]>
Commit | Line | Data |
---|---|---|
1da177e4 LT |
1 | /proc/sys/net/ipv4/* Variables: |
2 | ||
3 | ip_forward - BOOLEAN | |
4 | 0 - disabled (default) | |
5 | not 0 - enabled | |
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 | |
14 | default 64 | |
15 | ||
16 | ip_no_pmtu_disc - BOOLEAN | |
17 | Disable Path MTU Discovery. | |
18 | default FALSE | |
19 | ||
20 | min_pmtu - INTEGER | |
21 | default 562 - minimum discovered Path MTU | |
22 | ||
23 | mtu_expires - INTEGER | |
24 | Time, in seconds, that cached PMTU information is kept. | |
25 | ||
26 | min_adv_mss - INTEGER | |
27 | The advertised MSS depends on the first hop route MTU, but will | |
28 | never be lower than this setting. | |
29 | ||
30 | IP Fragmentation: | |
31 | ||
32 | ipfrag_high_thresh - INTEGER | |
33 | Maximum memory used to reassemble IP fragments. When | |
34 | ipfrag_high_thresh bytes of memory is allocated for this purpose, | |
35 | the fragment handler will toss packets until ipfrag_low_thresh | |
36 | is reached. | |
37 | ||
38 | ipfrag_low_thresh - INTEGER | |
39 | See ipfrag_high_thresh | |
40 | ||
41 | ipfrag_time - INTEGER | |
42 | Time in seconds to keep an IP fragment in memory. | |
43 | ||
44 | ipfrag_secret_interval - INTEGER | |
45 | Regeneration interval (in seconds) of the hash secret (or lifetime | |
46 | for the hash secret) for IP fragments. | |
47 | Default: 600 | |
48 | ||
49 | INET peer storage: | |
50 | ||
51 | inet_peer_threshold - INTEGER | |
52 | The approximate size of the storage. Starting from this threshold | |
53 | entries will be thrown aggressively. This threshold also determines | |
54 | entries' time-to-live and time intervals between garbage collection | |
55 | passes. More entries, less time-to-live, less GC interval. | |
56 | ||
57 | inet_peer_minttl - INTEGER | |
58 | Minimum time-to-live of entries. Should be enough to cover fragment | |
59 | time-to-live on the reassembling side. This minimum time-to-live is | |
60 | guaranteed if the pool size is less than inet_peer_threshold. | |
61 | Measured in jiffies(1). | |
62 | ||
63 | inet_peer_maxttl - INTEGER | |
64 | Maximum time-to-live of entries. Unused entries will expire after | |
65 | this period of time if there is no memory pressure on the pool (i.e. | |
66 | when the number of entries in the pool is very small). | |
67 | Measured in jiffies(1). | |
68 | ||
69 | inet_peer_gc_mintime - INTEGER | |
70 | Minimum interval between garbage collection passes. This interval is | |
71 | in effect under high memory pressure on the pool. | |
72 | Measured in jiffies(1). | |
73 | ||
74 | inet_peer_gc_maxtime - INTEGER | |
75 | Minimum interval between garbage collection passes. This interval is | |
76 | in effect under low (or absent) memory pressure on the pool. | |
77 | Measured in jiffies(1). | |
78 | ||
79 | TCP variables: | |
80 | ||
9772efb9 SH |
81 | tcp_abc - INTEGER |
82 | Controls Appropriate Byte Count defined in RFC3465. If set to | |
83 | 0 then does congestion avoid once per ack. 1 is conservative | |
84 | value, and 2 is more agressive. | |
85 | ||
1da177e4 LT |
86 | tcp_syn_retries - INTEGER |
87 | Number of times initial SYNs for an active TCP connection attempt | |
88 | will be retransmitted. Should not be higher than 255. Default value | |
89 | is 5, which corresponds to ~180seconds. | |
90 | ||
91 | tcp_synack_retries - INTEGER | |
92 | Number of times SYNACKs for a passive TCP connection attempt will | |
93 | be retransmitted. Should not be higher than 255. Default value | |
94 | is 5, which corresponds to ~180seconds. | |
95 | ||
96 | tcp_keepalive_time - INTEGER | |
97 | How often TCP sends out keepalive messages when keepalive is enabled. | |
98 | Default: 2hours. | |
99 | ||
100 | tcp_keepalive_probes - INTEGER | |
101 | How many keepalive probes TCP sends out, until it decides that the | |
102 | connection is broken. Default value: 9. | |
103 | ||
104 | tcp_keepalive_intvl - INTEGER | |
105 | How frequently the probes are send out. Multiplied by | |
106 | tcp_keepalive_probes it is time to kill not responding connection, | |
107 | after probes started. Default value: 75sec i.e. connection | |
108 | will be aborted after ~11 minutes of retries. | |
109 | ||
110 | tcp_retries1 - INTEGER | |
111 | How many times to retry before deciding that something is wrong | |
112 | and it is necessary to report this suspicion to network layer. | |
113 | Minimal RFC value is 3, it is default, which corresponds | |
114 | to ~3sec-8min depending on RTO. | |
