Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | /proc/sys/net/ipv4/* Variables: |
2 | ||
3 | ip_forward - BOOLEAN | |
4 | 0 - disabled (default) | |
e18f5feb | 5 | not 0 - enabled |
1da177e4 LT |
6 | |
7 | Forward Packets between interfaces. | |
8 | ||
9 | This variable is special, its change resets all configuration | |
10 | parameters to their default state (RFC1122 for hosts, RFC1812 | |
11 | for routers) | |
12 | ||
13 | ip_default_ttl - INTEGER | |
cc6f02dd ED |
14 | Default value of TTL field (Time To Live) for outgoing (but not |
15 | forwarded) IP packets. Should be between 1 and 255 inclusive. | |
16 | Default: 64 (as recommended by RFC1700) | |
1da177e4 | 17 | |
cd174e67 HFS |
18 | ip_no_pmtu_disc - INTEGER |
19 | Disable Path MTU Discovery. If enabled in mode 1 and a | |
188b04d5 HFS |
20 | fragmentation-required ICMP is received, the PMTU to this |
21 | destination will be set to min_pmtu (see below). You will need | |
22 | to raise min_pmtu to the smallest interface MTU on your system | |
23 | manually if you want to avoid locally generated fragments. | |
cd174e67 HFS |
24 | |
25 | In mode 2 incoming Path MTU Discovery messages will be | |
26 | discarded. Outgoing frames are handled the same as in mode 1, | |
27 | implicitly setting IP_PMTUDISC_DONT on every created socket. | |
28 | ||
8ed1dc44 HFS |
29 | Mode 3 is a hardend pmtu discover mode. The kernel will only |
30 | accept fragmentation-needed errors if the underlying protocol | |
31 | can verify them besides a plain socket lookup. Current | |
32 | protocols for which pmtu events will be honored are TCP, SCTP | |
33 | and DCCP as they verify e.g. the sequence number or the | |
34 | association. This mode should not be enabled globally but is | |
35 | only intended to secure e.g. name servers in namespaces where | |
36 | TCP path mtu must still work but path MTU information of other | |
37 | protocols should be discarded. If enabled globally this mode | |
38 | could break other protocols. | |
39 | ||
40 | Possible values: 0-3 | |
188b04d5 | 41 | Default: FALSE |
1da177e4 LT |
42 | |
43 | min_pmtu - INTEGER | |
20db93c3 | 44 | default 552 - minimum discovered Path MTU |
1da177e4 | 45 | |
f87c10a8 HFS |
46 | ip_forward_use_pmtu - BOOLEAN |
47 | By default we don't trust protocol path MTUs while forwarding | |
48 | because they could be easily forged and can lead to unwanted | |
49 | fragmentation by the router. | |
50 | You only need to enable this if you have user-space software | |
51 | which tries to discover path mtus by itself and depends on the | |
52 | kernel honoring this information. This is normally not the | |
53 | case. | |
54 | Default: 0 (disabled) | |
55 | Possible values: | |
56 | 0 - disabled | |
57 | 1 - enabled | |
58 | ||
219b5f29 LV |
59 | fwmark_reflect - BOOLEAN |
60 | Controls the fwmark of kernel-generated IPv4 reply packets that are not | |
61 | associated with a socket for example, TCP RSTs or ICMP echo replies). | |
62 | If unset, these packets have a fwmark of zero. If set, they have the | |
63 | fwmark of the packet they are replying to. | |
64 | Default: 0 | |
65 | ||
cbaf087a BG |
66 | route/max_size - INTEGER |
67 | Maximum number of routes allowed in the kernel. Increase | |
68 | this when using large numbers of interfaces and/or routes. | |
25050c63 AS |
69 | From linux kernel 3.6 onwards, this is deprecated for ipv4 |
70 | as route cache is no longer used. | |
cbaf087a | 71 | |
2724680b YH |
72 | neigh/default/gc_thresh1 - INTEGER |
73 | Minimum number of entries to keep. Garbage collector will not | |
74 | purge entries if there are fewer than this number. | |
b66c66dc | 75 | Default: 128 |
2724680b | 76 | |
a3d12146 | 77 | neigh/default/gc_thresh2 - INTEGER |
78 | Threshold when garbage collector becomes more aggressive about | |
79 | purging entries. Entries older than 5 seconds will be cleared | |
80 | when over this number. | |
81 | Default: 512 | |
82 | ||
cbaf087a BG |
83 | neigh/default/gc_thresh3 - INTEGER |
84 | Maximum number of neighbor entries allowed. Increase this | |
85 | when using large numbers of interfaces and when communicating | |
86 | with large numbers of directly-connected peers. | |
cc868028 | 87 | Default: 1024 |
cbaf087a | 88 | |
8b5c171b ED |
89 | neigh/default/unres_qlen_bytes - INTEGER |
90 | The maximum number of bytes which may be used by packets | |
91 | queued for each unresolved address by other network layers. | |
92 | (added in linux 3.3) | |
3b09adcb | 93 | Setting negative value is meaningless and will return error. |
cc868028 | 94 | Default: 65536 Bytes(64KB) |
8b5c171b ED |
95 | |
96 | neigh/default/unres_qlen - INTEGER | |
97 | The maximum number of packets which may be queued for each | |
98 | unresolved address by other network layers. | |
99 | (deprecated in linux 3.3) : use unres_qlen_bytes instead. | |
cc868028 | 100 | Prior to linux 3.3, the default value is 3 which may cause |
5d248c49 | 101 | unexpected packet loss. The current default value is calculated |
cc868028 SW |
102 | according to default value of unres_qlen_bytes and true size of |
103 | packet. | |
104 | Default: 31 | |
8b5c171b | 105 | |
1da177e4 LT |
106 | mtu_expires - INTEGER |
107 | Time, in seconds, that cached PMTU information is kept. | |
108 | ||
109 | min_adv_mss - INTEGER | |
110 | The advertised MSS depends on the first hop route MTU, but will | |
111 | never be lower than this setting. | |
112 | ||
113 | IP Fragmentation: | |
114 | ||
115 | ipfrag_high_thresh - INTEGER | |
e18f5feb | 116 | Maximum memory used to reassemble IP fragments. When |
1da177e4 LT |
117 | ipfrag_high_thresh bytes of memory is allocated for this purpose, |
118 | the fragment handler will toss packets until ipfrag_low_thresh | |
1bab4c75 NA |
119 | is reached. This also serves as a maximum limit to namespaces |
120 | different from the initial one. | |
e18f5feb | 121 | |
1da177e4 | 122 | ipfrag_low_thresh - INTEGER |
b13d3cbf FW |
123 | Maximum memory used to reassemble IP fragments before the kernel |
124 | begins to remove incomplete fragment queues to free up resources. | |
125 | The kernel still accepts new fragments for defragmentation. | |
1da177e4 LT |
126 | |
127 | ipfrag_time - INTEGER | |
e18f5feb | 128 | Time in seconds to keep an IP fragment in memory. |
1da177e4 | 129 | |
89cee8b1 | 130 | ipfrag_max_dist - INTEGER |
e18f5feb JDB |
131 | ipfrag_max_dist is a non-negative integer value which defines the |
132 | maximum "disorder" which is allowed among fragments which share a | |
133 | common IP source address. Note that reordering of packets is | |
134 | not unusual, but if a large number of fragments arrive from a source | |
135 | IP address while a particular fragment queue remains incomplete, it | |
136 | probably indicates that one or more fragments belonging to that queue | |
137 | have been lost. When ipfrag_max_dist is positive, an additional check | |
138 | is done on fragments before they are added to a reassembly queue - if | |
139 | ipfrag_max_dist (or more) fragments have arrived from a particular IP | |
140 | address between additions to any IP fragment queue using that source | |
141 | address, it's presumed that one or more fragments in the queue are | |
142 | lost. The existing fragment queue will be dropped, and a new one | |
89cee8b1 HX |
143 | started. An ipfrag_max_dist value of zero disables this check. |
144 | ||
145 | Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can | |
146 | result in unnecessarily dropping fragment queues when normal | |
e18f5feb JDB |
147 | reordering of packets occurs, which could lead to poor application |
148 | performance. Using a very large value, e.g. 50000, increases the | |
149 | likelihood of incorrectly reassembling IP fragments that originate | |
89cee8b1 HX |
150 | from different IP datagrams, which could result in data corruption. |
151 | Default: 64 | |
152 | ||
1da177e4 LT |
153 | INET peer storage: |
154 | ||
155 | inet_peer_threshold - INTEGER | |
e18f5feb | 156 | The approximate size of the storage. Starting from this threshold |
1da177e4 LT |
157 | entries will be thrown aggressively. This threshold also determines |
158 | entries' time-to-live and time intervals between garbage collection | |
159 | passes. More entries, less time-to-live, less GC interval. | |
160 | ||
161 | inet_peer_minttl - INTEGER | |
162 | Minimum time-to-live of entries. Should be enough to cover fragment | |
163 | time-to-live on the reassembling side. This minimum time-to-live is | |
164 | guaranteed if the pool size is less than inet_peer_threshold. | |
77a538d5 | 165 | Measured in seconds. |
1da177e4 LT |
166 | |
167 | inet_peer_maxttl - INTEGER | |
168 | Maximum time-to-live of entries. Unused entries will expire after | |
169 | this period of time if there is no memory pressure on the pool (i.e. | |
170 | when the number of entries in the pool is very small). | |
77a538d5 | 171 | Measured in seconds. |
1da177e4 | 172 | |
e18f5feb | 173 | TCP variables: |
1da177e4 | 174 | |
ef56e622 SH |
175 | somaxconn - INTEGER |
176 | Limit of socket listen() backlog, known in userspace as SOMAXCONN. | |
177 | Defaults to 128. See also tcp_max_syn_backlog for additional tuning | |
178 | for TCP sockets. | |
179 | ||
ef56e622 SH |
180 | tcp_abort_on_overflow - BOOLEAN |
181 | If listening service is too slow to accept new connections, | |
182 | reset them. Default state is FALSE. It means that if overflow | |
183 | occurred due to a burst, connection will recover. Enable this | |
184 | option _only_ if you are really sure that listening daemon | |
185 | cannot be tuned to accept connections faster. Enabling this | |
186 | option can harm clients of your server. | |
1da177e4 | 187 | |
ef56e622 SH |
188 | tcp_adv_win_scale - INTEGER |
189 | Count buffering overhead as bytes/2^tcp_adv_win_scale | |
190 | (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale), | |
191 | if it is <= 0. | |
0147fc05 | 192 | Possible values are [-31, 31], inclusive. |
b49960a0 | 193 | Default: 1 |
1da177e4 | 194 | |
ef56e622 SH |
195 | tcp_allowed_congestion_control - STRING |
196 | Show/set the congestion control choices available to non-privileged | |
197 | processes. The list is a subset of those listed in | |
198 | tcp_available_congestion_control. | |
199 | Default is "reno" and the default setting (tcp_congestion_control). | |
1da177e4 | 200 | |
ef56e622 SH |
201 | tcp_app_win - INTEGER |
202 | Reserve max(window/2^tcp_app_win, mss) of window for application | |
203 | buffer. Value 0 is special, it means that nothing is reserved. | |
204 | Default: 31 | |
1da177e4 | 205 | |
f54b3111 ED |
206 | tcp_autocorking - BOOLEAN |
207 | Enable TCP auto corking : | |
208 | When applications do consecutive small write()/sendmsg() system calls, | |
209 | we try to coalesce these small writes as much as possible, to lower | |
210 | total amount of sent packets. This is done if at least one prior | |
211 | packet for the flow is waiting in Qdisc queues or device transmit | |
212 | queue. Applications can still use TCP_CORK for optimal behavior | |
213 | when they know how/when to uncork their sockets. | |
214 | Default : 1 | |
215 | ||
ef56e622 SH |
