요약·해설과 원문, 전문 번역을 서로 분리했습니다. API 이름, symbol, source path는 원문 표기를 사용합니다.
1. 요약·해설
원문의 핵심 논리와 kernel programming 관점의 보충 설명입니다. 아래의 전문 번역과는 별도로 작성했습니다.
2. 영어 원문 전체
번역 기준이 된 Linux v6.18.37 원문입니다. 줄 번호는 이 버전의 파일 좌표입니다.
원문 전체 펼치기
1
.. SPDX-License-Identifier: GPL-2.0
3
=========
4
IP Sysctl
5
=========
7
/proc/sys/net/ipv4/* Variables
8
==============================
10
ip_forward - BOOLEAN
11
Forward Packets between interfaces.
13
This variable is special, its change resets all configuration
14
parameters to their default state (RFC1122 for hosts, RFC1812
15
for routers)
17
Possible values:
19
- 0 (disabled)
20
- 1 (enabled)
22
Default: 0 (disabled)
24
ip_default_ttl - INTEGER
25
Default value of TTL field (Time To Live) for outgoing (but not
26
forwarded) IP packets. Should be between 1 and 255 inclusive.
27
Default: 64 (as recommended by RFC1700)
29
ip_no_pmtu_disc - INTEGER
30
Disable Path MTU Discovery. If enabled in mode 1 and a
31
fragmentation-required ICMP is received, the PMTU to this
32
destination will be set to the smallest of the old MTU to
33
this destination and min_pmtu (see below). You will need
34
to raise min_pmtu to the smallest interface MTU on your system
35
manually if you want to avoid locally generated fragments.
37
In mode 2 incoming Path MTU Discovery messages will be
38
discarded. Outgoing frames are handled the same as in mode 1,
39
implicitly setting IP_PMTUDISC_DONT on every created socket.
41
Mode 3 is a hardened pmtu discover mode. The kernel will only
42
accept fragmentation-needed errors if the underlying protocol
43
can verify them besides a plain socket lookup. Current
44
protocols for which pmtu events will be honored are TCP and
45
SCTP as they verify e.g. the sequence number or the
46
association. This mode should not be enabled globally but is
47
only intended to secure e.g. name servers in namespaces where
48
TCP path mtu must still work but path MTU information of other
49
protocols should be discarded. If enabled globally this mode
50
could break other protocols.
52
Possible values: 0-3
54
Default: FALSE
56
min_pmtu - INTEGER
57
default 552 - minimum Path MTU. Unless this is changed manually,
58
each cached pmtu will never be lower than this setting.
60
ip_forward_use_pmtu - BOOLEAN
61
By default we don't trust protocol path MTUs while forwarding
62
because they could be easily forged and can lead to unwanted
63
fragmentation by the router.
64
You only need to enable this if you have user-space software
65
which tries to discover path mtus by itself and depends on the
66
kernel honoring this information. This is normally not the
67
case.
69
Possible values:
71
- 0 (disabled)
72
- 1 (enabled)
74
Default: 0 (disabled)
76
fwmark_reflect - BOOLEAN
77
Controls the fwmark of kernel-generated IPv4 reply packets that are not
78
associated with a socket for example, TCP RSTs or ICMP echo replies).
79
If disabled, these packets have a fwmark of zero. If enabled, they have the
80
fwmark of the packet they are replying to.
82
Possible values:
84
- 0 (disabled)
85
- 1 (enabled)
87
Default: 0 (disabled)
89
fib_multipath_use_neigh - BOOLEAN
90
Use status of existing neighbor entry when determining nexthop for
91
multipath routes. If disabled, neighbor information is not used and
92
packets could be directed to a failed nexthop. Only valid for kernels
93
built with CONFIG_IP_ROUTE_MULTIPATH enabled.
95
Possible values:
97
- 0 (disabled)
98
- 1 (enabled)
100
Default: 0 (disabled)
102
fib_multipath_hash_policy - INTEGER
103
Controls which hash policy to use for multipath routes. Only valid
104
for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
106
Default: 0 (Layer 3)
108
Possible values:
110
- 0 - Layer 3
111
- 1 - Layer 4
112
- 2 - Layer 3 or inner Layer 3 if present
113
- 3 - Custom multipath hash. Fields used for multipath hash calculation
114
are determined by fib_multipath_hash_fields sysctl
116
fib_multipath_hash_fields - UNSIGNED INTEGER
117
When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
118
fields used for multipath hash calculation are determined by this
119
sysctl.
121
This value is a bitmask which enables various fields for multipath hash
122
calculation.
124
Possible fields are:
126
====== ============================
127
0x0001 Source IP address
128
0x0002 Destination IP address
129
0x0004 IP protocol
130
0x0008 Unused (Flow Label)
131
0x0010 Source port
132
0x0020 Destination port
133
0x0040 Inner source IP address
134
0x0080 Inner destination IP address
135
0x0100 Inner IP protocol
136
0x0200 Inner Flow Label
137
0x0400 Inner source port
138
0x0800 Inner destination port
139
====== ============================
141
Default: 0x0007 (source IP, destination IP and IP protocol)
143
fib_multipath_hash_seed - UNSIGNED INTEGER
144
The seed value used when calculating hash for multipath routes. Applies
145
to both IPv4 and IPv6 datapath. Only present for kernels built with
146
CONFIG_IP_ROUTE_MULTIPATH enabled.
148
When set to 0, the seed value used for multipath routing defaults to an
149
internal random-generated one.
151
The actual hashing algorithm is not specified -- there is no guarantee
152
that a next hop distribution effected by a given seed will keep stable
153
across kernel versions.
155
Default: 0 (random)
157
fib_sync_mem - UNSIGNED INTEGER
158
Amount of dirty memory from fib entries that can be backlogged before
159
synchronize_rcu is forced.
161
Default: 512kB Minimum: 64kB Maximum: 64MB
163
ip_forward_update_priority - INTEGER
164
Whether to update SKB priority from "TOS" field in IPv4 header after it
165
is forwarded. The new SKB priority is mapped from TOS field value
166
according to an rt_tos2priority table (see e.g. man tc-prio).
168
Default: 1 (Update priority.)
170
Possible values:
172
- 0 - Do not update priority.
173
- 1 - Update priority.
175
route/max_size - INTEGER
176
Maximum number of routes allowed in the kernel. Increase
177
this when using large numbers of interfaces and/or routes.
179
From linux kernel 3.6 onwards, this is deprecated for ipv4
180
as route cache is no longer used.
182
From linux kernel 6.3 onwards, this is deprecated for ipv6
183
as garbage collection manages cached route entries.
185
neigh/default/gc_thresh1 - INTEGER
186
Minimum number of entries to keep. Garbage collector will not
187
purge entries if there are fewer than this number.
189
Default: 128
191
neigh/default/gc_thresh2 - INTEGER
192
Threshold when garbage collector becomes more aggressive about
193
purging entries. Entries older than 5 seconds will be cleared
194
when over this number.
196
Default: 512
198
neigh/default/gc_thresh3 - INTEGER
199
Maximum number of non-PERMANENT neighbor entries allowed. Increase
200
this when using large numbers of interfaces and when communicating
201
with large numbers of directly-connected peers.
203
Default: 1024
205
neigh/default/unres_qlen_bytes - INTEGER
206
The maximum number of bytes which may be used by packets
207
queued for each unresolved address by other network layers.
208
(added in linux 3.3)
210
Setting negative value is meaningless and will return error.
212
Default: SK_WMEM_DEFAULT, (same as net.core.wmem_default).
214
Exact value depends on architecture and kernel options,
215
but should be enough to allow queuing 256 packets
216
of medium size.
218
neigh/default/unres_qlen - INTEGER
219
The maximum number of packets which may be queued for each
220
unresolved address by other network layers.
222
(deprecated in linux 3.3) : use unres_qlen_bytes instead.
224
Prior to linux 3.3, the default value is 3 which may cause
225
unexpected packet loss. The current default value is calculated
226
according to default value of unres_qlen_bytes and true size of
227
packet.
229
Default: 101
231
neigh/default/interval_probe_time_ms - INTEGER
232
The probe interval for neighbor entries with NTF_MANAGED flag,
233
the min value is 1.
235
Default: 5000
237
mtu_expires - INTEGER
238
Time, in seconds, that cached PMTU information is kept.
240
min_adv_mss - INTEGER
241
The advertised MSS depends on the first hop route MTU, but will
242
never be lower than this setting.
244
fib_notify_on_flag_change - INTEGER
245
Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
246
RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
248
After installing a route to the kernel, user space receives an
249
acknowledgment, which means the route was installed in the kernel,
250
but not necessarily in hardware.
251
It is also possible for a route already installed in hardware to change
252
its action and therefore its flags. For example, a host route that is
253
trapping packets can be "promoted" to perform decapsulation following
254
the installation of an IPinIP/VXLAN tunnel.
255
The notifications will indicate to user-space the state of the route.
257
Default: 0 (Do not emit notifications.)
259
Possible values:
261
- 0 - Do not emit notifications.
262
- 1 - Emit notifications.
263
- 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
265
IP Fragmentation:
267
ipfrag_high_thresh - LONG INTEGER
268
Maximum memory used to reassemble IP fragments.
270
ipfrag_low_thresh - LONG INTEGER
271
(Obsolete since linux-4.17)
272
Maximum memory used to reassemble IP fragments before the kernel
273
begins to remove incomplete fragment queues to free up resources.
274
The kernel still accepts new fragments for defragmentation.
276
ipfrag_time - INTEGER
277
Time in seconds to keep an IP fragment in memory.
279
ipfrag_max_dist - INTEGER
280
ipfrag_max_dist is a non-negative integer value which defines the
281
maximum "disorder" which is allowed among fragments which share a
282
common IP source address. Note that reordering of packets is
283
not unusual, but if a large number of fragments arrive from a source
284
IP address while a particular fragment queue remains incomplete, it
285
probably indicates that one or more fragments belonging to that queue
286
have been lost. When ipfrag_max_dist is positive, an additional check
287
is done on fragments before they are added to a reassembly queue - if
288
ipfrag_max_dist (or more) fragments have arrived from a particular IP
289
address between additions to any IP fragment queue using that source
290
address, it's presumed that one or more fragments in the queue are
291
lost. The existing fragment queue will be dropped, and a new one
292
started. An ipfrag_max_dist value of zero disables this check.
294
Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
295
result in unnecessarily dropping fragment queues when normal
296
reordering of packets occurs, which could lead to poor application
297
performance. Using a very large value, e.g. 50000, increases the
298
likelihood of incorrectly reassembling IP fragments that originate
299
from different IP datagrams, which could result in data corruption.
300
Default: 64
302
bc_forwarding - INTEGER
303
bc_forwarding enables the feature described in rfc1812#section-5.3.5.2
304
and rfc2644. It allows the router to forward directed broadcast.
305
To enable this feature, the 'all' entry and the input interface entry
306
should be set to 1.
307
Default: 0
309
INET peer storage
310
=================
312
inet_peer_threshold - INTEGER
313
The approximate size of the storage. Starting from this threshold
314
entries will be thrown aggressively. This threshold also determines
315
entries' time-to-live and time intervals between garbage collection
316
passes. More entries, less time-to-live, less GC interval.
318
inet_peer_minttl - INTEGER
319
Minimum time-to-live of entries. Should be enough to cover fragment
320
time-to-live on the reassembling side. This minimum time-to-live is
321
guaranteed if the pool size is less than inet_peer_threshold.
322
Measured in seconds.
324
inet_peer_maxttl - INTEGER
325
Maximum time-to-live of entries. Unused entries will expire after
326
this period of time if there is no memory pressure on the pool (i.e.
327
when the number of entries in the pool is very small).
328
Measured in seconds.
330
TCP variables
331
=============
333
somaxconn - INTEGER
334
Limit of socket listen() backlog, known in userspace as SOMAXCONN.
335
Defaults to 4096. (Was 128 before linux-5.4)
336
See also tcp_max_syn_backlog for additional tuning for TCP sockets.
338
tcp_abort_on_overflow - BOOLEAN
339
If listening service is too slow to accept new connections,
340
reset them. Default state is FALSE. It means that if overflow
341
occurred due to a burst, connection will recover. Enable this
342
option _only_ if you are really sure that listening daemon
343
cannot be tuned to accept connections faster. Enabling this
344
option can harm clients of your server.
346
tcp_adv_win_scale - INTEGER
347
Obsolete since linux-6.6
348
Count buffering overhead as bytes/2^tcp_adv_win_scale
349
(if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
350
if it is <= 0.
352
Possible values are [-31, 31], inclusive.
354
Default: 1
356
tcp_allowed_congestion_control - STRING
357
Show/set the congestion control choices available to non-privileged
358
processes. The list is a subset of those listed in
359
tcp_available_congestion_control.
361
Default is "reno" and the default setting (tcp_congestion_control).
363
tcp_app_win - INTEGER
364
Reserve max(window/2^tcp_app_win, mss) of window for application
365
buffer. Value 0 is special, it means that nothing is reserved.
367
Possible values are [0, 31], inclusive.
369
Default: 31
371
tcp_autocorking - BOOLEAN
372
Enable TCP auto corking :
373
When applications do consecutive small write()/sendmsg() system calls,
374
we try to coalesce these small writes as much as possible, to lower
375
total amount of sent packets. This is done if at least one prior
376
packet for the flow is waiting in Qdisc queues or device transmit
377
queue. Applications can still use TCP_CORK for optimal behavior
378
when they know how/when to uncork their sockets.
380
Possible values:
382
- 0 (disabled)
383
- 1 (enabled)
385
Default: 1 (enabled)
387
tcp_available_congestion_control - STRING
388
Shows the available congestion control choices that are registered.
389
More congestion control algorithms may be available as modules,
390
but not loaded.
392
tcp_base_mss - INTEGER
393
The initial value of search_low to be used by the packetization layer
394
Path MTU discovery (MTU probing). If MTU probing is enabled,
395
this is the initial MSS used by the connection.
397
tcp_mtu_probe_floor - INTEGER
398
If MTU probing is enabled this caps the minimum MSS used for search_low
399
for the connection.
401
Default : 48
403
tcp_min_snd_mss - INTEGER
404
TCP SYN and SYNACK messages usually advertise an ADVMSS option,
405
as described in RFC 1122 and RFC 6691.
407
If this ADVMSS option is smaller than tcp_min_snd_mss,
408
it is silently capped to tcp_min_snd_mss.
410
Default : 48 (at least 8 bytes of payload per segment)
412
tcp_congestion_control - STRING
413
Set the congestion control algorithm to be used for new
414
connections. The algorithm "reno" is always available, but
415
additional choices may be available based on kernel configuration.
416
Default is set as part of kernel configuration.
417
For passive connections, the listener congestion control choice
418
is inherited.
420
[see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
422
tcp_dsack - BOOLEAN
423
Allows TCP to send "duplicate" SACKs.
425
Possible values:
427
- 0 (disabled)
428
- 1 (enabled)
430
Default: 1 (enabled)
432
tcp_early_retrans - INTEGER
433
Tail loss probe (TLP) converts RTOs occurring due to tail
434
losses into fast recovery (RFC8985). Note that
435
TLP requires RACK to function properly (see tcp_recovery below)
437
Possible values:
439
- 0 disables TLP
440
- 3 or 4 enables TLP
442
Default: 3
444
tcp_ecn - INTEGER
445
Control use of Explicit Congestion Notification (ECN) by TCP.
446
ECN is used only when both ends of the TCP connection indicate support
447
for it. This feature is useful in avoiding losses due to congestion by
448
allowing supporting routers to signal congestion before having to drop
449
packets. A host that supports ECN both sends ECN at the IP layer and
450
feeds back ECN at the TCP layer. The highest variant of ECN feedback
451
that both peers support is chosen by the ECN negotiation (Accurate ECN,
452
ECN, or no ECN).
454
The highest negotiated variant for incoming connection requests
455
and the highest variant requested by outgoing connection
456
attempts:
458
===== ==================== ====================
459
Value Incoming connections Outgoing connections
460
===== ==================== ====================
461
0 No ECN No ECN
462
1 ECN ECN
463
2 ECN No ECN
464
3 AccECN AccECN
465
4 AccECN ECN
466
5 AccECN No ECN
467
===== ==================== ====================
469
Default: 2
471
tcp_ecn_option - INTEGER
472
Control Accurate ECN (AccECN) option sending when AccECN has been
473
successfully negotiated during handshake. Send logic inhibits
474
sending AccECN options regarless of this setting when no AccECN
475
option has been seen for the reverse direction.
477
Possible values are:
479
= ============================================================
480
0 Never send AccECN option. This also disables sending AccECN
481
option in SYN/ACK during handshake.
482
1 Send AccECN option sparingly according to the minimum option
483
rules outlined in draft-ietf-tcpm-accurate-ecn.
484
2 Send AccECN option on every packet whenever it fits into TCP
485
option space.
486
= ============================================================
488
Default: 2
490
tcp_ecn_option_beacon - INTEGER
491
Control Accurate ECN (AccECN) option sending frequency per RTT and it
492
takes effect only when tcp_ecn_option is set to 2.
494
Default: 3 (AccECN will be send at least 3 times per RTT)
496
tcp_ecn_fallback - BOOLEAN
497
If the kernel detects that ECN connection misbehaves, enable fall
498
back to non-ECN. Currently, this knob implements the fallback
499
from RFC3168, section 6.1.1.1., but we reserve that in future,
500
additional detection mechanisms could be implemented under this
501
knob. The value is not used, if tcp_ecn or per route (or congestion
502
control) ECN settings are disabled.
504
Possible values:
506
- 0 (disabled)
507
- 1 (enabled)
509
Default: 1 (enabled)
511
tcp_fack - BOOLEAN
512
This is a legacy option, it has no effect anymore.
514
tcp_fin_timeout - INTEGER
515
The length of time an orphaned (no longer referenced by any
516
application) connection will remain in the FIN_WAIT_2 state
517
before it is aborted at the local end. While a perfectly
518
valid "receive only" state for an un-orphaned connection, an
519
orphaned connection in FIN_WAIT_2 state could otherwise wait
520
forever for the remote to close its end of the connection.
522
Cf. tcp_max_orphans
524
Default: 60 seconds
526
tcp_frto - INTEGER
527
Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
528
F-RTO is an enhanced recovery algorithm for TCP retransmission
529
timeouts. It is particularly beneficial in networks where the
530
RTT fluctuates (e.g., wireless). F-RTO is sender-side only
531
modification. It does not require any support from the peer.
533
By default it's enabled with a non-zero value. 0 disables F-RTO.
535
tcp_fwmark_accept - BOOLEAN
536
If enabled, incoming connections to listening sockets that do not have a
537
socket mark will set the mark of the accepting socket to the fwmark of
538
the incoming SYN packet. This will cause all packets on that connection
539
(starting from the first SYNACK) to be sent with that fwmark. The
540
listening socket's mark is unchanged. Listening sockets that already
541
have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
542
unaffected.
544
Possible values:
546
- 0 (disabled)
547
- 1 (enabled)
549
Default: 0 (disabled)
551
tcp_invalid_ratelimit - INTEGER
552
Limit the maximal rate for sending duplicate acknowledgments
553
in response to incoming TCP packets that are for an existing
554
connection but that are invalid due to any of these reasons:
556
(a) out-of-window sequence number,
557
(b) out-of-window acknowledgment number, or
558
(c) PAWS (Protection Against Wrapped Sequence numbers) check failure
560
This can help mitigate simple "ack loop" DoS attacks, wherein
561
a buggy or malicious middlebox or man-in-the-middle can
562
rewrite TCP header fields in manner that causes each endpoint
563
to think that the other is sending invalid TCP segments, thus
564
causing each side to send an unterminating stream of duplicate
565
acknowledgments for invalid segments.
567
Using 0 disables rate-limiting of dupacks in response to
568
invalid segments; otherwise this value specifies the minimal
569
space between sending such dupacks, in milliseconds.
571
Default: 500 (milliseconds).
573
tcp_keepalive_time - INTEGER
574
How often TCP sends out keepalive messages when keepalive is enabled.
575
Default: 2hours.
577
tcp_keepalive_probes - INTEGER
578
How many keepalive probes TCP sends out, until it decides that the
579
connection is broken. Default value: 9.
581
tcp_keepalive_intvl - INTEGER
582
How frequently the probes are send out. Multiplied by
583
tcp_keepalive_probes it is time to kill not responding connection,
584
after probes started. Default value: 75sec i.e. connection
585
will be aborted after ~11 minutes of retries.
587
tcp_l3mdev_accept - BOOLEAN
588
Enables child sockets to inherit the L3 master device index.
589
Enabling this option allows a "global" listen socket to work
590
across L3 master domains (e.g., VRFs) with connected sockets
591
derived from the listen socket to be bound to the L3 domain in
592
which the packets originated. Only valid when the kernel was
593
compiled with CONFIG_NET_L3_MASTER_DEV.
595
Possible values:
597
- 0 (disabled)
598
- 1 (enabled)
600
Default: 0 (disabled)
602
tcp_low_latency - BOOLEAN
603
This is a legacy option, it has no effect anymore.
605
tcp_max_orphans - INTEGER
606
Maximal number of TCP sockets not attached to any user file handle,
607
held by system. If this number is exceeded orphaned connections are
608
reset immediately and warning is printed. This limit exists
609
only to prevent simple DoS attacks, you _must_ not rely on this
610
or lower the limit artificially, but rather increase it
611
(probably, after increasing installed memory),
612
if network conditions require more than default value,
613
and tune network services to linger and kill such states
614
more aggressively. Let me to remind again: each orphan eats
615
up to ~64K of unswappable memory.
617
tcp_max_syn_backlog - INTEGER
618
Maximal number of remembered connection requests (SYN_RECV),
619
which have not received an acknowledgment from connecting client.
621
This is a per-listener limit.
623
The minimal value is 128 for low memory machines, and it will
624
increase in proportion to the memory of machine.
626
If server suffers from overload, try increasing this number.
628
Remember to also check /proc/sys/net/core/somaxconn
629
A SYN_RECV request socket consumes about 304 bytes of memory.
631
tcp_max_tw_buckets - INTEGER
632
Maximal number of timewait sockets held by system simultaneously.
633
If this number is exceeded time-wait socket is immediately destroyed
634
and warning is printed. This limit exists only to prevent
635
simple DoS attacks, you _must_ not lower the limit artificially,
636
but rather increase it (probably, after increasing installed memory),
637
if network conditions require more than default value.
639
tcp_mem - vector of 3 INTEGERs: min, pressure, max
640
min: below this number of pages TCP is not bothered about its
641
memory appetite.
643
pressure: when amount of memory allocated by TCP exceeds this number
644
of pages, TCP moderates its memory consumption and enters memory
645
pressure mode, which is exited when memory consumption falls
646
under "min".
648
max: number of pages allowed for queueing by all TCP sockets.
650
Defaults are calculated at boot time from amount of available
651
memory.
653
tcp_min_rtt_wlen - INTEGER
654
The window length of the windowed min filter to track the minimum RTT.
655
A shorter window lets a flow more quickly pick up new (higher)
656
minimum RTT when it is moved to a longer path (e.g., due to traffic
657
engineering). A longer window makes the filter more resistant to RTT
658
inflations such as transient congestion. The unit is seconds.
660
Possible values: 0 - 86400 (1 day)
662
Default: 300
664
tcp_moderate_rcvbuf - BOOLEAN
665
If enabled, TCP performs receive buffer auto-tuning, attempting to
666
automatically size the buffer (no greater than tcp_rmem[2]) to
667
match the size required by the path for full throughput.
669
Possible values:
671
- 0 (disabled)
672
- 1 (enabled)
674
Default: 1 (enabled)
676
tcp_mtu_probing - INTEGER
677
Controls TCP Packetization-Layer Path MTU Discovery. Takes three
678
values:
680
- 0 - Disabled
681
- 1 - Disabled by default, enabled when an ICMP black hole detected
682
- 2 - Always enabled, use initial MSS of tcp_base_mss.
684
tcp_probe_interval - UNSIGNED INTEGER
685
Controls how often to start TCP Packetization-Layer Path MTU
686
Discovery reprobe. The default is reprobing every 10 minutes as
687
per RFC4821.
689
tcp_probe_threshold - INTEGER
690
Controls when TCP Packetization-Layer Path MTU Discovery probing
691
will stop in respect to the width of search range in bytes. Default
692
is 8 bytes.
694
tcp_no_metrics_save - BOOLEAN
695
By default, TCP saves various connection metrics in the route cache
696
when the connection closes, so that connections established in the
697
near future can use these to set initial conditions. Usually, this
698
increases overall performance, but may sometimes cause performance
699
degradation. If enabled, TCP will not cache metrics on closing
700
connections.
702
Possible values:
704
- 0 (disabled)
705
- 1 (enabled)
707
Default: 0 (disabled)
709
tcp_no_ssthresh_metrics_save - BOOLEAN
710
Controls whether TCP saves ssthresh metrics in the route cache.
711
If enabled, ssthresh metrics are disabled.
713
Possible values:
715
- 0 (disabled)
716
- 1 (enabled)
718
Default: 1 (enabled)
720
tcp_orphan_retries - INTEGER
721
This value influences the timeout of a locally closed TCP connection,
722
when RTO retransmissions remain unacknowledged.
723
See tcp_retries2 for more details.
725
The default value is 8.
727
If your machine is a loaded WEB server,
728
you should think about lowering this value, such sockets
729
may consume significant resources. Cf. tcp_max_orphans.
731
tcp_recovery - INTEGER
732
This value is a bitmap to enable various experimental loss recovery
733
features.
735
========= =============================================================
736
RACK: 0x1 enables RACK loss detection, for fast detection of lost
737
retransmissions and tail drops, and resilience to
738
reordering. currently, setting this bit to 0 has no
739
effect, since RACK is the only supported loss detection
740
algorithm.
742
RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
744
RACK: 0x4 disables RACK's DUPACK threshold heuristic
745
========= =============================================================
747
Default: 0x1
749
tcp_reflect_tos - BOOLEAN
750
For listening sockets, reuse the DSCP value of the initial SYN message
751
for outgoing packets. This allows to have both directions of a TCP
752
stream to use the same DSCP value, assuming DSCP remains unchanged for
753
the lifetime of the connection.
755
This options affects both IPv4 and IPv6.
757
Possible values:
759
- 0 (disabled)
760
- 1 (enabled)
762
Default: 0 (disabled)
764
tcp_reordering - INTEGER
765
Initial reordering level of packets in a TCP stream.
766
TCP stack can then dynamically adjust flow reordering level
767
between this initial value and tcp_max_reordering
769
Default: 3
771
tcp_max_reordering - INTEGER
772
Maximal reordering level of packets in a TCP stream.
773
300 is a fairly conservative value, but you might increase it
774
if paths are using per packet load balancing (like bonding rr mode)
776
Default: 300
778
tcp_retrans_collapse - BOOLEAN
779
Bug-to-bug compatibility with some broken printers.
780
On retransmit try to send bigger packets to work around bugs in
781
certain TCP stacks.
783
Possible values:
785
- 0 (disabled)
786
- 1 (enabled)
788
Default: 1 (enabled)
790
tcp_retries1 - INTEGER
791
This value influences the time, after which TCP decides, that
792
something is wrong due to unacknowledged RTO retransmissions,
793
and reports this suspicion to the network layer.
794
See tcp_retries2 for more details.
796
RFC 1122 recommends at least 3 retransmissions, which is the
797
default.
799
tcp_retries2 - INTEGER
800
This value influences the timeout of an alive TCP connection,
801
when RTO retransmissions remain unacknowledged.
802
Given a value of N, a hypothetical TCP connection following
803
exponential backoff with an initial RTO of TCP_RTO_MIN would
804
retransmit N times before killing the connection at the (N+1)th RTO.
806
The default value of 15 yields a hypothetical timeout of 924.6
807
seconds and is a lower bound for the effective timeout.
808
TCP will effectively time out at the first RTO which exceeds the
809
hypothetical timeout.
810
If tcp_rto_max_ms is decreased, it is recommended to also
811
change tcp_retries2.
813
RFC 1122 recommends at least 100 seconds for the timeout,
814
which corresponds to a value of at least 8.
816
tcp_rfc1337 - BOOLEAN
817
If enabled, the TCP stack behaves conforming to RFC1337. If unset,
818
we are not conforming to RFC, but prevent TCP TIME_WAIT
819
assassination.
821
Possible values:
823
- 0 (disabled)
824
- 1 (enabled)
826
Default: 0 (disabled)
828
tcp_rmem - vector of 3 INTEGERs: min, default, max
829
min: Minimal size of receive buffer used by TCP sockets.
830
It is guaranteed to each TCP socket, even under moderate memory
831
pressure.
833
Default: 4K
835
default: initial size of receive buffer used by TCP sockets.
836
This value overrides net.core.rmem_default used by other protocols.
837
Default: 131072 bytes.
838
This value results in initial window of 65535.
840
max: maximal size of receive buffer allowed for automatically
841
selected receiver buffers for TCP socket.
842
Calling setsockopt() with SO_RCVBUF disables
843
automatic tuning of that socket's receive buffer size, in which
844
case this value is ignored.
845
Default: between 131072 and 32MB, depending on RAM size.
847
tcp_sack - BOOLEAN
848
Enable select acknowledgments (SACKS).
850
Possible values:
852
- 0 (disabled)
853
- 1 (enabled)
855
Default: 1 (enabled)
857
tcp_comp_sack_delay_ns - LONG INTEGER
858
TCP tries to reduce number of SACK sent, using a timer
859
based on 5% of SRTT, capped by this sysctl, in nano seconds.
860
The default is 1ms, based on TSO autosizing period.
862
Default : 1,000,000 ns (1 ms)
864
tcp_comp_sack_slack_ns - LONG INTEGER
865
This sysctl control the slack used when arming the
866
timer used by SACK compression. This gives extra time
867
for small RTT flows, and reduces system overhead by allowing
868
opportunistic reduction of timer interrupts.
870
Default : 100,000 ns (100 us)
872
tcp_comp_sack_nr - INTEGER
873
Max number of SACK that can be compressed.
874
Using 0 disables SACK compression.
876
Default : 44
878
tcp_backlog_ack_defer - BOOLEAN
879
If enabled, user thread processing socket backlog tries sending
880
one ACK for the whole queue. This helps to avoid potential
881
long latencies at end of a TCP socket syscall.
883
Possible values:
885
- 0 (disabled)
886
- 1 (enabled)
888
Default: 1 (enabled)
890
tcp_slow_start_after_idle - BOOLEAN
891
If enabled, provide RFC2861 behavior and time out the congestion
892
window after an idle period. An idle period is defined at
893
the current RTO. If unset, the congestion window will not
894
be timed out after an idle period.
896
Possible values:
898
- 0 (disabled)
899
- 1 (enabled)
901
Default: 1 (enabled)
903
tcp_stdurg - BOOLEAN
904
Use the Host requirements interpretation of the TCP urgent pointer field.
905
Most hosts use the older BSD interpretation, so if enabled,
906
Linux might not communicate correctly with them.
908
Possible values:
910
- 0 (disabled)
911
- 1 (enabled)
913
Default: 0 (disabled)
915
tcp_synack_retries - INTEGER
916
Number of times SYNACKs for a passive TCP connection attempt will
917
be retransmitted. Should not be higher than 255. Default value
918
is 5, which corresponds to 31seconds till the last retransmission
919
with the current initial RTO of 1second. With this the final timeout
920
for a passive TCP connection will happen after 63seconds.
922
tcp_syncookies - INTEGER
923
Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
924
Send out syncookies when the syn backlog queue of a socket
925
overflows. This is to prevent against the common 'SYN flood attack'
926
Default: 1
928
Note, that syncookies is fallback facility.
929
It MUST NOT be used to help highly loaded servers to stand
930
against legal connection rate. If you see SYN flood warnings
931
in your logs, but investigation shows that they occur
932
because of overload with legal connections, you should tune
933
another parameters until this warning disappear.
934
See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
936
syncookies seriously violate TCP protocol, do not allow
937
to use TCP extensions, can result in serious degradation
938
of some services (f.e. SMTP relaying), visible not by you,
939
but your clients and relays, contacting you. While you see
940
SYN flood warnings in logs not being really flooded, your server
941
is seriously misconfigured.
943
If you want to test which effects syncookies have to your
944
network connections you can set this knob to 2 to enable
945
unconditionally generation of syncookies.
947
tcp_migrate_req - BOOLEAN
948
The incoming connection is tied to a specific listening socket when
949
the initial SYN packet is received during the three-way handshake.
950
When a listener is closed, in-flight request sockets during the
951
handshake and established sockets in the accept queue are aborted.
953
If the listener has SO_REUSEPORT enabled, other listeners on the
954
same port should have been able to accept such connections. This
955
option makes it possible to migrate such child sockets to another
956
listener after close() or shutdown().
958
The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should
959
usually be used to define the policy to pick an alive listener.
960
Otherwise, the kernel will randomly pick an alive listener only if
961
this option is enabled.
963
Note that migration between listeners with different settings may
964
crash applications. Let's say migration happens from listener A to
965
B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from
966
the requests migrated from A. To avoid such a situation, cancel
967
migration by returning SK_DROP in the type of eBPF program, or
968
disable this option.
970
Possible values:
972
- 0 (disabled)
973
- 1 (enabled)
975
Default: 0 (disabled)
977
tcp_fastopen - INTEGER
978
Enable TCP Fast Open (RFC7413) to send and accept data in the opening
979
SYN packet.
981
The client support is enabled by flag 0x1 (on by default). The client
982
then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
983
rather than connect() to send data in SYN.
985
The server support is enabled by flag 0x2 (off by default). Then
986
either enable for all listeners with another flag (0x400) or
987
enable individual listeners via TCP_FASTOPEN socket option with
988
the option value being the length of the syn-data backlog.
990
The values (bitmap) are
992
===== ======== ======================================================
993
0x1 (client) enables sending data in the opening SYN on the client.
994
0x2 (server) enables the server support, i.e., allowing data in
995
a SYN packet to be accepted and passed to the
996
application before 3-way handshake finishes.
997
0x4 (client) send data in the opening SYN regardless of cookie
998
availability and without a cookie option.
999
0x200 (server) accept data-in-SYN w/o any cookie option present.