115 | ||
116 | tcp_retries2 - INTEGER | |
117 | How may times to retry before killing alive TCP connection. | |
118 | RFC1122 says that the limit should be longer than 100 sec. | |
119 | It is too small number. Default value 15 corresponds to ~13-30min | |
120 | depending on RTO. | |
121 | ||
122 | tcp_orphan_retries - INTEGER | |
123 | How may times to retry before killing TCP connection, closed | |
124 | by our side. Default value 7 corresponds to ~50sec-16min | |
125 | depending on RTO. If you machine is loaded WEB server, | |
126 | you should think about lowering this value, such sockets | |
127 | may consume significant resources. Cf. tcp_max_orphans. | |
128 | ||
129 | tcp_fin_timeout - INTEGER | |
130 | Time to hold socket in state FIN-WAIT-2, if it was closed | |
131 | by our side. Peer can be broken and never close its side, | |
132 | or even died unexpectedly. Default value is 60sec. | |
133 | Usual value used in 2.2 was 180 seconds, you may restore | |
134 | it, but remember that if your machine is even underloaded WEB server, | |
135 | you risk to overflow memory with kilotons of dead sockets, | |
136 | FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1, | |
137 | because they eat maximum 1.5K of memory, but they tend | |
138 | to live longer. Cf. tcp_max_orphans. | |
139 | ||
140 | tcp_max_tw_buckets - INTEGER | |
141 | Maximal number of timewait sockets held by system simultaneously. | |
142 | If this number is exceeded time-wait socket is immediately destroyed | |
143 | and warning is printed. This limit exists only to prevent | |
144 | simple DoS attacks, you _must_ not lower the limit artificially, | |
145 | but rather increase it (probably, after increasing installed memory), | |
146 | if network conditions require more than default value. | |
147 | ||
148 | tcp_tw_recycle - BOOLEAN | |
149 | Enable fast recycling TIME-WAIT sockets. Default value is 0. | |
150 | It should not be changed without advice/request of technical | |
151 | experts. | |
152 | ||
153 | tcp_tw_reuse - BOOLEAN | |
154 | Allow to reuse TIME-WAIT sockets for new connections when it is | |
155 | safe from protocol viewpoint. Default value is 0. | |
156 | It should not be changed without advice/request of technical | |
157 | experts. | |
158 | ||
159 | tcp_max_orphans - INTEGER | |
160 | Maximal number of TCP sockets not attached to any user file handle, | |
161 | held by system. If this number is exceeded orphaned connections are | |
162 | reset immediately and warning is printed. This limit exists | |
163 | only to prevent simple DoS attacks, you _must_ not rely on this | |
164 | or lower the limit artificially, but rather increase it | |
165 | (probably, after increasing installed memory), | |
166 | if network conditions require more than default value, | |
167 | and tune network services to linger and kill such states | |
168 | more aggressively. Let me to remind again: each orphan eats | |
169 | up to ~64K of unswappable memory. | |
170 | ||
171 | tcp_abort_on_overflow - BOOLEAN | |
172 | If listening service is too slow to accept new connections, | |
173 | reset them. Default state is FALSE. It means that if overflow | |
174 | occurred due to a burst, connection will recover. Enable this | |
175 | option _only_ if you are really sure that listening daemon | |
176 | cannot be tuned to accept connections faster. Enabling this | |
177 | option can harm clients of your server. | |
178 | ||
179 | tcp_syncookies - BOOLEAN | |
180 | Only valid when the kernel was compiled with CONFIG_SYNCOOKIES | |
181 | Send out syncookies when the syn backlog queue of a socket | |
182 | overflows. This is to prevent against the common 'syn flood attack' | |
183 | Default: FALSE | |
184 | ||
185 | Note, that syncookies is fallback facility. | |
186 | It MUST NOT be used to help highly loaded servers to stand | |
187 | against legal connection rate. If you see synflood warnings | |
188 | in your logs, but investigation shows that they occur | |
189 | because of overload with legal connections, you should tune | |
190 | another parameters until this warning disappear. | |
191 | See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow. | |
192 | ||
193 | syncookies seriously violate TCP protocol, do not allow | |
194 | to use TCP extensions, can result in serious degradation | |
195 | of some services (f.e. SMTP relaying), visible not by you, | |
196 | but your clients and relays, contacting you. While you see | |
197 | synflood warnings in logs not being really flooded, your server | |