216 | tcp_available_congestion_control - STRING |
217 | Shows the available congestion control choices that are registered. | |
218 | More congestion control algorithms may be available as modules, | |
219 | but not loaded. | |
1da177e4 | 220 | |
71599cd1 | 221 | tcp_base_mss - INTEGER |
4edc2f34 SH |
222 | The initial value of search_low to be used by the packetization layer |
223 | Path MTU discovery (MTU probing). If MTU probing is enabled, | |
224 | this is the initial MSS used by the connection. | |
71599cd1 | 225 | |
ef56e622 SH |
226 | tcp_congestion_control - STRING |
227 | Set the congestion control algorithm to be used for new | |
228 | connections. The algorithm "reno" is always available, but | |
229 | additional choices may be available based on kernel configuration. | |
230 | Default is set as part of kernel configuration. | |
d8a6e65f ED |
231 | For passive connections, the listener congestion control choice |
232 | is inherited. | |
233 | [see setsockopt(listenfd, SOL_TCP, TCP_CONGESTION, "name" ...) ] | |
1da177e4 | 234 | |
ef56e622 SH |
235 | tcp_dsack - BOOLEAN |
236 | Allows TCP to send "duplicate" SACKs. | |
1da177e4 | 237 | |
eed530b6 YC |
238 | tcp_early_retrans - INTEGER |
239 | Enable Early Retransmit (ER), per RFC 5827. ER lowers the threshold | |
240 | for triggering fast retransmit when the amount of outstanding data is | |
241 | small and when no previously unsent data can be transmitted (such | |
6ba8a3b1 | 242 | that limited transmit could be used). Also controls the use of |
3dd17ede | 243 | Tail loss probe (TLP) that converts RTOs occurring due to tail |
6ba8a3b1 | 244 | losses into fast recovery (draft-dukkipati-tcpm-tcp-loss-probe-01). |
eed530b6 YC |
245 | Possible values: |
246 | 0 disables ER | |
247 | 1 enables ER | |
248 | 2 enables ER but delays fast recovery and fast retransmit | |
249 | by a fourth of RTT. This mitigates connection falsely | |
250 | recovers when network has a small degree of reordering | |
251 | (less than 3 packets). | |
6ba8a3b1 ND |
252 | 3 enables delayed ER and TLP. |
253 | 4 enables TLP only. | |
254 | Default: 3 | |
eed530b6 | 255 | |
34a6ef38 | 256 | tcp_ecn - INTEGER |
7e3a2dc5 RJ |
257 | Control use of Explicit Congestion Notification (ECN) by TCP. |
258 | ECN is used only when both ends of the TCP connection indicate | |
259 | support for it. This feature is useful in avoiding losses due | |
260 | to congestion by allowing supporting routers to signal | |
261 | congestion before having to drop packets. | |
255cac91 | 262 | Possible values are: |
7e3a2dc5 | 263 | 0 Disable ECN. Neither initiate nor accept ECN. |
3d55b323 VS |
264 | 1 Enable ECN when requested by incoming connections and |
265 | also request ECN on outgoing connection attempts. | |
266 | 2 Enable ECN when requested by incoming connections | |
7e3a2dc5 | 267 | but do not request ECN on outgoing connections. |
255cac91 | 268 | Default: 2 |
ef56e622 SH |
269 | |
270 | tcp_fack - BOOLEAN | |
271 | Enable FACK congestion avoidance and fast retransmission. | |
272 | The value is not used, if tcp_sack is not enabled. | |
1da177e4 LT |
273 | |
274 | tcp_fin_timeout - INTEGER | |
d825da2e RJ |
275 | The length of time an orphaned (no longer referenced by any |
276 | application) connection will remain in the FIN_WAIT_2 state | |
277 | before it is aborted at the local end. While a perfectly | |
278 | valid "receive only" state for an un-orphaned connection, an | |
279 | orphaned connection in FIN_WAIT_2 state could otherwise wait | |
280 | forever for the remote to close its end of the connection. | |
281 | Cf. tcp_max_orphans | |
282 | Default: 60 seconds | |
1da177e4 | 283 | |
89808060 | 284 | tcp_frto - INTEGER |
e33099f9 | 285 | Enables Forward RTO-Recovery (F-RTO) defined in RFC5682. |
cd99889c | 286 | F-RTO is an enhanced recovery algorithm for TCP retransmission |
e33099f9 YC |
287 | timeouts. It is particularly beneficial in networks where the |
288 | RTT fluctuates (e.g., wireless). F-RTO is sender-side only | |
289 | modification. It does not require any support from the peer. | |
290 | ||
291 | By default it's enabled with a non-zero value. 0 disables F-RTO. | |
1da177e4 | 292 | |
032ee423 NC |
293 | tcp_invalid_ratelimit - INTEGER |
294 | Limit the maximal rate for sending duplicate acknowledgments | |
295 | in response to incoming TCP packets that are for an existing | |
296 | connection but that are invalid due to any of these reasons: | |
297 | ||
298 | (a) out-of-window sequence number, | |
299 | (b) out-of-window acknowledgment number, or | |
300 | (c) PAWS (Protection Against Wrapped Sequence numbers) check failure | |
301 | ||
302 | This can help mitigate simple "ack loop" DoS attacks, wherein | |
303 | a buggy or malicious middlebox or man-in-the-middle can | |
304 | rewrite TCP header fields in manner that causes each endpoint | |
305 | to think that the other is sending invalid TCP segments, thus | |
306 | causing each side to send an unterminating stream of duplicate | |
307 | acknowledgments for invalid segments. | |
308 | ||
309 | Using 0 disables rate-limiting of dupacks in response to | |
310 | invalid segments; otherwise this value specifies the minimal | |
311 | space between sending such dupacks, in milliseconds. | |
312 | ||
313 | Default: 500 (milliseconds). | |
314 | ||
ef56e622 SH |
315 | tcp_keepalive_time - INTEGER |
316 | How often TCP sends out keepalive messages when keepalive is enabled. | |
317 | Default: 2hours. | |
1da177e4 | 318 | |
ef56e622 SH |
319 | tcp_keepalive_probes - INTEGER |
320 | How many keepalive probes TCP sends out, until it decides that the | |
321 | connection is broken. Default value: 9. | |
322 | ||
323 | tcp_keepalive_intvl - INTEGER | |
324 | How frequently the probes are send out. Multiplied by | |
325 | tcp_keepalive_probes it is time to kill not responding connection, | |
326 | after probes started. Default value: 75sec i.e. connection | |
327 | will be aborted after ~11 minutes of retries. | |
328 | ||
329 | tcp_low_latency - BOOLEAN | |
330 | If set, the TCP stack makes decisions that prefer lower | |
331 | latency as opposed to higher throughput. By default, this | |
332 | option is not set meaning that higher throughput is preferred. | |
333 | An example of an application where this default should be | |
334 | changed would be a Beowulf compute cluster. | |
335 | Default: 0 | |
1da177e4 LT |
336 | |
337 | tcp_max_orphans - INTEGER | |
338 | Maximal number of TCP sockets not attached to any user file handle, | |
339 | held by system. If this number is exceeded orphaned connections are | |
340 | reset immediately and warning is printed. This limit exists | |
341 | only to prevent simple DoS attacks, you _must_ not rely on this | |
342 | or lower the limit artificially, but rather increase it | |
343 | (probably, after increasing installed memory), | |
344 | if network conditions require more than default value, | |
345 | and tune network services to linger and kill such states | |
346 | more aggressively. Let me to remind again: each orphan eats | |
347 | up to ~64K of unswappable memory. | |
348 | ||
1da177e4 | 349 | tcp_max_syn_backlog - INTEGER |
99b53bdd PP |
350 | Maximal number of remembered connection requests, which have not |
351 | received an acknowledgment from connecting client. | |
352 | The minimal value is 128 for low memory machines, and it will | |
353 | increase in proportion to the memory of machine. | |
354 | If server suffers from overload, try increasing this number. | |
1da177e4 | 355 | |
ef56e622 SH |
356 | tcp_max_tw_buckets - INTEGER |
357 | Maximal number of timewait sockets held by system simultaneously. | |
358 | If this number is exceeded time-wait socket is immediately destroyed | |
359 | and warning is printed. This limit exists only to prevent | |
360 | simple DoS attacks, you _must_ not lower the limit artificially, | |
361 | but rather increase it (probably, after increasing installed memory), | |
362 | if network conditions require more than default value. | |
1da177e4 | 363 | |
ef56e622 SH |
364 | tcp_mem - vector of 3 INTEGERs: min, pressure, max |
365 | min: below this number of pages TCP is not bothered about its | |
366 | memory appetite. | |
1da177e4 | 367 | |
ef56e622 SH |
368 | pressure: when amount of memory allocated by TCP exceeds this number |
369 | of pages, TCP moderates its memory consumption and enters memory | |
370 | pressure mode, which is exited when memory consumption falls | |
371 | under "min". | |
1da177e4 | 372 | |
ef56e622 | 373 | max: number of pages allowed for queueing by all TCP sockets. |
1da177e4 | 374 | |
ef56e622 SH |
375 | Defaults are calculated at boot time from amount of available |
376 | memory. | |
1da177e4 | 377 | |
71599cd1 | 378 | tcp_moderate_rcvbuf - BOOLEAN |
4edc2f34 | 379 | If set, TCP performs receive buffer auto-tuning, attempting to |
71599cd1 JH |
380 | automatically size the buffer (no greater than tcp_rmem[2]) to |
381 | match the size required by the path for full throughput. Enabled by | |
382 | default. | |
383 | ||
384 | tcp_mtu_probing - INTEGER | |
385 | Controls TCP Packetization-Layer Path MTU Discovery. Takes three | |
386 | values: | |
387 | 0 - Disabled | |
388 | 1 - Disabled by default, enabled when an ICMP black hole detected | |
389 | 2 - Always enabled, use initial MSS of tcp_base_mss. | |
390 | ||
fab42760 FD |
391 | tcp_probe_interval - INTEGER |
392 | Controls how often to start TCP Packetization-Layer Path MTU | |
393 | Discovery reprobe. The default is reprobing every 10 minutes as | |
394 | per RFC4821. | |
395 | ||
396 | tcp_probe_threshold - INTEGER | |
397 | Controls when TCP Packetization-Layer Path MTU Discovery probing | |
398 | will stop in respect to the width of search range in bytes. Default | |
399 | is 8 bytes. | |
400 | ||
71599cd1 JH |
401 | tcp_no_metrics_save - BOOLEAN |
402 | By default, TCP saves various connection metrics in the route cache | |
403 | when the connection closes, so that connections established in the | |
404 | near future can use these to set initial conditions. Usually, this | |
405 | increases overall performance, but may sometimes cause performance | |
0f035b8e | 406 | degradation. If set, TCP will not cache metrics on closing |
71599cd1 JH |
407 | connections. |
408 | ||
ef56e622 | 409 | tcp_orphan_retries - INTEGER |
5d789229 DL |
410 | This value influences the timeout of a locally closed TCP connection, |
411 | when RTO retransmissions remain unacknowledged. | |
412 | See tcp_retries2 for more details. | |
413 | ||
06b8fc5d | 414 | The default value is 8. |
5d789229 | 415 | If your machine is a loaded WEB server, |
ef56e622 SH |
416 | you should think about lowering this value, such sockets |
417 | may consume significant resources. Cf. tcp_max_orphans. | |
1da177e4 LT |
418 | |
419 | tcp_reordering - INTEGER | |
dca145ff ED |
420 | Initial reordering level of packets in a TCP stream. |
421 | TCP stack can then dynamically adjust flow reordering level | |
422 | between this initial value and tcp_max_reordering | |
e18f5feb | 423 | Default: 3 |
1da177e4 | 424 | |
dca145ff ED |
425 | tcp_max_reordering - INTEGER |