1000
0x400 (server) enable all listeners to support Fast Open by
1001
default without explicit TCP_FASTOPEN socket option.
1002
===== ======== ======================================================
1004
Default: 0x1
1006
Note that additional client or server features are only
1007
effective if the basic support (0x1 and 0x2) are enabled respectively.
1009
tcp_fastopen_blackhole_timeout_sec - INTEGER
1010
Initial time period in second to disable Fastopen on active TCP sockets
1011
when a TFO firewall blackhole issue happens.
1012
This time period will grow exponentially when more blackhole issues
1013
get detected right after Fastopen is re-enabled and will reset to
1014
initial value when the blackhole issue goes away.
1015
0 to disable the blackhole detection.
1017
By default, it is set to 0 (feature is disabled).
1019
tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
1020
The list consists of a primary key and an optional backup key. The
1021
primary key is used for both creating and validating cookies, while the
1022
optional backup key is only used for validating cookies. The purpose of
1023
the backup key is to maximize TFO validation when keys are rotated.
1025
A randomly chosen primary key may be configured by the kernel if
1026
the tcp_fastopen sysctl is set to 0x400 (see above), or if the
1027
TCP_FASTOPEN setsockopt() optname is set and a key has not been
1028
previously configured via sysctl. If keys are configured via
1029
setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
1030
per-socket keys will be used instead of any keys that are specified via
1031
sysctl.
1033
A key is specified as 4 8-digit hexadecimal integers which are separated
1034
by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
1035
omitted. A primary and a backup key may be specified by separating them
1036
by a comma. If only one key is specified, it becomes the primary key and
1037
any previously configured backup keys are removed.
1039
tcp_syn_retries - INTEGER
1040
Number of times initial SYNs for an active TCP connection attempt
1041
will be retransmitted. Should not be higher than 127. Default value
1042
is 6, which corresponds to 67seconds (with tcp_syn_linear_timeouts = 4)
1043
till the last retransmission with the current initial RTO of 1second.
1044
With this the final timeout for an active TCP connection attempt
1045
will happen after 131seconds.
1047
tcp_timestamps - INTEGER
1048
Enable timestamps as defined in RFC1323.
1050
- 0: Disabled.
1051
- 1: Enable timestamps as defined in RFC1323 and use random offset for
1052
each connection rather than only using the current time.
1053
- 2: Like 1, but without random offsets.
1055
Default: 1
1057
tcp_min_tso_segs - INTEGER
1058
Minimal number of segments per TSO frame.
1060
Since linux-3.12, TCP does an automatic sizing of TSO frames,
1061
depending on flow rate, instead of filling 64Kbytes packets.
1062
For specific usages, it's possible to force TCP to build big
1063
TSO frames. Note that TCP stack might split too big TSO packets
1064
if available window is too small.
1066
Default: 2
1068
tcp_tso_rtt_log - INTEGER
1069
Adjustment of TSO packet sizes based on min_rtt
1071
Starting from linux-5.18, TCP autosizing can be tweaked
1072
for flows having small RTT.
1074
Old autosizing was splitting the pacing budget to send 1024 TSO
1075
per second.
1077
tso_packet_size = sk->sk_pacing_rate / 1024;
1079
With the new mechanism, we increase this TSO sizing using:
1081
distance = min_rtt_usec / (2^tcp_tso_rtt_log)
1082
tso_packet_size += gso_max_size >> distance;
1084
This means that flows between very close hosts can use bigger
1085
TSO packets, reducing their cpu costs.
1087
If you want to use the old autosizing, set this sysctl to 0.
1089
Default: 9 (2^9 = 512 usec)
1091
tcp_pacing_ss_ratio - INTEGER
1092
sk->sk_pacing_rate is set by TCP stack using a ratio applied
1093
to current rate. (current_rate = cwnd * mss / srtt)
1094
If TCP is in slow start, tcp_pacing_ss_ratio is applied
1095
to let TCP probe for bigger speeds, assuming cwnd can be
1096
doubled every other RTT.
1098
Default: 200
1100
tcp_pacing_ca_ratio - INTEGER
1101
sk->sk_pacing_rate is set by TCP stack using a ratio applied
1102
to current rate. (current_rate = cwnd * mss / srtt)
1103
If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
1104
is applied to conservatively probe for bigger throughput.
1106
Default: 120
1108
tcp_syn_linear_timeouts - INTEGER
1109
The number of times for an active TCP connection to retransmit SYNs with
1110
a linear backoff timeout before defaulting to an exponential backoff
1111
timeout. This has no effect on SYNACK at the passive TCP side.
1113
With an initial RTO of 1 and tcp_syn_linear_timeouts = 4 we would
1114
expect SYN RTOs to be: 1, 1, 1, 1, 1, 2, 4, ... (4 linear timeouts,
1115
and the first exponential backoff using 2^0 * initial_RTO).
1116
Default: 4
1118
tcp_tso_win_divisor - INTEGER
1119
This allows control over what percentage of the congestion window
1120
can be consumed by a single TSO frame.
1121
The setting of this parameter is a choice between burstiness and
1122
building larger TSO frames.
1124
Default: 3
1126
tcp_tw_reuse - INTEGER
1127
Enable reuse of TIME-WAIT sockets for new connections when it is
1128
safe from protocol viewpoint.
1130
- 0 - disable
1131
- 1 - global enable
1132
- 2 - enable for loopback traffic only
1134
It should not be changed without advice/request of technical
1135
experts.
1137
Default: 2
1139
tcp_tw_reuse_delay - UNSIGNED INTEGER
1140
The delay in milliseconds before a TIME-WAIT socket can be reused by a
1141
new connection, if TIME-WAIT socket reuse is enabled. The actual reuse
1142
threshold is within [N, N+1] range, where N is the requested delay in
1143
milliseconds, to ensure the delay interval is never shorter than the
1144
configured value.
1146
This setting contains an assumption about the other TCP timestamp clock
1147
tick interval. It should not be set to a value lower than the peer's
1148
clock tick for PAWS (Protection Against Wrapped Sequence numbers)
1149
mechanism work correctly for the reused connection.
1151
Default: 1000 (milliseconds)
1153
tcp_window_scaling - BOOLEAN
1154
Enable window scaling as defined in RFC1323.
1156
Possible values:
1158
- 0 (disabled)
1159
- 1 (enabled)
1161
Default: 1 (enabled)
1163
tcp_shrink_window - BOOLEAN
1164
This changes how the TCP receive window is calculated.
1166
RFC 7323, section 2.4, says there are instances when a retracted
1167
window can be offered, and that TCP implementations MUST ensure
1168
that they handle a shrinking window, as specified in RFC 1122.
1170
Possible values:
1172
- 0 (disabled) - The window is never shrunk.
1173
- 1 (enabled) - The window is shrunk when necessary to remain within
1174
the memory limit set by autotuning (sk_rcvbuf).
1175
This only occurs if a non-zero receive window
1176
scaling factor is also in effect.
1178
Default: 0 (disabled)
1180
tcp_wmem - vector of 3 INTEGERs: min, default, max
1181
min: Amount of memory reserved for send buffers for TCP sockets.
1182
Each TCP socket has rights to use it due to fact of its birth.
1184
Default: 4K
1186
default: initial size of send buffer used by TCP sockets. This
1187
value overrides net.core.wmem_default used by other protocols.
1189
It is usually lower than net.core.wmem_default.
1191
Default: 16K
1193
max: Maximal amount of memory allowed for automatically tuned
1194
send buffers for TCP sockets. This value does not override
1195
net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
1196
automatic tuning of that socket's send buffer size, in which case
1197
this value is ignored.
1199
Default: between 64K and 4MB, depending on RAM size.
1201
tcp_notsent_lowat - UNSIGNED INTEGER
1202
A TCP socket can control the amount of unsent bytes in its write queue,
1203
thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
1204
reports POLLOUT events if the amount of unsent bytes is below a per
1205
socket value, and if the write queue is not full. sendmsg() will
1206
also not add new buffers if the limit is hit.
1208
This global variable controls the amount of unsent data for
1209
sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
1210
to the global variable has immediate effect.
1212
Default: UINT_MAX (0xFFFFFFFF)
1214
tcp_workaround_signed_windows - BOOLEAN
1215
If enabled, assume no receipt of a window scaling option means the
1216
remote TCP is broken and treats the window as a signed quantity.
1217
If disabled, assume the remote TCP is not broken even if we do
1218
not receive a window scaling option from them.
1220
Possible values:
1222
- 0 (disabled)
1223
- 1 (enabled)
1225
Default: 0 (disabled)
1227
tcp_thin_linear_timeouts - BOOLEAN
1228
Enable dynamic triggering of linear timeouts for thin streams.
1229
If enabled, a check is performed upon retransmission by timeout to
1230
determine if the stream is thin (less than 4 packets in flight).
1231
As long as the stream is found to be thin, up to 6 linear
1232
timeouts may be performed before exponential backoff mode is
1233
initiated. This improves retransmission latency for
1234
non-aggressive thin streams, often found to be time-dependent.
1235
For more information on thin streams, see
1236
Documentation/networking/tcp-thin.rst
1238
Possible values:
1240
- 0 (disabled)
1241
- 1 (enabled)
1243
Default: 0 (disabled)
1245
tcp_limit_output_bytes - INTEGER
1246
Controls TCP Small Queue limit per tcp socket.
1247
TCP bulk sender tends to increase packets in flight until it
1248
gets losses notifications. With SNDBUF autotuning, this can
1249
result in a large amount of packets queued on the local machine
1250
(e.g.: qdiscs, CPU backlog, or device) hurting latency of other
1251
flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes
1252
limits the number of bytes on qdisc or device to reduce artificial
1253
RTT/cwnd and reduce bufferbloat.
1255
Default: 4194304 (4 MB)
1257
tcp_challenge_ack_limit - INTEGER
1258
Limits number of Challenge ACK sent per second, as recommended
1259
in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
1260
Note that this per netns rate limit can allow some side channel
1261
attacks and probably should not be enabled.
1262
TCP stack implements per TCP socket limits anyway.
1263
Default: INT_MAX (unlimited)
1265
tcp_ehash_entries - INTEGER
1266
Show the number of hash buckets for TCP sockets in the current
1267
networking namespace.
1269
A negative value means the networking namespace does not own its
1270
hash buckets and shares the initial networking namespace's one.
1272
tcp_child_ehash_entries - INTEGER
1273
Control the number of hash buckets for TCP sockets in the child
1274
networking namespace, which must be set before clone() or unshare().
1276
If the value is not 0, the kernel uses a value rounded up to 2^n
1277
as the actual hash bucket size. 0 is a special value, meaning
1278
the child networking namespace will share the initial networking
1279
namespace's hash buckets.
1281
Note that the child will use the global one in case the kernel
1282
fails to allocate enough memory. In addition, the global hash
1283
buckets are spread over available NUMA nodes, but the allocation
1284
of the child hash table depends on the current process's NUMA
1285
policy, which could result in performance differences.
1287
Note also that the default value of tcp_max_tw_buckets and
1288
tcp_max_syn_backlog depend on the hash bucket size.
1290
Possible values: 0, 2^n (n: 0 - 24 (16Mi))
1292
Default: 0
1294
tcp_plb_enabled - BOOLEAN
1295
If enabled and the underlying congestion control (e.g. DCTCP) supports
1296
and enables PLB feature, TCP PLB (Protective Load Balancing) is
1297
enabled. PLB is described in the following paper:
1298
https://doi.org/10.1145/3544216.3544226. Based on PLB parameters,
1299
upon sensing sustained congestion, TCP triggers a change in
1300
flow label field for outgoing IPv6 packets. A change in flow label
1301
field potentially changes the path of outgoing packets for switches
1302
that use ECMP/WCMP for routing.
1304
PLB changes socket txhash which results in a change in IPv6 Flow Label
1305
field, and currently no-op for IPv4 headers. It is possible
1306
to apply PLB for IPv4 with other network header fields (e.g. TCP
1307
or IPv4 options) or using encapsulation where outer header is used
1308
by switches to determine next hop. In either case, further host
1309
and switch side changes will be needed.
1311
If enabled, PLB assumes that congestion signal (e.g. ECN) is made
1312
available and used by congestion control module to estimate a
1313
congestion measure (e.g. ce_ratio). PLB needs a congestion measure to
1314
make repathing decisions.
1316
Possible values:
1318
- 0 (disabled)
1319
- 1 (enabled)
1321
Default: 0 (disabled)
1323
tcp_plb_idle_rehash_rounds - INTEGER
1324
Number of consecutive congested rounds (RTT) seen after which
1325
a rehash can be performed, given there are no packets in flight.
1326
This is referred to as M in PLB paper:
1327
https://doi.org/10.1145/3544216.3544226.
1329
Possible Values: 0 - 31
1331
Default: 3
1333
tcp_plb_rehash_rounds - INTEGER
1334
Number of consecutive congested rounds (RTT) seen after which
1335
a forced rehash can be performed. Be careful when setting this
1336
parameter, as a small value increases the risk of retransmissions.
1337
This is referred to as N in PLB paper:
1338
https://doi.org/10.1145/3544216.3544226.
1340
Possible Values: 0 - 31
1342
Default: 12
1344
tcp_plb_suspend_rto_sec - INTEGER
1345
Time, in seconds, to suspend PLB in event of an RTO. In order to avoid
1346
having PLB repath onto a connectivity "black hole", after an RTO a TCP
1347
connection suspends PLB repathing for a random duration between 1x and
1348
2x of this parameter. Randomness is added to avoid concurrent rehashing
1349
of multiple TCP connections. This should be set corresponding to the
1350
amount of time it takes to repair a failed link.
1352
Possible Values: 0 - 255
1354
Default: 60
1356
tcp_plb_cong_thresh - INTEGER
1357
Fraction of packets marked with congestion over a round (RTT) to
1358
tag that round as congested. This is referred to as K in the PLB paper:
1359
https://doi.org/10.1145/3544216.3544226.
1361
The 0-1 fraction range is mapped to 0-256 range to avoid floating
1362
point operations. For example, 128 means that if at least 50% of
1363
the packets in a round were marked as congested then the round
1364
will be tagged as congested.
1366
Setting threshold to 0 means that PLB repaths every RTT regardless
1367
of congestion. This is not intended behavior for PLB and should be
1368
used only for experimentation purpose.
1370
Possible Values: 0 - 256
1372
Default: 128
1374
tcp_pingpong_thresh - INTEGER
1375
The number of estimated data replies sent for estimated incoming data
1376
requests that must happen before TCP considers that a connection is a
1377
"ping-pong" (request-response) connection for which delayed
1378
acknowledgments can provide benefits.
1380
This threshold is 1 by default, but some applications may need a higher
1381
threshold for optimal performance.
1383
Possible Values: 1 - 255
1385
Default: 1
1387
tcp_rto_min_us - INTEGER
1388
Minimal TCP retransmission timeout (in microseconds). Note that the
1389
rto_min route option has the highest precedence for configuring this
1390
setting, followed by the TCP_BPF_RTO_MIN and TCP_RTO_MIN_US socket
1391
options, followed by this tcp_rto_min_us sysctl.
1393
The recommended practice is to use a value less or equal to 200000
1394
microseconds.
1396
Possible Values: 1 - INT_MAX
1398
Default: 200000
1400
tcp_rto_max_ms - INTEGER
1401
Maximal TCP retransmission timeout (in ms).
1402
Note that TCP_RTO_MAX_MS socket option has higher precedence.
1404
When changing tcp_rto_max_ms, it is important to understand
1405
that tcp_retries2 might need a change.
1407
Possible Values: 1000 - 120,000
1409
Default: 120,000
1411
UDP variables
1412
=============
1414
udp_l3mdev_accept - BOOLEAN
1415
Enabling this option allows a "global" bound socket to work
1416
across L3 master domains (e.g., VRFs) with packets capable of
1417
being received regardless of the L3 domain in which they
1418
originated. Only valid when the kernel was compiled with
1419
CONFIG_NET_L3_MASTER_DEV.
1421
Possible values:
1423
- 0 (disabled)
1424
- 1 (enabled)
1426
Default: 0 (disabled)
1428
udp_mem - vector of 3 INTEGERs: min, pressure, max
1429
Number of pages allowed for queueing by all UDP sockets.
1431
min: Number of pages allowed for queueing by all UDP sockets.
1433
pressure: This value was introduced to follow format of tcp_mem.
1435
max: This value was introduced to follow format of tcp_mem.
1437
Default is calculated at boot time from amount of available memory.
1439
udp_rmem_min - INTEGER
1440
Minimal size of receive buffer used by UDP sockets in moderation.
1441
Each UDP socket is able to use the size for receiving data, even if
1442
total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
1444
Default: 4K
1446
udp_wmem_min - INTEGER
1447
UDP does not have tx memory accounting and this tunable has no effect.
1449
udp_hash_entries - INTEGER
1450
Show the number of hash buckets for UDP sockets in the current
1451
networking namespace.
1453
A negative value means the networking namespace does not own its
1454
hash buckets and shares the initial networking namespace's one.
1456
udp_child_hash_entries - INTEGER
1457
Control the number of hash buckets for UDP sockets in the child
1458
networking namespace, which must be set before clone() or unshare().
1460
If the value is not 0, the kernel uses a value rounded up to 2^n
1461
as the actual hash bucket size. 0 is a special value, meaning
1462
the child networking namespace will share the initial networking
1463
namespace's hash buckets.
1465
Note that the child will use the global one in case the kernel
1466
fails to allocate enough memory. In addition, the global hash
1467
buckets are spread over available NUMA nodes, but the allocation
1468
of the child hash table depends on the current process's NUMA
1469
policy, which could result in performance differences.
1471
Possible values: 0, 2^n (n: 7 (128) - 16 (64K))
1473
Default: 0
1476
RAW variables
1477
=============
1479
raw_l3mdev_accept - BOOLEAN
1480
Enabling this option allows a "global" bound socket to work
1481
across L3 master domains (e.g., VRFs) with packets capable of
1482
being received regardless of the L3 domain in which they
1483
originated. Only valid when the kernel was compiled with
1484
CONFIG_NET_L3_MASTER_DEV.
1486
Possible values:
1488
- 0 (disabled)
1489
- 1 (enabled)
1491
Default: 1 (enabled)
1493
CIPSOv4 Variables
1494
=================
1496
cipso_cache_enable - BOOLEAN
1497
If enabled, enable additions to and lookups from the CIPSO label mapping
1498
cache. If disabled, additions are ignored and lookups always result in a
1499
miss. However, regardless of the setting the cache is still
1500
invalidated when required when means you can safely toggle this on and
1501
off and the cache will always be "safe".
1503
Possible values:
1505
- 0 (disabled)
1506
- 1 (enabled)
1508
Default: 1 (enabled)
1510
cipso_cache_bucket_size - INTEGER
1511
The CIPSO label cache consists of a fixed size hash table with each
1512
hash bucket containing a number of cache entries. This variable limits
1513
the number of entries in each hash bucket; the larger the value is, the
1514
more CIPSO label mappings that can be cached. When the number of
1515
entries in a given hash bucket reaches this limit adding new entries
1516
causes the oldest entry in the bucket to be removed to make room.
1518
Default: 10
1520
cipso_rbm_optfmt - BOOLEAN
1521
Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
1522
the CIPSO draft specification (see Documentation/netlabel for details).
1523
This means that when set the CIPSO tag will be padded with empty
1524
categories in order to make the packet data 32-bit aligned.
1526
Possible values:
1528
- 0 (disabled)
1529
- 1 (enabled)
1531
Default: 0 (disabled)
1533
cipso_rbm_strictvalid - BOOLEAN
1534
If enabled, do a very strict check of the CIPSO option when
1535
ip_options_compile() is called. If disabled, relax the checks done during
1536
ip_options_compile(). Either way is "safe" as errors are caught else
1537
where in the CIPSO processing code but setting this to 0 (False) should
1538
result in less work (i.e. it should be faster) but could cause problems
1539
with other implementations that require strict checking.
1541
Possible values:
1543
- 0 (disabled)
1544
- 1 (enabled)
1546
Default: 0 (disabled)
1548
IP Variables
1549
============
1551
ip_local_port_range - 2 INTEGERS
1552
Defines the local port range that is used by TCP and UDP to
1553
choose the local port. The first number is the first, the
1554
second the last local port number.
1555
If possible, it is better these numbers have different parity
1556
(one even and one odd value).
1557
Must be greater than or equal to ip_unprivileged_port_start.
1558
The default values are 32768 and 60999 respectively.
1560
ip_local_reserved_ports - list of comma separated ranges
1561
Specify the ports which are reserved for known third-party
1562
applications. These ports will not be used by automatic port
1563
assignments (e.g. when calling connect() or bind() with port
1564
number 0). Explicit port allocation behavior is unchanged.
1566
The format used for both input and output is a comma separated
1567
list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
1568
10). Writing to the file will clear all previously reserved
1569
ports and update the current list with the one given in the
1570
input.
1572
Note that ip_local_port_range and ip_local_reserved_ports
1573
settings are independent and both are considered by the kernel
1574
when determining which ports are available for automatic port
1575
assignments.
1577
You can reserve ports which are not in the current
1578
ip_local_port_range, e.g.::
1580
$ cat /proc/sys/net/ipv4/ip_local_port_range
1581
32000 60999
1582
$ cat /proc/sys/net/ipv4/ip_local_reserved_ports
1583
8080,9148
1585
although this is redundant. However such a setting is useful
1586
if later the port range is changed to a value that will
1587
include the reserved ports. Also keep in mind, that overlapping
1588
of these ranges may affect probability of selecting ephemeral
1589
ports which are right after block of reserved ports.
1591
Default: Empty
1593
ip_unprivileged_port_start - INTEGER
1594
This is a per-namespace sysctl. It defines the first
1595
unprivileged port in the network namespace. Privileged ports
1596
require root or CAP_NET_BIND_SERVICE in order to bind to them.
1597
To disable all privileged ports, set this to 0. They must not
1598
overlap with the ip_local_port_range.
1600
Default: 1024
1602
ip_nonlocal_bind - BOOLEAN
1603
If enabled, allows processes to bind() to non-local IP addresses,
1604
which can be quite useful - but may break some applications.
1606
Possible values:
1608
- 0 (disabled)
1609
- 1 (enabled)
1611
Default: 0 (disabled)
1613
ip_autobind_reuse - BOOLEAN
1614
By default, bind() does not select the ports automatically even if
1615
the new socket and all sockets bound to the port have SO_REUSEADDR.
1616
ip_autobind_reuse allows bind() to reuse the port and this is useful
1617
when you use bind()+connect(), but may break some applications.
1618
The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
1619
option should only be set by experts.
1621
Possible values:
1623
- 0 (disabled)
1624
- 1 (enabled)
1626
Default: 0 (disabled)
1628
ip_dynaddr - INTEGER
1629
If set non-zero, enables support for dynamic addresses.
1630
If set to a non-zero value larger than 1, a kernel log
1631
message will be printed when dynamic address rewriting
1632
occurs.
1634
Default: 0
1636
ip_early_demux - BOOLEAN
1637
Optimize input packet processing down to one demux for
1638
certain kinds of local sockets. Currently we only do this
1639
for established TCP and connected UDP sockets.
1641
It may add an additional cost for pure routing workloads that
1642
reduces overall throughput, in such case you should disable it.
1644
Possible values:
1646
- 0 (disabled)
1647
- 1 (enabled)
1649
Default: 1 (enabled)
1651
ping_group_range - 2 INTEGERS
1652
Restrict ICMP_PROTO datagram sockets to users in the group range.
1653
The default is "1 0", meaning, that nobody (not even root) may
1654
create ping sockets. Setting it to "100 100" would grant permissions
1655
to the single group. "0 4294967294" would enable it for the world, "100
1656
4294967294" would enable it for the users, but not daemons.
1658
tcp_early_demux - BOOLEAN
1659
Enable early demux for established TCP sockets.
1661
Possible values:
1663
- 0 (disabled)
1664
- 1 (enabled)
1666
Default: 1 (enabled)
1668
udp_early_demux - BOOLEAN
1669
Enable early demux for connected UDP sockets. Disable this if
1670
your system could experience more unconnected load.
1672
Possible values:
1674
- 0 (disabled)
1675
- 1 (enabled)
1677
Default: 1 (enabled)
1679
icmp_echo_ignore_all - BOOLEAN
1680
If enabled, then the kernel will ignore all ICMP ECHO
1681
requests sent to it.
1683
Possible values:
1685
- 0 (disabled)
1686
- 1 (enabled)
1688
Default: 0 (disabled)
1690
icmp_echo_enable_probe - BOOLEAN
1691
If enabled, then the kernel will respond to RFC 8335 PROBE
1692
requests sent to it.
1694
Possible values:
1696
- 0 (disabled)
1697
- 1 (enabled)
1699
Default: 0 (disabled)
1701
icmp_echo_ignore_broadcasts - BOOLEAN
1702
If enabled, then the kernel will ignore all ICMP ECHO and
1703
TIMESTAMP requests sent to it via broadcast/multicast.
1705
Possible values:
1707
- 0 (disabled)
1708
- 1 (enabled)
1710
Default: 1 (enabled)
1712
icmp_ratelimit - INTEGER
1713
Limit the maximal rates for sending ICMP packets whose type matches
1714
icmp_ratemask (see below) to specific targets.
1715
0 to disable any limiting,
1716
otherwise the minimal space between responses in milliseconds.
1717
Note that another sysctl, icmp_msgs_per_sec limits the number
1718
of ICMP packets sent on all targets.
1720
Default: 1000
1722
icmp_msgs_per_sec - INTEGER
1723
Limit maximal number of ICMP packets sent per second from this host.
1724
Only messages whose type matches icmp_ratemask (see below) are
1725
controlled by this limit. For security reasons, the precise count
1726
of messages per second is randomized.
1728
Default: 1000
1730
icmp_msgs_burst - INTEGER
1731
icmp_msgs_per_sec controls number of ICMP packets sent per second,
1732
while icmp_msgs_burst controls the burst size of these packets.
1733
For security reasons, the precise burst size is randomized.
1735
Default: 50
1737
icmp_ratemask - INTEGER
1738
Mask made of ICMP types for which rates are being limited.
1740
Significant bits: IHGFEDCBA9876543210
1742
Default mask: 0000001100000011000 (6168)
1744
Bit definitions (see include/linux/icmp.h):
1746
= =========================
1747
0 Echo Reply
1748
3 Destination Unreachable [1]_
1749
4 Source Quench [1]_
1750
5 Redirect
1751
8 Echo Request
1752
B Time Exceeded [1]_
1753
C Parameter Problem [1]_
1754
D Timestamp Request
1755
E Timestamp Reply
1756
F Info Request
1757
G Info Reply
1758
H Address Mask Request
1759
I Address Mask Reply
1760
= =========================
1762
.. [1] These are rate limited by default (see default mask above)
1764
icmp_ignore_bogus_error_responses - BOOLEAN
1765
Some routers violate RFC1122 by sending bogus responses to broadcast
1766
frames. Such violations are normally logged via a kernel warning.
1767
If enabled, the kernel will not give such warnings, which
1768
will avoid log file clutter.
1770
Possible values:
1772
- 0 (disabled)
1773
- 1 (enabled)
1775
Default: 1 (enabled)
1777
icmp_errors_use_inbound_ifaddr - BOOLEAN
1779
If disabled, icmp error messages are sent with the primary address of
1780
the exiting interface.
1782
If enabled, the message will be sent with the primary address of
1783
the interface that received the packet that caused the icmp error.
1784
This is the behaviour many network administrators will expect from
1785
a router. And it can make debugging complicated network layouts
1786
much easier.
1788
Note that if no primary address exists for the interface selected,
1789
then the primary address of the first non-loopback interface that
1790
has one will be used regardless of this setting.
1792
Possible values:
1794
- 0 (disabled)
1795
- 1 (enabled)
1797
Default: 0 (disabled)
1799
igmp_max_memberships - INTEGER
1800
Change the maximum number of multicast groups we can subscribe to.
1801
Default: 20
1803
Theoretical maximum value is bounded by having to send a membership
1804
report in a single datagram (i.e. the report can't span multiple
1805
datagrams, or risk confusing the switch and leaving groups you don't
1806
intend to).
1808
The number of supported groups 'M' is bounded by the number of group
1809
report entries you can fit into a single datagram of 65535 bytes.
1811
M = 65536-sizeof (ip header)/(sizeof(Group record))
1813
Group records are variable length, with a minimum of 12 bytes.
1814
So net.ipv4.igmp_max_memberships should not be set higher than:
1816
(65536-24) / 12 = 5459
1818
The value 5459 assumes no IP header options, so in practice
1819
this number may be lower.
1821
igmp_max_msf - INTEGER
1822
Maximum number of addresses allowed in the source filter list for a
1823
multicast group.
1825
Default: 10
1827
igmp_qrv - INTEGER
1828
Controls the IGMP query robustness variable (see RFC2236 8.1).
1830
Default: 2 (as specified by RFC2236 8.1)
1832
Minimum: 1 (as specified by RFC6636 4.5)
1834
force_igmp_version - INTEGER
1835
- 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
1836
allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
1837
Present timer expires.
1838
- 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
1839
receive IGMPv2/v3 query.
1840
- 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
1841
IGMPv1 query message. Will reply report if receive IGMPv3 query.
1842
- 3 - Enforce to use IGMP version 3. The same react with default 0.
1844
.. note::
1846
this is not the same with force_mld_version because IGMPv3 RFC3376
1847
Security Considerations does not have clear description that we could
1848
ignore other version messages completely as MLDv2 RFC3810. So make
1849
this value as default 0 is recommended.
1851
``conf/interface/*``
1852
changes special settings per interface (where
1853
interface" is the name of your network interface)
1855
``conf/all/*``
1856
is special, changes the settings for all interfaces
1858
log_martians - BOOLEAN
1859
Log packets with impossible addresses to kernel log.
1860
log_martians for the interface will be enabled if at least one of
1861
conf/{all,interface}/log_martians is set to TRUE,
1862
it will be disabled otherwise
1864
accept_redirects - BOOLEAN
1865
Accept ICMP redirect messages.
1866
accept_redirects for the interface will be enabled if:
1868
- both conf/{all,interface}/accept_redirects are TRUE in the case
1869
forwarding for the interface is enabled
1871
or
1873
- at least one of conf/{all,interface}/accept_redirects is TRUE in the
1874
case forwarding for the interface is disabled
1876
accept_redirects for the interface will be disabled otherwise
1878
default:
1880
- TRUE (host)
1881
- FALSE (router)
1883
forwarding - BOOLEAN
1884
Enable IP forwarding on this interface. This controls whether packets
1885
received _on_ this interface can be forwarded.
1887
mc_forwarding - BOOLEAN
1888
Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
1889
and a multicast routing daemon is required.
1890
conf/all/mc_forwarding must also be set to TRUE to enable multicast
1891
routing for the interface
1893
medium_id - INTEGER
1894
Integer value used to differentiate the devices by the medium they
1895
are attached to. Two devices can have different id values when
1896
the broadcast packets are received only on one of them.
1897
The default value 0 means that the device is the only interface
1898
to its medium, value of -1 means that medium is not known.
1900
Currently, it is used to change the proxy_arp behavior:
1901
the proxy_arp feature is enabled for packets forwarded between
1902
two devices attached to different media.
1904
proxy_arp - BOOLEAN
1905
Do proxy arp.
1907
proxy_arp for the interface will be enabled if at least one of
1908
conf/{all,interface}/proxy_arp is set to TRUE,
1909
it will be disabled otherwise
1911
proxy_arp_pvlan - BOOLEAN
1912
Private VLAN proxy arp.
1914
Basically allow proxy arp replies back to the same interface
1915
(from which the ARP request/solicitation was received).
1917
This is done to support (ethernet) switch features, like RFC
1918
3069, where the individual ports are NOT allowed to
1919
communicate with each other, but they are allowed to talk to
1920
the upstream router. As described in RFC 3069, it is possible
1921
to allow these hosts to communicate through the upstream
1922
router by proxy_arp'ing. Don't need to be used together with
1923
proxy_arp.
1925
This technology is known by different names:
1927
- In RFC 3069 it is called VLAN Aggregation.
1928
- Cisco and Allied Telesyn call it Private VLAN.
1929
- Hewlett-Packard call it Source-Port filtering or port-isolation.
1930
- Ericsson call it MAC-Forced Forwarding (RFC Draft).
1932
proxy_delay - INTEGER
1933
Delay proxy response.
1935
Delay response to a neighbor solicitation when proxy_arp
1936
or proxy_ndp is enabled. A random value between [0, proxy_delay)
1937
will be chosen, setting to zero means reply with no delay.
1938
Value in jiffies. Defaults to 80.
1940
shared_media - BOOLEAN
1941
Send(router) or accept(host) RFC1620 shared media redirects.
1942
Overrides secure_redirects.
1944
shared_media for the interface will be enabled if at least one of
1945
conf/{all,interface}/shared_media is set to TRUE,
1946
it will be disabled otherwise
1948
default TRUE
1950
secure_redirects - BOOLEAN
1951
Accept ICMP redirect messages only to gateways listed in the
1952
interface's current gateway list. Even if disabled, RFC1122 redirect
1953
rules still apply.
1955
Overridden by shared_media.
1957
secure_redirects for the interface will be enabled if at least one of
1958
conf/{all,interface}/secure_redirects is set to TRUE,
1959
it will be disabled otherwise
1961
default TRUE
1963
send_redirects - BOOLEAN
1964
Send redirects, if router.
1966
send_redirects for the interface will be enabled if at least one of
1967
conf/{all,interface}/send_redirects is set to TRUE,
1968
it will be disabled otherwise
1970
Default: TRUE
1972
bootp_relay - BOOLEAN
1973
Accept packets with source address 0.b.c.d destined
1974
not to this host as local ones. It is supposed, that
1975
BOOTP relay daemon will catch and forward such packets.
1976
conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
1977
for the interface
1979
default FALSE
1981
Not Implemented Yet.
1983
accept_source_route - BOOLEAN
1984
Accept packets with SRR option.