198 | is seriously misconfigured. | |
199 | ||
200 | tcp_stdurg - BOOLEAN | |
201 | Use the Host requirements interpretation of the TCP urg pointer field. | |
202 | Most hosts use the older BSD interpretation, so if you turn this on | |
203 | Linux might not communicate correctly with them. | |
204 | Default: FALSE | |
205 | ||
206 | tcp_max_syn_backlog - INTEGER | |
207 | Maximal number of remembered connection requests, which are | |
208 | still did not receive an acknowledgment from connecting client. | |
209 | Default value is 1024 for systems with more than 128Mb of memory, | |
210 | and 128 for low memory machines. If server suffers of overload, | |
211 | try to increase this number. | |
212 | ||
213 | tcp_window_scaling - BOOLEAN | |
214 | Enable window scaling as defined in RFC1323. | |
215 | ||
216 | tcp_timestamps - BOOLEAN | |
217 | Enable timestamps as defined in RFC1323. | |
218 | ||
219 | tcp_sack - BOOLEAN | |
220 | Enable select acknowledgments (SACKS). | |
221 | ||
222 | tcp_fack - BOOLEAN | |
223 | Enable FACK congestion avoidance and fast retransmission. | |
224 | The value is not used, if tcp_sack is not enabled. | |
225 | ||
226 | tcp_dsack - BOOLEAN | |
227 | Allows TCP to send "duplicate" SACKs. | |
228 | ||
229 | tcp_ecn - BOOLEAN | |
230 | Enable Explicit Congestion Notification in TCP. | |
231 | ||
232 | tcp_reordering - INTEGER | |
233 | Maximal reordering of packets in a TCP stream. | |
234 | Default: 3 | |
235 | ||
236 | tcp_retrans_collapse - BOOLEAN | |
237 | Bug-to-bug compatibility with some broken printers. | |
238 | On retransmit try to send bigger packets to work around bugs in | |
239 | certain TCP stacks. | |
240 | ||
241 | tcp_wmem - vector of 3 INTEGERs: min, default, max | |
242 | min: Amount of memory reserved for send buffers for TCP socket. | |
243 | Each TCP socket has rights to use it due to fact of its birth. | |
244 | Default: 4K | |
245 | ||
246 | default: Amount of memory allowed for send buffers for TCP socket | |
247 | by default. This value overrides net.core.wmem_default used | |
248 | by other protocols, it is usually lower than net.core.wmem_default. | |
249 | Default: 16K | |
250 | ||
251 | max: Maximal amount of memory allowed for automatically selected | |
252 | send buffers for TCP socket. This value does not override | |
253 | net.core.wmem_max, "static" selection via SO_SNDBUF does not use this. | |
254 | Default: 128K | |
255 | ||
256 | tcp_rmem - vector of 3 INTEGERs: min, default, max | |
257 | min: Minimal size of receive buffer used by TCP sockets. | |
258 | It is guaranteed to each TCP socket, even under moderate memory | |
259 | pressure. | |
260 | Default: 8K | |
261 | ||
262 | default: default size of receive buffer used by TCP sockets. | |
263 | This value overrides net.core.rmem_default used by other protocols. | |
264 | Default: 87380 bytes. This value results in window of 65535 with | |
265 | default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit | |
266 | less for default tcp_app_win. See below about these variables. | |
267 | ||
268 | max: maximal size of receive buffer allowed for automatically | |
269 | selected receiver buffers for TCP socket. This value does not override | |
270 | net.core.rmem_max, "static" selection via SO_RCVBUF does not use this. | |
271 | Default: 87380*2 bytes. | |
272 | ||
273 | tcp_mem - vector of 3 INTEGERs: min, pressure, max | |
274 | low: below this number of pages TCP is not bothered about its | |
275 | memory appetite. | |
276 | ||
277 | pressure: when amount of memory allocated by TCP exceeds this number | |
278 | of pages, TCP moderates its memory consumption and enters memory | |
279 | pressure mode, which is exited when memory consumption falls | |
280 | under "low". | |
281 | ||
282 | high: number of pages allowed for queueing by all TCP sockets. | |
283 | ||
284 | Defaults are calculated at boot time from amount of available | |
285 | memory. | |
286 | ||
287 | tcp_app_win - INTEGER | |
288 | Reserve max(window/2^tcp_app_win, mss) of window for application | |
289 | buffer. Value 0 is special, it means that nothing is reserved. | |
290 | Default: 31 | |
291 | ||
292 | tcp_adv_win_scale - INTEGER | |
293 | Count buffering overhead as bytes/2^tcp_adv_win_scale | |
294 | (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale), | |
295 | if it is <= 0. | |
296 | Default: 2 | |
297 | ||
298 | tcp_rfc1337 - BOOLEAN | |
299 | If set, the TCP stack behaves conforming to RFC1337. If unset, | |