426 | Maximal reordering level of packets in a TCP stream. | |
427 | 300 is a fairly conservative value, but you might increase it | |
428 | if paths are using per packet load balancing (like bonding rr mode) | |
429 | Default: 300 | |
430 | ||
1da177e4 LT |
431 | tcp_retrans_collapse - BOOLEAN |
432 | Bug-to-bug compatibility with some broken printers. | |
433 | On retransmit try to send bigger packets to work around bugs in | |
434 | certain TCP stacks. | |
435 | ||
ef56e622 | 436 | tcp_retries1 - INTEGER |
5d789229 DL |
437 | This value influences the time, after which TCP decides, that |
438 | something is wrong due to unacknowledged RTO retransmissions, | |
439 | and reports this suspicion to the network layer. | |
440 | See tcp_retries2 for more details. | |
441 | ||
442 | RFC 1122 recommends at least 3 retransmissions, which is the | |
443 | default. | |
1da177e4 | 444 | |
ef56e622 | 445 | tcp_retries2 - INTEGER |
5d789229 DL |
446 | This value influences the timeout of an alive TCP connection, |
447 | when RTO retransmissions remain unacknowledged. | |
448 | Given a value of N, a hypothetical TCP connection following | |
449 | exponential backoff with an initial RTO of TCP_RTO_MIN would | |
450 | retransmit N times before killing the connection at the (N+1)th RTO. | |
451 | ||
452 | The default value of 15 yields a hypothetical timeout of 924.6 | |
453 | seconds and is a lower bound for the effective timeout. | |
454 | TCP will effectively time out at the first RTO which exceeds the | |
455 | hypothetical timeout. | |
456 | ||
457 | RFC 1122 recommends at least 100 seconds for the timeout, | |
458 | which corresponds to a value of at least 8. | |
1da177e4 | 459 | |
ef56e622 SH |
460 | tcp_rfc1337 - BOOLEAN |
461 | If set, the TCP stack behaves conforming to RFC1337. If unset, | |
462 | we are not conforming to RFC, but prevent TCP TIME_WAIT | |
463 | assassination. | |
464 | Default: 0 | |
1da177e4 LT |
465 | |
466 | tcp_rmem - vector of 3 INTEGERs: min, default, max | |
467 | min: Minimal size of receive buffer used by TCP sockets. | |
468 | It is guaranteed to each TCP socket, even under moderate memory | |
469 | pressure. | |
6539fefd | 470 | Default: 1 page |
1da177e4 | 471 | |
53025f5e | 472 | default: initial size of receive buffer used by TCP sockets. |
1da177e4 LT |
473 | This value overrides net.core.rmem_default used by other protocols. |
474 | Default: 87380 bytes. This value results in window of 65535 with | |
475 | default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit | |
476 | less for default tcp_app_win. See below about these variables. | |
477 | ||
478 | max: maximal size of receive buffer allowed for automatically | |
479 | selected receiver buffers for TCP socket. This value does not override | |
53025f5e BF |
480 | net.core.rmem_max. Calling setsockopt() with SO_RCVBUF disables |
481 | automatic tuning of that socket's receive buffer size, in which | |
482 | case this value is ignored. | |
b49960a0 | 483 | Default: between 87380B and 6MB, depending on RAM size. |
1da177e4 | 484 | |
ef56e622 SH |
485 | tcp_sack - BOOLEAN |
486 | Enable select acknowledgments (SACKS). | |
1da177e4 | 487 | |
ef56e622 SH |
488 | tcp_slow_start_after_idle - BOOLEAN |
489 | If set, provide RFC2861 behavior and time out the congestion | |
490 | window after an idle period. An idle period is defined at | |
491 | the current RTO. If unset, the congestion window will not | |
492 | be timed out after an idle period. | |
493 | Default: 1 | |
1da177e4 | 494 | |
ef56e622 | 495 | tcp_stdurg - BOOLEAN |
4edc2f34 | 496 | Use the Host requirements interpretation of the TCP urgent pointer field. |
ef56e622 SH |
497 | Most hosts use the older BSD interpretation, so if you turn this on |
498 | Linux might not communicate correctly with them. | |
499 | Default: FALSE | |
1da177e4 | 500 | |
ef56e622 SH |
501 | tcp_synack_retries - INTEGER |
502 | Number of times SYNACKs for a passive TCP connection attempt will | |
503 | be retransmitted. Should not be higher than 255. Default value | |
6c9ff979 AB |
504 | is 5, which corresponds to 31seconds till the last retransmission |
505 | with the current initial RTO of 1second. With this the final timeout | |
506 | for a passive TCP connection will happen after 63seconds. | |
1da177e4 | 507 | |
ef56e622 | 508 | tcp_syncookies - BOOLEAN |
a3c910d2 | 509 | Only valid when the kernel was compiled with CONFIG_SYN_COOKIES |
ef56e622 | 510 | Send out syncookies when the syn backlog queue of a socket |
4edc2f34 | 511 | overflows. This is to prevent against the common 'SYN flood attack' |
a3c910d2 | 512 | Default: 1 |
1da177e4 | 513 | |
ef56e622 SH |
514 | Note, that syncookies is fallback facility. |
515 | It MUST NOT be used to help highly loaded servers to stand | |
4edc2f34 | 516 | against legal connection rate. If you see SYN flood warnings |
ef56e622 SH |
517 | in your logs, but investigation shows that they occur |
518 | because of overload with legal connections, you should tune | |
519 | another parameters until this warning disappear. | |
520 | See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow. | |
1da177e4 | 521 | |
ef56e622 SH |
522 | syncookies seriously violate TCP protocol, do not allow |
523 | to use TCP extensions, can result in serious degradation | |
524 | of some services (f.e. SMTP relaying), visible not by you, | |
525 | but your clients and relays, contacting you. While you see | |
4edc2f34 | 526 | SYN flood warnings in logs not being really flooded, your server |
ef56e622 | 527 | is seriously misconfigured. |
1da177e4 | 528 | |
5ad37d5d HFS |
529 | If you want to test which effects syncookies have to your |
530 | network connections you can set this knob to 2 to enable | |
531 | unconditionally generation of syncookies. | |
532 | ||
cf60af03 YC |
533 | tcp_fastopen - INTEGER |
534 | Enable TCP Fast Open feature (draft-ietf-tcpm-fastopen) to send data | |
535 | in the opening SYN packet. To use this feature, the client application | |
10467163 JC |
536 | must use sendmsg() or sendto() with MSG_FASTOPEN flag rather than |
537 | connect() to perform a TCP handshake automatically. | |
538 | ||
539 | The values (bitmap) are | |
0d41cca4 | 540 | 1: Enables sending data in the opening SYN on the client w/ MSG_FASTOPEN. |
10467163 JC |
541 | 2: Enables TCP Fast Open on the server side, i.e., allowing data in |
542 | a SYN packet to be accepted and passed to the application before | |
543 | 3-way hand shake finishes. | |
544 | 4: Send data in the opening SYN regardless of cookie availability and | |
545 | without a cookie option. | |
546 | 0x100: Accept SYN data w/o validating the cookie. | |
547 | 0x200: Accept data-in-SYN w/o any cookie option present. | |
548 | 0x400/0x800: Enable Fast Open on all listeners regardless of the | |
549 | TCP_FASTOPEN socket option. The two different flags designate two | |
550 | different ways of setting max_qlen without the TCP_FASTOPEN socket | |
551 | option. | |
cf60af03 | 552 | |
0d41cca4 | 553 | Default: 1 |
cf60af03 | 554 | |
10467163 JC |
555 | Note that the client & server side Fast Open flags (1 and 2 |
556 | respectively) must be also enabled before the rest of flags can take | |
557 | effect. | |
558 | ||
559 | See include/net/tcp.h and the code for more details. | |
560 | ||
ef56e622 SH |
561 | tcp_syn_retries - INTEGER |
562 | Number of times initial SYNs for an active TCP connection attempt | |
563 | will be retransmitted. Should not be higher than 255. Default value | |
3b09adcb | 564 | is 6, which corresponds to 63seconds till the last retransmission |
6c9ff979 AB |
565 | with the current initial RTO of 1second. With this the final timeout |
566 | for an active TCP connection attempt will happen after 127seconds. | |
ef56e622 SH |
567 | |
568 | tcp_timestamps - BOOLEAN | |
569 | Enable timestamps as defined in RFC1323. | |
1da177e4 | 570 | |
95bd09eb ED |
571 | tcp_min_tso_segs - INTEGER |
572 | Minimal number of segments per TSO frame. | |
573 | Since linux-3.12, TCP does an automatic sizing of TSO frames, | |
574 | depending on flow rate, instead of filling 64Kbytes packets. | |
575 | For specific usages, it's possible to force TCP to build big | |
576 | TSO frames. Note that TCP stack might split too big TSO packets | |
577 | if available window is too small. | |
578 | Default: 2 | |
579 | ||
1da177e4 | 580 | tcp_tso_win_divisor - INTEGER |
ef56e622 SH |
581 | This allows control over what percentage of the congestion window |
582 | can be consumed by a single TSO frame. | |
583 | The setting of this parameter is a choice between burstiness and | |
584 | building larger TSO frames. | |
585 | Default: 3 | |
1da177e4 | 586 | |
ef56e622 SH |
587 | tcp_tw_recycle - BOOLEAN |
588 | Enable fast recycling TIME-WAIT sockets. Default value is 0. | |
589 | It should not be changed without advice/request of technical | |
590 | experts. | |
1da177e4 | 591 | |
ef56e622 SH |
592 | tcp_tw_reuse - BOOLEAN |
593 | Allow to reuse TIME-WAIT sockets for new connections when it is | |
594 | safe from protocol viewpoint. Default value is 0. | |
595 | It should not be changed without advice/request of technical | |
596 | experts. | |
ce7bc3bf | 597 | |
ef56e622 SH |
598 | tcp_window_scaling - BOOLEAN |
599 | Enable window scaling as defined in RFC1323. | |
3ff825b2 | 600 | |
ef56e622 | 601 | tcp_wmem - vector of 3 INTEGERs: min, default, max |
53025f5e | 602 | min: Amount of memory reserved for send buffers for TCP sockets. |
ef56e622 | 603 | Each TCP socket has rights to use it due to fact of its birth. |
6539fefd | 604 | Default: 1 page |
9d7bcfc6 | 605 | |
53025f5e BF |
606 | default: initial size of send buffer used by TCP sockets. This |
607 | value overrides net.core.wmem_default used by other protocols. | |
608 | It is usually lower than net.core.wmem_default. | |
ef56e622 SH |
609 | Default: 16K |
610 | ||
53025f5e BF |
611 | max: Maximal amount of memory allowed for automatically tuned |
612 | send buffers for TCP sockets. This value does not override | |
613 | net.core.wmem_max. Calling setsockopt() with SO_SNDBUF disables | |
614 | automatic tuning of that socket's send buffer size, in which case | |
615 | this value is ignored. | |
616 | Default: between 64K and 4MB, depending on RAM size. | |
1da177e4 | 617 | |
c9bee3b7 ED |
618 | tcp_notsent_lowat - UNSIGNED INTEGER |
619 | A TCP socket can control the amount of unsent bytes in its write queue, | |
620 | thanks to TCP_NOTSENT_LOWAT socket option. poll()/select()/epoll() | |
621 | reports POLLOUT events if the amount of unsent bytes is below a per | |
622 | socket value, and if the write queue is not full. sendmsg() will | |
623 | also not add new buffers if the limit is hit. | |
624 | ||
625 | This global variable controls the amount of unsent data for | |
626 | sockets not using TCP_NOTSENT_LOWAT. For these sockets, a change | |
627 | to the global variable has immediate effect. | |
628 | ||
629 | Default: UINT_MAX (0xFFFFFFFF) | |
630 | ||
15d99e02 RJ |
631 | tcp_workaround_signed_windows - BOOLEAN |
632 | If set, assume no receipt of a window scaling option means the | |
633 | remote TCP is broken and treats the window as a signed quantity. | |