1985
conf/all/accept_source_route must also be set to TRUE to accept packets
1986
with SRR option on the interface
1988
default
1990
- TRUE (router)
1991
- FALSE (host)
1993
accept_local - BOOLEAN
1994
Accept packets with local source addresses. In combination with
1995
suitable routing, this can be used to direct packets between two
1996
local interfaces over the wire and have them accepted properly.
1997
default FALSE
1999
route_localnet - BOOLEAN
2000
Do not consider loopback addresses as martian source or destination
2001
while routing. This enables the use of 127/8 for local routing purposes.
2003
default FALSE
2005
rp_filter - INTEGER
2006
- 0 - No source validation.
2007
- 1 - Strict mode as defined in RFC3704 Strict Reverse Path
2008
Each incoming packet is tested against the FIB and if the interface
2009
is not the best reverse path the packet check will fail.
2010
By default failed packets are discarded.
2011
- 2 - Loose mode as defined in RFC3704 Loose Reverse Path
2012
Each incoming packet's source address is also tested against the FIB
2013
and if the source address is not reachable via any interface
2014
the packet check will fail.
2016
Current recommended practice in RFC3704 is to enable strict mode
2017
to prevent IP spoofing from DDos attacks. If using asymmetric routing
2018
or other complicated routing, then loose mode is recommended.
2020
The max value from conf/{all,interface}/rp_filter is used
2021
when doing source validation on the {interface}.
2023
Default value is 0. Note that some distributions enable it
2024
in startup scripts.
2026
src_valid_mark - BOOLEAN
2027
- 0 - The fwmark of the packet is not included in reverse path
2028
route lookup. This allows for asymmetric routing configurations
2029
utilizing the fwmark in only one direction, e.g., transparent
2030
proxying.
2032
- 1 - The fwmark of the packet is included in reverse path route
2033
lookup. This permits rp_filter to function when the fwmark is
2034
used for routing traffic in both directions.
2036
This setting also affects the utilization of fmwark when
2037
performing source address selection for ICMP replies, or
2038
determining addresses stored for the IPOPT_TS_TSANDADDR and
2039
IPOPT_RR IP options.
2041
The max value from conf/{all,interface}/src_valid_mark is used.
2043
Default value is 0.
2045
arp_filter - BOOLEAN
2046
- 1 - Allows you to have multiple network interfaces on the same
2047
subnet, and have the ARPs for each interface be answered
2048
based on whether or not the kernel would route a packet from
2049
the ARP'd IP out that interface (therefore you must use source
2050
based routing for this to work). In other words it allows control
2051
of which cards (usually 1) will respond to an arp request.
2053
- 0 - (default) The kernel can respond to arp requests with addresses
2054
from other interfaces. This may seem wrong but it usually makes
2055
sense, because it increases the chance of successful communication.
2056
IP addresses are owned by the complete host on Linux, not by
2057
particular interfaces. Only for more complex setups like load-
2058
balancing, does this behaviour cause problems.
2060
arp_filter for the interface will be enabled if at least one of
2061
conf/{all,interface}/arp_filter is set to TRUE,
2062
it will be disabled otherwise
2064
arp_announce - INTEGER
2065
Define different restriction levels for announcing the local
2066
source IP address from IP packets in ARP requests sent on
2067
interface:
2069
- 0 - (default) Use any local address, configured on any interface
2070
- 1 - Try to avoid local addresses that are not in the target's
2071
subnet for this interface. This mode is useful when target
2072
hosts reachable via this interface require the source IP
2073
address in ARP requests to be part of their logical network
2074
configured on the receiving interface. When we generate the
2075
request we will check all our subnets that include the
2076
target IP and will preserve the source address if it is from
2077
such subnet. If there is no such subnet we select source
2078
address according to the rules for level 2.
2079
- 2 - Always use the best local address for this target.
2080
In this mode we ignore the source address in the IP packet
2081
and try to select local address that we prefer for talks with
2082
the target host. Such local address is selected by looking
2083
for primary IP addresses on all our subnets on the outgoing
2084
interface that include the target IP address. If no suitable
2085
local address is found we select the first local address
2086
we have on the outgoing interface or on all other interfaces,
2087
with the hope we will receive reply for our request and
2088
even sometimes no matter the source IP address we announce.
2090
The max value from conf/{all,interface}/arp_announce is used.
2092
Increasing the restriction level gives more chance for
2093
receiving answer from the resolved target while decreasing
2094
the level announces more valid sender's information.
2096
arp_ignore - INTEGER
2097
Define different modes for sending replies in response to
2098
received ARP requests that resolve local target IP addresses:
2100
- 0 - (default): reply for any local target IP address, configured
2101
on any interface
2102
- 1 - reply only if the target IP address is local address
2103
configured on the incoming interface
2104
- 2 - reply only if the target IP address is local address
2105
configured on the incoming interface and both with the
2106
sender's IP address are part from same subnet on this interface
2107
- 3 - do not reply for local addresses configured with scope host,
2108
only resolutions for global and link addresses are replied
2109
- 4-7 - reserved
2110
- 8 - do not reply for all local addresses
2112
The max value from conf/{all,interface}/arp_ignore is used
2113
when ARP request is received on the {interface}
2115
arp_notify - BOOLEAN
2116
Define mode for notification of address and device changes.
2118
== ==========================================================
2119
0 (default): do nothing
2120
1 Generate gratuitous arp requests when device is brought up
2121
or hardware address changes.
2122
== ==========================================================
2124
arp_accept - INTEGER
2125
Define behavior for accepting gratuitous ARP (garp) frames from devices
2126
that are not already present in the ARP table:
2128
- 0 - don't create new entries in the ARP table
2129
- 1 - create new entries in the ARP table
2130
- 2 - create new entries only if the source IP address is in the same
2131
subnet as an address configured on the interface that received the
2132
garp message.
2134
Both replies and requests type gratuitous arp will trigger the
2135
ARP table to be updated, if this setting is on.
2137
If the ARP table already contains the IP address of the
2138
gratuitous arp frame, the arp table will be updated regardless
2139
if this setting is on or off.
2141
arp_evict_nocarrier - BOOLEAN
2142
Clears the ARP cache on NOCARRIER events. This option is important for
2143
wireless devices where the ARP cache should not be cleared when roaming
2144
between access points on the same network. In most cases this should
2145
remain as the default (1).
2147
Possible values:
2149
- 0 (disabled) - Do not clear ARP cache on NOCARRIER events
2150
- 1 (enabled) - Clear the ARP cache on NOCARRIER events
2152
Default: 1 (enabled)
2154
mcast_solicit - INTEGER
2155
The maximum number of multicast probes in INCOMPLETE state,
2156
when the associated hardware address is unknown. Defaults
2157
to 3.
2159
ucast_solicit - INTEGER
2160
The maximum number of unicast probes in PROBE state, when
2161
the hardware address is being reconfirmed. Defaults to 3.
2163
app_solicit - INTEGER
2164
The maximum number of probes to send to the user space ARP daemon
2165
via netlink before dropping back to multicast probes (see
2166
mcast_resolicit). Defaults to 0.
2168
mcast_resolicit - INTEGER
2169
The maximum number of multicast probes after unicast and
2170
app probes in PROBE state. Defaults to 0.
2172
disable_policy - BOOLEAN
2173
Disable IPSEC policy (SPD) for this interface
2175
Possible values:
2177
- 0 (disabled)
2178
- 1 (enabled)
2180
Default: 0 (disabled)
2182
disable_xfrm - BOOLEAN
2183
Disable IPSEC encryption on this interface, whatever the policy
2185
Possible values:
2187
- 0 (disabled)
2188
- 1 (enabled)
2190
Default: 0 (disabled)
2192
igmpv2_unsolicited_report_interval - INTEGER
2193
The interval in milliseconds in which the next unsolicited
2194
IGMPv1 or IGMPv2 report retransmit will take place.
2196
Default: 10000 (10 seconds)
2198
igmpv3_unsolicited_report_interval - INTEGER
2199
The interval in milliseconds in which the next unsolicited
2200
IGMPv3 report retransmit will take place.
2202
Default: 1000 (1 seconds)
2204
ignore_routes_with_linkdown - BOOLEAN
2205
Ignore routes whose link is down when performing a FIB lookup.
2207
Possible values:
2209
- 0 (disabled)
2210
- 1 (enabled)
2212
Default: 0 (disabled)
2214
promote_secondaries - BOOLEAN
2215
When a primary IP address is removed from this interface
2216
promote a corresponding secondary IP address instead of
2217
removing all the corresponding secondary IP addresses.
2219
Possible values:
2221
- 0 (disabled)
2222
- 1 (enabled)
2224
Default: 0 (disabled)
2226
drop_unicast_in_l2_multicast - BOOLEAN
2227
Drop any unicast IP packets that are received in link-layer
2228
multicast (or broadcast) frames.
2230
This behavior (for multicast) is actually a SHOULD in RFC
2231
1122, but is disabled by default for compatibility reasons.
2233
Possible values:
2235
- 0 (disabled)
2236
- 1 (enabled)
2238
Default: 0 (disabled)
2240
drop_gratuitous_arp - BOOLEAN
2241
Drop all gratuitous ARP frames, for example if there's a known
2242
good ARP proxy on the network and such frames need not be used
2243
(or in the case of 802.11, must not be used to prevent attacks.)
2245
Possible values:
2247
- 0 (disabled)
2248
- 1 (enabled)
2250
Default: 0 (disabled)
2253
tag - INTEGER
2254
Allows you to write a number, which can be used as required.
2256
Default value is 0.
2258
xfrm4_gc_thresh - INTEGER
2259
(Obsolete since linux-4.14)
2260
The threshold at which we will start garbage collecting for IPv4
2261
destination cache entries. At twice this value the system will
2262
refuse new allocations.
2264
igmp_link_local_mcast_reports - BOOLEAN
2265
Enable IGMP reports for link local multicast groups in the
2266
224.0.0.X range.
2268
Default TRUE
2270
Alexey Kuznetsov.
2273
Updated by:
2275
- Andi Kleen
2277
- Nicolas Delon
2283
/proc/sys/net/ipv6/* Variables
2284
==============================
2286
IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
2287
apply to IPv6 [XXX?].
2289
bindv6only - BOOLEAN
2290
Default value for IPV6_V6ONLY socket option,
2291
which restricts use of the IPv6 socket to IPv6 communication
2292
only.
2294
Possible values:
2296
- 0 (disabled) - enable IPv4-mapped address feature
2297
- 1 (enabled) - disable IPv4-mapped address feature
2299
Default: 0 (disabled)
2301
flowlabel_consistency - BOOLEAN
2302
Protect the consistency (and unicity) of flow label.
2303
You have to disable it to use IPV6_FL_F_REFLECT flag on the
2304
flow label manager.
2306
Possible values:
2308
- 0 (disabled)
2309
- 1 (enabled)
2311
Default: 1 (enabled)
2313
auto_flowlabels - INTEGER
2314
Automatically generate flow labels based on a flow hash of the
2315
packet. This allows intermediate devices, such as routers, to
2316
identify packet flows for mechanisms like Equal Cost Multipath
2317
Routing (see RFC 6438).
2319
= ===========================================================
2320
0 automatic flow labels are completely disabled
2321
1 automatic flow labels are enabled by default, they can be
2322
disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
2323
socket option
2324
2 automatic flow labels are allowed, they may be enabled on a
2325
per socket basis using the IPV6_AUTOFLOWLABEL socket option
2326
3 automatic flow labels are enabled and enforced, they cannot
2327
be disabled by the socket option
2328
= ===========================================================
2330
Default: 1
2332
flowlabel_state_ranges - BOOLEAN
2333
Split the flow label number space into two ranges. 0-0x7FFFF is
2334
reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
2335
is reserved for stateless flow labels as described in RFC6437.
2337
Possible values:
2339
- 0 (disabled)
2340
- 1 (enabled)
2342
Default: 1 (enabled)
2345
flowlabel_reflect - INTEGER
2346
Control flow label reflection. Needed for Path MTU
2347
Discovery to work with Equal Cost Multipath Routing in anycast
2348
environments. See RFC 7690 and:
2349
https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
2351
This is a bitmask.
2353
- 1: enabled for established flows
2355
Note that this prevents automatic flowlabel changes, as done
2356
in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
2357
and "tcp: Change txhash on every SYN and RTO retransmit"
2359
- 2: enabled for TCP RESET packets (no active listener)
2360
If set, a RST packet sent in response to a SYN packet on a closed
2361
port will reflect the incoming flow label.
2363
- 4: enabled for ICMPv6 echo reply messages.
2365
Default: 0
2367
fib_multipath_hash_policy - INTEGER
2368
Controls which hash policy to use for multipath routes.
2370
Default: 0 (Layer 3)
2372
Possible values:
2374
- 0 - Layer 3 (source and destination addresses plus flow label)
2375
- 1 - Layer 4 (standard 5-tuple)
2376
- 2 - Layer 3 or inner Layer 3 if present
2377
- 3 - Custom multipath hash. Fields used for multipath hash calculation
2378
are determined by fib_multipath_hash_fields sysctl
2380
fib_multipath_hash_fields - UNSIGNED INTEGER
2381
When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
2382
fields used for multipath hash calculation are determined by this
2383
sysctl.
2385
This value is a bitmask which enables various fields for multipath hash
2386
calculation.
2388
Possible fields are:
2390
====== ============================
2391
0x0001 Source IP address
2392
0x0002 Destination IP address
2393
0x0004 IP protocol
2394
0x0008 Flow Label
2395
0x0010 Source port
2396
0x0020 Destination port
2397
0x0040 Inner source IP address
2398
0x0080 Inner destination IP address
2399
0x0100 Inner IP protocol
2400
0x0200 Inner Flow Label
2401
0x0400 Inner source port
2402
0x0800 Inner destination port
2403
====== ============================
2405
Default: 0x0007 (source IP, destination IP and IP protocol)
2407
anycast_src_echo_reply - BOOLEAN
2408
Controls the use of anycast addresses as source addresses for ICMPv6
2409
echo reply
2411
Possible values:
2413
- 0 (disabled)
2414
- 1 (enabled)
2416
Default: 0 (disabled)
2419
idgen_delay - INTEGER
2420
Controls the delay in seconds after which time to retry
2421
privacy stable address generation if a DAD conflict is
2422
detected.
2424
Default: 1 (as specified in RFC7217)
2426
idgen_retries - INTEGER
2427
Controls the number of retries to generate a stable privacy
2428
address if a DAD conflict is detected.
2430
Default: 3 (as specified in RFC7217)
2432
mld_qrv - INTEGER
2433
Controls the MLD query robustness variable (see RFC3810 9.1).
2435
Default: 2 (as specified by RFC3810 9.1)
2437
Minimum: 1 (as specified by RFC6636 4.5)
2439
max_dst_opts_number - INTEGER
2440
Maximum number of non-padding TLVs allowed in a Destination
2441
options extension header. If this value is less than zero
2442
then unknown options are disallowed and the number of known
2443
TLVs allowed is the absolute value of this number.
2445
Default: 8
2447
max_hbh_opts_number - INTEGER
2448
Maximum number of non-padding TLVs allowed in a Hop-by-Hop
2449
options extension header. If this value is less than zero
2450
then unknown options are disallowed and the number of known
2451
TLVs allowed is the absolute value of this number.
2453
Default: 8
2455
max_dst_opts_length - INTEGER
2456
Maximum length allowed for a Destination options extension
2457
header.
2459
Default: INT_MAX (unlimited)
2461
max_hbh_length - INTEGER
2462
Maximum length allowed for a Hop-by-Hop options extension
2463
header.
2465
Default: INT_MAX (unlimited)
2467
skip_notify_on_dev_down - BOOLEAN
2468
Controls whether an RTM_DELROUTE message is generated for routes
2469
removed when a device is taken down or deleted. IPv4 does not
2470
generate this message; IPv6 does by default. Setting this sysctl
2471
to true skips the message, making IPv4 and IPv6 on par in relying
2472
on userspace caches to track link events and evict routes.
2474
Possible values:
2476
- 0 (disabled) - generate the message
2477
- 1 (enabled) - skip generating the message
2479
Default: 0 (disabled)
2481
nexthop_compat_mode - BOOLEAN
2482
New nexthop API provides a means for managing nexthops independent of
2483
prefixes. Backwards compatibility with old route format is enabled by
2484
default which means route dumps and notifications contain the new
2485
nexthop attribute but also the full, expanded nexthop definition.
2486
Further, updates or deletes of a nexthop configuration generate route
2487
notifications for each fib entry using the nexthop. Once a system
2488
understands the new API, this sysctl can be disabled to achieve full
2489
performance benefits of the new API by disabling the nexthop expansion
2490
and extraneous notifications.
2492
Note that as a backward-compatible mode, dumping of modern features
2493
might be incomplete or wrong. For example, resilient groups will not be
2494
shown as such, but rather as just a list of next hops. Also weights that
2495
do not fit into 8 bits will show incorrectly.
2497
Default: true (backward compat mode)
2499
fib_notify_on_flag_change - INTEGER
2500
Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
2501
RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
2503
After installing a route to the kernel, user space receives an
2504
acknowledgment, which means the route was installed in the kernel,
2505
but not necessarily in hardware.
2506
It is also possible for a route already installed in hardware to change
2507
its action and therefore its flags. For example, a host route that is
2508
trapping packets can be "promoted" to perform decapsulation following
2509
the installation of an IPinIP/VXLAN tunnel.
2510
The notifications will indicate to user-space the state of the route.
2512
Default: 0 (Do not emit notifications.)
2514
Possible values:
2516
- 0 - Do not emit notifications.
2517
- 1 - Emit notifications.
2518
- 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
2520
ioam6_id - INTEGER
2521
Define the IOAM id of this node. Uses only 24 bits out of 32 in total.
2523
Possible value range:
2525
- Min: 0
2526
- Max: 0xFFFFFF
2528
Default: 0xFFFFFF
2530
ioam6_id_wide - LONG INTEGER
2531
Define the wide IOAM id of this node. Uses only 56 bits out of 64 in
2532
total. Can be different from ioam6_id.
2534
Possible value range:
2536
- Min: 0
2537
- Max: 0xFFFFFFFFFFFFFF
2539
Default: 0xFFFFFFFFFFFFFF
2541
IPv6 Fragmentation:
2543
ip6frag_high_thresh - INTEGER
2544
Maximum memory used to reassemble IPv6 fragments. When
2545
ip6frag_high_thresh bytes of memory is allocated for this purpose,
2546
the fragment handler will toss packets until ip6frag_low_thresh
2547
is reached.
2549
ip6frag_low_thresh - INTEGER
2550
See ip6frag_high_thresh
2552
ip6frag_time - INTEGER
2553
Time in seconds to keep an IPv6 fragment in memory.
2555
``conf/default/*``:
2556
Change the interface-specific default settings.
2558
These settings would be used during creating new interfaces.
2561
``conf/all/*``:
2562
Change all the interface-specific settings.
2564
[XXX: Other special features than forwarding?]
2566
conf/all/disable_ipv6 - BOOLEAN
2567
Changing this value is same as changing ``conf/default/disable_ipv6``
2568
setting and also all per-interface ``disable_ipv6`` settings to the same
2569
value.
2571
Reading this value does not have any particular meaning. It does not say
2572
whether IPv6 support is enabled or disabled. Returned value can be 1
2573
also in the case when some interface has ``disable_ipv6`` set to 0 and
2574
has configured IPv6 addresses.
2576
conf/all/forwarding - BOOLEAN
2577
Enable global IPv6 forwarding between all interfaces.
2579
IPv4 and IPv6 work differently here; the ``force_forwarding`` flag must
2580
be used to control which interfaces may forward packets.
2582
This also sets all interfaces' Host/Router setting
2583
'forwarding' to the specified value. See below for details.
2585
This referred to as global forwarding.
2587
proxy_ndp - BOOLEAN
2588
Do proxy ndp.
2590
Possible values:
2592
- 0 (disabled)
2593
- 1 (enabled)
2595
Default: 0 (disabled)
2597
force_forwarding - BOOLEAN
2598
Enable forwarding on this interface only -- regardless of the setting on
2599
``conf/all/forwarding``. When setting ``conf.all.forwarding`` to 0,
2600
the ``force_forwarding`` flag will be reset on all interfaces.
2602
fwmark_reflect - BOOLEAN
2603
Controls the fwmark of kernel-generated IPv6 reply packets that are not
2604
associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
2605
If disabled, these packets have a fwmark of zero. If enabled, they have the
2606
fwmark of the packet they are replying to.
2608
Possible values:
2610
- 0 (disabled)
2611
- 1 (enabled)
2613
Default: 0 (disabled)
2615
``conf/interface/*``:
2616
Change special settings per interface.
2618
The functional behaviour for certain settings is different
2619
depending on whether local forwarding is enabled or not.
2621
accept_ra - INTEGER
2622
Accept Router Advertisements; autoconfigure using them.
2624
It also determines whether or not to transmit Router
2625
Solicitations. If and only if the functional setting is to
2626
accept Router Advertisements, Router Solicitations will be
2627
transmitted.
2629
Possible values are:
2631
== ===========================================================
2632
0 Do not accept Router Advertisements.
2633
1 Accept Router Advertisements if forwarding is disabled.
2634
2 Overrule forwarding behaviour. Accept Router Advertisements
2635
even if forwarding is enabled.
2636
== ===========================================================
2638
Functional default:
2640
- enabled if local forwarding is disabled.
2641
- disabled if local forwarding is enabled.
2643
accept_ra_defrtr - BOOLEAN
2644
Learn default router in Router Advertisement.
2646
Functional default:
2648
- enabled if accept_ra is enabled.
2649
- disabled if accept_ra is disabled.
2651
ra_defrtr_metric - UNSIGNED INTEGER
2652
Route metric for default route learned in Router Advertisement. This value
2653
will be assigned as metric for the default route learned via IPv6 Router
2654
Advertisement. Takes affect only if accept_ra_defrtr is enabled.
2656
Possible values:
2657
1 to 0xFFFFFFFF
2659
Default: IP6_RT_PRIO_USER i.e. 1024.
2661
accept_ra_from_local - BOOLEAN
2662
Accept RA with source-address that is found on local machine
2663
if the RA is otherwise proper and able to be accepted.
2665
Default is to NOT accept these as it may be an un-intended
2666
network loop.
2668
Functional default:
2670
- enabled if accept_ra_from_local is enabled
2671
on a specific interface.
2672
- disabled if accept_ra_from_local is disabled
2673
on a specific interface.
2675
accept_ra_min_hop_limit - INTEGER
2676
Minimum hop limit Information in Router Advertisement.
2678
Hop limit Information in Router Advertisement less than this
2679
variable shall be ignored.
2681
Default: 1
2683
accept_ra_min_lft - INTEGER
2684
Minimum acceptable lifetime value in Router Advertisement.
2686
RA sections with a lifetime less than this value shall be
2687
ignored. Zero lifetimes stay unaffected.
2689
Default: 0
2691
accept_ra_pinfo - BOOLEAN
2692
Learn Prefix Information in Router Advertisement.
2694
Functional default:
2696
- enabled if accept_ra is enabled.
2697
- disabled if accept_ra is disabled.
2699
ra_honor_pio_life - BOOLEAN
2700
Whether to use RFC4862 Section 5.5.3e to determine the valid
2701
lifetime of an address matching a prefix sent in a Router
2702
Advertisement Prefix Information Option.
2704
Possible values:
2706
- 0 (disabled) - RFC4862 section 5.5.3e is used to determine
2707
the valid lifetime of the address.
2708
- 1 (enabled) - the PIO valid lifetime will always be honored.
2710
Default: 0 (disabled)
2712
ra_honor_pio_pflag - BOOLEAN
2713
The Prefix Information Option P-flag indicates the network can
2714
allocate a unique IPv6 prefix per client using DHCPv6-PD.
2715
This sysctl can be enabled when a userspace DHCPv6-PD client
2716
is running to cause the P-flag to take effect: i.e. the
2717
P-flag suppresses any effects of the A-flag within the same
2718
PIO. For a given PIO, P=1 and A=1 is treated as A=0.
2720
Possible values:
2722
- 0 (disabled) - the P-flag is ignored.
2723
- 1 (enabled) - the P-flag will disable SLAAC autoconfiguration
2724
for the given Prefix Information Option.
2726
Default: 0 (disabled)
2728
accept_ra_rt_info_min_plen - INTEGER
2729
Minimum prefix length of Route Information in RA.
2731
Route Information w/ prefix smaller than this variable shall
2732
be ignored.
2734
Functional default:
2736
* 0 if accept_ra_rtr_pref is enabled.
2737
* -1 if accept_ra_rtr_pref is disabled.
2739
accept_ra_rt_info_max_plen - INTEGER
2740
Maximum prefix length of Route Information in RA.
2742
Route Information w/ prefix larger than this variable shall
2743
be ignored.
2745
Functional default:
2747
* 0 if accept_ra_rtr_pref is enabled.
2748
* -1 if accept_ra_rtr_pref is disabled.
2750
accept_ra_rtr_pref - BOOLEAN
2751
Accept Router Preference in RA.
2753
Functional default:
2755
- enabled if accept_ra is enabled.
2756
- disabled if accept_ra is disabled.
2758
accept_ra_mtu - BOOLEAN
2759
Apply the MTU value specified in RA option 5 (RFC4861). If
2760
disabled, the MTU specified in the RA will be ignored.
2762
Functional default:
2764
- enabled if accept_ra is enabled.
2765
- disabled if accept_ra is disabled.
2767
accept_redirects - BOOLEAN
2768
Accept Redirects.
2770
Functional default:
2772
- enabled if local forwarding is disabled.
2773
- disabled if local forwarding is enabled.
2775
accept_source_route - INTEGER
2776
Accept source routing (routing extension header).
2778
- >= 0: Accept only routing header type 2.
2779
- < 0: Do not accept routing header.
2781
Default: 0
2783
autoconf - BOOLEAN
2784
Autoconfigure addresses using Prefix Information in Router
2785
Advertisements.
2787
Functional default:
2789
- enabled if accept_ra_pinfo is enabled.
2790
- disabled if accept_ra_pinfo is disabled.
2792
dad_transmits - INTEGER
2793
The amount of Duplicate Address Detection probes to send.
2795
Default: 1
2797
forwarding - INTEGER
2798
Configure interface-specific Host/Router behaviour.
2800
.. note::
2802
It is recommended to have the same setting on all
2803
interfaces; mixed router/host scenarios are rather uncommon.
2805
Possible values are:
2807
- 0 Forwarding disabled
2808
- 1 Forwarding enabled
2810
**FALSE (0)**:
2812
By default, Host behaviour is assumed. This means:
2814
1. IsRouter flag is not set in Neighbour Advertisements.
2815
2. If accept_ra is TRUE (default), transmit Router
2816
Solicitations.
2817
3. If accept_ra is TRUE (default), accept Router
2818
Advertisements (and do autoconfiguration).
2819
4. If accept_redirects is TRUE (default), accept Redirects.
2821
**TRUE (1)**:
2823
If local forwarding is enabled, Router behaviour is assumed.
2824
This means exactly the reverse from the above:
2826
1. IsRouter flag is set in Neighbour Advertisements.
2827
2. Router Solicitations are not sent unless accept_ra is 2.
2828
3. Router Advertisements are ignored unless accept_ra is 2.
2829
4. Redirects are ignored.
2831
Default: 0 (disabled) if global forwarding is disabled (default),
2832
otherwise 1 (enabled).
2834
hop_limit - INTEGER
2835
Default Hop Limit to set.
2837
Default: 64
2839
mtu - INTEGER
2840
Default Maximum Transfer Unit
2842
Default: 1280 (IPv6 required minimum)
2844
ip_nonlocal_bind - BOOLEAN
2845
If enabled, allows processes to bind() to non-local IPv6 addresses,
2846
which can be quite useful - but may break some applications.
2848
Possible values:
2850
- 0 (disabled)
2851
- 1 (enabled)
2853
Default: 0 (disabled)
2855
router_probe_interval - INTEGER
2856
Minimum interval (in seconds) between Router Probing described
2857
in RFC4191.
2859
Default: 60
2861
router_solicitation_delay - INTEGER
2862
Number of seconds to wait after interface is brought up
2863
before sending Router Solicitations.
2865
Default: 1
2867
router_solicitation_interval - INTEGER
2868
Number of seconds to wait between Router Solicitations.
2870
Default: 4
2872
router_solicitations - INTEGER
2873
Number of Router Solicitations to send until assuming no
2874
routers are present.
2876
Default: 3
2878
use_oif_addrs_only - BOOLEAN
2879
When enabled, the candidate source addresses for destinations
2880
routed via this interface are restricted to the set of addresses
2881
configured on this interface (vis. RFC 6724, section 4).
2883
Possible values:
2885
- 0 (disabled)
2886
- 1 (enabled)
2888
Default: 0 (disabled)
2890
use_tempaddr - INTEGER
2891
Preference for Privacy Extensions (RFC3041).
2893
* <= 0 : disable Privacy Extensions
2894
* == 1 : enable Privacy Extensions, but prefer public
2895
addresses over temporary addresses.
2896
* > 1 : enable Privacy Extensions and prefer temporary
2897
addresses over public addresses.
2899
Default:
2901
* 0 (for most devices)
2902
* -1 (for point-to-point devices and loopback devices)
2904
temp_valid_lft - INTEGER
2905
valid lifetime (in seconds) for temporary addresses. If less than the
2906
minimum required lifetime (typically 5-7 seconds), temporary addresses
2907
will not be created.
2909
Default: 172800 (2 days)
2911
temp_prefered_lft - INTEGER
2912
Preferred lifetime (in seconds) for temporary addresses. If
2913
temp_prefered_lft is less than the minimum required lifetime (typically
2914
5-7 seconds), the preferred lifetime is the minimum required. If
2915
temp_prefered_lft is greater than temp_valid_lft, the preferred lifetime
2916
is temp_valid_lft.
2918
Default: 86400 (1 day)
2920
keep_addr_on_down - INTEGER
2921
Keep all IPv6 addresses on an interface down event. If set static
2922
global addresses with no expiration time are not flushed.
2924
* >0 : enabled
2925
* 0 : system default
2926
* <0 : disabled
2928
Default: 0 (addresses are removed)
2930
max_desync_factor - INTEGER
2931
Maximum value for DESYNC_FACTOR, which is a random value
2932
that ensures that clients don't synchronize with each
2933
other and generate new addresses at exactly the same time.
2934
value is in seconds.
2936
Default: 600
2938
regen_min_advance - INTEGER
2939
How far in advance (in seconds), at minimum, to create a new temporary
2940
address before the current one is deprecated. This value is added to
2941
the amount of time that may be required for duplicate address detection
2942
to determine when to create a new address. Linux permits setting this
2943
value to less than the default of 2 seconds, but a value less than 2
2944
does not conform to RFC 8981.
2946
Default: 2
2948
regen_max_retry - INTEGER
2949
Number of attempts before give up attempting to generate
2950
valid temporary addresses.
2952
Default: 5
2954
max_addresses - INTEGER
2955
Maximum number of autoconfigured addresses per interface. Setting
2956
to zero disables the limitation. It is not recommended to set this
2957
value too large (or to zero) because it would be an easy way to
2958
crash the kernel by allowing too many addresses to be created.
2960
Default: 16
2962
disable_ipv6 - BOOLEAN
2963
Disable IPv6 operation. If accept_dad is set to 2, this value
2964
will be dynamically set to TRUE if DAD fails for the link-local
2965
address.
2967
Default: FALSE (enable IPv6 operation)
2969
When this value is changed from 1 to 0 (IPv6 is being enabled),
2970
it will dynamically create a link-local address on the given
2971
interface and start Duplicate Address Detection, if necessary.
2973
When this value is changed from 0 to 1 (IPv6 is being disabled),
2974
it will dynamically delete all addresses and routes on the given
2975
interface. From now on it will not possible to add addresses/routes
2976
to the selected interface.
2978
accept_dad - INTEGER
2979
Whether to accept DAD (Duplicate Address Detection).
2981
== ==============================================================
2982
0 Disable DAD
2983
1 Enable DAD (default)
2984
2 Enable DAD, and disable IPv6 operation if MAC-based duplicate
2985
link-local address has been found.
2986
== ==============================================================
2988
DAD operation and mode on a given interface will be selected according
2989
to the maximum value of conf/{all,interface}/accept_dad.
2991
force_tllao - BOOLEAN
2992
Enable sending the target link-layer address option even when
2993
responding to a unicast neighbor solicitation.
2995
Default: FALSE
2997
Quoting from RFC 2461, section 4.4, Target link-layer address:
2999
"The option MUST be included for multicast solicitations in order to
3000
avoid infinite Neighbor Solicitation "recursion" when the peer node
3001
does not have a cache entry to return a Neighbor Advertisements
3002
message. When responding to unicast solicitations, the option can be
3003
omitted since the sender of the solicitation has the correct link-
3004
layer address; otherwise it would not have be able to send the unicast
3005
solicitation in the first place. However, including the link-layer
3006
address in this case adds little overhead and eliminates a potential
3007
race condition where the sender deletes the cached link-layer address
3008
prior to receiving a response to a previous solicitation."
3010
ndisc_notify - BOOLEAN
3011
Define mode for notification of address and device changes.
3013
Possible values:
3015
- 0 (disabled) - do nothing
3016
- 1 (enabled) - Generate unsolicited neighbour advertisements when device is brought
3017
up or hardware address changes.
3019
Default: 0 (disabled)
3021
ndisc_tclass - INTEGER
3022
The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
3023
Discovery (Router Solicitation, Router Advertisement, Neighbor
3024
Solicitation, Neighbor Advertisement, Redirect) messages.
3025
These 8 bits can be interpreted as 6 high order bits holding the DSCP
3026
value and 2 low order bits representing ECN (which you probably want
3027
to leave cleared).
3029
* 0 - (default)
3031
ndisc_evict_nocarrier - BOOLEAN
3032
Clears the neighbor discovery table on NOCARRIER events. This option is
3033
important for wireless devices where the neighbor discovery cache should
3034
not be cleared when roaming between access points on the same network.
3035
In most cases this should remain as the default (1).
3037
Possible values:
3039
- 0 (disabled) - Do not clear neighbor discovery cache on NOCARRIER events.
3040
- 1 (enabled) - Clear neighbor discover cache on NOCARRIER events.