300 | we are not conforming to RFC, but prevent TCP TIME_WAIT | |
301 | assassination. | |
302 | Default: 0 | |
303 | ||
304 | tcp_low_latency - BOOLEAN | |
305 | If set, the TCP stack makes decisions that prefer lower | |
306 | latency as opposed to higher throughput. By default, this | |
307 | option is not set meaning that higher throughput is preferred. | |
308 | An example of an application where this default should be | |
309 | changed would be a Beowulf compute cluster. | |
310 | Default: 0 | |
311 | ||
1da177e4 LT |
312 | tcp_tso_win_divisor - INTEGER |
313 | This allows control over what percentage of the congestion window | |
314 | can be consumed by a single TSO frame. | |
315 | The setting of this parameter is a choice between burstiness and | |
316 | building larger TSO frames. | |
e83b8605 | 317 | Default: 3 |
1da177e4 LT |
318 | |
319 | tcp_frto - BOOLEAN | |
320 | Enables F-RTO, an enhanced recovery algorithm for TCP retransmission | |
321 | timeouts. It is particularly beneficial in wireless environments | |
322 | where packet loss is typically due to random radio interference | |
323 | rather than intermediate router congestion. | |
324 | ||
9d7bcfc6 SH |
325 | tcp_congestion_control - STRING |
326 | Set the congestion control algorithm to be used for new | |
327 | connections. The algorithm "reno" is always available, but | |
328 | additional choices may be available based on kernel configuration. | |
329 | ||
1da177e4 LT |
330 | somaxconn - INTEGER |
331 | Limit of socket listen() backlog, known in userspace as SOMAXCONN. | |
332 | Defaults to 128. See also tcp_max_syn_backlog for additional tuning | |
333 | for TCP sockets. | |
334 | ||
335 | IP Variables: | |
336 | ||
337 | ip_local_port_range - 2 INTEGERS | |
338 | Defines the local port range that is used by TCP and UDP to | |
339 | choose the local port. The first number is the first, the | |
340 | second the last local port number. Default value depends on | |
341 | amount of memory available on the system: | |
342 | > 128Mb 32768-61000 | |
343 | < 128Mb 1024-4999 or even less. | |
344 | This number defines number of active connections, which this | |
345 | system can issue simultaneously to systems not supporting | |
346 | TCP extensions (timestamps). With tcp_tw_recycle enabled | |
347 | (i.e. by default) range 1024-4999 is enough to issue up to | |
348 | 2000 connections per second to systems supporting timestamps. | |
349 | ||
350 | ip_nonlocal_bind - BOOLEAN | |
351 | If set, allows processes to bind() to non-local IP addresses, | |
352 | which can be quite useful - but may break some applications. | |
353 | Default: 0 | |
354 | ||
355 | ip_dynaddr - BOOLEAN | |
356 | If set non-zero, enables support for dynamic addresses. | |
357 | If set to a non-zero value larger than 1, a kernel log | |
358 | message will be printed when dynamic address rewriting | |
359 | occurs. | |
360 | Default: 0 | |
361 | ||
362 | icmp_echo_ignore_all - BOOLEAN | |
7ce31246 DM |
363 | If set non-zero, then the kernel will ignore all ICMP ECHO |
364 | requests sent to it. | |
365 | Default: 0 | |
366 | ||
1da177e4 | 367 | icmp_echo_ignore_broadcasts - BOOLEAN |
7ce31246 DM |
368 | If set non-zero, then the kernel will ignore all ICMP ECHO and |
369 | TIMESTAMP requests sent to it via broadcast/multicast. | |
370 | Default: 1 | |
1da177e4 LT |
371 | |
372 | icmp_ratelimit - INTEGER | |
373 | Limit the maximal rates for sending ICMP packets whose type matches | |
374 | icmp_ratemask (see below) to specific targets. | |
375 | 0 to disable any limiting, otherwise the maximal rate in jiffies(1) | |
376 | Default: 100 | |
377 | ||
378 | icmp_ratemask - INTEGER | |
379 | Mask made of ICMP types for which rates are being limited. | |
380 | Significant bits: IHGFEDCBA9876543210 | |
381 | Default mask: 0000001100000011000 (6168) | |
382 | ||
383 | Bit definitions (see include/linux/icmp.h): | |
384 | 0 Echo Reply | |
385 | 3 Destination Unreachable * | |
386 | 4 Source Quench * | |
387 | 5 Redirect | |
388 | 8 Echo Request | |
389 | B Time Exceeded * | |
390 | C Parameter Problem * | |
391 | D Timestamp Request | |
392 | E Timestamp Reply | |
393 | F Info Request | |
394 | G Info Reply | |
395 | H Address Mask Request | |
396 | I Address Mask Reply | |
397 | ||
398 | * These are rate limited by default (see default mask above) | |
399 | ||
400 | icmp_ignore_bogus_error_responses - BOOLEAN | |
401 | Some routers violate RFC1122 by sending bogus responses to broadcast | |