634 | If unset, assume the remote TCP is not broken even if we do | |
635 | not receive a window scaling option from them. | |
636 | Default: 0 | |
637 | ||
36e31b0a AP |
638 | tcp_thin_linear_timeouts - BOOLEAN |
639 | Enable dynamic triggering of linear timeouts for thin streams. | |
640 | If set, a check is performed upon retransmission by timeout to | |
641 | determine if the stream is thin (less than 4 packets in flight). | |
642 | As long as the stream is found to be thin, up to 6 linear | |
643 | timeouts may be performed before exponential backoff mode is | |
644 | initiated. This improves retransmission latency for | |
645 | non-aggressive thin streams, often found to be time-dependent. | |
646 | For more information on thin streams, see | |
647 | Documentation/networking/tcp-thin.txt | |
648 | Default: 0 | |
649 | ||
7e380175 AP |
650 | tcp_thin_dupack - BOOLEAN |
651 | Enable dynamic triggering of retransmissions after one dupACK | |
652 | for thin streams. If set, a check is performed upon reception | |
653 | of a dupACK to determine if the stream is thin (less than 4 | |
654 | packets in flight). As long as the stream is found to be thin, | |
655 | data is retransmitted on the first received dupACK. This | |
656 | improves retransmission latency for non-aggressive thin | |
657 | streams, often found to be time-dependent. | |
658 | For more information on thin streams, see | |
659 | Documentation/networking/tcp-thin.txt | |
660 | Default: 0 | |
661 | ||
46d3ceab ED |
662 | tcp_limit_output_bytes - INTEGER |
663 | Controls TCP Small Queue limit per tcp socket. | |
664 | TCP bulk sender tends to increase packets in flight until it | |
665 | gets losses notifications. With SNDBUF autotuning, this can | |
666 | result in a large amount of packets queued in qdisc/device | |
667 | on the local machine, hurting latency of other flows, for | |
668 | typical pfifo_fast qdiscs. | |
669 | tcp_limit_output_bytes limits the number of bytes on qdisc | |
670 | or device to reduce artificial RTT/cwnd and reduce bufferbloat. | |
46d3ceab ED |
671 | Default: 131072 |
672 | ||
282f23c6 ED |
673 | tcp_challenge_ack_limit - INTEGER |
674 | Limits number of Challenge ACK sent per second, as recommended | |
675 | in RFC 5961 (Improving TCP's Robustness to Blind In-Window Attacks) | |
676 | Default: 100 | |
677 | ||
95766fff HA |
678 | UDP variables: |
679 | ||
680 | udp_mem - vector of 3 INTEGERs: min, pressure, max | |
681 | Number of pages allowed for queueing by all UDP sockets. | |
682 | ||
683 | min: Below this number of pages UDP is not bothered about its | |
684 | memory appetite. When amount of memory allocated by UDP exceeds | |
685 | this number, UDP starts to moderate memory usage. | |
686 | ||
687 | pressure: This value was introduced to follow format of tcp_mem. | |
688 | ||
689 | max: Number of pages allowed for queueing by all UDP sockets. | |
690 | ||
691 | Default is calculated at boot time from amount of available memory. | |
692 | ||
693 | udp_rmem_min - INTEGER | |
694 | Minimal size of receive buffer used by UDP sockets in moderation. | |
695 | Each UDP socket is able to use the size for receiving data, even if | |
696 | total pages of UDP sockets exceed udp_mem pressure. The unit is byte. | |
6539fefd | 697 | Default: 1 page |
95766fff HA |
698 | |
699 | udp_wmem_min - INTEGER | |
700 | Minimal size of send buffer used by UDP sockets in moderation. | |
701 | Each UDP socket is able to use the size for sending data, even if | |
702 | total pages of UDP sockets exceed udp_mem pressure. The unit is byte. | |
6539fefd | 703 | Default: 1 page |
95766fff | 704 | |
8802f616 PM |
705 | CIPSOv4 Variables: |
706 | ||
707 | cipso_cache_enable - BOOLEAN | |
708 | If set, enable additions to and lookups from the CIPSO label mapping | |
709 | cache. If unset, additions are ignored and lookups always result in a | |
710 | miss. However, regardless of the setting the cache is still | |
711 | invalidated when required when means you can safely toggle this on and | |
712 | off and the cache will always be "safe". | |
713 | Default: 1 | |
714 | ||
715 | cipso_cache_bucket_size - INTEGER | |
716 | The CIPSO label cache consists of a fixed size hash table with each | |
717 | hash bucket containing a number of cache entries. This variable limits | |
718 | the number of entries in each hash bucket; the larger the value the | |
719 | more CIPSO label mappings that can be cached. When the number of | |
720 | entries in a given hash bucket reaches this limit adding new entries | |
721 | causes the oldest entry in the bucket to be removed to make room. | |
722 | Default: 10 | |
723 | ||
724 | cipso_rbm_optfmt - BOOLEAN | |
725 | Enable the "Optimized Tag 1 Format" as defined in section 3.4.2.6 of | |
726 | the CIPSO draft specification (see Documentation/netlabel for details). | |
727 | This means that when set the CIPSO tag will be padded with empty | |
728 | categories in order to make the packet data 32-bit aligned. | |
729 | Default: 0 | |
730 | ||
731 | cipso_rbm_structvalid - BOOLEAN | |
732 | If set, do a very strict check of the CIPSO option when | |
733 | ip_options_compile() is called. If unset, relax the checks done during | |
734 | ip_options_compile(). Either way is "safe" as errors are caught else | |
735 | where in the CIPSO processing code but setting this to 0 (False) should | |
736 | result in less work (i.e. it should be faster) but could cause problems | |
737 | with other implementations that require strict checking. | |
738 | Default: 0 | |
739 | ||
1da177e4 LT |
740 | IP Variables: |
741 | ||
742 | ip_local_port_range - 2 INTEGERS | |
743 | Defines the local port range that is used by TCP and UDP to | |
e18f5feb | 744 | choose the local port. The first number is the first, the |
5d6bd861 FLVC |
745 | second the last local port number. The default values are |
746 | 32768 and 61000 respectively. | |
1da177e4 | 747 | |
e3826f1e AW |
748 | ip_local_reserved_ports - list of comma separated ranges |
749 | Specify the ports which are reserved for known third-party | |
750 | applications. These ports will not be used by automatic port | |
751 | assignments (e.g. when calling connect() or bind() with port | |
752 | number 0). Explicit port allocation behavior is unchanged. | |
753 | ||
754 | The format used for both input and output is a comma separated | |
755 | list of ranges (e.g. "1,2-4,10-10" for ports 1, 2, 3, 4 and | |
756 | 10). Writing to the file will clear all previously reserved | |
757 | ports and update the current list with the one given in the | |
758 | input. | |
759 | ||
760 | Note that ip_local_port_range and ip_local_reserved_ports | |
761 | settings are independent and both are considered by the kernel | |
762 | when determining which ports are available for automatic port | |
763 | assignments. | |
764 | ||
765 | You can reserve ports which are not in the current | |
766 | ip_local_port_range, e.g.: | |
767 | ||
768 | $ cat /proc/sys/net/ipv4/ip_local_port_range | |
769 | 32000 61000 | |
770 | $ cat /proc/sys/net/ipv4/ip_local_reserved_ports | |
771 | 8080,9148 | |
772 | ||
773 | although this is redundant. However such a setting is useful | |
774 | if later the port range is changed to a value that will | |
775 | include the reserved ports. | |
776 | ||
777 | Default: Empty | |
778 | ||
1da177e4 LT |
779 | ip_nonlocal_bind - BOOLEAN |
780 | If set, allows processes to bind() to non-local IP addresses, | |
781 | which can be quite useful - but may break some applications. | |
782 | Default: 0 | |
783 | ||
784 | ip_dynaddr - BOOLEAN | |
785 | If set non-zero, enables support for dynamic addresses. | |
786 | If set to a non-zero value larger than 1, a kernel log | |
787 | message will be printed when dynamic address rewriting | |
788 | occurs. | |
789 | Default: 0 | |
790 | ||
e3d73bce CW |
791 | ip_early_demux - BOOLEAN |
792 | Optimize input packet processing down to one demux for | |
793 | certain kinds of local sockets. Currently we only do this | |
794 | for established TCP sockets. | |
795 | ||
796 | It may add an additional cost for pure routing workloads that | |
797 | reduces overall throughput, in such case you should disable it. | |
798 | Default: 1 | |
799 | ||
1da177e4 | 800 | icmp_echo_ignore_all - BOOLEAN |
7ce31246 DM |
801 | If set non-zero, then the kernel will ignore all ICMP ECHO |
802 | requests sent to it. | |
803 | Default: 0 | |
804 | ||
1da177e4 | 805 | icmp_echo_ignore_broadcasts - BOOLEAN |
7ce31246 DM |
806 | If set non-zero, then the kernel will ignore all ICMP ECHO and |
807 | TIMESTAMP requests sent to it via broadcast/multicast. | |
808 | Default: 1 | |
1da177e4 LT |
809 | |
810 | icmp_ratelimit - INTEGER | |
811 | Limit the maximal rates for sending ICMP packets whose type matches | |
812 | icmp_ratemask (see below) to specific targets. | |
6dbf4bca SH |
813 | 0 to disable any limiting, |
814 | otherwise the minimal space between responses in milliseconds. | |
4cdf507d ED |
815 | Note that another sysctl, icmp_msgs_per_sec limits the number |
816 | of ICMP packets sent on all targets. | |
6dbf4bca | 817 | Default: 1000 |
1da177e4 | 818 | |
4cdf507d ED |
819 | icmp_msgs_per_sec - INTEGER |
820 | Limit maximal number of ICMP packets sent per second from this host. | |
821 | Only messages whose type matches icmp_ratemask (see below) are | |
822 | controlled by this limit. | |
6dbf4bca | 823 | Default: 1000 |
1da177e4 | 824 | |
4cdf507d ED |
825 | icmp_msgs_burst - INTEGER |
826 | icmp_msgs_per_sec controls number of ICMP packets sent per second, | |
827 | while icmp_msgs_burst controls the burst size of these packets. | |
828 | Default: 50 | |
829 | ||
1da177e4 LT |
830 | icmp_ratemask - INTEGER |
831 | Mask made of ICMP types for which rates are being limited. | |
832 | Significant bits: IHGFEDCBA9876543210 | |
833 | Default mask: 0000001100000011000 (6168) | |
834 | ||
835 | Bit definitions (see include/linux/icmp.h): | |
836 | 0 Echo Reply | |
837 | 3 Destination Unreachable * | |
838 | 4 Source Quench * | |
839 | 5 Redirect | |
840 | 8 Echo Request | |
841 | B Time Exceeded * | |
842 | C Parameter Problem * | |
843 | D Timestamp Request | |
844 | E Timestamp Reply | |
845 | F Info Request | |
846 | G Info Reply | |
847 | H Address Mask Request | |
848 | I Address Mask Reply | |
849 | ||
850 | * These are rate limited by default (see default mask above) | |
851 | ||
852 | icmp_ignore_bogus_error_responses - BOOLEAN | |
853 | Some routers violate RFC1122 by sending bogus responses to broadcast | |
854 | frames. Such violations are normally logged via a kernel warning. | |
855 | If this is set to TRUE, the kernel will not give such warnings, which | |
856 | will avoid log file clutter. | |
e8b265e8 | 857 | Default: 1 |
1da177e4 | 858 | |
95f7daf1 H |
859 | icmp_errors_use_inbound_ifaddr - BOOLEAN |
860 | ||
861 | If zero, icmp error messages are sent with the primary address of | |
862 | the exiting interface. | |
e18f5feb | 863 | |
95f7daf1 H |
864 | If non-zero, the message will be sent with the primary address of |
865 | the interface that received the packet that caused the icmp error. | |
866 | This is the behaviour network many administrators will expect from | |
