
2924	The number of retransmissions that will be attempted on a given path
2925	before traffic is redirected to an alternate transport (should one
2926	exist).  Note this is distinct from path_max_retrans, as a path that
2927	passes the pf_retrans threshold can still be used.  Its only
2928	deprioritized when a transmission path is selected by the stack.  This
2929	setting is primarily used to enable fast failover mechanisms without
2930	having to reduce path_max_retrans to a very low value.  See:
2931	http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
2932	for details.  Note also that a value of pf_retrans > path_max_retrans
2933	disables this feature. Since both pf_retrans and path_max_retrans can
2934	be changed by userspace application, a variable pf_enable is used to
2935	disable pf state.
2936
2937	Default: 0
2938
2939ps_retrans - INTEGER
2940	Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
2941	from section-5 "Primary Path Switchover" in rfc7829.  The primary path
2942	will be changed to another active path when the path error counter on
2943	the old primary path exceeds PSMR, so that "the SCTP sender is allowed
2944	to continue data transmission on a new working path even when the old
2945	primary destination address becomes active again".   Note this feature
2946	is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
2947	and its value can't be less than 'pf_retrans' when changing by sysctl.
2948
2949	Default: 0xffff
2950
2951rto_initial - INTEGER
2952	The initial round trip timeout value in milliseconds that will be used
2953	in calculating round trip times.  This is the initial time interval
2954	for retransmissions.
2955
2956	Default: 3000
2957
2958rto_max - INTEGER
2959	The maximum value (in milliseconds) of the round trip timeout.  This
2960	is the largest time interval that can elapse between retransmissions.
2961
2962	Default: 60000
2963
2964rto_min - INTEGER
2965	The minimum value (in milliseconds) of the round trip timeout.  This
2966	is the smallest time interval the can elapse between retransmissions.
2967
2968	Default: 1000
2969
2970hb_interval - INTEGER
2971	The interval (in milliseconds) between HEARTBEAT chunks.  These chunks
2972	are sent at the specified interval on idle paths to probe the state of
2973	a given path between 2 associations.
2974
2975	Default: 30000
2976
2977sack_timeout - INTEGER
2978	The amount of time (in milliseconds) that the implementation will wait
2979	to send a SACK.
2980
2981	Default: 200
2982
2983valid_cookie_life - INTEGER
2984	The default lifetime of the SCTP cookie (in milliseconds).  The cookie
2985	is used during association establishment.
2986
2987	Default: 60000
2988
2989cookie_preserve_enable - BOOLEAN
2990	Enable or disable the ability to extend the lifetime of the SCTP cookie
2991	that is used during the establishment phase of SCTP association
2992
2993	- 1: Enable cookie lifetime extension.
2994	- 0: Disable
2995
2996	Default: 1
2997
2998cookie_hmac_alg - STRING
2999	Select the hmac algorithm used when generating the cookie value sent by
3000	a listening sctp socket to a connecting client in the INIT-ACK chunk.
3001	Valid values are:
3002
3003	* md5
3004	* sha1
3005	* none
3006
3007	Ability to assign md5 or sha1 as the selected alg is predicated on the
3008	configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and
3009	CONFIG_CRYPTO_SHA1).
3010
3011	Default: Dependent on configuration.  MD5 if available, else SHA1 if
3012	available, else none.
3013
3014rcvbuf_policy - INTEGER
3015	Determines if the receive buffer is attributed to the socket or to
3016	association.   SCTP supports the capability to create multiple
3017	associations on a single socket.  When using this capability, it is
3018	possible that a single stalled association that's buffering a lot
3019	of data may block other associations from delivering their data by
3020	consuming all of the receive buffer space.  To work around this,
3021	the rcvbuf_policy could be set to attribute the receiver buffer space
3022	to each association instead of the socket.  This prevents the described
3023	blocking.
3024
3025	- 1: rcvbuf space is per association
3026	- 0: rcvbuf space is per socket
3027
3028	Default: 0
3029
3030sndbuf_policy - INTEGER
3031	Similar to rcvbuf_policy above, this applies to send buffer space.
3032
3033	- 1: Send buffer is tracked per association
3034	- 0: Send buffer is tracked per socket.
3035
3036	Default: 0
3037
3038sctp_mem - vector of 3 INTEGERs: min, pressure, max
3039	Number of pages allowed for queueing by all SCTP sockets.
3040
3041	min: Below this number of pages SCTP is not bothered about its
3042	memory appetite. When amount of memory allocated by SCTP exceeds
3043	this number, SCTP starts to moderate memory usage.
3044
3045	pressure: This value was introduced to follow format of tcp_mem.
3046
3047	max: Number of pages allowed for queueing by all SCTP sockets.
3048
3049	Default is calculated at boot time from amount of available memory.
3050
3051sctp_rmem - vector of 3 INTEGERs: min, default, max
3052	Only the first value ("min") is used, "default" and "max" are
3053	ignored.