3042
Default: 1 (enabled)
3044
mldv1_unsolicited_report_interval - INTEGER
3045
The interval in milliseconds in which the next unsolicited
3046
MLDv1 report retransmit will take place.
3048
Default: 10000 (10 seconds)
3050
mldv2_unsolicited_report_interval - INTEGER
3051
The interval in milliseconds in which the next unsolicited
3052
MLDv2 report retransmit will take place.
3054
Default: 1000 (1 second)
3056
force_mld_version - INTEGER
3057
* 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
3058
* 1 - Enforce to use MLD version 1
3059
* 2 - Enforce to use MLD version 2
3061
suppress_frag_ndisc - INTEGER
3062
Control RFC 6980 (Security Implications of IPv6 Fragmentation
3063
with IPv6 Neighbor Discovery) behavior:
3065
* 1 - (default) discard fragmented neighbor discovery packets
3066
* 0 - allow fragmented neighbor discovery packets
3068
optimistic_dad - BOOLEAN
3069
Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
3071
Optimistic Duplicate Address Detection for the interface will be enabled
3072
if at least one of conf/{all,interface}/optimistic_dad is set to 1,
3073
it will be disabled otherwise.
3075
Possible values:
3077
- 0 (disabled)
3078
- 1 (enabled)
3080
Default: 0 (disabled)
3083
use_optimistic - BOOLEAN
3084
If enabled, do not classify optimistic addresses as deprecated during
3085
source address selection. Preferred addresses will still be chosen
3086
before optimistic addresses, subject to other ranking in the source
3087
address selection algorithm.
3089
This will be enabled if at least one of
3090
conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
3092
Possible values:
3094
- 0 (disabled)
3095
- 1 (enabled)
3097
Default: 0 (disabled)
3099
stable_secret - IPv6 address
3100
This IPv6 address will be used as a secret to generate IPv6
3101
addresses for link-local addresses and autoconfigured
3102
ones. All addresses generated after setting this secret will
3103
be stable privacy ones by default. This can be changed via the
3104
addrgenmode ip-link. conf/default/stable_secret is used as the
3105
secret for the namespace, the interface specific ones can
3106
overwrite that. Writes to conf/all/stable_secret are refused.
3108
It is recommended to generate this secret during installation
3109
of a system and keep it stable after that.
3111
By default the stable secret is unset.
3113
addr_gen_mode - INTEGER
3114
Defines how link-local and autoconf addresses are generated.
3116
= =================================================================
3117
0 generate address based on EUI64 (default)
3118
1 do no generate a link-local address, use EUI64 for addresses
3119
generated from autoconf
3120
2 generate stable privacy addresses, using the secret from
3121
stable_secret (RFC7217)
3122
3 generate stable privacy addresses, using a random secret if unset
3123
= =================================================================
3125
drop_unicast_in_l2_multicast - BOOLEAN
3126
Drop any unicast IPv6 packets that are received in link-layer
3127
multicast (or broadcast) frames.
3129
Possible values:
3131
- 0 (disabled)
3132
- 1 (enabled)
3134
Default: 0 (disabled)
3136
drop_unsolicited_na - BOOLEAN
3137
Drop all unsolicited neighbor advertisements, for example if there's
3138
a known good NA proxy on the network and such frames need not be used
3139
(or in the case of 802.11, must not be used to prevent attacks.)
3141
Possible values:
3143
- 0 (disabled)
3144
- 1 (enabled)
3146
Default: 0 (disabled).
3148
accept_untracked_na - INTEGER
3149
Define behavior for accepting neighbor advertisements from devices that
3150
are absent in the neighbor cache:
3152
- 0 - (default) Do not accept unsolicited and untracked neighbor
3153
advertisements.
3155
- 1 - Add a new neighbor cache entry in STALE state for routers on
3156
receiving a neighbor advertisement (either solicited or unsolicited)
3157
with target link-layer address option specified if no neighbor entry
3158
is already present for the advertised IPv6 address. Without this knob,
3159
NAs received for untracked addresses (absent in neighbor cache) are
3160
silently ignored.
3162
This is as per router-side behavior documented in RFC9131.
3164
This has lower precedence than drop_unsolicited_na.
3166
This will optimize the return path for the initial off-link
3167
communication that is initiated by a directly connected host, by
3168
ensuring that the first-hop router which turns on this setting doesn't
3169
have to buffer the initial return packets to do neighbor-solicitation.
3170
The prerequisite is that the host is configured to send unsolicited
3171
neighbor advertisements on interface bringup. This setting should be
3172
used in conjunction with the ndisc_notify setting on the host to
3173
satisfy this prerequisite.
3175
- 2 - Extend option (1) to add a new neighbor cache entry only if the
3176
source IP address is in the same subnet as an address configured on
3177
the interface that received the neighbor advertisement.
3179
enhanced_dad - BOOLEAN
3180
Include a nonce option in the IPv6 neighbor solicitation messages used for
3181
duplicate address detection per RFC7527. A received DAD NS will only signal
3182
a duplicate address if the nonce is different. This avoids any false
3183
detection of duplicates due to loopback of the NS messages that we send.
3184
The nonce option will be sent on an interface unless both of
3185
conf/{all,interface}/enhanced_dad are set to FALSE.
3187
Possible values:
3189
- 0 (disabled)
3190
- 1 (enabled)
3192
Default: 1 (enabled)
3194
``icmp/*``:
3195
===========
3197
ratelimit - INTEGER
3198
Limit the maximal rates for sending ICMPv6 messages to a particular
3199
peer.
3201
0 to disable any limiting,
3202
otherwise the space between responses in milliseconds.
3204
Default: 100
3206
ratemask - list of comma separated ranges
3207
For ICMPv6 message types matching the ranges in the ratemask, limit
3208
the sending of the message according to ratelimit parameter.
3210
The format used for both input and output is a comma separated
3211
list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
3212
129). Writing to the file will clear all previous ranges of ICMPv6
3213
message types and update the current list with the input.
3215
Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
3216
for numerical values of ICMPv6 message types, e.g. echo request is 128
3217
and echo reply is 129.
3219
Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
3221
echo_ignore_all - BOOLEAN
3222
If enabled, then the kernel will ignore all ICMP ECHO
3223
requests sent to it over the IPv6 protocol.
3225
Possible values:
3227
- 0 (disabled)
3228
- 1 (enabled)
3230
Default: 0 (disabled)
3232
echo_ignore_multicast - BOOLEAN
3233
If enabled, then the kernel will ignore all ICMP ECHO
3234
requests sent to it over the IPv6 protocol via multicast.
3236
Possible values:
3238
- 0 (disabled)
3239
- 1 (enabled)
3241
Default: 0 (disabled)
3243
echo_ignore_anycast - BOOLEAN
3244
If enabled, then the kernel will ignore all ICMP ECHO
3245
requests sent to it over the IPv6 protocol destined to anycast address.
3247
Possible values:
3249
- 0 (disabled)
3250
- 1 (enabled)
3252
Default: 0 (disabled)
3254
error_anycast_as_unicast - BOOLEAN
3255
If enabled, then the kernel will respond with ICMP Errors
3256
resulting from requests sent to it over the IPv6 protocol destined
3257
to anycast address essentially treating anycast as unicast.
3259
Possible values:
3261
- 0 (disabled)
3262
- 1 (enabled)
3264
Default: 0 (disabled)
3266
xfrm6_gc_thresh - INTEGER
3267
(Obsolete since linux-4.14)
3268
The threshold at which we will start garbage collecting for IPv6
3269
destination cache entries. At twice this value the system will
3270
refuse new allocations.
3273
IPv6 Update by:
3274
Pekka Savola <[email protected]>
3275
YOSHIFUJI Hideaki / USAGI Project <[email protected]>
3278
/proc/sys/net/bridge/* Variables:
3279
=================================
3281
bridge-nf-call-arptables - BOOLEAN
3283
Possible values:
3285
- 0 (disabled) - disable this.
3286
- 1 (enabled) - pass bridged ARP traffic to arptables' FORWARD chain.
3288
Default: 1 (enabled)
3290
bridge-nf-call-iptables - BOOLEAN
3292
Possible values:
3294
- 0 (disabled) - disable this.
3295
- 1 (enabled) - pass bridged IPv4 traffic to iptables' chains.
3297
Default: 1 (enabled)
3299
bridge-nf-call-ip6tables - BOOLEAN
3301
Possible values:
3303
- 0 (disabled) - disable this.
3304
- 1 (enabled) - pass bridged IPv6 traffic to ip6tables' chains.
3306
Default: 1 (enabled)
3308
bridge-nf-filter-vlan-tagged - BOOLEAN
3310
Possible values:
3312
- 0 (disabled) - disable this.
3313
- 1 (enabled) - pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables
3315
Default: 0 (disabled)
3317
bridge-nf-filter-pppoe-tagged - BOOLEAN
3319
Possible values:
3321
- 0 (disabled) - disable this.
3322
- 1 (enabled) - pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
3324
Default: 0 (disabled)
3326
bridge-nf-pass-vlan-input-dev - BOOLEAN
3327
- 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
3328
interface on the bridge and set the netfilter input device to the
3329
vlan. This allows use of e.g. "iptables -i br0.1" and makes the
3330
REDIRECT target work with vlan-on-top-of-bridge interfaces. When no
3331
matching vlan interface is found, or this switch is off, the input
3332
device is set to the bridge interface.
3334
- 0: disable bridge netfilter vlan interface lookup.
3336
Default: 0
3338
``proc/sys/net/sctp/*`` Variables:
3339
==================================
3341
addip_enable - BOOLEAN
3342
Enable or disable extension of Dynamic Address Reconfiguration
3343
(ADD-IP) functionality specified in RFC5061. This extension provides
3344
the ability to dynamically add and remove new addresses for the SCTP
3345
associations.
3347
Possible values:
3349
- 0 (disabled) - disable extension.
3350
- 1 (enabled) - enable extension
3352
Default: 0 (disabled)
3354
pf_enable - INTEGER
3355
Enable or disable pf (pf is short for potentially failed) state. A value
3356
of pf_retrans > path_max_retrans also disables pf state. That is, one of
3357
both pf_enable and pf_retrans > path_max_retrans can disable pf state.
3358
Since pf_retrans and path_max_retrans can be changed by userspace
3359
application, sometimes user expects to disable pf state by the value of
3360
pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
3361
or path_max_retrans is changed by the user application, this pf state is
3362
enabled. As such, it is necessary to add this to dynamically enable
3363
and disable pf state. See:
3364
https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
3365
details.
3367
Possible values:
3369
- 1: Enable pf.
3370
- 0: Disable pf.
3372
Default: 1
3374
pf_expose - INTEGER
3375
Unset or enable/disable pf (pf is short for potentially failed) state
3376
exposure. Applications can control the exposure of the PF path state
3377
in the SCTP_PEER_ADDR_CHANGE event and access of SCTP_PF-state
3378
transport info via SCTP_GET_PEER_ADDR_INFO sockopt.
3380
Possible values:
3382
- 0: Unset pf state exposure (compatible with old applications). No
3383
event will be sent but the transport info can be queried.
3384
- 1: Disable pf state exposure. No event will be sent and trying to
3385
obtain transport info will return -EACCESS.
3386
- 2: Enable pf state exposure. The event will be sent for a transport
3387
becoming SCTP_PF state and transport info can be obtained.
3389
Default: 0
3391
addip_noauth_enable - BOOLEAN
3392
Dynamic Address Reconfiguration (ADD-IP) requires the use of
3393
authentication to protect the operations of adding or removing new
3394
addresses. This requirement is mandated so that unauthorized hosts
3395
would not be able to hijack associations. However, older
3396
implementations may not have implemented this requirement while
3397
allowing the ADD-IP extension. For reasons of interoperability,
3398
we provide this variable to control the enforcement of the
3399
authentication requirement.
3401
== ===============================================================
3402
1 Allow ADD-IP extension to be used without authentication. This
3403
should only be set in a closed environment for interoperability
3404
with older implementations.
3406
0 Enforce the authentication requirement
3407
== ===============================================================
3409
Default: 0
3411
auth_enable - BOOLEAN
3412
Enable or disable Authenticated Chunks extension. This extension
3413
provides the ability to send and receive authenticated chunks and is
3414
required for secure operation of Dynamic Address Reconfiguration
3415
(ADD-IP) extension.
3417
Possible values:
3419
- 0 (disabled) - disable extension.
3420
- 1 (enabled) - enable extension
3422
Default: 0 (disabled)
3424
prsctp_enable - BOOLEAN
3425
Enable or disable the Partial Reliability extension (RFC3758) which
3426
is used to notify peers that a given DATA should no longer be expected.
3428
Possible values:
3430
- 0 (disabled) - disable extension.
3431
- 1 (enabled) - enable extension
3433
Default: 1 (enabled)
3435
max_burst - INTEGER
3436
The limit of the number of new packets that can be initially sent. It
3437
controls how bursty the generated traffic can be.
3439
Default: 4
3441
association_max_retrans - INTEGER
3442
Set the maximum number for retransmissions that an association can
3443
attempt deciding that the remote end is unreachable. If this value
3444
is exceeded, the association is terminated.
3446
Default: 10
3448
max_init_retransmits - INTEGER
3449
The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
3450
that an association will attempt before declaring the destination
3451
unreachable and terminating.
3453
Default: 8
3455
path_max_retrans - INTEGER
3456
The maximum number of retransmissions that will be attempted on a given
3457
path. Once this threshold is exceeded, the path is considered
3458
unreachable, and new traffic will use a different path when the
3459
association is multihomed.
3461
Default: 5
3463
pf_retrans - INTEGER
3464
The number of retransmissions that will be attempted on a given path
3465
before traffic is redirected to an alternate transport (should one
3466
exist). Note this is distinct from path_max_retrans, as a path that
3467
passes the pf_retrans threshold can still be used. Its only
3468
deprioritized when a transmission path is selected by the stack. This
3469
setting is primarily used to enable fast failover mechanisms without
3470
having to reduce path_max_retrans to a very low value. See:
3471
http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
3472
for details. Note also that a value of pf_retrans > path_max_retrans
3473
disables this feature. Since both pf_retrans and path_max_retrans can
3474
be changed by userspace application, a variable pf_enable is used to
3475
disable pf state.
3477
Default: 0
3479
ps_retrans - INTEGER
3480
Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
3481
from section-5 "Primary Path Switchover" in rfc7829. The primary path
3482
will be changed to another active path when the path error counter on
3483
the old primary path exceeds PSMR, so that "the SCTP sender is allowed
3484
to continue data transmission on a new working path even when the old
3485
primary destination address becomes active again". Note this feature
3486
is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
3487
and its value can't be less than 'pf_retrans' when changing by sysctl.
3489
Default: 0xffff
3491
rto_initial - INTEGER
3492
The initial round trip timeout value in milliseconds that will be used
3493
in calculating round trip times. This is the initial time interval
3494
for retransmissions.
3496
Default: 3000
3498
rto_max - INTEGER
3499
The maximum value (in milliseconds) of the round trip timeout. This
3500
is the largest time interval that can elapse between retransmissions.
3502
Default: 60000
3504
rto_min - INTEGER
3505
The minimum value (in milliseconds) of the round trip timeout. This
3506
is the smallest time interval the can elapse between retransmissions.
3508
Default: 1000
3510
hb_interval - INTEGER
3511
The interval (in milliseconds) between HEARTBEAT chunks. These chunks
3512
are sent at the specified interval on idle paths to probe the state of
3513
a given path between 2 associations.
3515
Default: 30000
3517
sack_timeout - INTEGER
3518
The amount of time (in milliseconds) that the implementation will wait
3519
to send a SACK.
3521
Default: 200
3523
valid_cookie_life - INTEGER
3524
The default lifetime of the SCTP cookie (in milliseconds). The cookie
3525
is used during association establishment.
3527
Default: 60000
3529
cookie_preserve_enable - BOOLEAN
3530
Enable or disable the ability to extend the lifetime of the SCTP cookie
3531
that is used during the establishment phase of SCTP association
3533
Possible values:
3535
- 0 (disabled) - disable.
3536
- 1 (enabled) - enable cookie lifetime extension.
3538
Default: 1 (enabled)
3540
cookie_hmac_alg - STRING
3541
Select the hmac algorithm used when generating the cookie value sent by
3542
a listening sctp socket to a connecting client in the INIT-ACK chunk.
3543
Valid values are:
3545
* sha256
3546
* none
3548
Default: sha256
3550
rcvbuf_policy - INTEGER
3551
Determines if the receive buffer is attributed to the socket or to
3552
association. SCTP supports the capability to create multiple
3553
associations on a single socket. When using this capability, it is
3554
possible that a single stalled association that's buffering a lot
3555
of data may block other associations from delivering their data by
3556
consuming all of the receive buffer space. To work around this,
3557
the rcvbuf_policy could be set to attribute the receiver buffer space
3558
to each association instead of the socket. This prevents the described
3559
blocking.
3561
- 1: rcvbuf space is per association
3562
- 0: rcvbuf space is per socket
3564
Default: 0
3566
sndbuf_policy - INTEGER
3567
Similar to rcvbuf_policy above, this applies to send buffer space.
3569
- 1: Send buffer is tracked per association
3570
- 0: Send buffer is tracked per socket.
3572
Default: 0
3574
sctp_mem - vector of 3 INTEGERs: min, pressure, max
3575
Number of pages allowed for queueing by all SCTP sockets.
3577
* min: Below this number of pages SCTP is not bothered about its
3578
memory usage. When amount of memory allocated by SCTP exceeds
3579
this number, SCTP starts to moderate memory usage.
3580
* pressure: This value was introduced to follow format of tcp_mem.
3581
* max: Maximum number of allowed pages.
3583
Default is calculated at boot time from amount of available memory.
3585
sctp_rmem - vector of 3 INTEGERs: min, default, max
3586
Only the first value ("min") is used, "default" and "max" are
3587
ignored.
3589
* min: Minimal size of receive buffer used by SCTP socket.
3590
It is guaranteed to each SCTP socket (but not association) even
3591
under moderate memory pressure.
3593
Default: 4K
3595
sctp_wmem - vector of 3 INTEGERs: min, default, max
3596
Only the first value ("min") is used, "default" and "max" are
3597
ignored.
3599
* min: Minimum size of send buffer that can be used by SCTP sockets.
3600
It is guaranteed to each SCTP socket (but not association) even
3601
under moderate memory pressure.
3603
Default: 4K
3605
addr_scope_policy - INTEGER
3606
Control IPv4 address scoping (see
3607
https://datatracker.ietf.org/doc/draft-stewart-tsvwg-sctp-ipv4/00/
3608
for details).
3610
- 0 - Disable IPv4 address scoping
3611
- 1 - Enable IPv4 address scoping
3612
- 2 - Follow draft but allow IPv4 private addresses
3613
- 3 - Follow draft but allow IPv4 link local addresses
3615
Default: 1
3617
udp_port - INTEGER
3618
The listening port for the local UDP tunneling sock. Normally it's
3619
using the IANA-assigned UDP port number 9899 (sctp-tunneling).
3621
This UDP sock is used for processing the incoming UDP-encapsulated
3622
SCTP packets (from RFC6951), and shared by all applications in the
3623
same net namespace. This UDP sock will be closed when the value is
3624
set to 0.
3626
The value will also be used to set the src port of the UDP header
3627
for the outgoing UDP-encapsulated SCTP packets. For the dest port,
3628
please refer to 'encap_port' below.
3630
Default: 0
3632
encap_port - INTEGER
3633
The default remote UDP encapsulation port.
3635
This value is used to set the dest port of the UDP header for the
3636
outgoing UDP-encapsulated SCTP packets by default. Users can also
3637
change the value for each sock/asoc/transport by using setsockopt.
3638
For further information, please refer to RFC6951.
3640
Note that when connecting to a remote server, the client should set
3641
this to the port that the UDP tunneling sock on the peer server is
3642
listening to and the local UDP tunneling sock on the client also
3643
must be started. On the server, it would get the encap_port from
3644
the incoming packet's source port.
3646
Default: 0
3648
plpmtud_probe_interval - INTEGER
3649
The time interval (in milliseconds) for the PLPMTUD probe timer,
3650
which is configured to expire after this period to receive an
3651
acknowledgment to a probe packet. This is also the time interval
3652
between the probes for the current pmtu when the probe search
3653
is done.
3655
PLPMTUD will be disabled when 0 is set, and other values for it
3656
must be >= 5000.
3658
Default: 0
3660
reconf_enable - BOOLEAN
3661
Enable or disable extension of Stream Reconfiguration functionality
3662
specified in RFC6525. This extension provides the ability to "reset"
3663
a stream, and it includes the Parameters of "Outgoing/Incoming SSN
3664
Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams".
3666
Possible values:
3668
- 0 (disabled) - Disable extension.
3669
- 1 (enabled) - Enable extension.
3671
Default: 0 (disabled)
3673
intl_enable - BOOLEAN
3674
Enable or disable extension of User Message Interleaving functionality
3675
specified in RFC8260. This extension allows the interleaving of user
3676
messages sent on different streams. With this feature enabled, I-DATA
3677
chunk will replace DATA chunk to carry user messages if also supported
3678
by the peer. Note that to use this feature, one needs to set this option
3679
to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2
3680
and SCTP_INTERLEAVING_SUPPORTED to 1.
3682
Possible values:
3684
- 0 (disabled) - Disable extension.
3685
- 1 (enabled) - Enable extension.
3687
Default: 0 (disabled)
3689
ecn_enable - BOOLEAN
3690
Control use of Explicit Congestion Notification (ECN) by SCTP.
3691
Like in TCP, ECN is used only when both ends of the SCTP connection
3692
indicate support for it. This feature is useful in avoiding losses
3693
due to congestion by allowing supporting routers to signal congestion
3694
before having to drop packets.
3696
Possible values:
3698
- 0 (disabled) - Disable ecn.
3699
- 1 (enabled) - Enable ecn.
3701
Default: 1 (enabled)
3703
l3mdev_accept - BOOLEAN
3704
Enabling this option allows a "global" bound socket to work
3705
across L3 master domains (e.g., VRFs) with packets capable of
3706
being received regardless of the L3 domain in which they
3707
originated. Only valid when the kernel was compiled with
3708
CONFIG_NET_L3_MASTER_DEV.
3710
Possible values:
3712
- 0 (disabled)
3713
- 1 (enabled)
3715
Default: 1 (enabled)
3718
``/proc/sys/net/core/*``
3719
========================
3721
Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
3724
``/proc/sys/net/unix/*``
3725
========================
3727
max_dgram_qlen - INTEGER
3728
The maximum length of dgram socket receive queue
3730
Default: 10
3. 한국어 전문 번역
영어 원문의 문단 순서와 의미를 유지한 전체 번역입니다. 코드, 함수명, symbol과 URL은 원문 표기를 유지합니다.
문서 범위
1-6이 문서는 Linux IP 계층이 `/proc/sys/net` 아래에 제공하는 sysctl 변수를 설명합니다. IPv4·IPv6의 전달과 경로 선택뿐 아니라 TCP, UDP, RAW, CIPSO, ICMP, IGMP/MLD, bridge netfilter, SCTP, Unix datagram queue까지 네트워크 namespace별 조정 지점을 한데 모은 참조 문서입니다.
.. SPDX-License-Identifier: GPL-2.0
=========
IP Sysctl
=========
IPv4 전달과 경로 MTU
7-88`ip_forward`는 interface 사이의 IPv4 packet 전달을 켭니다. 값을 바꾸면 RFC 1122 host 또는 RFC 1812 router 기본값으로 IPv4 구성 변수가 다시 초기화되므로 단순한 한 bit 변경으로 보아서는 안 됩니다. 기본은 0이고 `ip_default_ttl`의 기본 TTL은 64입니다.
`ip_no_pmtu_disc`는 Path MTU Discovery 동작을 0~3으로 고릅니다. 0은 정상 PMTUD, 1은 수신 PMTU를 `min_pmtu`로 제한, 2는 들어온 PMTU를 폐기하고 outgoing packet의 DF도 세우지 않는 모드, 3은 TCP·SCTP처럼 sequence로 검증할 수 있는 protocol만 hardened PMTU update를 허용하는 모드입니다. mode 2와 3은 namespace 전체에 적용하면 다른 traffic을 망가뜨릴 수 있어 보호가 필요한 제한된 환경에만 씁니다. `min_pmtu` 기본값은 552입니다.
`ip_forward_use_pmtu`는 forwarding 시 protocol PMTU 정보를 신뢰할지 정합니다. protocol spoofing 위험 때문에 기본은 0입니다. `fwmark_reflect`를 켜면 socket과 연결되지 않은 TCP RST, ICMP echo reply 같은 kernel 생성 응답이 원 packet의 fwmark를 이어받습니다.
/proc/sys/net/ipv4/* Variables
==============================
ip_forward - BOOLEAN
Forward Packets between interfaces.
This variable is special, its change resets all configuration
parameters to their default state (RFC1122 for hosts, RFC1812
for routers)
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
ip_default_ttl - INTEGER
Default value of TTL field (Time To Live) for outgoing (but not
forwarded) IP packets. Should be between 1 and 255 inclusive.
Default: 64 (as recommended by RFC1700)
ip_no_pmtu_disc - INTEGER
Disable Path MTU Discovery. If enabled in mode 1 and a
fragmentation-required ICMP is received, the PMTU to this
destination will be set to the smallest of the old MTU to
this destination and min_pmtu (see below). You will need
to raise min_pmtu to the smallest interface MTU on your system
manually if you want to avoid locally generated fragments.
In mode 2 incoming Path MTU Discovery messages will be
discarded. Outgoing frames are handled the same as in mode 1,
implicitly setting IP_PMTUDISC_DONT on every created socket.
Mode 3 is a hardened pmtu discover mode. The kernel will only
accept fragmentation-needed errors if the underlying protocol
can verify them besides a plain socket lookup. Current
protocols for which pmtu events will be honored are TCP and
SCTP as they verify e.g. the sequence number or the
association. This mode should not be enabled globally but is
only intended to secure e.g. name servers in namespaces where
TCP path mtu must still work but path MTU information of other
protocols should be discarded. If enabled globally this mode
could break other protocols.
Possible values: 0-3
Default: FALSE
min_pmtu - INTEGER
default 552 - minimum Path MTU. Unless this is changed manually,
each cached pmtu will never be lower than this setting.
ip_forward_use_pmtu - BOOLEAN
By default we don't trust protocol path MTUs while forwarding
because they could be easily forged and can lead to unwanted
fragmentation by the router.
You only need to enable this if you have user-space software
which tries to discover path mtus by itself and depends on the
kernel honoring this information. This is normally not the
case.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
fwmark_reflect - BOOLEAN
Controls the fwmark of kernel-generated IPv4 reply packets that are not
associated with a socket for example, TCP RSTs or ICMP echo replies).
If disabled, these packets have a fwmark of zero. If enabled, they have the
fwmark of the packet they are replying to.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
IPv4 multipath hash와 FIB 동기화
89-174`fib_multipath_use_neigh`는 ECMP nexthop 선택에 neighbor 상태를 반영합니다. 기본 0은 route group 안의 nexthop을 neighbor 상태와 무관하게 사용할 수 있고, 1은 실패한 neighbor를 피합니다.
`fib_multipath_hash_policy`는 multipath hash를 0=L3, 1=L4 5-tuple, 2=encapsulation이 있으면 inner L3, 3=사용자 지정으로 고릅니다. mode 3의 `fib_multipath_hash_fields` bitmask는 source/destination IP(0x1/0x2), protocol(0x4), source/destination port(0x10/0x20), inner IP·protocol·port(0x40~0x800)를 조합하며 기본 0x7은 양쪽 IP와 protocol입니다. `fib_multipath_hash_seed`는 namespace별 seed이고 0이면 random입니다. 같은 seed라도 kernel version 사이에 hash 결과가 안정적이라고 가정하면 안 됩니다.
`fib_sync_mem`은 route entry를 fib notifier에 동기화하면서 `synchronize_rcu()`를 호출하기 전 쌓을 수 있는 dirty memory를 byte 단위로 제한합니다. 기본 512 KiB, 최소 64 KiB, 최대 64 MiB입니다. `ip_forward_update_priority`가 1이면 forwarding 중 IPv4 header priority를 TOS에서 갱신합니다.
fib_multipath_use_neigh - BOOLEAN
Use status of existing neighbor entry when determining nexthop for
multipath routes. If disabled, neighbor information is not used and
packets could be directed to a failed nexthop. Only valid for kernels
built with CONFIG_IP_ROUTE_MULTIPATH enabled.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
fib_multipath_hash_policy - INTEGER
Controls which hash policy to use for multipath routes. Only valid
for kernels built with CONFIG_IP_ROUTE_MULTIPATH enabled.
Default: 0 (Layer 3)
Possible values:
- 0 - Layer 3
- 1 - Layer 4
- 2 - Layer 3 or inner Layer 3 if present
- 3 - Custom multipath hash. Fields used for multipath hash calculation
are determined by fib_multipath_hash_fields sysctl
fib_multipath_hash_fields - UNSIGNED INTEGER
When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
fields used for multipath hash calculation are determined by this
sysctl.
This value is a bitmask which enables various fields for multipath hash
calculation.
Possible fields are:
====== ============================
0x0001 Source IP address
0x0002 Destination IP address
0x0004 IP protocol
0x0008 Unused (Flow Label)
0x0010 Source port
0x0020 Destination port
0x0040 Inner source IP address
0x0080 Inner destination IP address
0x0100 Inner IP protocol
0x0200 Inner Flow Label
0x0400 Inner source port
0x0800 Inner destination port
====== ============================
Default: 0x0007 (source IP, destination IP and IP protocol)
fib_multipath_hash_seed - UNSIGNED INTEGER
The seed value used when calculating hash for multipath routes. Applies
to both IPv4 and IPv6 datapath. Only present for kernels built with
CONFIG_IP_ROUTE_MULTIPATH enabled.
When set to 0, the seed value used for multipath routing defaults to an
internal random-generated one.
The actual hashing algorithm is not specified -- there is no guarantee
that a next hop distribution effected by a given seed will keep stable
across kernel versions.
Default: 0 (random)
fib_sync_mem - UNSIGNED INTEGER
Amount of dirty memory from fib entries that can be backlogged before
synchronize_rcu is forced.
Default: 512kB Minimum: 64kB Maximum: 64MB
ip_forward_update_priority - INTEGER
Whether to update SKB priority from "TOS" field in IPv4 header after it
is forwarded. The new SKB priority is mapped from TOS field value
according to an rt_tos2priority table (see e.g. man tc-prio).
Default: 1 (Update priority.)
Possible values:
- 0 - Do not update priority.
- 1 - Update priority.
Route cache, neighbor queue와 offload 알림
175-266IPv4의 `route/max_size`는 3.6부터 route cache가 없어져 더 이상 의미가 없고 IPv6에는 별도 의미가 있습니다. neighbor cache의 `gc_thresh1`, `gc_thresh2`, `gc_thresh3`는 각각 최소 보존 수, 5초 이상 cache가 유지될 때 허용하는 soft maximum, 강제 GC 기준입니다. IPv6에서는 `gc_thresh3`가 동작에 필요한 최소 permanent entry보다 작아지지 않습니다.
`unres_qlen_bytes`는 address resolution이 끝나지 않은 neighbor마다 보류할 packet의 총 byte를 제한하고 기본값은 `SK_WMEM_MAX`와 같습니다. `unres_qlen`은 호환용 packet 수 제한으로 기본 101이지만 packet 메모리 overhead를 계산하지 못하므로 byte 제한이 더 정확합니다. `interval_probe_time_ms`는 NTF_USE로 queue된 probe의 최소 간격이며 기본 5000 ms입니다.
`mtu_expires`는 PMTU cache 수명, `min_adv_mss`는 첫 MSS 광고의 최소값입니다. `fib_notify_on_flag_change`는 route의 `RTM_F_OFFLOAD`, `RTM_F_TRAP`, `RTM_F_OFFLOAD_FAILED` flag 변화 때 `RTM_NEWROUTE`를 보낼지 정하며 0=없음, 1=모두, 2=offload 실패만입니다. kernel 설치 ACK는 hardware 설치까지 보장하지 않으므로 이 알림이 offload 상태 추적에 쓰입니다.
route/max_size - INTEGER
Maximum number of routes allowed in the kernel. Increase
this when using large numbers of interfaces and/or routes.
From linux kernel 3.6 onwards, this is deprecated for ipv4
as route cache is no longer used.
From linux kernel 6.3 onwards, this is deprecated for ipv6
as garbage collection manages cached route entries.
neigh/default/gc_thresh1 - INTEGER
Minimum number of entries to keep. Garbage collector will not
purge entries if there are fewer than this number.
Default: 128
neigh/default/gc_thresh2 - INTEGER
Threshold when garbage collector becomes more aggressive about
purging entries. Entries older than 5 seconds will be cleared
when over this number.
Default: 512
neigh/default/gc_thresh3 - INTEGER
Maximum number of non-PERMANENT neighbor entries allowed. Increase
this when using large numbers of interfaces and when communicating
with large numbers of directly-connected peers.
Default: 1024
neigh/default/unres_qlen_bytes - INTEGER
The maximum number of bytes which may be used by packets
queued for each unresolved address by other network layers.
(added in linux 3.3)
Setting negative value is meaningless and will return error.
Default: SK_WMEM_DEFAULT, (same as net.core.wmem_default).
Exact value depends on architecture and kernel options,
but should be enough to allow queuing 256 packets
of medium size.
neigh/default/unres_qlen - INTEGER
The maximum number of packets which may be queued for each
unresolved address by other network layers.
(deprecated in linux 3.3) : use unres_qlen_bytes instead.
Prior to linux 3.3, the default value is 3 which may cause
unexpected packet loss. The current default value is calculated
according to default value of unres_qlen_bytes and true size of
packet.
Default: 101
neigh/default/interval_probe_time_ms - INTEGER
The probe interval for neighbor entries with NTF_MANAGED flag,
the min value is 1.
Default: 5000
mtu_expires - INTEGER
Time, in seconds, that cached PMTU information is kept.
min_adv_mss - INTEGER
The advertised MSS depends on the first hop route MTU, but will
never be lower than this setting.
fib_notify_on_flag_change - INTEGER
Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
After installing a route to the kernel, user space receives an
acknowledgment, which means the route was installed in the kernel,
but not necessarily in hardware.