867 | a router. And it can make debugging complicated network layouts | |
e18f5feb | 868 | much easier. |
95f7daf1 H |
869 | |
870 | Note that if no primary address exists for the interface selected, | |
871 | then the primary address of the first non-loopback interface that | |
d6bc8ac9 | 872 | has one will be used regardless of this setting. |
95f7daf1 H |
873 | |
874 | Default: 0 | |
875 | ||
1da177e4 LT |
876 | igmp_max_memberships - INTEGER |
877 | Change the maximum number of multicast groups we can subscribe to. | |
878 | Default: 20 | |
879 | ||
d67ef35f JE |
880 | Theoretical maximum value is bounded by having to send a membership |
881 | report in a single datagram (i.e. the report can't span multiple | |
882 | datagrams, or risk confusing the switch and leaving groups you don't | |
883 | intend to). | |
1da177e4 | 884 | |
d67ef35f JE |
885 | The number of supported groups 'M' is bounded by the number of group |
886 | report entries you can fit into a single datagram of 65535 bytes. | |
887 | ||
888 | M = 65536-sizeof (ip header)/(sizeof(Group record)) | |
889 | ||
890 | Group records are variable length, with a minimum of 12 bytes. | |
891 | So net.ipv4.igmp_max_memberships should not be set higher than: | |
892 | ||
893 | (65536-24) / 12 = 5459 | |
894 | ||
895 | The value 5459 assumes no IP header options, so in practice | |
896 | this number may be lower. | |
897 | ||
898 | conf/interface/* changes special settings per interface (where | |
899 | "interface" is the name of your network interface) | |
900 | ||
901 | conf/all/* is special, changes the settings for all interfaces | |
1da177e4 | 902 | |
a9fe8e29 HFS |
903 | igmp_qrv - INTEGER |
904 | Controls the IGMP query robustness variable (see RFC2236 8.1). | |
905 | Default: 2 (as specified by RFC2236 8.1) | |
906 | Minimum: 1 (as specified by RFC6636 4.5) | |
907 | ||
1da177e4 LT |
908 | log_martians - BOOLEAN |
909 | Log packets with impossible addresses to kernel log. | |
910 | log_martians for the interface will be enabled if at least one of | |
911 | conf/{all,interface}/log_martians is set to TRUE, | |
912 | it will be disabled otherwise | |
913 | ||
914 | accept_redirects - BOOLEAN | |
915 | Accept ICMP redirect messages. | |
916 | accept_redirects for the interface will be enabled if: | |
e18f5feb JDB |
917 | - both conf/{all,interface}/accept_redirects are TRUE in the case |
918 | forwarding for the interface is enabled | |
1da177e4 | 919 | or |
e18f5feb JDB |
920 | - at least one of conf/{all,interface}/accept_redirects is TRUE in the |
921 | case forwarding for the interface is disabled | |
1da177e4 LT |
922 | accept_redirects for the interface will be disabled otherwise |
923 | default TRUE (host) | |
924 | FALSE (router) | |
925 | ||
926 | forwarding - BOOLEAN | |
927 | Enable IP forwarding on this interface. | |
928 | ||
929 | mc_forwarding - BOOLEAN | |
930 | Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE | |
931 | and a multicast routing daemon is required. | |
e18f5feb JDB |
932 | conf/all/mc_forwarding must also be set to TRUE to enable multicast |
933 | routing for the interface | |
1da177e4 LT |
934 | |
935 | medium_id - INTEGER | |
936 | Integer value used to differentiate the devices by the medium they | |
937 | are attached to. Two devices can have different id values when | |
938 | the broadcast packets are received only on one of them. | |
939 | The default value 0 means that the device is the only interface | |
940 | to its medium, value of -1 means that medium is not known. | |
e18f5feb | 941 | |
1da177e4 LT |
942 | Currently, it is used to change the proxy_arp behavior: |
943 | the proxy_arp feature is enabled for packets forwarded between | |
944 | two devices attached to different media. | |
945 | ||
946 | proxy_arp - BOOLEAN | |
947 | Do proxy arp. | |
948 | proxy_arp for the interface will be enabled if at least one of | |
949 | conf/{all,interface}/proxy_arp is set to TRUE, | |
950 | it will be disabled otherwise | |
951 | ||
65324144 JDB |
952 | proxy_arp_pvlan - BOOLEAN |
953 | Private VLAN proxy arp. | |
954 | Basically allow proxy arp replies back to the same interface | |
955 | (from which the ARP request/solicitation was received). | |
956 | ||
957 | This is done to support (ethernet) switch features, like RFC | |
958 | 3069, where the individual ports are NOT allowed to | |
959 | communicate with each other, but they are allowed to talk to | |
960 | the upstream router. As described in RFC 3069, it is possible | |
961 | to allow these hosts to communicate through the upstream | |
962 | router by proxy_arp'ing. Don't need to be used together with | |
963 | proxy_arp. | |
964 | ||
965 | This technology is known by different names: | |
966 | In RFC 3069 it is called VLAN Aggregation. | |
967 | Cisco and Allied Telesyn call it Private VLAN. | |
968 | Hewlett-Packard call it Source-Port filtering or port-isolation. | |
969 | Ericsson call it MAC-Forced Forwarding (RFC Draft). | |
970 | ||
1da177e4 LT |
971 | shared_media - BOOLEAN |
972 | Send(router) or accept(host) RFC1620 shared media redirects. | |
973 | Overrides ip_secure_redirects. | |
974 | shared_media for the interface will be enabled if at least one of | |
975 | conf/{all,interface}/shared_media is set to TRUE, | |
976 | it will be disabled otherwise | |
977 | default TRUE | |
978 | ||
979 | secure_redirects - BOOLEAN | |
980 | Accept ICMP redirect messages only for gateways, | |
981 | listed in default gateway list. | |
982 | secure_redirects for the interface will be enabled if at least one of | |
983 | conf/{all,interface}/secure_redirects is set to TRUE, | |
984 | it will be disabled otherwise | |
985 | default TRUE | |
986 | ||
987 | send_redirects - BOOLEAN | |
988 | Send redirects, if router. | |
989 | send_redirects for the interface will be enabled if at least one of | |
990 | conf/{all,interface}/send_redirects is set to TRUE, | |
991 | it will be disabled otherwise | |
992 | Default: TRUE | |
993 | ||
994 | bootp_relay - BOOLEAN | |
995 | Accept packets with source address 0.b.c.d destined | |
996 | not to this host as local ones. It is supposed, that | |
997 | BOOTP relay daemon will catch and forward such packets. | |
998 | conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay | |
999 | for the interface | |
1000 | default FALSE | |
1001 | Not Implemented Yet. | |
1002 | ||
1003 | accept_source_route - BOOLEAN | |
1004 | Accept packets with SRR option. | |
1005 | conf/all/accept_source_route must also be set to TRUE to accept packets | |
1006 | with SRR option on the interface | |
1007 | default TRUE (router) | |
1008 | FALSE (host) | |
1009 | ||
8153a10c | 1010 | accept_local - BOOLEAN |
72b126a4 SB |
1011 | Accept packets with local source addresses. In combination with |
1012 | suitable routing, this can be used to direct packets between two | |
1013 | local interfaces over the wire and have them accepted properly. | |
8153a10c PM |
1014 | default FALSE |
1015 | ||
d0daebc3 TG |
1016 | route_localnet - BOOLEAN |
1017 | Do not consider loopback addresses as martian source or destination | |
1018 | while routing. This enables the use of 127/8 for local routing purposes. | |
1019 | default FALSE | |
1020 | ||
c1cf8422 | 1021 | rp_filter - INTEGER |
1da177e4 | 1022 | 0 - No source validation. |
c1cf8422 SH |
1023 | 1 - Strict mode as defined in RFC3704 Strict Reverse Path |
1024 | Each incoming packet is tested against the FIB and if the interface | |
1025 | is not the best reverse path the packet check will fail. | |
1026 | By default failed packets are discarded. | |
1027 | 2 - Loose mode as defined in RFC3704 Loose Reverse Path | |
1028 | Each incoming packet's source address is also tested against the FIB | |
1029 | and if the source address is not reachable via any interface | |
1030 | the packet check will fail. | |
1031 | ||
e18f5feb | 1032 | Current recommended practice in RFC3704 is to enable strict mode |
bf869c30 | 1033 | to prevent IP spoofing from DDos attacks. If using asymmetric routing |
e18f5feb | 1034 | or other complicated routing, then loose mode is recommended. |
c1cf8422 | 1035 | |
1f5865e7 SW |
1036 | The max value from conf/{all,interface}/rp_filter is used |
1037 | when doing source validation on the {interface}. | |
1da177e4 LT |
1038 | |
1039 | Default value is 0. Note that some distributions enable it | |
1040 | in startup scripts. | |
1041 | ||
1042 | arp_filter - BOOLEAN | |
1043 | 1 - Allows you to have multiple network interfaces on the same | |
1044 | subnet, and have the ARPs for each interface be answered | |
1045 | based on whether or not the kernel would route a packet from | |
1046 | the ARP'd IP out that interface (therefore you must use source | |
1047 | based routing for this to work). In other words it allows control | |
1048 | of which cards (usually 1) will respond to an arp request. | |
1049 | ||
1050 | 0 - (default) The kernel can respond to arp requests with addresses | |
1051 | from other interfaces. This may seem wrong but it usually makes | |
1052 | sense, because it increases the chance of successful communication. | |
1053 | IP addresses are owned by the complete host on Linux, not by | |
1054 | particular interfaces. Only for more complex setups like load- | |
1055 | balancing, does this behaviour cause problems. | |
1056 | ||
1057 | arp_filter for the interface will be enabled if at least one of | |
1058 | conf/{all,interface}/arp_filter is set to TRUE, | |
1059 | it will be disabled otherwise | |
1060 | ||
1061 | arp_announce - INTEGER | |
1062 | Define different restriction levels for announcing the local | |
1063 | source IP address from IP packets in ARP requests sent on | |
1064 | interface: | |
1065 | 0 - (default) Use any local address, configured on any interface | |
1066 | 1 - Try to avoid local addresses that are not in the target's | |
1067 | subnet for this interface. This mode is useful when target | |
1068 | hosts reachable via this interface require the source IP | |
1069 | address in ARP requests to be part of their logical network | |
1070 | configured on the receiving interface. When we generate the | |
1071 | request we will check all our subnets that include the | |
1072 | target IP and will preserve the source address if it is from | |
1073 | such subnet. If there is no such subnet we select source | |
1074 | address according to the rules for level 2. | |
1075 | 2 - Always use the best local address for this target. | |
1076 | In this mode we ignore the source address in the IP packet | |
1077 | and try to select local address that we prefer for talks with | |
1078 | the target host. Such local address is selected by looking | |
1079 | for primary IP addresses on all our subnets on the outgoing | |
1080 | interface that include the target IP address. If no suitable | |
1081 | local address is found we select the first local address | |
1082 | we have on the outgoing interface or on all other interfaces, | |
1083 | with the hope we will receive reply for our request and | |
1084 | even sometimes no matter the source IP address we announce. | |
1085 | ||
1086 | The max value from conf/{all,interface}/arp_announce is used. | |