402 | frames. Such violations are normally logged via a kernel warning. | |
403 | If this is set to TRUE, the kernel will not give such warnings, which | |
404 | will avoid log file clutter. | |
405 | Default: FALSE | |
406 | ||
407 | igmp_max_memberships - INTEGER | |
408 | Change the maximum number of multicast groups we can subscribe to. | |
409 | Default: 20 | |
410 | ||
411 | conf/interface/* changes special settings per interface (where "interface" is | |
412 | the name of your network interface) | |
413 | conf/all/* is special, changes the settings for all interfaces | |
414 | ||
415 | ||
416 | log_martians - BOOLEAN | |
417 | Log packets with impossible addresses to kernel log. | |
418 | log_martians for the interface will be enabled if at least one of | |
419 | conf/{all,interface}/log_martians is set to TRUE, | |
420 | it will be disabled otherwise | |
421 | ||
422 | accept_redirects - BOOLEAN | |
423 | Accept ICMP redirect messages. | |
424 | accept_redirects for the interface will be enabled if: | |
425 | - both conf/{all,interface}/accept_redirects are TRUE in the case forwarding | |
426 | for the interface is enabled | |
427 | or | |
428 | - at least one of conf/{all,interface}/accept_redirects is TRUE in the case | |
429 | forwarding for the interface is disabled | |
430 | accept_redirects for the interface will be disabled otherwise | |
431 | default TRUE (host) | |
432 | FALSE (router) | |
433 | ||
434 | forwarding - BOOLEAN | |
435 | Enable IP forwarding on this interface. | |
436 | ||
437 | mc_forwarding - BOOLEAN | |
438 | Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE | |
439 | and a multicast routing daemon is required. | |
440 | conf/all/mc_forwarding must also be set to TRUE to enable multicast routing | |
441 | for the interface | |
442 | ||
443 | medium_id - INTEGER | |
444 | Integer value used to differentiate the devices by the medium they | |
445 | are attached to. Two devices can have different id values when | |
446 | the broadcast packets are received only on one of them. | |
447 | The default value 0 means that the device is the only interface | |
448 | to its medium, value of -1 means that medium is not known. | |
449 | ||
450 | Currently, it is used to change the proxy_arp behavior: | |
451 | the proxy_arp feature is enabled for packets forwarded between | |
452 | two devices attached to different media. | |
453 | ||
454 | proxy_arp - BOOLEAN | |
455 | Do proxy arp. | |
456 | proxy_arp for the interface will be enabled if at least one of | |
457 | conf/{all,interface}/proxy_arp is set to TRUE, | |
458 | it will be disabled otherwise | |
459 | ||
460 | shared_media - BOOLEAN | |
461 | Send(router) or accept(host) RFC1620 shared media redirects. | |
462 | Overrides ip_secure_redirects. | |
463 | shared_media for the interface will be enabled if at least one of | |
464 | conf/{all,interface}/shared_media is set to TRUE, | |
465 | it will be disabled otherwise | |
466 | default TRUE | |
467 | ||
468 | secure_redirects - BOOLEAN | |
469 | Accept ICMP redirect messages only for gateways, | |
470 | listed in default gateway list. | |
471 | secure_redirects for the interface will be enabled if at least one of | |
472 | conf/{all,interface}/secure_redirects is set to TRUE, | |
473 | it will be disabled otherwise | |
474 | default TRUE | |
475 | ||
476 | send_redirects - BOOLEAN | |
477 | Send redirects, if router. | |
478 | send_redirects for the interface will be enabled if at least one of | |
479 | conf/{all,interface}/send_redirects is set to TRUE, | |
480 | it will be disabled otherwise | |
481 | Default: TRUE | |
482 | ||
483 | bootp_relay - BOOLEAN | |
484 | Accept packets with source address 0.b.c.d destined | |
485 | not to this host as local ones. It is supposed, that | |
486 | BOOTP relay daemon will catch and forward such packets. | |
487 | conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay | |
488 | for the interface | |
489 | default FALSE | |
490 | Not Implemented Yet. | |
491 | ||
492 | accept_source_route - BOOLEAN | |
493 | Accept packets with SRR option. | |
494 | conf/all/accept_source_route must also be set to TRUE to accept packets | |
495 | with SRR option on the interface | |
496 | default TRUE (router) | |
497 | FALSE (host) | |
498 | ||
499 | rp_filter - BOOLEAN | |
500 | 1 - do source validation by reversed path, as specified in RFC1812 | |
501 | Recommended option for single homed hosts and stub network | |
502 | routers. Could cause troubles for complicated (not loop free) | |