It is also possible for a route already installed in hardware to change
its action and therefore its flags. For example, a host route that is
trapping packets can be "promoted" to perform decapsulation following
the installation of an IPinIP/VXLAN tunnel.
The notifications will indicate to user-space the state of the route.
Default: 0 (Do not emit notifications.)
Possible values:
- 0 - Do not emit notifications.
- 1 - Emit notifications.
- 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
IP Fragmentation:
IPv4 fragment와 directed broadcast
267-308`ipfrag_high_thresh`는 IPv4 fragment 재조립에 쓸 최대 memory이고 `ipfrag_low_thresh`는 4.17부터 폐기되어 효과가 없습니다. `ipfrag_time`은 fragment를 보관하는 초 단위 시간입니다.
`ipfrag_max_dist`는 같은 source의 fragment 사이에서 다른 IP fragment가 몇 개까지 도착할 수 있는지를 제한해 fragment ID 재사용으로 잘못 합쳐지는 것을 줄입니다. 0은 검사를 끄고 기본 64는 reorder가 큰 환경에서 조기 폐기를 일으킬 수 있습니다. source별 분리 경로가 흔하면 더 크게 조정할 수 있지만 지나치게 크면 오조립 위험이 커집니다. `bc_forwarding`은 directed broadcast forwarding이며 `conf/all`과 ingress interface가 모두 켜져야 하고 기본은 0입니다.
ipfrag_high_thresh - LONG INTEGER
Maximum memory used to reassemble IP fragments.
ipfrag_low_thresh - LONG INTEGER
(Obsolete since linux-4.17)
Maximum memory used to reassemble IP fragments before the kernel
begins to remove incomplete fragment queues to free up resources.
The kernel still accepts new fragments for defragmentation.
ipfrag_time - INTEGER
Time in seconds to keep an IP fragment in memory.
ipfrag_max_dist - INTEGER
ipfrag_max_dist is a non-negative integer value which defines the
maximum "disorder" which is allowed among fragments which share a
common IP source address. Note that reordering of packets is
not unusual, but if a large number of fragments arrive from a source
IP address while a particular fragment queue remains incomplete, it
probably indicates that one or more fragments belonging to that queue
have been lost. When ipfrag_max_dist is positive, an additional check
is done on fragments before they are added to a reassembly queue - if
ipfrag_max_dist (or more) fragments have arrived from a particular IP
address between additions to any IP fragment queue using that source
address, it's presumed that one or more fragments in the queue are
lost. The existing fragment queue will be dropped, and a new one
started. An ipfrag_max_dist value of zero disables this check.
Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can
result in unnecessarily dropping fragment queues when normal
reordering of packets occurs, which could lead to poor application
performance. Using a very large value, e.g. 50000, increases the
likelihood of incorrectly reassembling IP fragments that originate
from different IP datagrams, which could result in data corruption.
Default: 64
bc_forwarding - INTEGER
bc_forwarding enables the feature described in rfc1812#section-5.3.5.2
and rfc2644. It allows the router to forward directed broadcast.
To enable this feature, the 'all' entry and the input interface entry
should be set to 1.
Default: 0
INET peer 저장소
309-329INET peer storage는 route 사이에서 오래 유지해야 하는 peer별 정보를 보관합니다. `inet_peer_threshold`를 넘으면 entry를 공격적으로 폐기하며 기본 threshold는 memory 규모에 따라 정해집니다. `inet_peer_minttl`과 `inet_peer_maxttl`은 각각 peer entry의 최소·최대 수명으로, 기본은 120초와 600초입니다. cache pressure가 커질수록 실제 TTL은 두 값 사이에서 짧아집니다.
INET peer storage
=================
inet_peer_threshold - INTEGER
The approximate size of the storage. Starting from this threshold
entries will be thrown aggressively. This threshold also determines
entries' time-to-live and time intervals between garbage collection
passes. More entries, less time-to-live, less GC interval.
inet_peer_minttl - INTEGER
Minimum time-to-live of entries. Should be enough to cover fragment
time-to-live on the reassembling side. This minimum time-to-live is
guaranteed if the pool size is less than inet_peer_threshold.
Measured in seconds.
inet_peer_maxttl - INTEGER
Maximum time-to-live of entries. Unused entries will expire after
this period of time if there is no memory pressure on the pool (i.e.
when the number of entries in the pool is very small).
Measured in seconds.
TCP listen, window와 congestion control
330-443`somaxconn`은 listen backlog의 상한으로 기본 4096입니다. `tcp_abort_on_overflow`를 켜면 application이 accept를 따라가지 못할 때 연결을 조용히 재시도하게 두는 대신 reset하지만 client에 해를 줄 수 있습니다. `tcp_adv_win_scale`는 6.6부터 폐기되었고, `tcp_app_win`은 receive window 중 application buffer로 남길 비율을 정합니다.
`tcp_allowed_congestion_control`은 unprivileged process가 고를 수 있는 congestion algorithm 목록이고 `tcp_available_congestion_control`은 등록된 전체 목록, `tcp_congestion_control`은 새 연결의 기본 algorithm입니다. `tcp_autocorking`은 같은 flow의 앞 packet이 qdisc나 transmit queue에 있을 때 작은 write를 합칩니다.
`tcp_base_mss`는 MTU probing의 search_low, `tcp_mtu_probe_floor`는 마지막 probe 하한, `tcp_min_snd_mss`는 수신 MSS에 관계없이 outgoing segment가 지켜야 할 최소값입니다. `tcp_dsack`은 RFC 2883 duplicate SACK을, `tcp_early_retrans`는 Tail Loss Probe를 포함한 early retransmit 수준을 제어하며 값 3이 기본 TLP 모드입니다.
TCP variables
=============
somaxconn - INTEGER
Limit of socket listen() backlog, known in userspace as SOMAXCONN.
Defaults to 4096. (Was 128 before linux-5.4)
See also tcp_max_syn_backlog for additional tuning for TCP sockets.
tcp_abort_on_overflow - BOOLEAN
If listening service is too slow to accept new connections,
reset them. Default state is FALSE. It means that if overflow
occurred due to a burst, connection will recover. Enable this
option _only_ if you are really sure that listening daemon
cannot be tuned to accept connections faster. Enabling this
option can harm clients of your server.
tcp_adv_win_scale - INTEGER
Obsolete since linux-6.6
Count buffering overhead as bytes/2^tcp_adv_win_scale
(if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale),
if it is <= 0.
Possible values are [-31, 31], inclusive.
Default: 1
tcp_allowed_congestion_control - STRING
Show/set the congestion control choices available to non-privileged
processes. The list is a subset of those listed in
tcp_available_congestion_control.
Default is "reno" and the default setting (tcp_congestion_control).
tcp_app_win - INTEGER
Reserve max(window/2^tcp_app_win, mss) of window for application
buffer. Value 0 is special, it means that nothing is reserved.
Possible values are [0, 31], inclusive.
Default: 31
tcp_autocorking - BOOLEAN
Enable TCP auto corking :
When applications do consecutive small write()/sendmsg() system calls,
we try to coalesce these small writes as much as possible, to lower
total amount of sent packets. This is done if at least one prior
packet for the flow is waiting in Qdisc queues or device transmit
queue. Applications can still use TCP_CORK for optimal behavior
when they know how/when to uncork their sockets.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_available_congestion_control - STRING
Shows the available congestion control choices that are registered.
More congestion control algorithms may be available as modules,
but not loaded.
tcp_base_mss - INTEGER
The initial value of search_low to be used by the packetization layer
Path MTU discovery (MTU probing). If MTU probing is enabled,
this is the initial MSS used by the connection.
tcp_mtu_probe_floor - INTEGER
If MTU probing is enabled this caps the minimum MSS used for search_low
for the connection.
Default : 48
tcp_min_snd_mss - INTEGER
TCP SYN and SYNACK messages usually advertise an ADVMSS option,
as described in RFC 1122 and RFC 6691.
If this ADVMSS option is smaller than tcp_min_snd_mss,
it is silently capped to tcp_min_snd_mss.
Default : 48 (at least 8 bytes of payload per segment)
tcp_congestion_control - STRING
Set the congestion control algorithm to be used for new
connections. The algorithm "reno" is always available, but
additional choices may be available based on kernel configuration.
Default is set as part of kernel configuration.
For passive connections, the listener congestion control choice
is inherited.
[see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ]
tcp_dsack - BOOLEAN
Allows TCP to send "duplicate" SACKs.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_early_retrans - INTEGER
Tail loss probe (TLP) converts RTOs occurring due to tail
losses into fast recovery (RFC8985). Note that
TLP requires RACK to function properly (see tcp_recovery below)
Possible values:
- 0 disables TLP
- 3 or 4 enables TLP
Default: 3
TCP ECN, 종료와 keepalive
444-603`tcp_ecn`은 incoming/outgoing ECN 협상을 0~5로 세분합니다. 0은 ECN을 요청하지도 수락하지도 않고, 1/2는 classic ECN을 양방향 또는 incoming 위주로, 3/4/5는 AccECN을 양방향·incoming·fallback 조합으로 사용합니다. 기본 2는 outgoing SYN에서는 요청하지 않지만 peer 요청은 받아들입니다. `tcp_ecn_option`은 AccECN option을 0=없음, 1=필요할 때, 2=가능하면 매 packet으로 보내고 `tcp_ecn_option_beacon`은 RTT당 최소 beacon 수입니다. `tcp_ecn_fallback`은 RFC 3168 fallback을 켭니다.
`tcp_fack`은 효과가 없는 legacy 변수입니다. `tcp_fin_timeout`은 orphaned FIN_WAIT_2 연결의 보존 시간(기본 60초), `tcp_frto`는 F-RTO 복구, `tcp_fwmark_accept`는 passive connection의 fwmark 상속을 제어합니다. `tcp_invalid_ratelimit`은 invalid duplicate ACK에 대한 challenge ACK 사이의 최소 간격으로 기본 500 ms이며 0은 제한을 끕니다.
keepalive는 `tcp_keepalive_time` 7200초, `tcp_keepalive_probes` 9회, `tcp_keepalive_intvl` 75초가 기본입니다. socket option으로 override할 수 있습니다. `tcp_l3mdev_accept`는 global bound TCP socket이 VRF 같은 L3 master domain을 가로질러 받을 수 있게 하고 기본 0입니다. `tcp_low_latency`는 4.14부터 효과가 없는 legacy 설정입니다.
tcp_ecn - INTEGER
Control use of Explicit Congestion Notification (ECN) by TCP.
ECN is used only when both ends of the TCP connection indicate support
for it. This feature is useful in avoiding losses due to congestion by
allowing supporting routers to signal congestion before having to drop
packets. A host that supports ECN both sends ECN at the IP layer and
feeds back ECN at the TCP layer. The highest variant of ECN feedback
that both peers support is chosen by the ECN negotiation (Accurate ECN,
ECN, or no ECN).
The highest negotiated variant for incoming connection requests
and the highest variant requested by outgoing connection
attempts:
===== ==================== ====================
Value Incoming connections Outgoing connections
===== ==================== ====================
0 No ECN No ECN
1 ECN ECN
2 ECN No ECN
3 AccECN AccECN
4 AccECN ECN
5 AccECN No ECN
===== ==================== ====================
Default: 2
tcp_ecn_option - INTEGER
Control Accurate ECN (AccECN) option sending when AccECN has been
successfully negotiated during handshake. Send logic inhibits
sending AccECN options regarless of this setting when no AccECN
option has been seen for the reverse direction.
Possible values are:
= ============================================================
0 Never send AccECN option. This also disables sending AccECN
option in SYN/ACK during handshake.
1 Send AccECN option sparingly according to the minimum option
rules outlined in draft-ietf-tcpm-accurate-ecn.
2 Send AccECN option on every packet whenever it fits into TCP
option space.
= ============================================================
Default: 2
tcp_ecn_option_beacon - INTEGER
Control Accurate ECN (AccECN) option sending frequency per RTT and it
takes effect only when tcp_ecn_option is set to 2.
Default: 3 (AccECN will be send at least 3 times per RTT)
tcp_ecn_fallback - BOOLEAN
If the kernel detects that ECN connection misbehaves, enable fall
back to non-ECN. Currently, this knob implements the fallback
from RFC3168, section 6.1.1.1., but we reserve that in future,
additional detection mechanisms could be implemented under this
knob. The value is not used, if tcp_ecn or per route (or congestion
control) ECN settings are disabled.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_fack - BOOLEAN
This is a legacy option, it has no effect anymore.
tcp_fin_timeout - INTEGER
The length of time an orphaned (no longer referenced by any
application) connection will remain in the FIN_WAIT_2 state
before it is aborted at the local end. While a perfectly
valid "receive only" state for an un-orphaned connection, an
orphaned connection in FIN_WAIT_2 state could otherwise wait
forever for the remote to close its end of the connection.
Cf. tcp_max_orphans
Default: 60 seconds
tcp_frto - INTEGER
Enables Forward RTO-Recovery (F-RTO) defined in RFC5682.
F-RTO is an enhanced recovery algorithm for TCP retransmission
timeouts. It is particularly beneficial in networks where the
RTT fluctuates (e.g., wireless). F-RTO is sender-side only
modification. It does not require any support from the peer.
By default it's enabled with a non-zero value. 0 disables F-RTO.
tcp_fwmark_accept - BOOLEAN
If enabled, incoming connections to listening sockets that do not have a
socket mark will set the mark of the accepting socket to the fwmark of
the incoming SYN packet. This will cause all packets on that connection
(starting from the first SYNACK) to be sent with that fwmark. The
listening socket's mark is unchanged. Listening sockets that already
have a fwmark set via setsockopt(SOL_SOCKET, SO_MARK, ...) are
unaffected.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_invalid_ratelimit - INTEGER
Limit the maximal rate for sending duplicate acknowledgments
in response to incoming TCP packets that are for an existing
connection but that are invalid due to any of these reasons:
(a) out-of-window sequence number,
(b) out-of-window acknowledgment number, or
(c) PAWS (Protection Against Wrapped Sequence numbers) check failure
This can help mitigate simple "ack loop" DoS attacks, wherein
a buggy or malicious middlebox or man-in-the-middle can
rewrite TCP header fields in manner that causes each endpoint
to think that the other is sending invalid TCP segments, thus
causing each side to send an unterminating stream of duplicate
acknowledgments for invalid segments.
Using 0 disables rate-limiting of dupacks in response to
invalid segments; otherwise this value specifies the minimal
space between sending such dupacks, in milliseconds.
Default: 500 (milliseconds).
tcp_keepalive_time - INTEGER
How often TCP sends out keepalive messages when keepalive is enabled.
Default: 2hours.
tcp_keepalive_probes - INTEGER
How many keepalive probes TCP sends out, until it decides that the
connection is broken. Default value: 9.
tcp_keepalive_intvl - INTEGER
How frequently the probes are send out. Multiplied by
tcp_keepalive_probes it is time to kill not responding connection,
after probes started. Default value: 75sec i.e. connection
will be aborted after ~11 minutes of retries.
tcp_l3mdev_accept - BOOLEAN
Enables child sockets to inherit the L3 master device index.
Enabling this option allows a "global" listen socket to work
across L3 master domains (e.g., VRFs) with connected sockets
derived from the listen socket to be bound to the L3 domain in
which the packets originated. Only valid when the kernel was
compiled with CONFIG_NET_L3_MASTER_DEV.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_low_latency - BOOLEAN
This is a legacy option, it has no effect anymore.TCP memory, PMTU와 metric cache
604-730`tcp_max_orphans`는 user file handle에 붙지 않은 TCP socket 수의 상한입니다. orphan 하나가 대략 64 KiB의 unswappable memory를 쓸 수 있어 과도한 증가는 DoS 위험을 키웁니다. `tcp_max_syn_backlog`은 아직 client ACK를 받지 못한 remembered connection request 수, `tcp_max_tw_buckets`는 TIME_WAIT socket 수의 전역 상한입니다.
`tcp_mem`의 min/pressure/max는 모든 TCP socket queue가 쓸 수 있는 page 수입니다. `tcp_min_rtt_wlen`은 min RTT filter window, `tcp_moderate_rcvbuf`는 receive buffer autotuning을 켭니다. `tcp_mtu_probing`은 0=끔, 1=ICMP black hole 감지 시 시작, 2=항상 시작이고 초기 MSS는 `tcp_base_mss`입니다. `tcp_probe_interval`은 probe 재시도 간격, `tcp_probe_threshold`는 search range를 멈출 임계값입니다.
`tcp_no_metrics_save`는 connection 종료 때 metric cache 저장을 막고 `tcp_no_ssthresh_metrics_save`는 ssthresh만 저장하지 않습니다. `tcp_orphan_retries`는 local TCP connection의 orphan probe 횟수이며 0이어도 최대 8회까지 재시도할 수 있어 즉시 종료를 뜻하지 않습니다.
tcp_max_orphans - INTEGER
Maximal number of TCP sockets not attached to any user file handle,
held by system. If this number is exceeded orphaned connections are
reset immediately and warning is printed. This limit exists
only to prevent simple DoS attacks, you _must_ not rely on this
or lower the limit artificially, but rather increase it
(probably, after increasing installed memory),
if network conditions require more than default value,
and tune network services to linger and kill such states
more aggressively. Let me to remind again: each orphan eats
up to ~64K of unswappable memory.
tcp_max_syn_backlog - INTEGER
Maximal number of remembered connection requests (SYN_RECV),
which have not received an acknowledgment from connecting client.
This is a per-listener limit.
The minimal value is 128 for low memory machines, and it will
increase in proportion to the memory of machine.
If server suffers from overload, try increasing this number.
Remember to also check /proc/sys/net/core/somaxconn
A SYN_RECV request socket consumes about 304 bytes of memory.
tcp_max_tw_buckets - INTEGER
Maximal number of timewait sockets held by system simultaneously.
If this number is exceeded time-wait socket is immediately destroyed
and warning is printed. This limit exists only to prevent
simple DoS attacks, you _must_ not lower the limit artificially,
but rather increase it (probably, after increasing installed memory),
if network conditions require more than default value.
tcp_mem - vector of 3 INTEGERs: min, pressure, max
min: below this number of pages TCP is not bothered about its
memory appetite.
pressure: when amount of memory allocated by TCP exceeds this number
of pages, TCP moderates its memory consumption and enters memory
pressure mode, which is exited when memory consumption falls
under "min".
max: number of pages allowed for queueing by all TCP sockets.
Defaults are calculated at boot time from amount of available
memory.
tcp_min_rtt_wlen - INTEGER
The window length of the windowed min filter to track the minimum RTT.
A shorter window lets a flow more quickly pick up new (higher)
minimum RTT when it is moved to a longer path (e.g., due to traffic
engineering). A longer window makes the filter more resistant to RTT
inflations such as transient congestion. The unit is seconds.
Possible values: 0 - 86400 (1 day)
Default: 300
tcp_moderate_rcvbuf - BOOLEAN
If enabled, TCP performs receive buffer auto-tuning, attempting to
automatically size the buffer (no greater than tcp_rmem[2]) to
match the size required by the path for full throughput.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_mtu_probing - INTEGER
Controls TCP Packetization-Layer Path MTU Discovery. Takes three
values:
- 0 - Disabled
- 1 - Disabled by default, enabled when an ICMP black hole detected
- 2 - Always enabled, use initial MSS of tcp_base_mss.
tcp_probe_interval - UNSIGNED INTEGER
Controls how often to start TCP Packetization-Layer Path MTU
Discovery reprobe. The default is reprobing every 10 minutes as
per RFC4821.
tcp_probe_threshold - INTEGER
Controls when TCP Packetization-Layer Path MTU Discovery probing
will stop in respect to the width of search range in bytes. Default
is 8 bytes.
tcp_no_metrics_save - BOOLEAN
By default, TCP saves various connection metrics in the route cache
when the connection closes, so that connections established in the
near future can use these to set initial conditions. Usually, this
increases overall performance, but may sometimes cause performance
degradation. If enabled, TCP will not cache metrics on closing
connections.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_no_ssthresh_metrics_save - BOOLEAN
Controls whether TCP saves ssthresh metrics in the route cache.
If enabled, ssthresh metrics are disabled.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_orphan_retries - INTEGER
This value influences the timeout of a locally closed TCP connection,
when RTO retransmissions remain unacknowledged.
See tcp_retries2 for more details.
The default value is 8.
If your machine is a loaded WEB server,
you should think about lowering this value, such sockets
may consume significant resources. Cf. tcp_max_orphans.
TCP RACK, reordering과 재전송
731-815`tcp_recovery`는 RACK loss detection option bitmap입니다. bit 0x1은 RACK, 0x2는 RACK의 static reordering window, 0x4는 DUPACK threshold로 reorder window를 끄는 동작을 비활성화합니다. 기본 0x1입니다. `tcp_reflect_tos`는 incoming SYN의 DSCP를 child socket에 반영하며 ECN bit는 반영하지 않습니다.
`tcp_reordering`은 packet reordering의 초기 추정값(기본 3), `tcp_max_reordering`은 최대 추정값(기본 300)입니다. `tcp_retrans_collapse`는 retransmit 때 full-sized packet으로 합치려 하지만 수신 측 bug 우회가 필요할 수 있습니다. `tcp_retries1`은 IP layer에 network 상태를 다시 평가시키기 전 재시도, `tcp_retries2`는 established connection을 포기하기 전 재시도이며 RFC 1122 기준상 최소 R2는 100초 이상에 해당해야 합니다.
tcp_recovery - INTEGER
This value is a bitmap to enable various experimental loss recovery
features.
========= =============================================================
RACK: 0x1 enables RACK loss detection, for fast detection of lost
retransmissions and tail drops, and resilience to
reordering. currently, setting this bit to 0 has no
effect, since RACK is the only supported loss detection
algorithm.
RACK: 0x2 makes RACK's reordering window static (min_rtt/4).
RACK: 0x4 disables RACK's DUPACK threshold heuristic
========= =============================================================
Default: 0x1
tcp_reflect_tos - BOOLEAN
For listening sockets, reuse the DSCP value of the initial SYN message
for outgoing packets. This allows to have both directions of a TCP
stream to use the same DSCP value, assuming DSCP remains unchanged for
the lifetime of the connection.
This options affects both IPv4 and IPv6.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_reordering - INTEGER
Initial reordering level of packets in a TCP stream.
TCP stack can then dynamically adjust flow reordering level
between this initial value and tcp_max_reordering
Default: 3
tcp_max_reordering - INTEGER
Maximal reordering level of packets in a TCP stream.
300 is a fairly conservative value, but you might increase it
if paths are using per packet load balancing (like bonding rr mode)
Default: 300
tcp_retrans_collapse - BOOLEAN
Bug-to-bug compatibility with some broken printers.
On retransmit try to send bigger packets to work around bugs in
certain TCP stacks.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_retries1 - INTEGER
This value influences the time, after which TCP decides, that
something is wrong due to unacknowledged RTO retransmissions,
and reports this suspicion to the network layer.
See tcp_retries2 for more details.
RFC 1122 recommends at least 3 retransmissions, which is the
default.
tcp_retries2 - INTEGER
This value influences the timeout of an alive TCP connection,
when RTO retransmissions remain unacknowledged.
Given a value of N, a hypothetical TCP connection following
exponential backoff with an initial RTO of TCP_RTO_MIN would
retransmit N times before killing the connection at the (N+1)th RTO.
The default value of 15 yields a hypothetical timeout of 924.6
seconds and is a lower bound for the effective timeout.
TCP will effectively time out at the first RTO which exceeds the
hypothetical timeout.
If tcp_rto_max_ms is decreased, it is recommended to also
change tcp_retries2.
RFC 1122 recommends at least 100 seconds for the timeout,
which corresponds to a value of at least 8.
TCP receive buffer, SACK와 SYN 보호
816-976`tcp_rfc1337`은 TIME_WAIT assassination으로부터 RFC 1337 방식으로 보호합니다. `tcp_rmem`의 min/default/max는 receive buffer의 최소 보장, 기본값, autotuning 최대값이며 `SO_RCVBUF`는 이를 override합니다. `tcp_sack`은 selective ACK을 켭니다.
SACK compression은 `tcp_comp_sack_delay_ns`만큼 ACK을 늦춰 압축하고 `tcp_comp_sack_slack_ns`는 timer slack, `tcp_comp_sack_nr`는 한 packet에 담을 최대 SACK 수입니다. `tcp_backlog_ack_defer`는 process context가 packet을 처리할 때 ACK 전송을 backlog 끝까지 미루고, `tcp_slow_start_after_idle`은 idle 뒤 congestion window를 timeout 기준으로 줄입니다. `tcp_stdurg`는 RFC 1122 urgent pointer 해석을 사용하지만 기본은 BSD 호환 방식입니다.
`tcp_synack_retries`는 passive SYN/ACK 재전송 횟수입니다. `tcp_syncookies`는 0=끔, 1=SYN backlog overflow 때만, 2=항상이며 정상 부하 확장용이 아니라 마지막 방어 수단입니다. cookies는 TCP option 정보를 잃을 수 있고 listener overload를 고치지 않습니다. `tcp_migrate_req`는 `SO_REUSEPORT` listener가 죽을 때 handshake request를 다른 listener로 옮기며 BPF program이 대상 socket을 선택할 수 있습니다.
tcp_rfc1337 - BOOLEAN
If enabled, the TCP stack behaves conforming to RFC1337. If unset,
we are not conforming to RFC, but prevent TCP TIME_WAIT
assassination.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_rmem - vector of 3 INTEGERs: min, default, max
min: Minimal size of receive buffer used by TCP sockets.
It is guaranteed to each TCP socket, even under moderate memory
pressure.
Default: 4K
default: initial size of receive buffer used by TCP sockets.
This value overrides net.core.rmem_default used by other protocols.
Default: 131072 bytes.
This value results in initial window of 65535.
max: maximal size of receive buffer allowed for automatically
selected receiver buffers for TCP socket.
Calling setsockopt() with SO_RCVBUF disables
automatic tuning of that socket's receive buffer size, in which
case this value is ignored.
Default: between 131072 and 32MB, depending on RAM size.
tcp_sack - BOOLEAN
Enable select acknowledgments (SACKS).
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_comp_sack_delay_ns - LONG INTEGER
TCP tries to reduce number of SACK sent, using a timer
based on 5% of SRTT, capped by this sysctl, in nano seconds.
The default is 1ms, based on TSO autosizing period.
Default : 1,000,000 ns (1 ms)
tcp_comp_sack_slack_ns - LONG INTEGER
This sysctl control the slack used when arming the
timer used by SACK compression. This gives extra time
for small RTT flows, and reduces system overhead by allowing
opportunistic reduction of timer interrupts.
Default : 100,000 ns (100 us)
tcp_comp_sack_nr - INTEGER
Max number of SACK that can be compressed.
Using 0 disables SACK compression.
Default : 44
tcp_backlog_ack_defer - BOOLEAN
If enabled, user thread processing socket backlog tries sending
one ACK for the whole queue. This helps to avoid potential
long latencies at end of a TCP socket syscall.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_slow_start_after_idle - BOOLEAN
If enabled, provide RFC2861 behavior and time out the congestion
window after an idle period. An idle period is defined at
the current RTO. If unset, the congestion window will not
be timed out after an idle period.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_stdurg - BOOLEAN
Use the Host requirements interpretation of the TCP urgent pointer field.
Most hosts use the older BSD interpretation, so if enabled,
Linux might not communicate correctly with them.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_synack_retries - INTEGER
Number of times SYNACKs for a passive TCP connection attempt will
be retransmitted. Should not be higher than 255. Default value
is 5, which corresponds to 31seconds till the last retransmission
with the current initial RTO of 1second. With this the final timeout
for a passive TCP connection will happen after 63seconds.
tcp_syncookies - INTEGER
Only valid when the kernel was compiled with CONFIG_SYN_COOKIES
Send out syncookies when the syn backlog queue of a socket
overflows. This is to prevent against the common 'SYN flood attack'
Default: 1
Note, that syncookies is fallback facility.
It MUST NOT be used to help highly loaded servers to stand
against legal connection rate. If you see SYN flood warnings
in your logs, but investigation shows that they occur
because of overload with legal connections, you should tune
another parameters until this warning disappear.
See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow.
syncookies seriously violate TCP protocol, do not allow
to use TCP extensions, can result in serious degradation
of some services (f.e. SMTP relaying), visible not by you,
but your clients and relays, contacting you. While you see
SYN flood warnings in logs not being really flooded, your server
is seriously misconfigured.
If you want to test which effects syncookies have to your
network connections you can set this knob to 2 to enable
unconditionally generation of syncookies.
tcp_migrate_req - BOOLEAN
The incoming connection is tied to a specific listening socket when
the initial SYN packet is received during the three-way handshake.
When a listener is closed, in-flight request sockets during the
handshake and established sockets in the accept queue are aborted.
If the listener has SO_REUSEPORT enabled, other listeners on the
same port should have been able to accept such connections. This
option makes it possible to migrate such child sockets to another
listener after close() or shutdown().
The BPF_SK_REUSEPORT_SELECT_OR_MIGRATE type of eBPF program should
usually be used to define the policy to pick an alive listener.
Otherwise, the kernel will randomly pick an alive listener only if
this option is enabled.
Note that migration between listeners with different settings may
crash applications. Let's say migration happens from listener A to
B, and only B has TCP_SAVE_SYN enabled. B cannot read SYN data from
the requests migrated from A. To avoid such a situation, cancel
migration by returning SK_DROP in the type of eBPF program, or
disable this option.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
TCP Fast Open, timestamp와 pacing
977-1125`tcp_fastopen`은 bitmap입니다. 0x1 client, 0x2 server, 0x4 cookie 없는 client, 0x200 cookie 없는 server, 0x400은 `TCP_FASTOPEN` socket option 없이 모든 listener를 켭니다. server 쪽은 listener 또는 0x400 설정이 추가로 필요합니다. `tcp_fastopen_blackhole_timeout_sec`는 TFO blackhole 감지 뒤 기능을 끄는 초기 timeout이고 연속 실패하면 지수 backoff합니다. `tcp_fastopen_key`는 최대 4개의 128-bit key를 지정하며 첫 key로 cookie를 만들고 나머지는 검증에만 씁니다.
`tcp_syn_retries`는 active SYN 재전송 횟수, `tcp_timestamps`는 0=끔, 1=random offset timestamp, 2=offset 없는 timestamp입니다. `tcp_min_tso_segs`는 TSO frame당 최소 segment 수이고 `tcp_tso_rtt_log`는 RTT에 따라 autosizing하는 packet 크기의 지수 기준을 바꿉니다.
`tcp_pacing_ss_ratio`와 `tcp_pacing_ca_ratio`는 slow start와 congestion avoidance에서 internal pacing rate를 current rate의 몇 %로 잡을지 정합니다. `tcp_syn_linear_timeouts` 횟수까지 SYN timeout을 선형으로 둔 뒤 exponential backoff하며 기본 4는 초기 RTO 1초에서 6번째 재전송까지 약 10초가 걸립니다. `tcp_tso_win_divisor`는 한 TSO frame이 congestion window의 어느 분수까지 쓸지 정합니다.
tcp_fastopen - INTEGER
Enable TCP Fast Open (RFC7413) to send and accept data in the opening
SYN packet.
The client support is enabled by flag 0x1 (on by default). The client
then must use sendmsg() or sendto() with the MSG_FASTOPEN flag,
rather than connect() to send data in SYN.
The server support is enabled by flag 0x2 (off by default). Then
either enable for all listeners with another flag (0x400) or
enable individual listeners via TCP_FASTOPEN socket option with
the option value being the length of the syn-data backlog.
The values (bitmap) are
===== ======== ======================================================
0x1 (client) enables sending data in the opening SYN on the client.
0x2 (server) enables the server support, i.e., allowing data in
a SYN packet to be accepted and passed to the
application before 3-way handshake finishes.
0x4 (client) send data in the opening SYN regardless of cookie
availability and without a cookie option.
0x200 (server) accept data-in-SYN w/o any cookie option present.
0x400 (server) enable all listeners to support Fast Open by
default without explicit TCP_FASTOPEN socket option.
===== ======== ======================================================
Default: 0x1
Note that additional client or server features are only
effective if the basic support (0x1 and 0x2) are enabled respectively.
tcp_fastopen_blackhole_timeout_sec - INTEGER
Initial time period in second to disable Fastopen on active TCP sockets
when a TFO firewall blackhole issue happens.
This time period will grow exponentially when more blackhole issues
get detected right after Fastopen is re-enabled and will reset to
initial value when the blackhole issue goes away.
0 to disable the blackhole detection.
By default, it is set to 0 (feature is disabled).
tcp_fastopen_key - list of comma separated 32-digit hexadecimal INTEGERs
The list consists of a primary key and an optional backup key. The
primary key is used for both creating and validating cookies, while the
optional backup key is only used for validating cookies. The purpose of
the backup key is to maximize TFO validation when keys are rotated.
A randomly chosen primary key may be configured by the kernel if
the tcp_fastopen sysctl is set to 0x400 (see above), or if the
TCP_FASTOPEN setsockopt() optname is set and a key has not been
previously configured via sysctl. If keys are configured via
setsockopt() by using the TCP_FASTOPEN_KEY optname, then those
per-socket keys will be used instead of any keys that are specified via
sysctl.
A key is specified as 4 8-digit hexadecimal integers which are separated
by a '-' as: xxxxxxxx-xxxxxxxx-xxxxxxxx-xxxxxxxx. Leading zeros may be
omitted. A primary and a backup key may be specified by separating them
by a comma. If only one key is specified, it becomes the primary key and
any previously configured backup keys are removed.
tcp_syn_retries - INTEGER
Number of times initial SYNs for an active TCP connection attempt
will be retransmitted. Should not be higher than 127. Default value
is 6, which corresponds to 67seconds (with tcp_syn_linear_timeouts = 4)
till the last retransmission with the current initial RTO of 1second.
With this the final timeout for an active TCP connection attempt
will happen after 131seconds.
tcp_timestamps - INTEGER
Enable timestamps as defined in RFC1323.
- 0: Disabled.