1087 | ||
1088 | Increasing the restriction level gives more chance for | |
1089 | receiving answer from the resolved target while decreasing | |
1090 | the level announces more valid sender's information. | |
1091 | ||
1092 | arp_ignore - INTEGER | |
1093 | Define different modes for sending replies in response to | |
1094 | received ARP requests that resolve local target IP addresses: | |
1095 | 0 - (default): reply for any local target IP address, configured | |
1096 | on any interface | |
1097 | 1 - reply only if the target IP address is local address | |
1098 | configured on the incoming interface | |
1099 | 2 - reply only if the target IP address is local address | |
1100 | configured on the incoming interface and both with the | |
1101 | sender's IP address are part from same subnet on this interface | |
1102 | 3 - do not reply for local addresses configured with scope host, | |
1103 | only resolutions for global and link addresses are replied | |
1104 | 4-7 - reserved | |
1105 | 8 - do not reply for all local addresses | |
1106 | ||
1107 | The max value from conf/{all,interface}/arp_ignore is used | |
1108 | when ARP request is received on the {interface} | |
1109 | ||
eefef1cf SH |
1110 | arp_notify - BOOLEAN |
1111 | Define mode for notification of address and device changes. | |
1112 | 0 - (default): do nothing | |
3f8dc236 | 1113 | 1 - Generate gratuitous arp requests when device is brought up |
eefef1cf SH |
1114 | or hardware address changes. |
1115 | ||
c1b1bce8 | 1116 | arp_accept - BOOLEAN |
6d955180 OP |
1117 | Define behavior for gratuitous ARP frames who's IP is not |
1118 | already present in the ARP table: | |
1119 | 0 - don't create new entries in the ARP table | |
1120 | 1 - create new entries in the ARP table | |
1121 | ||
1122 | Both replies and requests type gratuitous arp will trigger the | |
1123 | ARP table to be updated, if this setting is on. | |
1124 | ||
1125 | If the ARP table already contains the IP address of the | |
1126 | gratuitous arp frame, the arp table will be updated regardless | |
1127 | if this setting is on or off. | |
1128 | ||
89c69d3c YH |
1129 | mcast_solicit - INTEGER |
1130 | The maximum number of multicast probes in INCOMPLETE state, | |
1131 | when the associated hardware address is unknown. Defaults | |
1132 | to 3. | |
1133 | ||
1134 | ucast_solicit - INTEGER | |
1135 | The maximum number of unicast probes in PROBE state, when | |
1136 | the hardware address is being reconfirmed. Defaults to 3. | |
c1b1bce8 | 1137 | |
1da177e4 LT |
1138 | app_solicit - INTEGER |
1139 | The maximum number of probes to send to the user space ARP daemon | |
1140 | via netlink before dropping back to multicast probes (see | |
89c69d3c YH |
1141 | mcast_resolicit). Defaults to 0. |
1142 | ||
1143 | mcast_resolicit - INTEGER | |
1144 | The maximum number of multicast probes after unicast and | |
1145 | app probes in PROBE state. Defaults to 0. | |
1da177e4 LT |
1146 | |
1147 | disable_policy - BOOLEAN | |
1148 | Disable IPSEC policy (SPD) for this interface | |
1149 | ||
1150 | disable_xfrm - BOOLEAN | |
1151 | Disable IPSEC encryption on this interface, whatever the policy | |
1152 | ||
fc4eba58 HFS |
1153 | igmpv2_unsolicited_report_interval - INTEGER |
1154 | The interval in milliseconds in which the next unsolicited | |
1155 | IGMPv1 or IGMPv2 report retransmit will take place. | |
1156 | Default: 10000 (10 seconds) | |
1da177e4 | 1157 | |
fc4eba58 HFS |
1158 | igmpv3_unsolicited_report_interval - INTEGER |
1159 | The interval in milliseconds in which the next unsolicited | |
1160 | IGMPv3 report retransmit will take place. | |
1161 | Default: 1000 (1 seconds) | |
1da177e4 | 1162 | |
d922e1cb MS |
1163 | promote_secondaries - BOOLEAN |
1164 | When a primary IP address is removed from this interface | |
1165 | promote a corresponding secondary IP address instead of | |
1166 | removing all the corresponding secondary IP addresses. | |
1167 | ||
1168 | ||
1da177e4 LT |
1169 | tag - INTEGER |
1170 | Allows you to write a number, which can be used as required. | |
1171 | Default value is 0. | |
1172 | ||
1da177e4 LT |
1173 | Alexey Kuznetsov. |
1174 | kuznet@ms2.inr.ac.ru | |
1175 | ||
1176 | Updated by: | |
1177 | Andi Kleen | |
1178 | ak@muc.de | |
1179 | Nicolas Delon | |
1180 | delon.nicolas@wanadoo.fr | |
1181 | ||
1182 | ||
1183 | ||
1184 | ||
1185 | /proc/sys/net/ipv6/* Variables: | |
1186 | ||
1187 | IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also | |
1188 | apply to IPv6 [XXX?]. | |
1189 | ||
1190 | bindv6only - BOOLEAN | |
1191 | Default value for IPV6_V6ONLY socket option, | |
e18f5feb | 1192 | which restricts use of the IPv6 socket to IPv6 communication |
1da177e4 LT |
1193 | only. |
1194 | TRUE: disable IPv4-mapped address feature | |
1195 | FALSE: enable IPv4-mapped address feature | |
1196 | ||
d5c073ca | 1197 | Default: FALSE (as specified in RFC3493) |
1da177e4 | 1198 | |
6444f72b FF |
1199 | flowlabel_consistency - BOOLEAN |
1200 | Protect the consistency (and unicity) of flow label. | |
1201 | You have to disable it to use IPV6_FL_F_REFLECT flag on the | |
1202 | flow label manager. | |
1203 | TRUE: enabled | |
1204 | FALSE: disabled | |
1205 | Default: TRUE | |
1206 | ||
cb1ce2ef TH |
1207 | auto_flowlabels - BOOLEAN |
1208 | Automatically generate flow labels based based on a flow hash | |
1209 | of the packet. This allows intermediate devices, such as routers, | |
1210 | to idenfify packet flows for mechanisms like Equal Cost Multipath | |
1211 | Routing (see RFC 6438). | |
1212 | TRUE: enabled | |
1213 | FALSE: disabled | |
1214 | Default: false | |
1215 | ||
509aba3b FLB |
1216 | anycast_src_echo_reply - BOOLEAN |
1217 | Controls the use of anycast addresses as source addresses for ICMPv6 | |
1218 | echo reply | |
1219 | TRUE: enabled | |
1220 | FALSE: disabled | |
1221 | Default: FALSE | |
1222 | ||
9f0761c1 HFS |
1223 | idgen_delay - INTEGER |
1224 | Controls the delay in seconds after which time to retry | |
1225 | privacy stable address generation if a DAD conflict is | |
1226 | detected. | |
1227 | Default: 1 (as specified in RFC7217) | |
1228 | ||
1229 | idgen_retries - INTEGER | |
1230 | Controls the number of retries to generate a stable privacy | |
1231 | address if a DAD conflict is detected. | |
1232 | Default: 3 (as specified in RFC7217) | |
1233 | ||
2f711939 HFS |
1234 | mld_qrv - INTEGER |
1235 | Controls the MLD query robustness variable (see RFC3810 9.1). | |
1236 | Default: 2 (as specified by RFC3810 9.1) | |
1237 | Minimum: 1 (as specified by RFC6636 4.5) | |
1238 | ||
1da177e4 LT |
1239 | IPv6 Fragmentation: |
1240 | ||
1241 | ip6frag_high_thresh - INTEGER | |
e18f5feb | 1242 | Maximum memory used to reassemble IPv6 fragments. When |
1da177e4 LT |
1243 | ip6frag_high_thresh bytes of memory is allocated for this purpose, |
1244 | the fragment handler will toss packets until ip6frag_low_thresh | |
1245 | is reached. | |
e18f5feb | 1246 | |
1da177e4 | 1247 | ip6frag_low_thresh - INTEGER |
e18f5feb | 1248 | See ip6frag_high_thresh |
1da177e4 LT |
1249 | |
1250 | ip6frag_time - INTEGER | |
1251 | Time in seconds to keep an IPv6 fragment in memory. | |
1252 | ||
1da177e4 LT |
1253 | conf/default/*: |
1254 | Change the interface-specific default settings. | |
1255 | ||
1256 | ||
1257 | conf/all/*: | |
e18f5feb | 1258 | Change all the interface-specific settings. |
1da177e4 LT |
1259 | |
1260 | [XXX: Other special features than forwarding?] | |
1261 | ||
1262 | conf/all/forwarding - BOOLEAN | |
e18f5feb | 1263 | Enable global IPv6 forwarding between all interfaces. |
1da177e4 | 1264 | |
e18f5feb | 1265 | IPv4 and IPv6 work differently here; e.g. netfilter must be used |
1da177e4 LT |
1266 | to control which interfaces may forward packets and which not. |
1267 | ||
e18f5feb | 1268 | This also sets all interfaces' Host/Router setting |
1da177e4 LT |
1269 | 'forwarding' to the specified value. See below for details. |
1270 | ||
1271 | This referred to as global forwarding. | |
1272 | ||
fbea49e1 YH |
1273 | proxy_ndp - BOOLEAN |
1274 | Do proxy ndp. | |
1275 | ||
219b5f29 LV |
1276 | fwmark_reflect - BOOLEAN |
1277 | Controls the fwmark of kernel-generated IPv6 reply packets that are not | |
1278 | associated with a socket for example, TCP RSTs or ICMPv6 echo replies). | |
1279 | If unset, these packets have a fwmark of zero. If set, they have the | |
1280 | fwmark of the packet they are replying to. | |
1281 | Default: 0 | |
1282 | ||
1da177e4 LT |
1283 | conf/interface/*: |
1284 | Change special settings per interface. | |
1285 | ||
e18f5feb | 1286 | The functional behaviour for certain settings is different |
1da177e4 LT |
1287 | depending on whether local forwarding is enabled or not. |
1288 | ||
605b91c8 | 1289 | accept_ra - INTEGER |
1da177e4 | 1290 | Accept Router Advertisements; autoconfigure using them. |
e18f5feb | 1291 | |
026359bc TA |
1292 | It also determines whether or not to transmit Router |
1293 | Solicitations. If and only if the functional setting is to | |
1294 | accept Router Advertisements, Router Solicitations will be | |
1295 | transmitted. | |
1296 | ||
ae8abfa0 TG |
1297 | Possible values are: |
1298 | 0 Do not accept Router Advertisements. | |
1299 | 1 Accept Router Advertisements if forwarding is disabled. | |
1300 | 2 Overrule forwarding behaviour. Accept Router Advertisements | |
1301 | even if forwarding is enabled. | |
1302 | ||
1da177e4 LT |
1303 | Functional default: enabled if local forwarding is disabled. |
1304 | disabled if local forwarding is enabled. | |
1305 | ||
65f5c7c1 YH |
1306 | accept_ra_defrtr - BOOLEAN |
1307 | Learn default router in Router Advertisement. | |
1308 | ||
1309 | Functional default: enabled if accept_ra is enabled. | |
1310 | disabled if accept_ra is disabled. | |
1311 | ||
d9333196 BG |
1312 | accept_ra_from_local - BOOLEAN |
1313 | Accept RA with source-address that is found on local machine | |
1314 | if the RA is otherwise proper and able to be accepted. | |
1315 | Default is to NOT accept these as it may be an un-intended | |
1316 | network loop. | |
1317 | ||
1318 | Functional default: | |
1319 | enabled if accept_ra_from_local is enabled | |
1320 | on a specific interface. | |
1321 | disabled if accept_ra_from_local is disabled | |
1322 | on a specific interface. | |
1323 | ||
c4fd30eb | 1324 | accept_ra_pinfo - BOOLEAN |
2fe0ae78 | 1325 | Learn Prefix Information in Router Advertisement. |
c4fd30eb YH |
1326 | |
1327 | Functional default: enabled if accept_ra is enabled. | |
1328 | disabled if accept_ra is disabled. | |
1329 | ||
09c884d4 YH |
1330 | accept_ra_rt_info_max_plen - INTEGER |
1331 | Maximum prefix length of Route Information in RA. | |
1332 | ||
1333 | Route Information w/ prefix larger than or equal to this | |
1334 | variable shall be ignored. | |
1335 | ||
1336 | Functional default: 0 if accept_ra_rtr_pref is enabled. | |
1337 | -1 if accept_ra_rtr_pref is disabled. | |
1338 | ||
930d6ff2 YH |
1339 | accept_ra_rtr_pref - BOOLEAN |
1340 | Accept Router Preference in RA. | |
1341 | ||
1342 | Functional default: enabled if accept_ra is enabled. | |
1343 | disabled if accept_ra is disabled. | |