503 | networks running a slow unreliable protocol (sort of RIP), | |
504 | or using static routes. | |
505 | ||
506 | 0 - No source validation. | |
507 | ||
508 | conf/all/rp_filter must also be set to TRUE to do source validation | |
509 | on the interface | |
510 | ||
511 | Default value is 0. Note that some distributions enable it | |
512 | in startup scripts. | |
513 | ||
514 | arp_filter - BOOLEAN | |
515 | 1 - Allows you to have multiple network interfaces on the same | |
516 | subnet, and have the ARPs for each interface be answered | |
517 | based on whether or not the kernel would route a packet from | |
518 | the ARP'd IP out that interface (therefore you must use source | |
519 | based routing for this to work). In other words it allows control | |
520 | of which cards (usually 1) will respond to an arp request. | |
521 | ||
522 | 0 - (default) The kernel can respond to arp requests with addresses | |
523 | from other interfaces. This may seem wrong but it usually makes | |
524 | sense, because it increases the chance of successful communication. | |
525 | IP addresses are owned by the complete host on Linux, not by | |
526 | particular interfaces. Only for more complex setups like load- | |
527 | balancing, does this behaviour cause problems. | |
528 | ||
529 | arp_filter for the interface will be enabled if at least one of | |
530 | conf/{all,interface}/arp_filter is set to TRUE, | |
531 | it will be disabled otherwise | |
532 | ||
533 | arp_announce - INTEGER | |
534 | Define different restriction levels for announcing the local | |
535 | source IP address from IP packets in ARP requests sent on | |
536 | interface: | |
537 | 0 - (default) Use any local address, configured on any interface | |
538 | 1 - Try to avoid local addresses that are not in the target's | |
539 | subnet for this interface. This mode is useful when target | |
540 | hosts reachable via this interface require the source IP | |
541 | address in ARP requests to be part of their logical network | |
542 | configured on the receiving interface. When we generate the | |
543 | request we will check all our subnets that include the | |
544 | target IP and will preserve the source address if it is from | |
545 | such subnet. If there is no such subnet we select source | |
546 | address according to the rules for level 2. | |
547 | 2 - Always use the best local address for this target. | |
548 | In this mode we ignore the source address in the IP packet | |
549 | and try to select local address that we prefer for talks with | |
550 | the target host. Such local address is selected by looking | |
551 | for primary IP addresses on all our subnets on the outgoing | |
552 | interface that include the target IP address. If no suitable | |
553 | local address is found we select the first local address | |
554 | we have on the outgoing interface or on all other interfaces, | |
555 | with the hope we will receive reply for our request and | |
556 | even sometimes no matter the source IP address we announce. | |
557 | ||
558 | The max value from conf/{all,interface}/arp_announce is used. | |
559 | ||
560 | Increasing the restriction level gives more chance for | |
561 | receiving answer from the resolved target while decreasing | |
562 | the level announces more valid sender's information. | |
563 | ||
564 | arp_ignore - INTEGER | |
565 | Define different modes for sending replies in response to | |
566 | received ARP requests that resolve local target IP addresses: | |
567 | 0 - (default): reply for any local target IP address, configured | |
568 | on any interface | |
569 | 1 - reply only if the target IP address is local address | |
570 | configured on the incoming interface | |
571 | 2 - reply only if the target IP address is local address | |
572 | configured on the incoming interface and both with the | |
573 | sender's IP address are part from same subnet on this interface | |
574 | 3 - do not reply for local addresses configured with scope host, | |
575 | only resolutions for global and link addresses are replied | |
576 | 4-7 - reserved | |
577 | 8 - do not reply for all local addresses | |
578 | ||
579 | The max value from conf/{all,interface}/arp_ignore is used | |
580 | when ARP request is received on the {interface} | |
581 | ||
582 | app_solicit - INTEGER | |
583 | The maximum number of probes to send to the user space ARP daemon | |
584 | via netlink before dropping back to multicast probes (see | |
585 | mcast_solicit). Defaults to 0. | |