- 1: Enable timestamps as defined in RFC1323 and use random offset for
each connection rather than only using the current time.
- 2: Like 1, but without random offsets.
Default: 1
tcp_min_tso_segs - INTEGER
Minimal number of segments per TSO frame.
Since linux-3.12, TCP does an automatic sizing of TSO frames,
depending on flow rate, instead of filling 64Kbytes packets.
For specific usages, it's possible to force TCP to build big
TSO frames. Note that TCP stack might split too big TSO packets
if available window is too small.
Default: 2
tcp_tso_rtt_log - INTEGER
Adjustment of TSO packet sizes based on min_rtt
Starting from linux-5.18, TCP autosizing can be tweaked
for flows having small RTT.
Old autosizing was splitting the pacing budget to send 1024 TSO
per second.
tso_packet_size = sk->sk_pacing_rate / 1024;
With the new mechanism, we increase this TSO sizing using:
distance = min_rtt_usec / (2^tcp_tso_rtt_log)
tso_packet_size += gso_max_size >> distance;
This means that flows between very close hosts can use bigger
TSO packets, reducing their cpu costs.
If you want to use the old autosizing, set this sysctl to 0.
Default: 9 (2^9 = 512 usec)
tcp_pacing_ss_ratio - INTEGER
sk->sk_pacing_rate is set by TCP stack using a ratio applied
to current rate. (current_rate = cwnd * mss / srtt)
If TCP is in slow start, tcp_pacing_ss_ratio is applied
to let TCP probe for bigger speeds, assuming cwnd can be
doubled every other RTT.
Default: 200
tcp_pacing_ca_ratio - INTEGER
sk->sk_pacing_rate is set by TCP stack using a ratio applied
to current rate. (current_rate = cwnd * mss / srtt)
If TCP is in congestion avoidance phase, tcp_pacing_ca_ratio
is applied to conservatively probe for bigger throughput.
Default: 120
tcp_syn_linear_timeouts - INTEGER
The number of times for an active TCP connection to retransmit SYNs with
a linear backoff timeout before defaulting to an exponential backoff
timeout. This has no effect on SYNACK at the passive TCP side.
With an initial RTO of 1 and tcp_syn_linear_timeouts = 4 we would
expect SYN RTOs to be: 1, 1, 1, 1, 1, 2, 4, ... (4 linear timeouts,
and the first exponential backoff using 2^0 * initial_RTO).
Default: 4
tcp_tso_win_divisor - INTEGER
This allows control over what percentage of the congestion window
can be consumed by a single TSO frame.
The setting of this parameter is a choice between burstiness and
building larger TSO frames.
Default: 3
TCP TIME_WAIT, window와 send buffer
1126-1200`tcp_tw_reuse`는 TIME_WAIT socket 재사용을 0=끔, 1=전역, 2=loopback만으로 정하며 기본 2입니다. `tcp_tw_reuse_delay`는 재사용 전 최소 지연으로 기본 1000 ms이며 peer timestamp clock tick보다 작게 두면 안 됩니다.
`tcp_window_scaling`은 RFC 1323 window scaling을 사용합니다. `tcp_shrink_window`는 receive buffer pressure 때 window를 이미 광고한 크기보다 줄일 수 있게 하지만 peer가 shrink를 잘못 처리하면 문제가 생길 수 있습니다. `tcp_wmem`의 min/default/max는 send buffer memory 보장과 autotuning 범위이며 `SO_SNDBUF`가 override할 수 있습니다.
tcp_tw_reuse - INTEGER
Enable reuse of TIME-WAIT sockets for new connections when it is
safe from protocol viewpoint.
- 0 - disable
- 1 - global enable
- 2 - enable for loopback traffic only
It should not be changed without advice/request of technical
experts.
Default: 2
tcp_tw_reuse_delay - UNSIGNED INTEGER
The delay in milliseconds before a TIME-WAIT socket can be reused by a
new connection, if TIME-WAIT socket reuse is enabled. The actual reuse
threshold is within [N, N+1] range, where N is the requested delay in
milliseconds, to ensure the delay interval is never shorter than the
configured value.
This setting contains an assumption about the other TCP timestamp clock
tick interval. It should not be set to a value lower than the peer's
clock tick for PAWS (Protection Against Wrapped Sequence numbers)
mechanism work correctly for the reused connection.
Default: 1000 (milliseconds)
tcp_window_scaling - BOOLEAN
Enable window scaling as defined in RFC1323.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
tcp_shrink_window - BOOLEAN
This changes how the TCP receive window is calculated.
RFC 7323, section 2.4, says there are instances when a retracted
window can be offered, and that TCP implementations MUST ensure
that they handle a shrinking window, as specified in RFC 1122.
Possible values:
- 0 (disabled) - The window is never shrunk.
- 1 (enabled) - The window is shrunk when necessary to remain within
the memory limit set by autotuning (sk_rcvbuf).
This only occurs if a non-zero receive window
scaling factor is also in effect.
Default: 0 (disabled)
tcp_wmem - vector of 3 INTEGERs: min, default, max
min: Amount of memory reserved for send buffers for TCP sockets.
Each TCP socket has rights to use it due to fact of its birth.
Default: 4K
default: initial size of send buffer used by TCP sockets. This
value overrides net.core.wmem_default used by other protocols.
It is usually lower than net.core.wmem_default.
Default: 16K
max: Maximal amount of memory allowed for automatically tuned
send buffers for TCP sockets. This value does not override
net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables
automatic tuning of that socket's send buffer size, in which case
this value is ignored.
Default: between 64K and 4MB, depending on RAM size.
TCP unsent queue, hash와 PLB
1201-1373`tcp_notsent_lowat`는 `TCP_NOTSENT_LOWAT`을 따로 설정하지 않은 socket의 unsent byte 기준이며 기본 `UINT_MAX`는 제한 없음입니다. `tcp_workaround_signed_windows`는 window scaling 없는 잘못된 peer의 signed window를 우회합니다. `tcp_thin_linear_timeouts`는 얇은 stream에서 최대 6회의 RTO를 선형으로 두고, `tcp_limit_output_bytes`는 qdisc/device queue에 쌓을 socket별 byte를 제한해 bufferbloat를 줄입니다. `tcp_challenge_ack_limit`은 4.7부터 per-netns random 값으로 바뀌어 legacy global limit은 사실상 무한입니다.
`tcp_ehash_entries`는 현재 namespace TCP established hash bucket 수를 읽으며 초기 namespace에서만 boot parameter로 쓸 수 있습니다. `tcp_child_ehash_entries`는 child namespace의 hash 크기를 예약하며 0이면 initial namespace hash를 공유하고 양수는 2의 거듭제곱 bucket으로 반올림됩니다.
Protective Load Balancing은 congestion이 지속되는 ECMP path에서 flow hash를 바꿉니다. `tcp_plb_enabled`가 master switch이고 `tcp_plb_idle_rehash_rounds`는 idle 뒤 rehash, `tcp_plb_rehash_rounds`는 congestion round 수, `tcp_plb_suspend_rto_sec`는 RTO 뒤 PLB 중지 시간, `tcp_plb_cong_thresh`는 round가 congestion으로 판정될 CE 비율입니다. PLB는 ECN을 전제로 하며 기본은 꺼져 있습니다.
tcp_notsent_lowat - UNSIGNED INTEGER
A TCP socket can control the amount of unsent bytes in its write queue,
thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll()
reports POLLOUT events if the amount of unsent bytes is below a per
socket value, and if the write queue is not full. sendmsg() will
also not add new buffers if the limit is hit.
This global variable controls the amount of unsent data for
sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change
to the global variable has immediate effect.
Default: UINT_MAX (0xFFFFFFFF)
tcp_workaround_signed_windows - BOOLEAN
If enabled, assume no receipt of a window scaling option means the
remote TCP is broken and treats the window as a signed quantity.
If disabled, assume the remote TCP is not broken even if we do
not receive a window scaling option from them.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_thin_linear_timeouts - BOOLEAN
Enable dynamic triggering of linear timeouts for thin streams.
If enabled, a check is performed upon retransmission by timeout to
determine if the stream is thin (less than 4 packets in flight).
As long as the stream is found to be thin, up to 6 linear
timeouts may be performed before exponential backoff mode is
initiated. This improves retransmission latency for
non-aggressive thin streams, often found to be time-dependent.
For more information on thin streams, see
Documentation/networking/tcp-thin.rst
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_limit_output_bytes - INTEGER
Controls TCP Small Queue limit per tcp socket.
TCP bulk sender tends to increase packets in flight until it
gets losses notifications. With SNDBUF autotuning, this can
result in a large amount of packets queued on the local machine
(e.g.: qdiscs, CPU backlog, or device) hurting latency of other
flows, for typical pfifo_fast qdiscs. tcp_limit_output_bytes
limits the number of bytes on qdisc or device to reduce artificial
RTT/cwnd and reduce bufferbloat.
Default: 4194304 (4 MB)
tcp_challenge_ack_limit - INTEGER
Limits number of Challenge ACK sent per second, as recommended
in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks)
Note that this per netns rate limit can allow some side channel
attacks and probably should not be enabled.
TCP stack implements per TCP socket limits anyway.
Default: INT_MAX (unlimited)
tcp_ehash_entries - INTEGER
Show the number of hash buckets for TCP sockets in the current
networking namespace.
A negative value means the networking namespace does not own its
hash buckets and shares the initial networking namespace's one.
tcp_child_ehash_entries - INTEGER
Control the number of hash buckets for TCP sockets in the child
networking namespace, which must be set before clone() or unshare().
If the value is not 0, the kernel uses a value rounded up to 2^n
as the actual hash bucket size. 0 is a special value, meaning
the child networking namespace will share the initial networking
namespace's hash buckets.
Note that the child will use the global one in case the kernel
fails to allocate enough memory. In addition, the global hash
buckets are spread over available NUMA nodes, but the allocation
of the child hash table depends on the current process's NUMA
policy, which could result in performance differences.
Note also that the default value of tcp_max_tw_buckets and
tcp_max_syn_backlog depend on the hash bucket size.
Possible values: 0, 2^n (n: 0 - 24 (16Mi))
Default: 0
tcp_plb_enabled - BOOLEAN
If enabled and the underlying congestion control (e.g. DCTCP) supports
and enables PLB feature, TCP PLB (Protective Load Balancing) is
enabled. PLB is described in the following paper:
https://doi.org/10.1145/3544216.3544226. Based on PLB parameters,
upon sensing sustained congestion, TCP triggers a change in
flow label field for outgoing IPv6 packets. A change in flow label
field potentially changes the path of outgoing packets for switches
that use ECMP/WCMP for routing.
PLB changes socket txhash which results in a change in IPv6 Flow Label
field, and currently no-op for IPv4 headers. It is possible
to apply PLB for IPv4 with other network header fields (e.g. TCP
or IPv4 options) or using encapsulation where outer header is used
by switches to determine next hop. In either case, further host
and switch side changes will be needed.
If enabled, PLB assumes that congestion signal (e.g. ECN) is made
available and used by congestion control module to estimate a
congestion measure (e.g. ce_ratio). PLB needs a congestion measure to
make repathing decisions.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tcp_plb_idle_rehash_rounds - INTEGER
Number of consecutive congested rounds (RTT) seen after which
a rehash can be performed, given there are no packets in flight.
This is referred to as M in PLB paper:
https://doi.org/10.1145/3544216.3544226.
Possible Values: 0 - 31
Default: 3
tcp_plb_rehash_rounds - INTEGER
Number of consecutive congested rounds (RTT) seen after which
a forced rehash can be performed. Be careful when setting this
parameter, as a small value increases the risk of retransmissions.
This is referred to as N in PLB paper:
https://doi.org/10.1145/3544216.3544226.
Possible Values: 0 - 31
Default: 12
tcp_plb_suspend_rto_sec - INTEGER
Time, in seconds, to suspend PLB in event of an RTO. In order to avoid
having PLB repath onto a connectivity "black hole", after an RTO a TCP
connection suspends PLB repathing for a random duration between 1x and
2x of this parameter. Randomness is added to avoid concurrent rehashing
of multiple TCP connections. This should be set corresponding to the
amount of time it takes to repair a failed link.
Possible Values: 0 - 255
Default: 60
tcp_plb_cong_thresh - INTEGER
Fraction of packets marked with congestion over a round (RTT) to
tag that round as congested. This is referred to as K in the PLB paper:
https://doi.org/10.1145/3544216.3544226.
The 0-1 fraction range is mapped to 0-256 range to avoid floating
point operations. For example, 128 means that if at least 50% of
the packets in a round were marked as congested then the round
will be tagged as congested.
Setting threshold to 0 means that PLB repaths every RTT regardless
of congestion. This is not intended behavior for PLB and should be
used only for experimentation purpose.
Possible Values: 0 - 256
Default: 128
TCP ping-pong 판정과 RTO 범위
1374-1410`tcp_pingpong_thresh`는 request/response 교환 횟수가 기준 이하인 connection을 ping-pong으로 분류해 delayed ACK 동작에 반영합니다. 기본 1입니다. `tcp_rto_min_us`는 TCP RTO 하한을 microsecond로 정하며 0이면 route의 `rto_min`을 사용합니다. `tcp_rto_max_ms`는 RTO 상한으로 기본 120000 ms이고 1000 ms보다 작게 둘 수 없습니다.
tcp_pingpong_thresh - INTEGER
The number of estimated data replies sent for estimated incoming data
requests that must happen before TCP considers that a connection is a
"ping-pong" (request-response) connection for which delayed
acknowledgments can provide benefits.
This threshold is 1 by default, but some applications may need a higher
threshold for optimal performance.
Possible Values: 1 - 255
Default: 1
tcp_rto_min_us - INTEGER
Minimal TCP retransmission timeout (in microseconds). Note that the
rto_min route option has the highest precedence for configuring this
setting, followed by the TCP_BPF_RTO_MIN and TCP_RTO_MIN_US socket
options, followed by this tcp_rto_min_us sysctl.
The recommended practice is to use a value less or equal to 200000
microseconds.
Possible Values: 1 - INT_MAX
Default: 200000
tcp_rto_max_ms - INTEGER
Maximal TCP retransmission timeout (in ms).
Note that TCP_RTO_MAX_MS socket option has higher precedence.
When changing tcp_rto_max_ms, it is important to understand
that tcp_retries2 might need a change.
Possible Values: 1000 - 120,000
Default: 120,000
UDP 변수
1411-1475`udp_l3mdev_accept`를 켜면 global bound UDP socket이 VRF 같은 L3 master domain을 가로질러 packet을 받을 수 있으며 기본 0입니다. `udp_mem`은 전체 UDP socket queue의 min/pressure/max page 수, `udp_rmem_min`은 memory pressure에서도 socket마다 보장할 receive buffer입니다. `udp_wmem_min`은 문서상 효과가 없습니다.
`udp_hash_entries`는 현재 namespace의 UDP hash table bucket 수입니다. `udp_child_hash_entries`는 child namespace가 별도 hash table을 만들 때 요청할 bucket 수이고 0이면 initial namespace table을 공유합니다. 별도 table은 1024보다 작게 만들 수 없으며 2의 거듭제곱으로 반올림됩니다.
UDP variables
=============
udp_l3mdev_accept - BOOLEAN
Enabling this option allows a "global" bound socket to work
across L3 master domains (e.g., VRFs) with packets capable of
being received regardless of the L3 domain in which they
originated. Only valid when the kernel was compiled with
CONFIG_NET_L3_MASTER_DEV.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
udp_mem - vector of 3 INTEGERs: min, pressure, max
Number of pages allowed for queueing by all UDP sockets.
min: Number of pages allowed for queueing by all UDP sockets.
pressure: This value was introduced to follow format of tcp_mem.
max: This value was introduced to follow format of tcp_mem.
Default is calculated at boot time from amount of available memory.
udp_rmem_min - INTEGER
Minimal size of receive buffer used by UDP sockets in moderation.
Each UDP socket is able to use the size for receiving data, even if
total pages of UDP sockets exceed udp_mem pressure. The unit is byte.
Default: 4K
udp_wmem_min - INTEGER
UDP does not have tx memory accounting and this tunable has no effect.
udp_hash_entries - INTEGER
Show the number of hash buckets for UDP sockets in the current
networking namespace.
A negative value means the networking namespace does not own its
hash buckets and shares the initial networking namespace's one.
udp_child_hash_entries - INTEGER
Control the number of hash buckets for UDP sockets in the child
networking namespace, which must be set before clone() or unshare().
If the value is not 0, the kernel uses a value rounded up to 2^n
as the actual hash bucket size. 0 is a special value, meaning
the child networking namespace will share the initial networking
namespace's hash buckets.
Note that the child will use the global one in case the kernel
fails to allocate enough memory. In addition, the global hash
buckets are spread over available NUMA nodes, but the allocation
of the child hash table depends on the current process's NUMA
policy, which could result in performance differences.
Possible values: 0, 2^n (n: 7 (128) - 16 (64K))
Default: 0
RAW socket 변수
1476-1492`raw_l3mdev_accept`는 global bound RAW socket이 모든 L3 master domain의 packet을 받을 수 있게 합니다. kernel이 `CONFIG_NET_L3_MASTER_DEV`로 빌드되었을 때만 유효하며 기본 1입니다.
RAW variables
=============
raw_l3mdev_accept - BOOLEAN
Enabling this option allows a "global" bound socket to work
across L3 master domains (e.g., VRFs) with packets capable of
being received regardless of the L3 domain in which they
originated. Only valid when the kernel was compiled with
CONFIG_NET_L3_MASTER_DEV.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
CIPSOv4 변수
1493-1547`cipso_cache_enable`은 incoming packet의 DOI/level/category mapping cache를 켜고 기본 1입니다. `cipso_cache_bucket_size`는 hash bucket마다 보관할 entry 수로 기본 10이며 0은 사실상 cache를 끕니다.
`cipso_rbm_optfmt`는 restricted bitmap category를 가능한 경우 trailing zero 없이 optimized form으로 만들고 기본 0은 fixed 32-bit multiple 형식입니다. `cipso_rbm_strictvalid`는 incoming CIPSO option을 엄격히 검사합니다. 켜면 malformed option을 더 잘 막지만 처리량이 줄어 기본은 1입니다.
CIPSOv4 Variables
=================
cipso_cache_enable - BOOLEAN
If enabled, enable additions to and lookups from the CIPSO label mapping
cache. If disabled, additions are ignored and lookups always result in a
miss. However, regardless of the setting the cache is still
invalidated when required when means you can safely toggle this on and
off and the cache will always be "safe".
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
cipso_cache_bucket_size - INTEGER
The CIPSO label cache consists of a fixed size hash table with each
hash bucket containing a number of cache entries. This variable limits
the number of entries in each hash bucket; the larger the value is, the
more CIPSO label mappings that can be cached. When the number of
entries in a given hash bucket reaches this limit adding new entries
causes the oldest entry in the bucket to be removed to make room.
Default: 10
cipso_rbm_optfmt - BOOLEAN
Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of
the CIPSO draft specification (see Documentation/netlabel for details).
This means that when set the CIPSO tag will be padded with empty
categories in order to make the packet data 32-bit aligned.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
cipso_rbm_strictvalid - BOOLEAN
If enabled, do a very strict check of the CIPSO option when
ip_options_compile() is called. If disabled, relax the checks done during
ip_options_compile(). Either way is "safe" as errors are caught else
where in the CIPSO processing code but setting this to 0 (False) should
result in less work (i.e. it should be faster) but could cause problems
with other implementations that require strict checking.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
Local port, bind와 early demux
1548-1678`ip_local_port_range`는 TCP/UDP가 자동 선택할 local port의 첫·끝 값이며 기본 32768~60999입니다. 두 끝의 parity를 다르게 두는 편이 좋고 시작은 `ip_unprivileged_port_start` 이상이어야 합니다. `ip_local_reserved_ports`는 자동 할당에서 빼 둘 port/range 목록입니다. explicit bind에는 영향을 주지 않고 쓸 때 기존 목록 전체를 교체합니다. 두 range는 독립적으로 평가되며 예약 block 직후 port의 선택 확률에 영향을 줄 수 있습니다.
namespace별 `ip_unprivileged_port_start`는 capability 없이 bind 가능한 첫 port이고 기본 1024, 0이면 privileged port가 없습니다. `ip_nonlocal_bind`는 local address가 아닌 주소에 bind를 허용합니다. `ip_autobind_reuse`는 `SO_REUSEADDR` socket의 bind+connect 자동 port 재사용을 허용하지만 `IP_BIND_ADDRESS_NO_PORT`가 권장되며 전문가용입니다. `ip_dynaddr`는 dynamic address rewrite를 켜고 1보다 크면 rewrite를 log합니다.
`ip_early_demux`는 established TCP와 connected UDP의 local input demux를 앞당겨 socket workload를 줄이지만 순수 router에서는 추가 비용이 될 수 있습니다. `ping_group_range`는 ICMP datagram ping socket을 만들 수 있는 group ID 범위입니다. 기본 `1 0`은 root를 포함해 아무도 허용하지 않습니다. `tcp_early_demux`와 `udp_early_demux`로 protocol별 최적화를 끌 수 있습니다.
IP Variables
============
ip_local_port_range - 2 INTEGERS
Defines the local port range that is used by TCP and UDP to
choose the local port. The first number is the first, the
second the last local port number.
If possible, it is better these numbers have different parity
(one even and one odd value).
Must be greater than or equal to ip_unprivileged_port_start.
The default values are 32768 and 60999 respectively.
ip_local_reserved_ports - list of comma separated ranges
Specify the ports which are reserved for known third-party
applications. These ports will not be used by automatic port
assignments (e.g. when calling connect() or bind() with port
number 0). Explicit port allocation behavior is unchanged.
The format used for both input and output is a comma separated
list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and
10). Writing to the file will clear all previously reserved
ports and update the current list with the one given in the
input.
Note that ip_local_port_range and ip_local_reserved_ports
settings are independent and both are considered by the kernel
when determining which ports are available for automatic port
assignments.
You can reserve ports which are not in the current
ip_local_port_range, e.g.::
$ cat /proc/sys/net/ipv4/ip_local_port_range
32000 60999
$ cat /proc/sys/net/ipv4/ip_local_reserved_ports
8080,9148
although this is redundant. However such a setting is useful
if later the port range is changed to a value that will
include the reserved ports. Also keep in mind, that overlapping
of these ranges may affect probability of selecting ephemeral
ports which are right after block of reserved ports.
Default: Empty
ip_unprivileged_port_start - INTEGER
This is a per-namespace sysctl. It defines the first
unprivileged port in the network namespace. Privileged ports
require root or CAP_NET_BIND_SERVICE in order to bind to them.
To disable all privileged ports, set this to 0. They must not
overlap with the ip_local_port_range.
Default: 1024
ip_nonlocal_bind - BOOLEAN
If enabled, allows processes to bind() to non-local IP addresses,
which can be quite useful - but may break some applications.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
ip_autobind_reuse - BOOLEAN
By default, bind() does not select the ports automatically even if
the new socket and all sockets bound to the port have SO_REUSEADDR.
ip_autobind_reuse allows bind() to reuse the port and this is useful
when you use bind()+connect(), but may break some applications.
The preferred solution is to use IP_BIND_ADDRESS_NO_PORT and this
option should only be set by experts.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
ip_dynaddr - INTEGER
If set non-zero, enables support for dynamic addresses.
If set to a non-zero value larger than 1, a kernel log
message will be printed when dynamic address rewriting
occurs.
Default: 0
ip_early_demux - BOOLEAN
Optimize input packet processing down to one demux for
certain kinds of local sockets. Currently we only do this
for established TCP and connected UDP sockets.
It may add an additional cost for pure routing workloads that
reduces overall throughput, in such case you should disable it.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
ping_group_range - 2 INTEGERS
Restrict ICMP_PROTO datagram sockets to users in the group range.
The default is "1 0", meaning, that nobody (not even root) may
create ping sockets. Setting it to "100 100" would grant permissions
to the single group. "0 4294967294" would enable it for the world, "100
4294967294" would enable it for the users, but not daemons.
tcp_early_demux - BOOLEAN
Enable early demux for established TCP sockets.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
udp_early_demux - BOOLEAN
Enable early demux for connected UDP sockets. Disable this if
your system could experience more unconnected load.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
ICMPv4 rate와 응답 정책
1679-1798`icmp_echo_ignore_all`은 모든 IPv4 echo request를 무시합니다. `icmp_echo_enable_probe`는 RFC 8335 extended echo request에 응답할지 정하고 `icmp_echo_ignore_broadcasts`는 broadcast/multicast echo를 무시하며 기본 1입니다.
`icmp_ratelimit`은 `icmp_ratemask`에 해당하는 ICMP type 응답 사이의 최소 ms이고 0이면 제한이 없습니다. `icmp_msgs_per_sec`와 `icmp_msgs_burst`는 per-CPU global rate와 burst를 대략 제한하며 randomization 때문에 정확한 한도로 보아서는 안 됩니다. 기본 mask 6168은 Destination Unreachable, Source Quench, Time Exceeded, Parameter Problem에 제한을 적용합니다.
`icmp_ignore_bogus_error_responses`는 RFC 1122를 어기고 broadcast frame에 보낸 bogus error를 무시합니다. `icmp_errors_use_inbound_ifaddr`는 ICMP error source address를 packet이 들어온 interface의 primary address로 고릅니다. 기본 0은 packet을 보내는 interface의 primary address를 쓰며 router 식별과 traceroute 결과가 달라질 수 있습니다.
icmp_echo_ignore_all - BOOLEAN
If enabled, then the kernel will ignore all ICMP ECHO
requests sent to it.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
icmp_echo_enable_probe - BOOLEAN
If enabled, then the kernel will respond to RFC 8335 PROBE
requests sent to it.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
icmp_echo_ignore_broadcasts - BOOLEAN
If enabled, then the kernel will ignore all ICMP ECHO and
TIMESTAMP requests sent to it via broadcast/multicast.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
icmp_ratelimit - INTEGER
Limit the maximal rates for sending ICMP packets whose type matches
icmp_ratemask (see below) to specific targets.
0 to disable any limiting,
otherwise the minimal space between responses in milliseconds.
Note that another sysctl, icmp_msgs_per_sec limits the number
of ICMP packets sent on all targets.
Default: 1000
icmp_msgs_per_sec - INTEGER
Limit maximal number of ICMP packets sent per second from this host.
Only messages whose type matches icmp_ratemask (see below) are
controlled by this limit. For security reasons, the precise count
of messages per second is randomized.
Default: 1000
icmp_msgs_burst - INTEGER
icmp_msgs_per_sec controls number of ICMP packets sent per second,
while icmp_msgs_burst controls the burst size of these packets.
For security reasons, the precise burst size is randomized.
Default: 50
icmp_ratemask - INTEGER
Mask made of ICMP types for which rates are being limited.
Significant bits: IHGFEDCBA9876543210
Default mask: 0000001100000011000 (6168)
Bit definitions (see include/linux/icmp.h):
= =========================
0 Echo Reply
3 Destination Unreachable [1]_
4 Source Quench [1]_
5 Redirect
8 Echo Request
B Time Exceeded [1]_
C Parameter Problem [1]_
D Timestamp Request
E Timestamp Reply
F Info Request
G Info Reply
H Address Mask Request
I Address Mask Reply
= =========================
.. [1] These are rate limited by default (see default mask above)
icmp_ignore_bogus_error_responses - BOOLEAN
Some routers violate RFC1122 by sending bogus responses to broadcast
frames. Such violations are normally logged via a kernel warning.
If enabled, the kernel will not give such warnings, which
will avoid log file clutter.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
icmp_errors_use_inbound_ifaddr - BOOLEAN
If disabled, icmp error messages are sent with the primary address of
the exiting interface.
If enabled, the message will be sent with the primary address of
the interface that received the packet that caused the icmp error.
This is the behaviour many network administrators will expect from
a router. And it can make debugging complicated network layouts
much easier.
Note that if no primary address exists for the interface selected,
then the primary address of the first non-loopback interface that
has one will be used regardless of this setting.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
IGMP membership과 version
1799-1850`igmp_max_memberships`는 socket이 join할 수 있는 multicast group 수로 기본 20입니다. 보고 packet 하나가 담을 수 있는 최대치는 65535-byte IP packet 한계 때문에 대략 5459이고 너무 크게 하면 report가 여러 packet으로 나뉘어 switch가 처리하지 못할 수 있습니다. `igmp_max_msf`는 group별 source filter 수로 기본 10입니다.
`igmp_qrv`는 IGMP Query Robustness Variable이며 기본 2, 7보다 크면 IGMPv3 report의 3-bit QRV에 0으로 표시됩니다. `force_igmp_version`은 0=자동, 1=IGMPv1, 2=IGMPv2, 3=IGMPv3이고 특수 0/1 값의 legacy 차이를 문서에 따릅니다.
igmp_max_memberships - INTEGER
Change the maximum number of multicast groups we can subscribe to.
Default: 20
Theoretical maximum value is bounded by having to send a membership
report in a single datagram (i.e. the report can't span multiple
datagrams, or risk confusing the switch and leaving groups you don't
intend to).
The number of supported groups 'M' is bounded by the number of group
report entries you can fit into a single datagram of 65535 bytes.
M = 65536-sizeof (ip header)/(sizeof(Group record))
Group records are variable length, with a minimum of 12 bytes.
So net.ipv4.igmp_max_memberships should not be set higher than:
(65536-24) / 12 = 5459
The value 5459 assumes no IP header options, so in practice
this number may be lower.
igmp_max_msf - INTEGER
Maximum number of addresses allowed in the source filter list for a
multicast group.
Default: 10
igmp_qrv - INTEGER
Controls the IGMP query robustness variable (see RFC2236 8.1).
Default: 2 (as specified by RFC2236 8.1)
Minimum: 1 (as specified by RFC6636 4.5)
force_igmp_version - INTEGER
- 0 - (default) No enforcement of a IGMP version, IGMPv1/v2 fallback
allowed. Will back to IGMPv3 mode again if all IGMPv1/v2 Querier
Present timer expires.
- 1 - Enforce to use IGMP version 1. Will also reply IGMPv1 report if
receive IGMPv2/v3 query.
- 2 - Enforce to use IGMP version 2. Will fallback to IGMPv1 if receive
IGMPv1 query message. Will reply report if receive IGMPv3 query.
- 3 - Enforce to use IGMP version 3. The same react with default 0.
.. note::
this is not the same with force_mld_version because IGMPv3 RFC3376
Security Considerations does not have clear description that we could
ignore other version messages completely as MLDv2 RFC3810. So make
this value as default 0 is recommended.
IPv4 interface별 forwarding과 redirect
1851-2044`conf/interface/*`는 한 interface, `conf/all/*`은 모든 interface에 적용합니다. `log_martians`는 불가능한 source address를 log합니다. `accept_redirects`는 host에서는 all 또는 interface 중 하나가 true면, forwarding router에서는 둘 다 true일 때만 켜집니다. 기본은 host=true, router=false입니다. `forwarding`은 해당 ingress interface packet의 전달, `mc_forwarding`은 `CONFIG_MROUTE`와 daemon을 전제로 multicast routing을 켭니다.
`medium_id`는 interface가 공유하는 medium을 구분해 서로 다른 medium 사이 proxy ARP 판단에 쓰고 0=유일 interface, -1=미상입니다. `proxy_arp`는 proxy ARP, `proxy_arp_pvlan`은 같은 interface로 다시 응답해 RFC 3069 port isolation host가 upstream router를 통해 통신하게 합니다. `proxy_delay`는 proxy ARP/NDP 응답에 [0, 값) jiffy random delay를 둡니다.
`shared_media`는 RFC 1620 redirect를 보내거나 받아 `secure_redirects`보다 우선합니다. `secure_redirects`는 현재 gateway list의 gateway로 가는 redirect만 받고, `send_redirects`는 router가 redirect를 보냅니다. `bootp_relay`는 아직 구현되지 않았습니다. `accept_source_route`는 SRR option을 받고, `accept_local`은 local source packet을 wire에서 받아들이며, `route_localnet`은 127/8을 routing에서 martian으로 보지 않습니다.
`rp_filter`는 0=검증 없음, 1=best reverse path인 strict RFC 3704, 2=어느 interface로든 reachable하면 되는 loose mode입니다. spoofing 방지에는 strict가 좋지만 asymmetric routing에는 loose가 적합합니다. all/interface 중 큰 값을 씁니다. `src_valid_mark`를 켜면 reverse lookup과 ICMP source/일부 IP option address 선택에 packet fwmark를 포함해 양방향 policy routing과 맞춥니다.
``conf/interface/*``
changes special settings per interface (where
interface" is the name of your network interface)
``conf/all/*``
is special, changes the settings for all interfaces
log_martians - BOOLEAN
Log packets with impossible addresses to kernel log.
log_martians for the interface will be enabled if at least one of
conf/{all,interface}/log_martians is set to TRUE,
it will be disabled otherwise
accept_redirects - BOOLEAN
Accept ICMP redirect messages.
accept_redirects for the interface will be enabled if:
- both conf/{all,interface}/accept_redirects are TRUE in the case
forwarding for the interface is enabled
or
- at least one of conf/{all,interface}/accept_redirects is TRUE in the
case forwarding for the interface is disabled
accept_redirects for the interface will be disabled otherwise
default:
- TRUE (host)
- FALSE (router)
forwarding - BOOLEAN
Enable IP forwarding on this interface. This controls whether packets
received _on_ this interface can be forwarded.
mc_forwarding - BOOLEAN
Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE
and a multicast routing daemon is required.
conf/all/mc_forwarding must also be set to TRUE to enable multicast
routing for the interface
medium_id - INTEGER
Integer value used to differentiate the devices by the medium they
are attached to. Two devices can have different id values when
the broadcast packets are received only on one of them.
The default value 0 means that the device is the only interface
to its medium, value of -1 means that medium is not known.
Currently, it is used to change the proxy_arp behavior:
the proxy_arp feature is enabled for packets forwarded between
two devices attached to different media.
proxy_arp - BOOLEAN
Do proxy arp.
proxy_arp for the interface will be enabled if at least one of
conf/{all,interface}/proxy_arp is set to TRUE,
it will be disabled otherwise
proxy_arp_pvlan - BOOLEAN
Private VLAN proxy arp.