1344 | ||
c2943f14 HH |
1345 | accept_ra_mtu - BOOLEAN |
1346 | Apply the MTU value specified in RA option 5 (RFC4861). If | |
1347 | disabled, the MTU specified in the RA will be ignored. | |
1348 | ||
1349 | Functional default: enabled if accept_ra is enabled. | |
1350 | disabled if accept_ra is disabled. | |
1351 | ||
1da177e4 LT |
1352 | accept_redirects - BOOLEAN |
1353 | Accept Redirects. | |
1354 | ||
1355 | Functional default: enabled if local forwarding is disabled. | |
1356 | disabled if local forwarding is enabled. | |
1357 | ||
0bcbc926 YH |
1358 | accept_source_route - INTEGER |
1359 | Accept source routing (routing extension header). | |
1360 | ||
bb4dbf9e | 1361 | >= 0: Accept only routing header type 2. |
0bcbc926 YH |
1362 | < 0: Do not accept routing header. |
1363 | ||
1364 | Default: 0 | |
1365 | ||
1da177e4 | 1366 | autoconf - BOOLEAN |
e18f5feb | 1367 | Autoconfigure addresses using Prefix Information in Router |
1da177e4 LT |
1368 | Advertisements. |
1369 | ||
c4fd30eb YH |
1370 | Functional default: enabled if accept_ra_pinfo is enabled. |
1371 | disabled if accept_ra_pinfo is disabled. | |
1da177e4 LT |
1372 | |
1373 | dad_transmits - INTEGER | |
1374 | The amount of Duplicate Address Detection probes to send. | |
1375 | Default: 1 | |
e18f5feb | 1376 | |
605b91c8 | 1377 | forwarding - INTEGER |
e18f5feb | 1378 | Configure interface-specific Host/Router behaviour. |
1da177e4 | 1379 | |
e18f5feb | 1380 | Note: It is recommended to have the same setting on all |
1da177e4 LT |
1381 | interfaces; mixed router/host scenarios are rather uncommon. |
1382 | ||
ae8abfa0 TG |
1383 | Possible values are: |
1384 | 0 Forwarding disabled | |
1385 | 1 Forwarding enabled | |
ae8abfa0 TG |
1386 | |
1387 | FALSE (0): | |
1da177e4 LT |
1388 | |
1389 | By default, Host behaviour is assumed. This means: | |
1390 | ||
1391 | 1. IsRouter flag is not set in Neighbour Advertisements. | |
026359bc TA |
1392 | 2. If accept_ra is TRUE (default), transmit Router |
1393 | Solicitations. | |
e18f5feb | 1394 | 3. If accept_ra is TRUE (default), accept Router |
1da177e4 LT |
1395 | Advertisements (and do autoconfiguration). |
1396 | 4. If accept_redirects is TRUE (default), accept Redirects. | |
1397 | ||
ae8abfa0 | 1398 | TRUE (1): |
1da177e4 | 1399 | |
e18f5feb | 1400 | If local forwarding is enabled, Router behaviour is assumed. |
1da177e4 LT |
1401 | This means exactly the reverse from the above: |
1402 | ||
1403 | 1. IsRouter flag is set in Neighbour Advertisements. | |
026359bc | 1404 | 2. Router Solicitations are not sent unless accept_ra is 2. |
ae8abfa0 | 1405 | 3. Router Advertisements are ignored unless accept_ra is 2. |
1da177e4 LT |
1406 | 4. Redirects are ignored. |
1407 | ||
ae8abfa0 TG |
1408 | Default: 0 (disabled) if global forwarding is disabled (default), |
1409 | otherwise 1 (enabled). | |
1da177e4 LT |
1410 | |
1411 | hop_limit - INTEGER | |
1412 | Default Hop Limit to set. | |
1413 | Default: 64 | |
1414 | ||
1415 | mtu - INTEGER | |
1416 | Default Maximum Transfer Unit | |
1417 | Default: 1280 (IPv6 required minimum) | |
1418 | ||
52e16356 YH |
1419 | router_probe_interval - INTEGER |
1420 | Minimum interval (in seconds) between Router Probing described | |
1421 | in RFC4191. | |
1422 | ||
1423 | Default: 60 | |
1424 | ||
1da177e4 LT |
1425 | router_solicitation_delay - INTEGER |
1426 | Number of seconds to wait after interface is brought up | |
1427 | before sending Router Solicitations. | |
1428 | Default: 1 | |
1429 | ||
1430 | router_solicitation_interval - INTEGER | |
1431 | Number of seconds to wait between Router Solicitations. | |
1432 | Default: 4 | |
1433 | ||
1434 | router_solicitations - INTEGER | |
e18f5feb | 1435 | Number of Router Solicitations to send until assuming no |
1da177e4 LT |
1436 | routers are present. |
1437 | Default: 3 | |
1438 | ||
1439 | use_tempaddr - INTEGER | |
1440 | Preference for Privacy Extensions (RFC3041). | |
1441 | <= 0 : disable Privacy Extensions | |
1442 | == 1 : enable Privacy Extensions, but prefer public | |
1443 | addresses over temporary addresses. | |
1444 | > 1 : enable Privacy Extensions and prefer temporary | |
1445 | addresses over public addresses. | |
1446 | Default: 0 (for most devices) | |
1447 | -1 (for point-to-point devices and loopback devices) | |
1448 | ||
1449 | temp_valid_lft - INTEGER | |
1450 | valid lifetime (in seconds) for temporary addresses. | |
1451 | Default: 604800 (7 days) | |
1452 | ||
1453 | temp_prefered_lft - INTEGER | |
1454 | Preferred lifetime (in seconds) for temporary addresses. | |
1455 | Default: 86400 (1 day) | |
1456 | ||
1457 | max_desync_factor - INTEGER | |
1458 | Maximum value for DESYNC_FACTOR, which is a random value | |
e18f5feb | 1459 | that ensures that clients don't synchronize with each |
1da177e4 LT |
1460 | other and generate new addresses at exactly the same time. |
1461 | value is in seconds. | |
1462 | Default: 600 | |
e18f5feb | 1463 | |
1da177e4 LT |
1464 | regen_max_retry - INTEGER |
1465 | Number of attempts before give up attempting to generate | |
1466 | valid temporary addresses. | |
1467 | Default: 5 | |
1468 | ||
1469 | max_addresses - INTEGER | |
e79dc484 BH |
1470 | Maximum number of autoconfigured addresses per interface. Setting |
1471 | to zero disables the limitation. It is not recommended to set this | |
1472 | value too large (or to zero) because it would be an easy way to | |
1473 | crash the kernel by allowing too many addresses to be created. | |
1da177e4 LT |
1474 | Default: 16 |
1475 | ||
778d80be | 1476 | disable_ipv6 - BOOLEAN |
9bdd8d40 BH |
1477 | Disable IPv6 operation. If accept_dad is set to 2, this value |
1478 | will be dynamically set to TRUE if DAD fails for the link-local | |
1479 | address. | |
778d80be YH |
1480 | Default: FALSE (enable IPv6 operation) |
1481 | ||
56d417b1 BH |
1482 | When this value is changed from 1 to 0 (IPv6 is being enabled), |
1483 | it will dynamically create a link-local address on the given | |
1484 | interface and start Duplicate Address Detection, if necessary. | |
1485 | ||
1486 | When this value is changed from 0 to 1 (IPv6 is being disabled), | |
1487 | it will dynamically delete all address on the given interface. | |
1488 | ||
1b34be74 YH |
1489 | accept_dad - INTEGER |
1490 | Whether to accept DAD (Duplicate Address Detection). | |
1491 | 0: Disable DAD | |
1492 | 1: Enable DAD (default) | |
1493 | 2: Enable DAD, and disable IPv6 operation if MAC-based duplicate | |
1494 | link-local address has been found. | |
1495 | ||
f7734fdf OP |
1496 | force_tllao - BOOLEAN |
1497 | Enable sending the target link-layer address option even when | |
1498 | responding to a unicast neighbor solicitation. | |
1499 | Default: FALSE | |
1500 | ||
1501 | Quoting from RFC 2461, section 4.4, Target link-layer address: | |
1502 | ||
1503 | "The option MUST be included for multicast solicitations in order to | |
1504 | avoid infinite Neighbor Solicitation "recursion" when the peer node | |
1505 | does not have a cache entry to return a Neighbor Advertisements | |
1506 | message. When responding to unicast solicitations, the option can be | |
1507 | omitted since the sender of the solicitation has the correct link- | |
1508 | layer address; otherwise it would not have be able to send the unicast | |
1509 | solicitation in the first place. However, including the link-layer | |
1510 | address in this case adds little overhead and eliminates a potential | |
1511 | race condition where the sender deletes the cached link-layer address | |
1512 | prior to receiving a response to a previous solicitation." | |
1513 | ||
db2b620a HFS |
1514 | ndisc_notify - BOOLEAN |
1515 | Define mode for notification of address and device changes. | |
1516 | 0 - (default): do nothing | |
1517 | 1 - Generate unsolicited neighbour advertisements when device is brought | |
1518 | up or hardware address changes. | |
1519 | ||
fc4eba58 HFS |
1520 | mldv1_unsolicited_report_interval - INTEGER |
1521 | The interval in milliseconds in which the next unsolicited | |
1522 | MLDv1 report retransmit will take place. | |
1523 | Default: 10000 (10 seconds) | |
1524 | ||
1525 | mldv2_unsolicited_report_interval - INTEGER | |
1526 | The interval in milliseconds in which the next unsolicited | |
1527 | MLDv2 report retransmit will take place. | |
1528 | Default: 1000 (1 second) | |
1529 | ||
f2127810 DB |
1530 | force_mld_version - INTEGER |
1531 | 0 - (default) No enforcement of a MLD version, MLDv1 fallback allowed | |
1532 | 1 - Enforce to use MLD version 1 | |
1533 | 2 - Enforce to use MLD version 2 | |
1534 | ||
b800c3b9 HFS |
1535 | suppress_frag_ndisc - INTEGER |
1536 | Control RFC 6980 (Security Implications of IPv6 Fragmentation | |
1537 | with IPv6 Neighbor Discovery) behavior: | |
1538 | 1 - (default) discard fragmented neighbor discovery packets | |
1539 | 0 - allow fragmented neighbor discovery packets | |
1540 | ||
7fd2561e EK |
1541 | optimistic_dad - BOOLEAN |
1542 | Whether to perform Optimistic Duplicate Address Detection (RFC 4429). | |
1543 | 0: disabled (default) | |
1544 | 1: enabled | |
1545 | ||
1546 | use_optimistic - BOOLEAN | |
1547 | If enabled, do not classify optimistic addresses as deprecated during | |
1548 | source address selection. Preferred addresses will still be chosen | |
1549 | before optimistic addresses, subject to other ranking in the source | |
1550 | address selection algorithm. | |
1551 | 0: disabled (default) | |
1552 | 1: enabled | |
1553 | ||
9f0761c1 HFS |
1554 | stable_secret - IPv6 address |
1555 | This IPv6 address will be used as a secret to generate IPv6 | |
1556 | addresses for link-local addresses and autoconfigured | |
1557 | ones. All addresses generated after setting this secret will | |
1558 | be stable privacy ones by default. This can be changed via the | |
1559 | addrgenmode ip-link. conf/default/stable_secret is used as the | |
1560 | secret for the namespace, the interface specific ones can | |
1561 | overwrite that. Writes to conf/all/stable_secret are refused. | |
1562 | ||
1563 | It is recommended to generate this secret during installation | |
1564 | of a system and keep it stable after that. | |
1565 | ||
1566 | By default the stable secret is unset. | |
1567 | ||
1da177e4 LT |
1568 | icmp/*: |
1569 | ratelimit - INTEGER | |
1570 | Limit the maximal rates for sending ICMPv6 packets. | |
6dbf4bca SH |
1571 | 0 to disable any limiting, |
1572 | otherwise the minimal space between responses in milliseconds. | |
1573 | Default: 1000 | |
1da177e4 LT |
1574 | |
1575 | ||
1576 | IPv6 Update by: | |
1577 | Pekka Savola <pekkas@netcore.fi> | |
1578 | YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org> | |
1579 | ||
1580 | ||
1581 | /proc/sys/net/bridge/* Variables: | |
1582 | ||
1583 | bridge-nf-call-arptables - BOOLEAN | |
1584 | 1 : pass bridged ARP traffic to arptables' FORWARD chain. | |
1585 | 0 : disable this. | |
1586 | Default: 1 | |
1587 | ||
1588 | bridge-nf-call-iptables - BOOLEAN | |
1589 | 1 : pass bridged IPv4 traffic to iptables' chains. | |
1590 | 0 : disable this. | |
1591 | Default: 1 | |
1592 | ||
1593 | bridge-nf-call-ip6tables - BOOLEAN | |