586 | ||
587 | disable_policy - BOOLEAN | |
588 | Disable IPSEC policy (SPD) for this interface | |
589 | ||
590 | disable_xfrm - BOOLEAN | |
591 | Disable IPSEC encryption on this interface, whatever the policy | |
592 | ||
593 | ||
594 | ||
595 | tag - INTEGER | |
596 | Allows you to write a number, which can be used as required. | |
597 | Default value is 0. | |
598 | ||
599 | (1) Jiffie: internal timeunit for the kernel. On the i386 1/100s, on the | |
600 | Alpha 1/1024s. See the HZ define in /usr/include/asm/param.h for the exact | |
601 | value on your system. | |
602 | ||
603 | Alexey Kuznetsov. | |
604 | kuznet@ms2.inr.ac.ru | |
605 | ||
606 | Updated by: | |
607 | Andi Kleen | |
608 | ak@muc.de | |
609 | Nicolas Delon | |
610 | delon.nicolas@wanadoo.fr | |
611 | ||
612 | ||
613 | ||
614 | ||
615 | /proc/sys/net/ipv6/* Variables: | |
616 | ||
617 | IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also | |
618 | apply to IPv6 [XXX?]. | |
619 | ||
620 | bindv6only - BOOLEAN | |
621 | Default value for IPV6_V6ONLY socket option, | |
622 | which restricts use of the IPv6 socket to IPv6 communication | |
623 | only. | |
624 | TRUE: disable IPv4-mapped address feature | |
625 | FALSE: enable IPv4-mapped address feature | |
626 | ||
627 | Default: FALSE (as specified in RFC2553bis) | |
628 | ||
629 | IPv6 Fragmentation: | |
630 | ||
631 | ip6frag_high_thresh - INTEGER | |
632 | Maximum memory used to reassemble IPv6 fragments. When | |
633 | ip6frag_high_thresh bytes of memory is allocated for this purpose, | |
634 | the fragment handler will toss packets until ip6frag_low_thresh | |
635 | is reached. | |
636 | ||
637 | ip6frag_low_thresh - INTEGER | |
638 | See ip6frag_high_thresh | |
639 | ||
640 | ip6frag_time - INTEGER | |
641 | Time in seconds to keep an IPv6 fragment in memory. | |
642 | ||
643 | ip6frag_secret_interval - INTEGER | |
644 | Regeneration interval (in seconds) of the hash secret (or lifetime | |
645 | for the hash secret) for IPv6 fragments. | |
646 | Default: 600 | |
647 | ||
648 | conf/default/*: | |
649 | Change the interface-specific default settings. | |
650 | ||
651 | ||
652 | conf/all/*: | |
653 | Change all the interface-specific settings. | |
654 | ||
655 | [XXX: Other special features than forwarding?] | |
656 | ||
657 | conf/all/forwarding - BOOLEAN | |
658 | Enable global IPv6 forwarding between all interfaces. | |
659 | ||
660 | IPv4 and IPv6 work differently here; e.g. netfilter must be used | |
661 | to control which interfaces may forward packets and which not. | |
662 | ||
663 | This also sets all interfaces' Host/Router setting | |
664 | 'forwarding' to the specified value. See below for details. | |
665 | ||
666 | This referred to as global forwarding. | |
667 | ||
668 | conf/interface/*: | |
669 | Change special settings per interface. | |
670 | ||
671 | The functional behaviour for certain settings is different | |
672 | depending on whether local forwarding is enabled or not. | |
673 | ||
674 | accept_ra - BOOLEAN | |
675 | Accept Router Advertisements; autoconfigure using them. | |
676 | ||
677 | Functional default: enabled if local forwarding is disabled. | |
678 | disabled if local forwarding is enabled. | |
679 | ||
680 | accept_redirects - BOOLEAN | |
681 | Accept Redirects. | |
682 | ||
683 | Functional default: enabled if local forwarding is disabled. | |
684 | disabled if local forwarding is enabled. | |
685 | ||
686 | autoconf - BOOLEAN | |
687 | Autoconfigure addresses using Prefix Information in Router | |
688 | Advertisements. | |
689 | ||
690 | Functional default: enabled if accept_ra is enabled. | |
691 | disabled if accept_ra is disabled. | |
692 | ||
693 | dad_transmits - INTEGER | |
694 | The amount of Duplicate Address Detection probes to send. | |
695 | Default: 1 | |
696 | ||
697 | forwarding - BOOLEAN | |
698 | Configure interface-specific Host/Router behaviour. | |
699 | ||
700 | Note: It is recommended to have the same setting on all | |
701 | interfaces; mixed router/host scenarios are rather uncommon. | |
702 | ||
703 | FALSE: | |
704 | ||
705 | By default, Host behaviour is assumed. This means: | |
706 | ||
707 | 1. IsRouter flag is not set in Neighbour Advertisements. | |
708 | 2. Router Solicitations are being sent when necessary. | |
709 | 3. If accept_ra is TRUE (default), accept Router | |
710 | Advertisements (and do autoconfiguration). | |
711 | 4. If accept_redirects is TRUE (default), accept Redirects. | |
712 | ||
713 | TRUE: | |