3054
3055	min: Minimal size of receive buffer used by SCTP socket.
3056	It is guaranteed to each SCTP socket (but not association) even
3057	under moderate memory pressure.
3058
3059	Default: 4K
3060
3061sctp_wmem  - vector of 3 INTEGERs: min, default, max
3062	Only the first value ("min") is used, "default" and "max" are
3063	ignored.
3064
3065	min: Minimum size of send buffer that can be used by SCTP sockets.
3066	It is guaranteed to each SCTP socket (but not association) even
3067	under moderate memory pressure.
3068
3069	Default: 4K
3070
3071addr_scope_policy - INTEGER
3072	Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00
3073
3074	- 0   - Disable IPv4 address scoping
3075	- 1   - Enable IPv4 address scoping
3076	- 2   - Follow draft but allow IPv4 private addresses
3077	- 3   - Follow draft but allow IPv4 link local addresses
3078
3079	Default: 1
3080
3081udp_port - INTEGER
3082	The listening port for the local UDP tunneling sock. Normally it's
3083	using the IANA-assigned UDP port number 9899 (sctp-tunneling).
3084
3085	This UDP sock is used for processing the incoming UDP-encapsulated
3086	SCTP packets (from RFC6951), and shared by all applications in the
3087	same net namespace. This UDP sock will be closed when the value is
3088	set to 0.
3089
3090	The value will also be used to set the src port of the UDP header
3091	for the outgoing UDP-encapsulated SCTP packets. For the dest port,
3092	please refer to 'encap_port' below.
3093
3094	Default: 0
3095
3096encap_port - INTEGER
3097	The default remote UDP encapsulation port.
3098
3099	This value is used to set the dest port of the UDP header for the
3100	outgoing UDP-encapsulated SCTP packets by default. Users can also
3101	change the value for each sock/asoc/transport by using setsockopt.
3102	For further information, please refer to RFC6951.
3103
3104	Note that when connecting to a remote server, the client should set
3105	this to the port that the UDP tunneling sock on the peer server is
3106	listening to and the local UDP tunneling sock on the client also
3107	must be started. On the server, it would get the encap_port from
3108	the incoming packet's source port.
3109
3110	Default: 0
3111
3112plpmtud_probe_interval - INTEGER
3113        The time interval (in milliseconds) for the PLPMTUD probe timer,
3114        which is configured to expire after this period to receive an
3115        acknowledgment to a probe packet. This is also the time interval
3116        between the probes for the current pmtu when the probe search
3117        is done.
3118
3119        PLPMTUD will be disabled when 0 is set, and other values for it
3120        must be >= 5000.
3121
3122	Default: 0
3123
3124reconf_enable - BOOLEAN
3125        Enable or disable extension of Stream Reconfiguration functionality
3126        specified in RFC6525. This extension provides the ability to "reset"
3127        a stream, and it includes the Parameters of "Outgoing/Incoming SSN
3128        Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams".
3129
3130	- 1: Enable extension.
3131	- 0: Disable extension.
3132
3133	Default: 0
3134
3135intl_enable - BOOLEAN
3136        Enable or disable extension of User Message Interleaving functionality
3137        specified in RFC8260. This extension allows the interleaving of user
3138        messages sent on different streams. With this feature enabled, I-DATA
3139        chunk will replace DATA chunk to carry user messages if also supported
3140        by the peer. Note that to use this feature, one needs to set this option
3141        to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2
3142        and SCTP_INTERLEAVING_SUPPORTED to 1.
3143
3144	- 1: Enable extension.
3145	- 0: Disable extension.
3146
3147	Default: 0
3148
3149ecn_enable - BOOLEAN
3150        Control use of Explicit Congestion Notification (ECN) by SCTP.
3151        Like in TCP, ECN is used only when both ends of the SCTP connection
3152        indicate support for it. This feature is useful in avoiding losses
3153        due to congestion by allowing supporting routers to signal congestion
3154        before having to drop packets.
3155
3156        1: Enable ecn.
3157        0: Disable ecn.
3158
3159        Default: 1
3160
3161l3mdev_accept - BOOLEAN
3162	Enabling this option allows a "global" bound socket to work
3163	across L3 master domains (e.g., VRFs) with packets capable of
3164	being received regardless of the L3 domain in which they
3165	originated. Only valid when the kernel was compiled with
3166	CONFIG_NET_L3_MASTER_DEV.
3167
3168	Default: 1 (enabled)
3169
3170
3171``/proc/sys/net/core/*``
3172========================
3173
3174	Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
3175
3176
3177``/proc/sys/net/unix/*``
3178========================
3179
3180max_dgram_qlen - INTEGER
3181	The maximum length of dgram socket receive queue
3182
3183	Default: 10
3184
3185