1594 | 1 : pass bridged IPv6 traffic to ip6tables' chains. | |
1595 | 0 : disable this. | |
1596 | Default: 1 | |
1597 | ||
1598 | bridge-nf-filter-vlan-tagged - BOOLEAN | |
516299d2 MM |
1599 | 1 : pass bridged vlan-tagged ARP/IP/IPv6 traffic to {arp,ip,ip6}tables. |
1600 | 0 : disable this. | |
4981682c | 1601 | Default: 0 |
516299d2 MM |
1602 | |
1603 | bridge-nf-filter-pppoe-tagged - BOOLEAN | |
1604 | 1 : pass bridged pppoe-tagged IP/IPv6 traffic to {ip,ip6}tables. | |
1da177e4 | 1605 | 0 : disable this. |
4981682c | 1606 | Default: 0 |
1da177e4 | 1607 | |
4981682c PNA |
1608 | bridge-nf-pass-vlan-input-dev - BOOLEAN |
1609 | 1: if bridge-nf-filter-vlan-tagged is enabled, try to find a vlan | |
1610 | interface on the bridge and set the netfilter input device to the vlan. | |
1611 | This allows use of e.g. "iptables -i br0.1" and makes the REDIRECT | |
1612 | target work with vlan-on-top-of-bridge interfaces. When no matching | |
1613 | vlan interface is found, or this switch is off, the input device is | |
1614 | set to the bridge interface. | |
1615 | 0: disable bridge netfilter vlan interface lookup. | |
1616 | Default: 0 | |
1da177e4 | 1617 | |
32e8d494 VY |
1618 | proc/sys/net/sctp/* Variables: |
1619 | ||
1620 | addip_enable - BOOLEAN | |
1621 | Enable or disable extension of Dynamic Address Reconfiguration | |
1622 | (ADD-IP) functionality specified in RFC5061. This extension provides | |
1623 | the ability to dynamically add and remove new addresses for the SCTP | |
1624 | associations. | |
1625 | ||
1626 | 1: Enable extension. | |
1627 | ||
1628 | 0: Disable extension. | |
1629 | ||
1630 | Default: 0 | |
1631 | ||
1632 | addip_noauth_enable - BOOLEAN | |
1633 | Dynamic Address Reconfiguration (ADD-IP) requires the use of | |
1634 | authentication to protect the operations of adding or removing new | |
1635 | addresses. This requirement is mandated so that unauthorized hosts | |
1636 | would not be able to hijack associations. However, older | |
1637 | implementations may not have implemented this requirement while | |
1638 | allowing the ADD-IP extension. For reasons of interoperability, | |
1639 | we provide this variable to control the enforcement of the | |
1640 | authentication requirement. | |
1641 | ||
1642 | 1: Allow ADD-IP extension to be used without authentication. This | |
1643 | should only be set in a closed environment for interoperability | |
1644 | with older implementations. | |
1645 | ||
1646 | 0: Enforce the authentication requirement | |
1647 | ||
1648 | Default: 0 | |
1649 | ||
1650 | auth_enable - BOOLEAN | |
1651 | Enable or disable Authenticated Chunks extension. This extension | |
1652 | provides the ability to send and receive authenticated chunks and is | |
1653 | required for secure operation of Dynamic Address Reconfiguration | |
1654 | (ADD-IP) extension. | |
1655 | ||
1656 | 1: Enable this extension. | |
1657 | 0: Disable this extension. | |
1658 | ||
1659 | Default: 0 | |
1660 | ||
1661 | prsctp_enable - BOOLEAN | |
1662 | Enable or disable the Partial Reliability extension (RFC3758) which | |
1663 | is used to notify peers that a given DATA should no longer be expected. | |
1664 | ||
1665 | 1: Enable extension | |
1666 | 0: Disable | |
1667 | ||
1668 | Default: 1 | |
1669 | ||
1670 | max_burst - INTEGER | |
1671 | The limit of the number of new packets that can be initially sent. It | |
1672 | controls how bursty the generated traffic can be. | |
1673 | ||
1674 | Default: 4 | |
1675 | ||
1676 | association_max_retrans - INTEGER | |
1677 | Set the maximum number for retransmissions that an association can | |
1678 | attempt deciding that the remote end is unreachable. If this value | |
1679 | is exceeded, the association is terminated. | |
1680 | ||
1681 | Default: 10 | |
1682 | ||
1683 | max_init_retransmits - INTEGER | |
1684 | The maximum number of retransmissions of INIT and COOKIE-ECHO chunks | |
1685 | that an association will attempt before declaring the destination | |
1686 | unreachable and terminating. | |
1687 | ||
1688 | Default: 8 | |
1689 | ||
1690 | path_max_retrans - INTEGER | |
1691 | The maximum number of retransmissions that will be attempted on a given | |
1692 | path. Once this threshold is exceeded, the path is considered | |
1693 | unreachable, and new traffic will use a different path when the | |
1694 | association is multihomed. | |
1695 | ||
1696 | Default: 5 | |
1697 | ||
5aa93bcf NH |
1698 | pf_retrans - INTEGER |
1699 | The number of retransmissions that will be attempted on a given path | |
1700 | before traffic is redirected to an alternate transport (should one | |
1701 | exist). Note this is distinct from path_max_retrans, as a path that | |
1702 | passes the pf_retrans threshold can still be used. Its only | |
1703 | deprioritized when a transmission path is selected by the stack. This | |
1704 | setting is primarily used to enable fast failover mechanisms without | |
1705 | having to reduce path_max_retrans to a very low value. See: | |
1706 | http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt | |
1707 | for details. Note also that a value of pf_retrans > path_max_retrans | |
1708 | disables this feature | |
1709 | ||
1710 | Default: 0 | |
1711 | ||
32e8d494 VY |
1712 | rto_initial - INTEGER |
1713 | The initial round trip timeout value in milliseconds that will be used | |
1714 | in calculating round trip times. This is the initial time interval | |
1715 | for retransmissions. | |
1716 | ||
1717 | Default: 3000 | |
1da177e4 | 1718 | |
32e8d494 VY |
1719 | rto_max - INTEGER |
1720 | The maximum value (in milliseconds) of the round trip timeout. This | |
1721 | is the largest time interval that can elapse between retransmissions. | |
1722 | ||
1723 | Default: 60000 | |
1724 | ||
1725 | rto_min - INTEGER | |
1726 | The minimum value (in milliseconds) of the round trip timeout. This | |
1727 | is the smallest time interval the can elapse between retransmissions. | |
1728 | ||
1729 | Default: 1000 | |
1730 | ||
1731 | hb_interval - INTEGER | |
1732 | The interval (in milliseconds) between HEARTBEAT chunks. These chunks | |
1733 | are sent at the specified interval on idle paths to probe the state of | |
1734 | a given path between 2 associations. | |
1735 | ||
1736 | Default: 30000 | |
1737 | ||
1738 | sack_timeout - INTEGER | |
1739 | The amount of time (in milliseconds) that the implementation will wait | |
1740 | to send a SACK. | |
1741 | ||
1742 | Default: 200 | |
1743 | ||
1744 | valid_cookie_life - INTEGER | |
1745 | The default lifetime of the SCTP cookie (in milliseconds). The cookie | |
1746 | is used during association establishment. | |
1747 | ||
1748 | Default: 60000 | |
1749 | ||
1750 | cookie_preserve_enable - BOOLEAN | |
1751 | Enable or disable the ability to extend the lifetime of the SCTP cookie | |
1752 | that is used during the establishment phase of SCTP association | |
1753 | ||
1754 | 1: Enable cookie lifetime extension. | |
1755 | 0: Disable | |
1756 | ||
1757 | Default: 1 | |
1758 | ||
3c68198e NH |
1759 | cookie_hmac_alg - STRING |
1760 | Select the hmac algorithm used when generating the cookie value sent by | |
1761 | a listening sctp socket to a connecting client in the INIT-ACK chunk. | |
1762 | Valid values are: | |
1763 | * md5 | |
1764 | * sha1 | |
1765 | * none | |
1766 | Ability to assign md5 or sha1 as the selected alg is predicated on the | |
3b09adcb | 1767 | configuration of those algorithms at build time (CONFIG_CRYPTO_MD5 and |
3c68198e NH |
1768 | CONFIG_CRYPTO_SHA1). |
1769 | ||
1770 | Default: Dependent on configuration. MD5 if available, else SHA1 if | |
1771 | available, else none. | |
1772 | ||
32e8d494 VY |
1773 | rcvbuf_policy - INTEGER |
1774 | Determines if the receive buffer is attributed to the socket or to | |
1775 | association. SCTP supports the capability to create multiple | |
1776 | associations on a single socket. When using this capability, it is | |
1777 | possible that a single stalled association that's buffering a lot | |
1778 | of data may block other associations from delivering their data by | |
1779 | consuming all of the receive buffer space. To work around this, | |
1780 | the rcvbuf_policy could be set to attribute the receiver buffer space | |
1781 | to each association instead of the socket. This prevents the described | |
1782 | blocking. | |
1783 | ||
1784 | 1: rcvbuf space is per association | |
3b09adcb | 1785 | 0: rcvbuf space is per socket |
32e8d494 VY |
1786 | |
1787 | Default: 0 | |
1788 | ||
1789 | sndbuf_policy - INTEGER | |
1790 | Similar to rcvbuf_policy above, this applies to send buffer space. | |
1791 | ||
1792 | 1: Send buffer is tracked per association | |
1793 | 0: Send buffer is tracked per socket. | |
1794 | ||
1795 | Default: 0 | |
1796 | ||
1797 | sctp_mem - vector of 3 INTEGERs: min, pressure, max | |
1798 | Number of pages allowed for queueing by all SCTP sockets. | |
1799 | ||
1800 | min: Below this number of pages SCTP is not bothered about its | |
1801 | memory appetite. When amount of memory allocated by SCTP exceeds | |
1802 | this number, SCTP starts to moderate memory usage. | |
1803 | ||
1804 | pressure: This value was introduced to follow format of tcp_mem. | |
1805 | ||
1806 | max: Number of pages allowed for queueing by all SCTP sockets. | |
1807 | ||
1808 | Default is calculated at boot time from amount of available memory. | |
1809 | ||
1810 | sctp_rmem - vector of 3 INTEGERs: min, default, max | |
a6e1204b MM |
1811 | Only the first value ("min") is used, "default" and "max" are |
1812 | ignored. | |
1813 | ||
1814 | min: Minimal size of receive buffer used by SCTP socket. | |
1815 | It is guaranteed to each SCTP socket (but not association) even | |
1816 | under moderate memory pressure. | |
1817 | ||
1818 | Default: 1 page | |
32e8d494 VY |
1819 | |
1820 | sctp_wmem - vector of 3 INTEGERs: min, default, max | |
a6e1204b | 1821 | Currently this tunable has no effect. |
32e8d494 | 1822 | |
72388433 BD |
1823 | addr_scope_policy - INTEGER |
1824 | Control IPv4 address scoping - draft-stewart-tsvwg-sctp-ipv4-00 | |
1825 | ||
1826 | 0 - Disable IPv4 address scoping | |
1827 | 1 - Enable IPv4 address scoping | |
1828 | 2 - Follow draft but allow IPv4 private addresses | |
1829 | 3 - Follow draft but allow IPv4 link local addresses | |
1830 | ||
1831 | Default: 1 | |
1832 | ||
1da177e4 | 1833 | |
4edc2f34 | 1834 | /proc/sys/net/core/* |
c60f6aa8 | 1835 | Please see: Documentation/sysctl/net.txt for descriptions of these entries. |
705efc3b | 1836 | |
4edc2f34 SH |
1837 | |
1838 | /proc/sys/net/unix/* | |
705efc3b WT |
1839 | max_dgram_qlen - INTEGER |
1840 | The maximum length of dgram socket receive queue | |
1841 | ||
1842 | Default: 10 | |
1843 | ||
1844 | ||
1845 | UNDOCUMENTED: | |
4edc2f34 SH |
1846 | |
1847 | /proc/sys/net/irda/* | |
1848 | fast_poll_increase FIXME | |
1849 | warn_noreply_time FIXME | |
1850 | discovery_slots FIXME | |
1851 | slot_timeout FIXME | |
1852 | max_baud_rate FIXME | |
1853 | discovery_timeout FIXME | |
1854 | lap_keepalive_time FIXME | |
1855 | max_noreply_time FIXME | |
1856 | max_tx_data_size FIXME | |
1857 | max_tx_window FIXME | |
1858 | min_tx_turn_time FIXME |