Basically allow proxy arp replies back to the same interface
(from which the ARP request/solicitation was received).
This is done to support (ethernet) switch features, like RFC
3069, where the individual ports are NOT allowed to
communicate with each other, but they are allowed to talk to
the upstream router. As described in RFC 3069, it is possible
to allow these hosts to communicate through the upstream
router by proxy_arp'ing. Don't need to be used together with
proxy_arp.
This technology is known by different names:
- In RFC 3069 it is called VLAN Aggregation.
- Cisco and Allied Telesyn call it Private VLAN.
- Hewlett-Packard call it Source-Port filtering or port-isolation.
- Ericsson call it MAC-Forced Forwarding (RFC Draft).
proxy_delay - INTEGER
Delay proxy response.
Delay response to a neighbor solicitation when proxy_arp
or proxy_ndp is enabled. A random value between [0, proxy_delay)
will be chosen, setting to zero means reply with no delay.
Value in jiffies. Defaults to 80.
shared_media - BOOLEAN
Send(router) or accept(host) RFC1620 shared media redirects.
Overrides secure_redirects.
shared_media for the interface will be enabled if at least one of
conf/{all,interface}/shared_media is set to TRUE,
it will be disabled otherwise
default TRUE
secure_redirects - BOOLEAN
Accept ICMP redirect messages only to gateways listed in the
interface's current gateway list. Even if disabled, RFC1122 redirect
rules still apply.
Overridden by shared_media.
secure_redirects for the interface will be enabled if at least one of
conf/{all,interface}/secure_redirects is set to TRUE,
it will be disabled otherwise
default TRUE
send_redirects - BOOLEAN
Send redirects, if router.
send_redirects for the interface will be enabled if at least one of
conf/{all,interface}/send_redirects is set to TRUE,
it will be disabled otherwise
Default: TRUE
bootp_relay - BOOLEAN
Accept packets with source address 0.b.c.d destined
not to this host as local ones. It is supposed, that
BOOTP relay daemon will catch and forward such packets.
conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay
for the interface
default FALSE
Not Implemented Yet.
accept_source_route - BOOLEAN
Accept packets with SRR option.
conf/all/accept_source_route must also be set to TRUE to accept packets
with SRR option on the interface
default
- TRUE (router)
- FALSE (host)
accept_local - BOOLEAN
Accept packets with local source addresses. In combination with
suitable routing, this can be used to direct packets between two
local interfaces over the wire and have them accepted properly.
default FALSE
route_localnet - BOOLEAN
Do not consider loopback addresses as martian source or destination
while routing. This enables the use of 127/8 for local routing purposes.
default FALSE
rp_filter - INTEGER
- 0 - No source validation.
- 1 - Strict mode as defined in RFC3704 Strict Reverse Path
Each incoming packet is tested against the FIB and if the interface
is not the best reverse path the packet check will fail.
By default failed packets are discarded.
- 2 - Loose mode as defined in RFC3704 Loose Reverse Path
Each incoming packet's source address is also tested against the FIB
and if the source address is not reachable via any interface
the packet check will fail.
Current recommended practice in RFC3704 is to enable strict mode
to prevent IP spoofing from DDos attacks. If using asymmetric routing
or other complicated routing, then loose mode is recommended.
The max value from conf/{all,interface}/rp_filter is used
when doing source validation on the {interface}.
Default value is 0. Note that some distributions enable it
in startup scripts.
src_valid_mark - BOOLEAN
- 0 - The fwmark of the packet is not included in reverse path
route lookup. This allows for asymmetric routing configurations
utilizing the fwmark in only one direction, e.g., transparent
proxying.
- 1 - The fwmark of the packet is included in reverse path route
lookup. This permits rp_filter to function when the fwmark is
used for routing traffic in both directions.
This setting also affects the utilization of fmwark when
performing source address selection for ICMP replies, or
determining addresses stored for the IPOPT_TS_TSANDADDR and
IPOPT_RR IP options.
The max value from conf/{all,interface}/src_valid_mark is used.
Default value is 0.
IPv4 ARP 응답과 neighbor probe
2045-2191`arp_filter`는 같은 subnet의 여러 interface 중 route lookup상 outgoing interface인 card만 ARP에 응답하게 하며 source-based routing이 필요합니다. 기본 0에서는 Linux host 전체가 IP address를 소유한다는 모델에 따라 다른 interface 주소에도 응답할 수 있습니다. `arp_announce`는 ARP request sender address 선택을 0=아무 local address, 1=target subnet 밖 주소 회피, 2=outgoing interface에서 target에 가장 적합한 주소로 제한합니다.
`arp_ignore`는 reply 범위를 0=어느 local 주소든, 1=incoming interface 주소만, 2=같은 subnet까지, 3=host-scope 제외, 8=모두 무시로 강화합니다. 4~7은 예약입니다. `arp_notify`는 device up/MAC 변경 때 gratuitous ARP를 보내고 `arp_accept`는 cache에 없는 gratuitous ARP를 0=새 entry 없음, 1=허용, 2=같은 subnet만 허용으로 정합니다. 기존 entry는 설정과 무관하게 갱신됩니다.
`arp_evict_nocarrier`는 carrier가 사라질 때 ARP cache를 비우며 기본 1입니다. 같은 network의 AP 사이를 roaming하는 wireless 환경에서는 0이 유용할 수 있습니다. `mcast_solicit`, `ucast_solicit`, `app_solicit`, `mcast_resolicit`은 INCOMPLETE/PROBE 상태에서 multicast, unicast, userspace ARP daemon, 재multicast probe의 최대 횟수입니다. `disable_policy`와 `disable_xfrm`은 각각 interface의 IPsec SPD와 encryption을 끕니다.
arp_filter - BOOLEAN
- 1 - Allows you to have multiple network interfaces on the same
subnet, and have the ARPs for each interface be answered
based on whether or not the kernel would route a packet from
the ARP'd IP out that interface (therefore you must use source
based routing for this to work). In other words it allows control
of which cards (usually 1) will respond to an arp request.
- 0 - (default) The kernel can respond to arp requests with addresses
from other interfaces. This may seem wrong but it usually makes
sense, because it increases the chance of successful communication.
IP addresses are owned by the complete host on Linux, not by
particular interfaces. Only for more complex setups like load-
balancing, does this behaviour cause problems.
arp_filter for the interface will be enabled if at least one of
conf/{all,interface}/arp_filter is set to TRUE,
it will be disabled otherwise
arp_announce - INTEGER
Define different restriction levels for announcing the local
source IP address from IP packets in ARP requests sent on
interface:
- 0 - (default) Use any local address, configured on any interface
- 1 - Try to avoid local addresses that are not in the target's
subnet for this interface. This mode is useful when target
hosts reachable via this interface require the source IP
address in ARP requests to be part of their logical network
configured on the receiving interface. When we generate the
request we will check all our subnets that include the
target IP and will preserve the source address if it is from
such subnet. If there is no such subnet we select source
address according to the rules for level 2.
- 2 - Always use the best local address for this target.
In this mode we ignore the source address in the IP packet
and try to select local address that we prefer for talks with
the target host. Such local address is selected by looking
for primary IP addresses on all our subnets on the outgoing
interface that include the target IP address. If no suitable
local address is found we select the first local address
we have on the outgoing interface or on all other interfaces,
with the hope we will receive reply for our request and
even sometimes no matter the source IP address we announce.
The max value from conf/{all,interface}/arp_announce is used.
Increasing the restriction level gives more chance for
receiving answer from the resolved target while decreasing
the level announces more valid sender's information.
arp_ignore - INTEGER
Define different modes for sending replies in response to
received ARP requests that resolve local target IP addresses:
- 0 - (default): reply for any local target IP address, configured
on any interface
- 1 - reply only if the target IP address is local address
configured on the incoming interface
- 2 - reply only if the target IP address is local address
configured on the incoming interface and both with the
sender's IP address are part from same subnet on this interface
- 3 - do not reply for local addresses configured with scope host,
only resolutions for global and link addresses are replied
- 4-7 - reserved
- 8 - do not reply for all local addresses
The max value from conf/{all,interface}/arp_ignore is used
when ARP request is received on the {interface}
arp_notify - BOOLEAN
Define mode for notification of address and device changes.
== ==========================================================
0 (default): do nothing
1 Generate gratuitous arp requests when device is brought up
or hardware address changes.
== ==========================================================
arp_accept - INTEGER
Define behavior for accepting gratuitous ARP (garp) frames from devices
that are not already present in the ARP table:
- 0 - don't create new entries in the ARP table
- 1 - create new entries in the ARP table
- 2 - create new entries only if the source IP address is in the same
subnet as an address configured on the interface that received the
garp message.
Both replies and requests type gratuitous arp will trigger the
ARP table to be updated, if this setting is on.
If the ARP table already contains the IP address of the
gratuitous arp frame, the arp table will be updated regardless
if this setting is on or off.
arp_evict_nocarrier - BOOLEAN
Clears the ARP cache on NOCARRIER events. This option is important for
wireless devices where the ARP cache should not be cleared when roaming
between access points on the same network. In most cases this should
remain as the default (1).
Possible values:
- 0 (disabled) - Do not clear ARP cache on NOCARRIER events
- 1 (enabled) - Clear the ARP cache on NOCARRIER events
Default: 1 (enabled)
mcast_solicit - INTEGER
The maximum number of multicast probes in INCOMPLETE state,
when the associated hardware address is unknown. Defaults
to 3.
ucast_solicit - INTEGER
The maximum number of unicast probes in PROBE state, when
the hardware address is being reconfirmed. Defaults to 3.
app_solicit - INTEGER
The maximum number of probes to send to the user space ARP daemon
via netlink before dropping back to multicast probes (see
mcast_resolicit). Defaults to 0.
mcast_resolicit - INTEGER
The maximum number of multicast probes after unicast and
app probes in PROBE state. Defaults to 0.
disable_policy - BOOLEAN
Disable IPSEC policy (SPD) for this interface
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
disable_xfrm - BOOLEAN
Disable IPSEC encryption on this interface, whatever the policy
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
IPv4 interface 기타 정책
2192-2282`igmpv2_unsolicited_report_interval`은 IGMPv1/v2 report 재전송 간격 10초, `igmpv3_unsolicited_report_interval`은 IGMPv3 간격 1초가 기본입니다. `ignore_routes_with_linkdown`은 down link route를 FIB lookup에서 제외합니다. `promote_secondaries`는 primary address 삭제 때 관련 secondary를 함께 없애는 대신 하나를 primary로 승격합니다.
`drop_unicast_in_l2_multicast`는 L2 multicast/broadcast frame에 실린 unicast IPv4를 폐기하고 `drop_gratuitous_arp`는 모든 gratuitous ARP를 버립니다. 둘 다 compatibility 때문에 기본 0입니다. `tag`는 자유롭게 쓸 수 있는 정수 tag입니다. `xfrm4_gc_thresh`는 4.14부터 폐기되었습니다. `igmp_link_local_mcast_reports`는 224.0.0.X link-local group에도 IGMP report를 보내며 기본 true입니다. 마지막에는 원 작성자와 갱신자 credit이 기록되어 있습니다.
igmpv2_unsolicited_report_interval - INTEGER
The interval in milliseconds in which the next unsolicited
IGMPv1 or IGMPv2 report retransmit will take place.
Default: 10000 (10 seconds)
igmpv3_unsolicited_report_interval - INTEGER
The interval in milliseconds in which the next unsolicited
IGMPv3 report retransmit will take place.
Default: 1000 (1 seconds)
ignore_routes_with_linkdown - BOOLEAN
Ignore routes whose link is down when performing a FIB lookup.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
promote_secondaries - BOOLEAN
When a primary IP address is removed from this interface
promote a corresponding secondary IP address instead of
removing all the corresponding secondary IP addresses.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
drop_unicast_in_l2_multicast - BOOLEAN
Drop any unicast IP packets that are received in link-layer
multicast (or broadcast) frames.
This behavior (for multicast) is actually a SHOULD in RFC
1122, but is disabled by default for compatibility reasons.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
drop_gratuitous_arp - BOOLEAN
Drop all gratuitous ARP frames, for example if there's a known
good ARP proxy on the network and such frames need not be used
(or in the case of 802.11, must not be used to prevent attacks.)
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
tag - INTEGER
Allows you to write a number, which can be used as required.
Default value is 0.
xfrm4_gc_thresh - INTEGER
(Obsolete since linux-4.14)
The threshold at which we will start garbage collecting for IPv4
destination cache entries. At twice this value the system will
refuse new allocations.
igmp_link_local_mcast_reports - BOOLEAN
Enable IGMP reports for link local multicast groups in the
224.0.0.X range.
Default TRUE
Alexey Kuznetsov.
[email protected]
Updated by:
- Andi Kleen
[email protected]
- Nicolas Delon
[email protected]
IPv6 socket, flow label과 ECMP
2283-2406IPv6에는 별도 global `tcp_*`가 없고 `ipv4/` 아래 TCP 설정이 IPv6에도 적용됩니다. `bindv6only`는 `IPV6_V6ONLY` 기본값이며 0이면 IPv4-mapped address를 허용합니다. `flowlabel_consistency`는 flow label의 일관성과 유일성을 보호하고, `auto_flowlabels`는 packet flow hash로 label을 0=완전 끔, 1=기본 켬/socket에서 끌 수 있음, 2=기본 끔/socket에서 켤 수 있음, 3=강제 켬으로 정합니다.
`flowlabel_state_ranges`는 0~0x7ffff를 flow manager, 0x80000~0xfffff를 RFC 6437 stateless label에 예약합니다. `flowlabel_reflect`는 bit 1=established flow, 2=closed port TCP RST, 4=ICMPv6 echo reply에 incoming label을 반영해 anycast ECMP에서 PMTUD가 같은 path를 타게 합니다.
IPv6 `fib_multipath_hash_policy`는 0=L3 주소+flow label, 1=L4 5-tuple, 2=inner L3 우선, 3=custom입니다. custom `fib_multipath_hash_fields`는 outer/inner source·destination IP, protocol, flow label, port bit를 조합하고 기본 0x7은 source IP, destination IP, protocol입니다.
/proc/sys/net/ipv6/* Variables
==============================
IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also
apply to IPv6 [XXX?].
bindv6only - BOOLEAN
Default value for IPV6_V6ONLY socket option,
which restricts use of the IPv6 socket to IPv6 communication
only.
Possible values:
- 0 (disabled) - enable IPv4-mapped address feature
- 1 (enabled) - disable IPv4-mapped address feature
Default: 0 (disabled)
flowlabel_consistency - BOOLEAN
Protect the consistency (and unicity) of flow label.
You have to disable it to use IPV6_FL_F_REFLECT flag on the
flow label manager.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
auto_flowlabels - INTEGER
Automatically generate flow labels based on a flow hash of the
packet. This allows intermediate devices, such as routers, to
identify packet flows for mechanisms like Equal Cost Multipath
Routing (see RFC 6438).
= ===========================================================
0 automatic flow labels are completely disabled
1 automatic flow labels are enabled by default, they can be
disabled on a per socket basis using the IPV6_AUTOFLOWLABEL
socket option
2 automatic flow labels are allowed, they may be enabled on a
per socket basis using the IPV6_AUTOFLOWLABEL socket option
3 automatic flow labels are enabled and enforced, they cannot
be disabled by the socket option
= ===========================================================
Default: 1
flowlabel_state_ranges - BOOLEAN
Split the flow label number space into two ranges. 0-0x7FFFF is
reserved for the IPv6 flow manager facility, 0x80000-0xFFFFF
is reserved for stateless flow labels as described in RFC6437.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
flowlabel_reflect - INTEGER
Control flow label reflection. Needed for Path MTU
Discovery to work with Equal Cost Multipath Routing in anycast
environments. See RFC 7690 and:
https://tools.ietf.org/html/draft-wang-6man-flow-label-reflection-01
This is a bitmask.
- 1: enabled for established flows
Note that this prevents automatic flowlabel changes, as done
in "tcp: change IPv6 flow-label upon receiving spurious retransmission"
and "tcp: Change txhash on every SYN and RTO retransmit"
- 2: enabled for TCP RESET packets (no active listener)
If set, a RST packet sent in response to a SYN packet on a closed
port will reflect the incoming flow label.
- 4: enabled for ICMPv6 echo reply messages.
Default: 0
fib_multipath_hash_policy - INTEGER
Controls which hash policy to use for multipath routes.
Default: 0 (Layer 3)
Possible values:
- 0 - Layer 3 (source and destination addresses plus flow label)
- 1 - Layer 4 (standard 5-tuple)
- 2 - Layer 3 or inner Layer 3 if present
- 3 - Custom multipath hash. Fields used for multipath hash calculation
are determined by fib_multipath_hash_fields sysctl
fib_multipath_hash_fields - UNSIGNED INTEGER
When fib_multipath_hash_policy is set to 3 (custom multipath hash), the
fields used for multipath hash calculation are determined by this
sysctl.
This value is a bitmask which enables various fields for multipath hash
calculation.
Possible fields are:
====== ============================
0x0001 Source IP address
0x0002 Destination IP address
0x0004 IP protocol
0x0008 Flow Label
0x0010 Source port
0x0020 Destination port
0x0040 Inner source IP address
0x0080 Inner destination IP address
0x0100 Inner IP protocol
0x0200 Inner Flow Label
0x0400 Inner source port
0x0800 Inner destination port
====== ============================
Default: 0x0007 (source IP, destination IP and IP protocol)
IPv6 option 제한, route 알림과 fragment
2407-2586`anycast_src_echo_reply`는 ICMPv6 echo reply source로 anycast address를 쓸지 정합니다. `idgen_delay`와 `idgen_retries`는 RFC 7217 stable privacy address가 DAD 충돌했을 때 재시도 간격과 횟수로 기본 1초, 3회입니다. `mld_qrv`의 기본은 2입니다.
`max_dst_opts_number`와 `max_hbh_opts_number`는 Destination/Hop-by-Hop extension header의 non-padding TLV 수를 제한합니다. 음수면 unknown option을 금지하고 절댓값만큼 known TLV를 허용합니다. `max_dst_opts_length`와 `max_hbh_length`는 header 길이 상한이며 기본 `INT_MAX`입니다.
`skip_notify_on_dev_down`은 device down/delete로 route가 사라질 때 IPv6 `RTM_DELROUTE`를 생략해 IPv4와 같은 동작으로 만듭니다. `nexthop_compat_mode`는 새 nexthop API를 옛 route dump 형식으로 펼치고 변경 route마다 알림을 보내는 호환 모드입니다. 끄면 성능은 좋아지지만 old client 호환이 사라집니다. resilient group과 8-bit를 넘는 weight는 호환 dump에서 부정확할 수 있습니다. `fib_notify_on_flag_change`는 IPv4와 같은 0/1/2 offload flag 알림 정책입니다.
`ioam6_id`는 24-bit node ID, `ioam6_id_wide`는 56-bit wide ID입니다. IPv6 fragmentation의 `ip6frag_high_thresh`, 폐기 기준인 `ip6frag_low_thresh`, 보존 초인 `ip6frag_time`을 제공합니다. `conf/default/*`은 새 interface 기본값, `conf/all/*`은 모든 interface 설정입니다. `conf/all/disable_ipv6` 읽기는 전체 활성 여부를 확정하지 못하고, global `forwarding`은 모든 interface의 host/router forwarding을 함께 바꿉니다.
anycast_src_echo_reply - BOOLEAN
Controls the use of anycast addresses as source addresses for ICMPv6
echo reply
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
idgen_delay - INTEGER
Controls the delay in seconds after which time to retry
privacy stable address generation if a DAD conflict is
detected.
Default: 1 (as specified in RFC7217)
idgen_retries - INTEGER
Controls the number of retries to generate a stable privacy
address if a DAD conflict is detected.
Default: 3 (as specified in RFC7217)
mld_qrv - INTEGER
Controls the MLD query robustness variable (see RFC3810 9.1).
Default: 2 (as specified by RFC3810 9.1)
Minimum: 1 (as specified by RFC6636 4.5)
max_dst_opts_number - INTEGER
Maximum number of non-padding TLVs allowed in a Destination
options extension header. If this value is less than zero
then unknown options are disallowed and the number of known
TLVs allowed is the absolute value of this number.
Default: 8
max_hbh_opts_number - INTEGER
Maximum number of non-padding TLVs allowed in a Hop-by-Hop
options extension header. If this value is less than zero
then unknown options are disallowed and the number of known
TLVs allowed is the absolute value of this number.
Default: 8
max_dst_opts_length - INTEGER
Maximum length allowed for a Destination options extension
header.
Default: INT_MAX (unlimited)
max_hbh_length - INTEGER
Maximum length allowed for a Hop-by-Hop options extension
header.
Default: INT_MAX (unlimited)
skip_notify_on_dev_down - BOOLEAN
Controls whether an RTM_DELROUTE message is generated for routes
removed when a device is taken down or deleted. IPv4 does not
generate this message; IPv6 does by default. Setting this sysctl
to true skips the message, making IPv4 and IPv6 on par in relying
on userspace caches to track link events and evict routes.
Possible values:
- 0 (disabled) - generate the message
- 1 (enabled) - skip generating the message
Default: 0 (disabled)
nexthop_compat_mode - BOOLEAN
New nexthop API provides a means for managing nexthops independent of
prefixes. Backwards compatibility with old route format is enabled by
default which means route dumps and notifications contain the new
nexthop attribute but also the full, expanded nexthop definition.
Further, updates or deletes of a nexthop configuration generate route
notifications for each fib entry using the nexthop. Once a system
understands the new API, this sysctl can be disabled to achieve full
performance benefits of the new API by disabling the nexthop expansion
and extraneous notifications.
Note that as a backward-compatible mode, dumping of modern features
might be incomplete or wrong. For example, resilient groups will not be
shown as such, but rather as just a list of next hops. Also weights that
do not fit into 8 bits will show incorrectly.
Default: true (backward compat mode)
fib_notify_on_flag_change - INTEGER
Whether to emit RTM_NEWROUTE notifications whenever RTM_F_OFFLOAD/
RTM_F_TRAP/RTM_F_OFFLOAD_FAILED flags are changed.
After installing a route to the kernel, user space receives an
acknowledgment, which means the route was installed in the kernel,
but not necessarily in hardware.
It is also possible for a route already installed in hardware to change
its action and therefore its flags. For example, a host route that is
trapping packets can be "promoted" to perform decapsulation following
the installation of an IPinIP/VXLAN tunnel.
The notifications will indicate to user-space the state of the route.
Default: 0 (Do not emit notifications.)
Possible values:
- 0 - Do not emit notifications.
- 1 - Emit notifications.
- 2 - Emit notifications only for RTM_F_OFFLOAD_FAILED flag change.
ioam6_id - INTEGER
Define the IOAM id of this node. Uses only 24 bits out of 32 in total.
Possible value range:
- Min: 0
- Max: 0xFFFFFF
Default: 0xFFFFFF
ioam6_id_wide - LONG INTEGER
Define the wide IOAM id of this node. Uses only 56 bits out of 64 in
total. Can be different from ioam6_id.
Possible value range:
- Min: 0
- Max: 0xFFFFFFFFFFFFFF
Default: 0xFFFFFFFFFFFFFF
IPv6 Fragmentation:
ip6frag_high_thresh - INTEGER
Maximum memory used to reassemble IPv6 fragments. When
ip6frag_high_thresh bytes of memory is allocated for this purpose,
the fragment handler will toss packets until ip6frag_low_thresh
is reached.
ip6frag_low_thresh - INTEGER
See ip6frag_high_thresh
ip6frag_time - INTEGER
Time in seconds to keep an IPv6 fragment in memory.
``conf/default/*``:
Change the interface-specific default settings.
These settings would be used during creating new interfaces.
``conf/all/*``:
Change all the interface-specific settings.
[XXX: Other special features than forwarding?]
conf/all/disable_ipv6 - BOOLEAN
Changing this value is same as changing ``conf/default/disable_ipv6``
setting and also all per-interface ``disable_ipv6`` settings to the same
value.
Reading this value does not have any particular meaning. It does not say
whether IPv6 support is enabled or disabled. Returned value can be 1
also in the case when some interface has ``disable_ipv6`` set to 0 and
has configured IPv6 addresses.
conf/all/forwarding - BOOLEAN
Enable global IPv6 forwarding between all interfaces.
IPv4 and IPv6 work differently here; the ``force_forwarding`` flag must
be used to control which interfaces may forward packets.
This also sets all interfaces' Host/Router setting
'forwarding' to the specified value. See below for details.
This referred to as global forwarding.
IPv6 forwarding과 Router Advertisement
2587-2796`proxy_ndp`는 proxy Neighbor Discovery, `force_forwarding`은 global forwarding과 무관하게 해당 interface만 전달하게 합니다. global forwarding을 0으로 바꾸면 모든 force flag도 reset됩니다. IPv6 `fwmark_reflect`는 socket 없는 RST/ICMPv6 reply가 incoming fwmark를 상속하게 합니다.
interface별 `accept_ra`는 0=RA 거부, 1=forwarding이 꺼졌을 때 수락, 2=forwarding 중에도 강제 수락입니다. 이 값이 실제로 RA를 수락할 때만 Router Solicitation도 보냅니다. `accept_ra_defrtr`는 default router 학습, `ra_defrtr_metric`은 학습 route metric(기본 1024), `accept_ra_from_local`은 local source RA 허용, `accept_ra_min_hop_limit`과 `accept_ra_min_lft`는 받아들일 hop limit·lifetime 하한입니다.
`accept_ra_pinfo`는 Prefix Information 학습입니다. `ra_honor_pio_life`가 0이면 RFC 4862 5.5.3e lifetime 규칙, 1이면 PIO lifetime을 그대로 존중합니다. `ra_honor_pio_pflag`는 DHCPv6-PD client가 P=1인 PIO의 A flag 효과를 억제하게 합니다. `accept_ra_rt_info_min_plen/max_plen`은 Route Information prefix 길이 범위, `accept_ra_rtr_pref`는 router preference, `accept_ra_mtu`는 RA option 5 MTU 적용을 제어합니다.
`accept_redirects`는 host mode에서 기본 허용, forwarding에서는 기본 거부입니다. `accept_source_route`는 음수면 routing header를 모두 거부하고 0 이상이면 type 2만 받습니다. `autoconf`는 PIO 기반 address 자동 구성을, `dad_transmits`는 DAD probe 횟수를 정합니다.
proxy_ndp - BOOLEAN
Do proxy ndp.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
force_forwarding - BOOLEAN
Enable forwarding on this interface only -- regardless of the setting on
``conf/all/forwarding``. When setting ``conf.all.forwarding`` to 0,
the ``force_forwarding`` flag will be reset on all interfaces.
fwmark_reflect - BOOLEAN
Controls the fwmark of kernel-generated IPv6 reply packets that are not
associated with a socket for example, TCP RSTs or ICMPv6 echo replies).
If disabled, these packets have a fwmark of zero. If enabled, they have the
fwmark of the packet they are replying to.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
``conf/interface/*``:
Change special settings per interface.
The functional behaviour for certain settings is different
depending on whether local forwarding is enabled or not.
accept_ra - INTEGER
Accept Router Advertisements; autoconfigure using them.
It also determines whether or not to transmit Router
Solicitations. If and only if the functional setting is to
accept Router Advertisements, Router Solicitations will be
transmitted.
Possible values are:
== ===========================================================
0 Do not accept Router Advertisements.
1 Accept Router Advertisements if forwarding is disabled.
2 Overrule forwarding behaviour. Accept Router Advertisements
even if forwarding is enabled.
== ===========================================================
Functional default:
- enabled if local forwarding is disabled.
- disabled if local forwarding is enabled.
accept_ra_defrtr - BOOLEAN
Learn default router in Router Advertisement.
Functional default:
- enabled if accept_ra is enabled.
- disabled if accept_ra is disabled.
ra_defrtr_metric - UNSIGNED INTEGER
Route metric for default route learned in Router Advertisement. This value
will be assigned as metric for the default route learned via IPv6 Router
Advertisement. Takes affect only if accept_ra_defrtr is enabled.
Possible values:
1 to 0xFFFFFFFF
Default: IP6_RT_PRIO_USER i.e. 1024.
accept_ra_from_local - BOOLEAN
Accept RA with source-address that is found on local machine
if the RA is otherwise proper and able to be accepted.
Default is to NOT accept these as it may be an un-intended
network loop.
Functional default:
- enabled if accept_ra_from_local is enabled
on a specific interface.
- disabled if accept_ra_from_local is disabled
on a specific interface.
accept_ra_min_hop_limit - INTEGER
Minimum hop limit Information in Router Advertisement.
Hop limit Information in Router Advertisement less than this
variable shall be ignored.
Default: 1
accept_ra_min_lft - INTEGER
Minimum acceptable lifetime value in Router Advertisement.
RA sections with a lifetime less than this value shall be
ignored. Zero lifetimes stay unaffected.
Default: 0
accept_ra_pinfo - BOOLEAN
Learn Prefix Information in Router Advertisement.
Functional default:
- enabled if accept_ra is enabled.
- disabled if accept_ra is disabled.
ra_honor_pio_life - BOOLEAN
Whether to use RFC4862 Section 5.5.3e to determine the valid
lifetime of an address matching a prefix sent in a Router
Advertisement Prefix Information Option.
Possible values:
- 0 (disabled) - RFC4862 section 5.5.3e is used to determine
the valid lifetime of the address.
- 1 (enabled) - the PIO valid lifetime will always be honored.
Default: 0 (disabled)
ra_honor_pio_pflag - BOOLEAN
The Prefix Information Option P-flag indicates the network can
allocate a unique IPv6 prefix per client using DHCPv6-PD.
This sysctl can be enabled when a userspace DHCPv6-PD client
is running to cause the P-flag to take effect: i.e. the
P-flag suppresses any effects of the A-flag within the same
PIO. For a given PIO, P=1 and A=1 is treated as A=0.
Possible values:
- 0 (disabled) - the P-flag is ignored.
- 1 (enabled) - the P-flag will disable SLAAC autoconfiguration
for the given Prefix Information Option.
Default: 0 (disabled)
accept_ra_rt_info_min_plen - INTEGER
Minimum prefix length of Route Information in RA.
Route Information w/ prefix smaller than this variable shall
be ignored.
Functional default:
* 0 if accept_ra_rtr_pref is enabled.
* -1 if accept_ra_rtr_pref is disabled.
accept_ra_rt_info_max_plen - INTEGER
Maximum prefix length of Route Information in RA.
Route Information w/ prefix larger than this variable shall
be ignored.
Functional default:
* 0 if accept_ra_rtr_pref is enabled.
* -1 if accept_ra_rtr_pref is disabled.
accept_ra_rtr_pref - BOOLEAN
Accept Router Preference in RA.
Functional default:
- enabled if accept_ra is enabled.
- disabled if accept_ra is disabled.
accept_ra_mtu - BOOLEAN
Apply the MTU value specified in RA option 5 (RFC4861). If
disabled, the MTU specified in the RA will be ignored.
Functional default:
- enabled if accept_ra is enabled.
- disabled if accept_ra is disabled.
accept_redirects - BOOLEAN
Accept Redirects.
Functional default:
- enabled if local forwarding is disabled.
- disabled if local forwarding is enabled.
accept_source_route - INTEGER
Accept source routing (routing extension header).
- >= 0: Accept only routing header type 2.
- < 0: Do not accept routing header.
Default: 0
autoconf - BOOLEAN
Autoconfigure addresses using Prefix Information in Router
Advertisements.
Functional default:
- enabled if accept_ra_pinfo is enabled.
- disabled if accept_ra_pinfo is disabled.
dad_transmits - INTEGER
The amount of Duplicate Address Detection probes to send.
Default: 1
IPv6 host/router 상태와 privacy address
2797-2990interface `forwarding=0`은 host mode로 NA의 IsRouter flag를 지우고 RA/redirect를 수락하며 RS를 보냅니다. `forwarding=1`은 반대로 router mode가 되어 IsRouter를 세우고 `accept_ra=2`가 아니면 RS를 보내지 않고 RA를 무시하며 redirect도 무시합니다. 모든 interface에서 같은 mode를 쓰는 것이 권장됩니다. `hop_limit` 기본 64, IPv6 `mtu` 기본·최소는 1280입니다.
IPv6 `ip_nonlocal_bind`는 non-local IPv6 address bind를 허용합니다. Router probing/solicitation은 `router_probe_interval` 60초, `router_solicitation_delay` 1초, `router_solicitation_interval` 4초, `router_solicitations` 3회가 기본입니다. `use_oif_addrs_only`는 source candidate를 outgoing interface 주소로 제한합니다.
`use_tempaddr`는 0 이하=privacy extension 끔, 1=켜되 public 우선, 1보다 큼=temporary 우선입니다. `temp_valid_lft` 기본 2일, `temp_prefered_lft` 기본 1일이며 valid보다 길 수 없습니다. `keep_addr_on_down`은 interface down 때 static global 주소를 보존할지 정합니다. `max_desync_factor`는 host 간 주소 재생성 시점 분산, `regen_min_advance`는 deprecate 전에 새 temporary address를 만들 최소 선행 시간, `regen_max_retry`는 생성 시도 횟수입니다.
`max_addresses`는 interface별 autoconfigured address 상한이며 기본 16, 0은 무제한이지만 memory DoS 위험이 있습니다. `disable_ipv6`를 1로 바꾸면 해당 interface의 IPv6 주소와 route를 지우고 새 구성을 막으며 0으로 되돌리면 link-local address와 DAD를 시작합니다. `accept_dad`는 0=끔, 1=켬, 2=MAC 기반 link-local 중복이면 IPv6까지 끔이고 all/interface 중 큰 값을 씁니다.
forwarding - INTEGER
Configure interface-specific Host/Router behaviour.
.. note::
It is recommended to have the same setting on all
interfaces; mixed router/host scenarios are rather uncommon.
Possible values are:
- 0 Forwarding disabled
- 1 Forwarding enabled
**FALSE (0)**:
By default, Host behaviour is assumed. This means:
1. IsRouter flag is not set in Neighbour Advertisements.
2. If accept_ra is TRUE (default), transmit Router
Solicitations.
3. If accept_ra is TRUE (default), accept Router
Advertisements (and do autoconfiguration).