714 | ||
715 | If local forwarding is enabled, Router behaviour is assumed. | |
716 | This means exactly the reverse from the above: | |
717 | ||
718 | 1. IsRouter flag is set in Neighbour Advertisements. | |
719 | 2. Router Solicitations are not sent. | |
720 | 3. Router Advertisements are ignored. | |
721 | 4. Redirects are ignored. | |
722 | ||
723 | Default: FALSE if global forwarding is disabled (default), | |
724 | otherwise TRUE. | |
725 | ||
726 | hop_limit - INTEGER | |
727 | Default Hop Limit to set. | |
728 | Default: 64 | |
729 | ||
730 | mtu - INTEGER | |
731 | Default Maximum Transfer Unit | |
732 | Default: 1280 (IPv6 required minimum) | |
733 | ||
734 | router_solicitation_delay - INTEGER | |
735 | Number of seconds to wait after interface is brought up | |
736 | before sending Router Solicitations. | |
737 | Default: 1 | |
738 | ||
739 | router_solicitation_interval - INTEGER | |
740 | Number of seconds to wait between Router Solicitations. | |
741 | Default: 4 | |
742 | ||
743 | router_solicitations - INTEGER | |
744 | Number of Router Solicitations to send until assuming no | |
745 | routers are present. | |
746 | Default: 3 | |
747 | ||
748 | use_tempaddr - INTEGER | |
749 | Preference for Privacy Extensions (RFC3041). | |
750 | <= 0 : disable Privacy Extensions | |
751 | == 1 : enable Privacy Extensions, but prefer public | |
752 | addresses over temporary addresses. | |
753 | > 1 : enable Privacy Extensions and prefer temporary | |
754 | addresses over public addresses. | |
755 | Default: 0 (for most devices) | |
756 | -1 (for point-to-point devices and loopback devices) | |
757 | ||
758 | temp_valid_lft - INTEGER | |
759 | valid lifetime (in seconds) for temporary addresses. | |
760 | Default: 604800 (7 days) | |
761 | ||
762 | temp_prefered_lft - INTEGER | |
763 | Preferred lifetime (in seconds) for temporary addresses. | |
764 | Default: 86400 (1 day) | |
765 | ||
766 | max_desync_factor - INTEGER | |
767 | Maximum value for DESYNC_FACTOR, which is a random value | |
768 | that ensures that clients don't synchronize with each | |
769 | other and generate new addresses at exactly the same time. | |
770 | value is in seconds. | |
771 | Default: 600 | |
772 | ||
773 | regen_max_retry - INTEGER | |
774 | Number of attempts before give up attempting to generate | |
775 | valid temporary addresses. | |
776 | Default: 5 | |
777 | ||
778 | max_addresses - INTEGER | |
779 | Number of maximum addresses per interface. 0 disables limitation. | |
780 | It is recommended not set too large value (or 0) because it would | |
781 | be too easy way to crash kernel to allow to create too much of | |
782 | autoconfigured addresses. | |
783 | Default: 16 | |
784 | ||
785 | icmp/*: | |
786 | ratelimit - INTEGER | |
787 | Limit the maximal rates for sending ICMPv6 packets. | |
788 | 0 to disable any limiting, otherwise the maximal rate in jiffies(1) | |
789 | Default: 100 | |
790 | ||
791 | ||
792 | IPv6 Update by: | |
793 | Pekka Savola <pekkas@netcore.fi> | |
794 | YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org> | |
795 | ||
796 | ||
797 | /proc/sys/net/bridge/* Variables: | |
798 | ||
799 | bridge-nf-call-arptables - BOOLEAN | |
800 | 1 : pass bridged ARP traffic to arptables' FORWARD chain. | |
801 | 0 : disable this. | |
802 | Default: 1 | |
803 | ||
804 | bridge-nf-call-iptables - BOOLEAN | |
805 | 1 : pass bridged IPv4 traffic to iptables' chains. | |
806 | 0 : disable this. | |
807 | Default: 1 | |
808 | ||
809 | bridge-nf-call-ip6tables - BOOLEAN | |
810 | 1 : pass bridged IPv6 traffic to ip6tables' chains. | |
811 | 0 : disable this. | |
812 | Default: 1 | |
813 | ||
814 | bridge-nf-filter-vlan-tagged - BOOLEAN | |
815 | 1 : pass bridged vlan-tagged ARP/IP traffic to arptables/iptables. | |
816 | 0 : disable this. | |
817 | Default: 1 | |
818 | ||
819 | ||
820 | UNDOCUMENTED: | |
821 | ||
822 | dev_weight FIXME | |
823 | discovery_slots FIXME | |
824 | discovery_timeout FIXME | |
825 | fast_poll_increase FIXME | |
826 | ip6_queue_maxlen FIXME | |
827 | lap_keepalive_time FIXME | |
828 | lo_cong FIXME | |
829 | max_baud_rate FIXME | |
830 | max_dgram_qlen FIXME | |
831 | max_noreply_time FIXME | |
832 | max_tx_data_size FIXME | |
833 | max_tx_window FIXME | |
834 | min_tx_turn_time FIXME | |
835 | mod_cong FIXME | |
836 | no_cong FIXME | |
837 | no_cong_thresh FIXME | |
838 | slot_timeout FIXME | |
839 | warn_noreply_time FIXME | |
840 | ||
841 | $Id: ip-sysctl.txt,v 1.20 2001/12/13 09:00:18 davem Exp $ |