4. If accept_redirects is TRUE (default), accept Redirects.
**TRUE (1)**:
If local forwarding is enabled, Router behaviour is assumed.
This means exactly the reverse from the above:
1. IsRouter flag is set in Neighbour Advertisements.
2. Router Solicitations are not sent unless accept_ra is 2.
3. Router Advertisements are ignored unless accept_ra is 2.
4. Redirects are ignored.
Default: 0 (disabled) if global forwarding is disabled (default),
otherwise 1 (enabled).
hop_limit - INTEGER
Default Hop Limit to set.
Default: 64
mtu - INTEGER
Default Maximum Transfer Unit
Default: 1280 (IPv6 required minimum)
ip_nonlocal_bind - BOOLEAN
If enabled, allows processes to bind() to non-local IPv6 addresses,
which can be quite useful - but may break some applications.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
router_probe_interval - INTEGER
Minimum interval (in seconds) between Router Probing described
in RFC4191.
Default: 60
router_solicitation_delay - INTEGER
Number of seconds to wait after interface is brought up
before sending Router Solicitations.
Default: 1
router_solicitation_interval - INTEGER
Number of seconds to wait between Router Solicitations.
Default: 4
router_solicitations - INTEGER
Number of Router Solicitations to send until assuming no
routers are present.
Default: 3
use_oif_addrs_only - BOOLEAN
When enabled, the candidate source addresses for destinations
routed via this interface are restricted to the set of addresses
configured on this interface (vis. RFC 6724, section 4).
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
use_tempaddr - INTEGER
Preference for Privacy Extensions (RFC3041).
* <= 0 : disable Privacy Extensions
* == 1 : enable Privacy Extensions, but prefer public
addresses over temporary addresses.
* > 1 : enable Privacy Extensions and prefer temporary
addresses over public addresses.
Default:
* 0 (for most devices)
* -1 (for point-to-point devices and loopback devices)
temp_valid_lft - INTEGER
valid lifetime (in seconds) for temporary addresses. If less than the
minimum required lifetime (typically 5-7 seconds), temporary addresses
will not be created.
Default: 172800 (2 days)
temp_prefered_lft - INTEGER
Preferred lifetime (in seconds) for temporary addresses. If
temp_prefered_lft is less than the minimum required lifetime (typically
5-7 seconds), the preferred lifetime is the minimum required. If
temp_prefered_lft is greater than temp_valid_lft, the preferred lifetime
is temp_valid_lft.
Default: 86400 (1 day)
keep_addr_on_down - INTEGER
Keep all IPv6 addresses on an interface down event. If set static
global addresses with no expiration time are not flushed.
* >0 : enabled
* 0 : system default
* <0 : disabled
Default: 0 (addresses are removed)
max_desync_factor - INTEGER
Maximum value for DESYNC_FACTOR, which is a random value
that ensures that clients don't synchronize with each
other and generate new addresses at exactly the same time.
value is in seconds.
Default: 600
regen_min_advance - INTEGER
How far in advance (in seconds), at minimum, to create a new temporary
address before the current one is deprecated. This value is added to
the amount of time that may be required for duplicate address detection
to determine when to create a new address. Linux permits setting this
value to less than the default of 2 seconds, but a value less than 2
does not conform to RFC 8981.
Default: 2
regen_max_retry - INTEGER
Number of attempts before give up attempting to generate
valid temporary addresses.
Default: 5
max_addresses - INTEGER
Maximum number of autoconfigured addresses per interface. Setting
to zero disables the limitation. It is not recommended to set this
value too large (or to zero) because it would be an easy way to
crash the kernel by allowing too many addresses to be created.
Default: 16
disable_ipv6 - BOOLEAN
Disable IPv6 operation. If accept_dad is set to 2, this value
will be dynamically set to TRUE if DAD fails for the link-local
address.
Default: FALSE (enable IPv6 operation)
When this value is changed from 1 to 0 (IPv6 is being enabled),
it will dynamically create a link-local address on the given
interface and start Duplicate Address Detection, if necessary.
When this value is changed from 0 to 1 (IPv6 is being disabled),
it will dynamically delete all addresses and routes on the given
interface. From now on it will not possible to add addresses/routes
to the selected interface.
accept_dad - INTEGER
Whether to accept DAD (Duplicate Address Detection).
== ==============================================================
0 Disable DAD
1 Enable DAD (default)
2 Enable DAD, and disable IPv6 operation if MAC-based duplicate
link-local address has been found.
== ==============================================================
DAD operation and mode on a given interface will be selected according
to the maximum value of conf/{all,interface}/accept_dad.
IPv6 Neighbor Discovery, DAD와 address 생성
2991-3193`force_tllao`는 unicast Neighbor Solicitation 응답에도 target link-layer address option을 넣어 cache 삭제와 응답 사이 race를 줄입니다. `ndisc_notify`는 device up/MAC 변경 때 unsolicited NA를 보내고 `ndisc_tclass`는 ND message의 Traffic Class를 정합니다. `ndisc_evict_nocarrier`는 carrier 상실 때 ND cache를 비우며 wireless roaming에서는 0이 유용할 수 있습니다.
`mldv1_unsolicited_report_interval`과 `mldv2_unsolicited_report_interval` 기본은 10초와 1초입니다. `force_mld_version`은 0=자동, 1=MLDv1, 2=MLDv2입니다. `suppress_frag_ndisc` 기본 1은 RFC 6980에 따라 fragmented ND packet을 폐기합니다. `optimistic_dad`는 RFC 4429를 켜고 `use_optimistic`은 source 선택에서 optimistic address를 deprecated로 낮추지 않되 preferred보다는 뒤에 둡니다.
`stable_secret`은 RFC 7217 stable privacy address 생성 secret입니다. namespace 기본 secret은 `conf/default`, interface별 값은 이를 override하고 `conf/all` 쓰기는 거부됩니다. 설치 때 생성해 안정적으로 유지하는 것이 권장됩니다. `addr_gen_mode`는 0=EUI-64, 1=link-local 생성 안 함, 2=설정 secret의 stable privacy, 3=secret이 없으면 random secret의 stable privacy입니다.
`drop_unicast_in_l2_multicast`는 L2 multicast/broadcast에 담긴 unicast IPv6를 버리고 `drop_unsolicited_na`는 unsolicited NA를 버립니다. `accept_untracked_na`는 0=거부, 1=router에서 target LLA가 있는 untracked NA를 STALE entry로 추가, 2=source가 local subnet일 때만 추가합니다. drop 설정이 더 우선합니다. host의 `ndisc_notify`와 조합하면 off-link 통신의 첫 return packet을 위한 NS 지연을 줄일 수 있습니다. `enhanced_dad`는 RFC 7527 nonce를 넣어 자체 NS loopback을 중복으로 오인하지 않게 하며 기본 1입니다.
force_tllao - BOOLEAN
Enable sending the target link-layer address option even when
responding to a unicast neighbor solicitation.
Default: FALSE
Quoting from RFC 2461, section 4.4, Target link-layer address:
"The option MUST be included for multicast solicitations in order to
avoid infinite Neighbor Solicitation "recursion" when the peer node
does not have a cache entry to return a Neighbor Advertisements
message. When responding to unicast solicitations, the option can be
omitted since the sender of the solicitation has the correct link-
layer address; otherwise it would not have be able to send the unicast
solicitation in the first place. However, including the link-layer
address in this case adds little overhead and eliminates a potential
race condition where the sender deletes the cached link-layer address
prior to receiving a response to a previous solicitation."
ndisc_notify - BOOLEAN
Define mode for notification of address and device changes.
Possible values:
- 0 (disabled) - do nothing
- 1 (enabled) - Generate unsolicited neighbour advertisements when device is brought
up or hardware address changes.
Default: 0 (disabled)
ndisc_tclass - INTEGER
The IPv6 Traffic Class to use by default when sending IPv6 Neighbor
Discovery (Router Solicitation, Router Advertisement, Neighbor
Solicitation, Neighbor Advertisement, Redirect) messages.
These 8 bits can be interpreted as 6 high order bits holding the DSCP
value and 2 low order bits representing ECN (which you probably want
to leave cleared).
* 0 - (default)
ndisc_evict_nocarrier - BOOLEAN
Clears the neighbor discovery table on NOCARRIER events. This option is
important for wireless devices where the neighbor discovery cache should
not be cleared when roaming between access points on the same network.
In most cases this should remain as the default (1).
Possible values:
- 0 (disabled) - Do not clear neighbor discovery cache on NOCARRIER events.
- 1 (enabled) - Clear neighbor discover cache on NOCARRIER events.
Default: 1 (enabled)
mldv1_unsolicited_report_interval - INTEGER
The interval in milliseconds in which the next unsolicited
MLDv1 report retransmit will take place.
Default: 10000 (10 seconds)
mldv2_unsolicited_report_interval - INTEGER
The interval in milliseconds in which the next unsolicited
MLDv2 report retransmit will take place.
Default: 1000 (1 second)
force_mld_version - INTEGER
* 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed
* 1 - Enforce to use MLD version 1
* 2 - Enforce to use MLD version 2
suppress_frag_ndisc - INTEGER
Control RFC 6980 (Security Implications of IPv6 Fragmentation
with IPv6 Neighbor Discovery) behavior:
* 1 - (default) discard fragmented neighbor discovery packets
* 0 - allow fragmented neighbor discovery packets
optimistic_dad - BOOLEAN
Whether to perform Optimistic Duplicate Address Detection (RFC 4429).
Optimistic Duplicate Address Detection for the interface will be enabled
if at least one of conf/{all,interface}/optimistic_dad is set to 1,
it will be disabled otherwise.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
use_optimistic - BOOLEAN
If enabled, do not classify optimistic addresses as deprecated during
source address selection. Preferred addresses will still be chosen
before optimistic addresses, subject to other ranking in the source
address selection algorithm.
This will be enabled if at least one of
conf/{all,interface}/use_optimistic is set to 1, disabled otherwise.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
stable_secret - IPv6 address
This IPv6 address will be used as a secret to generate IPv6
addresses for link-local addresses and autoconfigured
ones. All addresses generated after setting this secret will
be stable privacy ones by default. This can be changed via the
addrgenmode ip-link. conf/default/stable_secret is used as the
secret for the namespace, the interface specific ones can
overwrite that. Writes to conf/all/stable_secret are refused.
It is recommended to generate this secret during installation
of a system and keep it stable after that.
By default the stable secret is unset.
addr_gen_mode - INTEGER
Defines how link-local and autoconf addresses are generated.
= =================================================================
0 generate address based on EUI64 (default)
1 do no generate a link-local address, use EUI64 for addresses
generated from autoconf
2 generate stable privacy addresses, using the secret from
stable_secret (RFC7217)
3 generate stable privacy addresses, using a random secret if unset
= =================================================================
drop_unicast_in_l2_multicast - BOOLEAN
Drop any unicast IPv6 packets that are received in link-layer
multicast (or broadcast) frames.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
drop_unsolicited_na - BOOLEAN
Drop all unsolicited neighbor advertisements, for example if there's
a known good NA proxy on the network and such frames need not be used
(or in the case of 802.11, must not be used to prevent attacks.)
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled).
accept_untracked_na - INTEGER
Define behavior for accepting neighbor advertisements from devices that
are absent in the neighbor cache:
- 0 - (default) Do not accept unsolicited and untracked neighbor
advertisements.
- 1 - Add a new neighbor cache entry in STALE state for routers on
receiving a neighbor advertisement (either solicited or unsolicited)
with target link-layer address option specified if no neighbor entry
is already present for the advertised IPv6 address. Without this knob,
NAs received for untracked addresses (absent in neighbor cache) are
silently ignored.
This is as per router-side behavior documented in RFC9131.
This has lower precedence than drop_unsolicited_na.
This will optimize the return path for the initial off-link
communication that is initiated by a directly connected host, by
ensuring that the first-hop router which turns on this setting doesn't
have to buffer the initial return packets to do neighbor-solicitation.
The prerequisite is that the host is configured to send unsolicited
neighbor advertisements on interface bringup. This setting should be
used in conjunction with the ndisc_notify setting on the host to
satisfy this prerequisite.
- 2 - Extend option (1) to add a new neighbor cache entry only if the
source IP address is in the same subnet as an address configured on
the interface that received the neighbor advertisement.
enhanced_dad - BOOLEAN
Include a nonce option in the IPv6 neighbor solicitation messages used for
duplicate address detection per RFC7527. A received DAD NS will only signal
a duplicate address if the nonce is different. This avoids any false
detection of duplicates due to loopback of the NS messages that we send.
The nonce option will be sent on an interface unless both of
conf/{all,interface}/enhanced_dad are set to FALSE.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
ICMPv6 제한과 echo 정책
3194-3277ICMPv6 `ratelimit`은 peer별 응답 사이의 최소 ms이고 0은 제한을 끕니다. `ratemask`는 `0-127,129` 같은 range 목록으로 제한할 ICMPv6 type을 정하며 쓰면 기존 목록을 교체합니다. 기본 `0-1,3-127`은 Packet Too Big을 제외한 ICMPv6 error를 제한합니다.
`echo_ignore_all`, `echo_ignore_multicast`, `echo_ignore_anycast`는 각각 모든, multicast 대상, anycast 대상 echo request를 무시합니다. `error_anycast_as_unicast`는 anycast 대상 요청에서 발생한 ICMP error에도 unicast처럼 응답합니다. 모두 기본 0입니다. `xfrm6_gc_thresh`는 4.14부터 폐기되었습니다. 마지막에는 IPv6 문서 갱신자 credit이 있습니다.
``icmp/*``:
===========
ratelimit - INTEGER
Limit the maximal rates for sending ICMPv6 messages to a particular
peer.
0 to disable any limiting,
otherwise the space between responses in milliseconds.
Default: 100
ratemask - list of comma separated ranges
For ICMPv6 message types matching the ranges in the ratemask, limit
the sending of the message according to ratelimit parameter.
The format used for both input and output is a comma separated
list of ranges (e.g. "0-127,129" for ICMPv6 message type 0 to 127 and
129). Writing to the file will clear all previous ranges of ICMPv6
message types and update the current list with the input.
Refer to: https://www.iana.org/assignments/icmpv6-parameters/icmpv6-parameters.xhtml
for numerical values of ICMPv6 message types, e.g. echo request is 128
and echo reply is 129.
Default: 0-1,3-127 (rate limit ICMPv6 errors except Packet Too Big)
echo_ignore_all - BOOLEAN
If enabled, then the kernel will ignore all ICMP ECHO
requests sent to it over the IPv6 protocol.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
echo_ignore_multicast - BOOLEAN
If enabled, then the kernel will ignore all ICMP ECHO
requests sent to it over the IPv6 protocol via multicast.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
echo_ignore_anycast - BOOLEAN
If enabled, then the kernel will ignore all ICMP ECHO
requests sent to it over the IPv6 protocol destined to anycast address.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
error_anycast_as_unicast - BOOLEAN
If enabled, then the kernel will respond with ICMP Errors
resulting from requests sent to it over the IPv6 protocol destined
to anycast address essentially treating anycast as unicast.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 0 (disabled)
xfrm6_gc_thresh - INTEGER
(Obsolete since linux-4.14)
The threshold at which we will start garbage collecting for IPv6
destination cache entries. At twice this value the system will
refuse new allocations.
IPv6 Update by:
Pekka Savola <[email protected]>
YOSHIFUJI Hideaki / USAGI Project <[email protected]>
Bridge netfilter 변수
3278-3337`bridge-nf-call-arptables`, `bridge-nf-call-iptables`, `bridge-nf-call-ip6tables`는 bridged ARP, IPv4, IPv6 traffic을 각 netfilter chain에 통과시키며 기본 1입니다. `bridge-nf-filter-vlan-tagged`와 `bridge-nf-filter-pppoe-tagged`는 VLAN/PPPoE tagged bridged traffic에도 filtering을 적용하며 기본 0입니다.
`bridge-nf-pass-vlan-input-dev`를 켜면 VLAN-tagged filtering 중 bridge 위의 matching VLAN interface를 찾아 netfilter input device로 사용합니다. 그러면 `iptables -i br0.1`과 VLAN-on-bridge의 REDIRECT가 동작합니다. match가 없거나 설정이 0이면 bridge interface를 씁니다.
/proc/sys/net/bridge/* Variables:
=================================
bridge-nf-call-arptables - BOOLEAN
Possible values:
- 0 (disabled) - disable this.
- 1 (enabled) - pass bridged ARP traffic to arptables' FORWARD chain.
Default: 1 (enabled)
bridge-nf-call-iptables - BOOLEAN
Possible values:
- 0 (disabled) - disable this.
- 1 (enabled) - pass bridged IPv4 traffic to iptables' chains.
Default: 1 (enabled)
bridge-nf-call-ip6tables - BOOLEAN
Possible values:
- 0 (disabled) - disable this.
- 1 (enabled) - pass bridged IPv6 traffic to ip6tables' chains.
Default: 1 (enabled)
bridge-nf-filter-vlan-tagged - BOOLEAN
Possible values:
- 0 (disabled) - disable this.
- 1 (enabled) - pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables
Default: 0 (disabled)
bridge-nf-filter-pppoe-tagged - BOOLEAN
Possible values:
- 0 (disabled) - disable this.
- 1 (enabled) - pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables.
Default: 0 (disabled)
bridge-nf-pass-vlan-input-dev - BOOLEAN
- 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan
interface on the bridge and set the netfilter input device to the
vlan. This allows use of e.g. "iptables -i br0.1" and makes the
REDIRECT target work with vlan-on-top-of-bridge interfaces. When no
matching vlan interface is found, or this switch is off, the input
device is set to the bridge interface.
- 0: disable bridge netfilter vlan interface lookup.
Default: 0
SCTP extension과 failover
3338-3509`addip_enable`은 RFC 5061 Dynamic Address Reconfiguration을, `auth_enable`은 ADD-IP 보호에 필요한 Authenticated Chunks를 켭니다. `addip_noauth_enable`은 구형 구현과의 폐쇄망 호환을 위해 인증 없는 ADD-IP를 허용하므로 일반 환경에서는 0을 유지해야 합니다. `prsctp_enable`은 RFC 3758 Partial Reliability를 켜고 기본 1입니다.
`pf_enable`은 Potentially Failed path 상태를 켭니다. `pf_retrans`가 `path_max_retrans`보다 크면 역시 PF를 끕니다. `pf_expose`는 application에 PF transition event와 `SCTP_GET_PEER_ADDR_INFO`를 0=event 없이 조회 허용, 1=둘 다 차단, 2=event와 조회 허용으로 노출합니다. `max_burst` 기본 4는 처음 보낼 새 packet burst를 제한합니다.
`association_max_retrans`는 association 전체 재전송 상한 10, `max_init_retransmits`는 INIT/COOKIE-ECHO 상한 8, `path_max_retrans`는 path unreachable 판정 상한 5입니다. `pf_retrans`는 alternate transport로 빨리 돌릴 threshold이고 path 자체는 계속 쓸 수 있습니다. `ps_retrans`는 RFC 7829 Primary Switchover threshold이며 기본 0xffff로 기능을 끄고 `pf_retrans`보다 작게 설정할 수 없습니다.
SCTP RTO의 `rto_initial`, `rto_max`, `rto_min` 기본값은 각각 3000, 60000, 1000 ms입니다.
``proc/sys/net/sctp/*`` Variables:
==================================
addip_enable - BOOLEAN
Enable or disable extension of Dynamic Address Reconfiguration
(ADD-IP) functionality specified in RFC5061. This extension provides
the ability to dynamically add and remove new addresses for the SCTP
associations.
Possible values:
- 0 (disabled) - disable extension.
- 1 (enabled) - enable extension
Default: 0 (disabled)
pf_enable - INTEGER
Enable or disable pf (pf is short for potentially failed) state. A value
of pf_retrans > path_max_retrans also disables pf state. That is, one of
both pf_enable and pf_retrans > path_max_retrans can disable pf state.
Since pf_retrans and path_max_retrans can be changed by userspace
application, sometimes user expects to disable pf state by the value of
pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
or path_max_retrans is changed by the user application, this pf state is
enabled. As such, it is necessary to add this to dynamically enable
and disable pf state. See:
https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
details.
Possible values:
- 1: Enable pf.
- 0: Disable pf.
Default: 1
pf_expose - INTEGER
Unset or enable/disable pf (pf is short for potentially failed) state
exposure. Applications can control the exposure of the PF path state
in the SCTP_PEER_ADDR_CHANGE event and access of SCTP_PF-state
transport info via SCTP_GET_PEER_ADDR_INFO sockopt.
Possible values:
- 0: Unset pf state exposure (compatible with old applications). No
event will be sent but the transport info can be queried.
- 1: Disable pf state exposure. No event will be sent and trying to
obtain transport info will return -EACCESS.
- 2: Enable pf state exposure. The event will be sent for a transport
becoming SCTP_PF state and transport info can be obtained.
Default: 0
addip_noauth_enable - BOOLEAN
Dynamic Address Reconfiguration (ADD-IP) requires the use of
authentication to protect the operations of adding or removing new
addresses. This requirement is mandated so that unauthorized hosts
would not be able to hijack associations. However, older
implementations may not have implemented this requirement while
allowing the ADD-IP extension. For reasons of interoperability,
we provide this variable to control the enforcement of the
authentication requirement.
== ===============================================================
1 Allow ADD-IP extension to be used without authentication. This
should only be set in a closed environment for interoperability
with older implementations.
0 Enforce the authentication requirement
== ===============================================================
Default: 0
auth_enable - BOOLEAN
Enable or disable Authenticated Chunks extension. This extension
provides the ability to send and receive authenticated chunks and is
required for secure operation of Dynamic Address Reconfiguration
(ADD-IP) extension.
Possible values:
- 0 (disabled) - disable extension.
- 1 (enabled) - enable extension
Default: 0 (disabled)
prsctp_enable - BOOLEAN
Enable or disable the Partial Reliability extension (RFC3758) which
is used to notify peers that a given DATA should no longer be expected.
Possible values:
- 0 (disabled) - disable extension.
- 1 (enabled) - enable extension
Default: 1 (enabled)
max_burst - INTEGER
The limit of the number of new packets that can be initially sent. It
controls how bursty the generated traffic can be.
Default: 4
association_max_retrans - INTEGER
Set the maximum number for retransmissions that an association can
attempt deciding that the remote end is unreachable. If this value
is exceeded, the association is terminated.
Default: 10
max_init_retransmits - INTEGER
The maximum number of retransmissions of INIT and COOKIE-ECHO chunks
that an association will attempt before declaring the destination
unreachable and terminating.
Default: 8
path_max_retrans - INTEGER
The maximum number of retransmissions that will be attempted on a given
path. Once this threshold is exceeded, the path is considered
unreachable, and new traffic will use a different path when the
association is multihomed.
Default: 5
pf_retrans - INTEGER
The number of retransmissions that will be attempted on a given path
before traffic is redirected to an alternate transport (should one
exist). Note this is distinct from path_max_retrans, as a path that
passes the pf_retrans threshold can still be used. Its only
deprioritized when a transmission path is selected by the stack. This
setting is primarily used to enable fast failover mechanisms without
having to reduce path_max_retrans to a very low value. See:
http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
for details. Note also that a value of pf_retrans > path_max_retrans
disables this feature. Since both pf_retrans and path_max_retrans can
be changed by userspace application, a variable pf_enable is used to
disable pf state.
Default: 0
ps_retrans - INTEGER
Primary.Switchover.Max.Retrans (PSMR), it's a tunable parameter coming
from section-5 "Primary Path Switchover" in rfc7829. The primary path
will be changed to another active path when the path error counter on
the old primary path exceeds PSMR, so that "the SCTP sender is allowed
to continue data transmission on a new working path even when the old
primary destination address becomes active again". Note this feature
is disabled by initializing 'ps_retrans' per netns as 0xffff by default,
and its value can't be less than 'pf_retrans' when changing by sysctl.
Default: 0xffff
rto_initial - INTEGER
The initial round trip timeout value in milliseconds that will be used
in calculating round trip times. This is the initial time interval
for retransmissions.
Default: 3000
rto_max - INTEGER
The maximum value (in milliseconds) of the round trip timeout. This
is the largest time interval that can elapse between retransmissions.
Default: 60000
rto_min - INTEGER
The minimum value (in milliseconds) of the round trip timeout. This
is the smallest time interval the can elapse between retransmissions.
Default: 1000
SCTP timer, memory와 encapsulation
3510-3717`hb_interval`은 idle path HEARTBEAT 간격 30000 ms, `sack_timeout`은 SACK 지연 200 ms, `valid_cookie_life`는 association cookie 수명 60000 ms입니다. `cookie_preserve_enable`은 cookie lifetime extension을 허용하고 `cookie_hmac_alg`는 기본 `sha256` 또는 `none`을 고릅니다.
`rcvbuf_policy`와 `sndbuf_policy`는 buffer accounting을 0=socket별, 1=association별로 정합니다. 한 socket의 stalled association이 다른 association을 막는 문제에는 association별 receive 정책이 유용합니다. `sctp_mem`은 전체 queue page의 min/pressure/max, `sctp_rmem`과 `sctp_wmem`은 첫 min 값만 사용해 memory pressure에서도 socket별 4 KiB를 보장하고 default/max 값은 무시합니다.
`addr_scope_policy`는 IPv4 scope를 0=끔, 1=draft 준수, 2=private address 허용, 3=link-local 허용으로 정합니다. `udp_port`는 RFC 6951 incoming SCTP-over-UDP를 받는 local shared tunnel port이자 outgoing source port이고 0이면 socket을 닫습니다. `encap_port`는 outgoing destination port의 기본값이며 socket/association/transport별 option으로 override할 수 있습니다.
`plpmtud_probe_interval`은 PLPMTUD probe ACK timer와 완료 후 current PMTU 재probe 간격입니다. 0은 기능을 끄고 다른 값은 최소 5000 ms입니다. `reconf_enable`은 RFC 6525 Stream Reconfiguration, `intl_enable`은 RFC 8260 User Message Interleaving을 켭니다. interleaving은 sysctl 외에 `SCTP_FRAGMENT_INTERLEAVE=2`와 `SCTP_INTERLEAVING_SUPPORTED=1`도 필요합니다. `ecn_enable`은 peer 양쪽이 지원할 때 SCTP ECN을 사용하고 기본 1입니다. `l3mdev_accept`는 global bound SCTP socket이 VRF를 가로질러 받게 하며 기본 1입니다.
hb_interval - INTEGER
The interval (in milliseconds) between HEARTBEAT chunks. These chunks
are sent at the specified interval on idle paths to probe the state of
a given path between 2 associations.
Default: 30000
sack_timeout - INTEGER
The amount of time (in milliseconds) that the implementation will wait
to send a SACK.
Default: 200
valid_cookie_life - INTEGER
The default lifetime of the SCTP cookie (in milliseconds). The cookie
is used during association establishment.
Default: 60000
cookie_preserve_enable - BOOLEAN
Enable or disable the ability to extend the lifetime of the SCTP cookie
that is used during the establishment phase of SCTP association
Possible values:
- 0 (disabled) - disable.
- 1 (enabled) - enable cookie lifetime extension.
Default: 1 (enabled)
cookie_hmac_alg - STRING
Select the hmac algorithm used when generating the cookie value sent by
a listening sctp socket to a connecting client in the INIT-ACK chunk.
Valid values are:
* sha256
* none
Default: sha256
rcvbuf_policy - INTEGER
Determines if the receive buffer is attributed to the socket or to
association. SCTP supports the capability to create multiple
associations on a single socket. When using this capability, it is
possible that a single stalled association that's buffering a lot
of data may block other associations from delivering their data by
consuming all of the receive buffer space. To work around this,
the rcvbuf_policy could be set to attribute the receiver buffer space
to each association instead of the socket. This prevents the described
blocking.
- 1: rcvbuf space is per association
- 0: rcvbuf space is per socket
Default: 0
sndbuf_policy - INTEGER
Similar to rcvbuf_policy above, this applies to send buffer space.
- 1: Send buffer is tracked per association
- 0: Send buffer is tracked per socket.
Default: 0
sctp_mem - vector of 3 INTEGERs: min, pressure, max
Number of pages allowed for queueing by all SCTP sockets.
* min: Below this number of pages SCTP is not bothered about its
memory usage. When amount of memory allocated by SCTP exceeds
this number, SCTP starts to moderate memory usage.
* pressure: This value was introduced to follow format of tcp_mem.
* max: Maximum number of allowed pages.
Default is calculated at boot time from amount of available memory.
sctp_rmem - vector of 3 INTEGERs: min, default, max
Only the first value ("min") is used, "default" and "max" are
ignored.
* min: Minimal size of receive buffer used by SCTP socket.
It is guaranteed to each SCTP socket (but not association) even
under moderate memory pressure.
Default: 4K
sctp_wmem - vector of 3 INTEGERs: min, default, max
Only the first value ("min") is used, "default" and "max" are
ignored.
* min: Minimum size of send buffer that can be used by SCTP sockets.
It is guaranteed to each SCTP socket (but not association) even
under moderate memory pressure.
Default: 4K
addr_scope_policy - INTEGER
Control IPv4 address scoping (see
https://datatracker.ietf.org/doc/draft-stewart-tsvwg-sctp-ipv4/00/
for details).
- 0 - Disable IPv4 address scoping
- 1 - Enable IPv4 address scoping
- 2 - Follow draft but allow IPv4 private addresses
- 3 - Follow draft but allow IPv4 link local addresses
Default: 1
udp_port - INTEGER
The listening port for the local UDP tunneling sock. Normally it's
using the IANA-assigned UDP port number 9899 (sctp-tunneling).
This UDP sock is used for processing the incoming UDP-encapsulated
SCTP packets (from RFC6951), and shared by all applications in the
same net namespace. This UDP sock will be closed when the value is
set to 0.
The value will also be used to set the src port of the UDP header
for the outgoing UDP-encapsulated SCTP packets. For the dest port,
please refer to 'encap_port' below.
Default: 0
encap_port - INTEGER
The default remote UDP encapsulation port.
This value is used to set the dest port of the UDP header for the
outgoing UDP-encapsulated SCTP packets by default. Users can also
change the value for each sock/asoc/transport by using setsockopt.
For further information, please refer to RFC6951.
Note that when connecting to a remote server, the client should set
this to the port that the UDP tunneling sock on the peer server is
listening to and the local UDP tunneling sock on the client also
must be started. On the server, it would get the encap_port from
the incoming packet's source port.
Default: 0
plpmtud_probe_interval - INTEGER
The time interval (in milliseconds) for the PLPMTUD probe timer,
which is configured to expire after this period to receive an
acknowledgment to a probe packet. This is also the time interval
between the probes for the current pmtu when the probe search
is done.
PLPMTUD will be disabled when 0 is set, and other values for it
must be >= 5000.
Default: 0
reconf_enable - BOOLEAN
Enable or disable extension of Stream Reconfiguration functionality
specified in RFC6525. This extension provides the ability to "reset"
a stream, and it includes the Parameters of "Outgoing/Incoming SSN
Reset", "SSN/TSN Reset" and "Add Outgoing/Incoming Streams".
Possible values:
- 0 (disabled) - Disable extension.
- 1 (enabled) - Enable extension.
Default: 0 (disabled)
intl_enable - BOOLEAN
Enable or disable extension of User Message Interleaving functionality
specified in RFC8260. This extension allows the interleaving of user
messages sent on different streams. With this feature enabled, I-DATA
chunk will replace DATA chunk to carry user messages if also supported
by the peer. Note that to use this feature, one needs to set this option
to 1 and also needs to set socket options SCTP_FRAGMENT_INTERLEAVE to 2
and SCTP_INTERLEAVING_SUPPORTED to 1.
Possible values:
- 0 (disabled) - Disable extension.
- 1 (enabled) - Enable extension.
Default: 0 (disabled)
ecn_enable - BOOLEAN
Control use of Explicit Congestion Notification (ECN) by SCTP.
Like in TCP, ECN is used only when both ends of the SCTP connection
indicate support for it. This feature is useful in avoiding losses
due to congestion by allowing supporting routers to signal congestion
before having to drop packets.
Possible values:
- 0 (disabled) - Disable ecn.
- 1 (enabled) - Enable ecn.
Default: 1 (enabled)
l3mdev_accept - BOOLEAN
Enabling this option allows a "global" bound socket to work
across L3 master domains (e.g., VRFs) with packets capable of
being received regardless of the L3 domain in which they
originated. Only valid when the kernel was compiled with
CONFIG_NET_L3_MASTER_DEV.
Possible values:
- 0 (disabled)
- 1 (enabled)
Default: 1 (enabled)
Core 참조와 Unix datagram queue
3718-3731`/proc/sys/net/core/*` 항목은 `Documentation/admin-guide/sysctl/net.rst`에서 설명합니다. `/proc/sys/net/unix/max_dgram_qlen`은 Unix datagram socket receive queue의 최대 길이이며 기본 10입니다.
``/proc/sys/net/core/*``
========================
Please see: Documentation/admin-guide/sysctl/net.rst for descriptions of these entries.
``/proc/sys/net/unix/*``
========================
max_dgram_qlen - INTEGER
The maximum length of dgram socket receive queue
Default: 10
요약·해설
ip-sysctl.rst:1-3731이 문서는 Linux network stack의 동작을 runtime에 바꾸는 sysctl의 의미와 기본값을 protocol별로 정리합니다. 같은 이름이라도 global, `conf/all`, `conf/default`, `conf/interface`의 결합 규칙이 다르고, forwarding처럼 값을 바꾸며 다른 기본값을 초기화하는 항목도 있어 설정 단위를 함께 읽어야 합니다.
긴 sysctl 목록을 운영 목적에 따라 분류했습니다.
새 interface 기본값과 현재 interface 설정의 역할을 구분해야 합니다.
성능 관련 숫자는 무조건 크게 하는 값이 아닙니다. queue와 memory 상한 증가는 burst 흡수와 동시 연결 수를 늘리지만 memory pressure와 지연을 키울 수 있고, PMTU·ECN·reordering·SYN 보호 설정은 peer와 middlebox의 실제 동작을 전제로 합니다. 변경 전 namespace 범위, socket option override, kernel build option, 기본값과 폐기 여부를 확인해야 합니다.