Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 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 | |
Herbert Xu | 89cee8b | 2005-12-13 23:14:27 -0800 | [diff] [blame^] | 49 | ipfrag_max_dist - INTEGER |
| 50 | ipfrag_max_dist is a non-negative integer value which defines the |
| 51 | maximum "disorder" which is allowed among fragments which share a |
| 52 | common IP source address. Note that reordering of packets is |
| 53 | not unusual, but if a large number of fragments arrive from a source |
| 54 | IP address while a particular fragment queue remains incomplete, it |
| 55 | probably indicates that one or more fragments belonging to that queue |
| 56 | have been lost. When ipfrag_max_dist is positive, an additional check |
| 57 | is done on fragments before they are added to a reassembly queue - if |
| 58 | ipfrag_max_dist (or more) fragments have arrived from a particular IP |
| 59 | address between additions to any IP fragment queue using that source |
| 60 | address, it's presumed that one or more fragments in the queue are |
| 61 | lost. The existing fragment queue will be dropped, and a new one |
| 62 | started. An ipfrag_max_dist value of zero disables this check. |
| 63 | |
| 64 | Using a very small value, e.g. 1 or 2, for ipfrag_max_dist can |
| 65 | result in unnecessarily dropping fragment queues when normal |
| 66 | reordering of packets occurs, which could lead to poor application |
| 67 | performance. Using a very large value, e.g. 50000, increases the |
| 68 | likelihood of incorrectly reassembling IP fragments that originate |
| 69 | from different IP datagrams, which could result in data corruption. |
| 70 | Default: 64 |
| 71 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 72 | INET peer storage: |
| 73 | |
| 74 | inet_peer_threshold - INTEGER |
| 75 | The approximate size of the storage. Starting from this threshold |
| 76 | entries will be thrown aggressively. This threshold also determines |
| 77 | entries' time-to-live and time intervals between garbage collection |
| 78 | passes. More entries, less time-to-live, less GC interval. |
| 79 | |
| 80 | inet_peer_minttl - INTEGER |
| 81 | Minimum time-to-live of entries. Should be enough to cover fragment |
| 82 | time-to-live on the reassembling side. This minimum time-to-live is |
| 83 | guaranteed if the pool size is less than inet_peer_threshold. |
| 84 | Measured in jiffies(1). |
| 85 | |
| 86 | inet_peer_maxttl - INTEGER |
| 87 | Maximum time-to-live of entries. Unused entries will expire after |
| 88 | this period of time if there is no memory pressure on the pool (i.e. |
| 89 | when the number of entries in the pool is very small). |
| 90 | Measured in jiffies(1). |
| 91 | |
| 92 | inet_peer_gc_mintime - INTEGER |
| 93 | Minimum interval between garbage collection passes. This interval is |
| 94 | in effect under high memory pressure on the pool. |
| 95 | Measured in jiffies(1). |
| 96 | |
| 97 | inet_peer_gc_maxtime - INTEGER |
| 98 | Minimum interval between garbage collection passes. This interval is |
| 99 | in effect under low (or absent) memory pressure on the pool. |
| 100 | Measured in jiffies(1). |
| 101 | |
| 102 | TCP variables: |
| 103 | |
Stephen Hemminger | 9772efb | 2005-11-10 17:09:53 -0800 | [diff] [blame] | 104 | tcp_abc - INTEGER |
| 105 | Controls Appropriate Byte Count defined in RFC3465. If set to |
| 106 | 0 then does congestion avoid once per ack. 1 is conservative |
| 107 | value, and 2 is more agressive. |
| 108 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 109 | tcp_syn_retries - INTEGER |
| 110 | Number of times initial SYNs for an active TCP connection attempt |
| 111 | will be retransmitted. Should not be higher than 255. Default value |
| 112 | is 5, which corresponds to ~180seconds. |
| 113 | |
| 114 | tcp_synack_retries - INTEGER |
| 115 | Number of times SYNACKs for a passive TCP connection attempt will |
| 116 | be retransmitted. Should not be higher than 255. Default value |
| 117 | is 5, which corresponds to ~180seconds. |
| 118 | |
| 119 | tcp_keepalive_time - INTEGER |
| 120 | How often TCP sends out keepalive messages when keepalive is enabled. |
| 121 | Default: 2hours. |
| 122 | |
| 123 | tcp_keepalive_probes - INTEGER |
| 124 | How many keepalive probes TCP sends out, until it decides that the |
| 125 | connection is broken. Default value: 9. |
| 126 | |
| 127 | tcp_keepalive_intvl - INTEGER |
| 128 | How frequently the probes are send out. Multiplied by |
| 129 | tcp_keepalive_probes it is time to kill not responding connection, |
| 130 | after probes started. Default value: 75sec i.e. connection |
| 131 | will be aborted after ~11 minutes of retries. |
| 132 | |
| 133 | tcp_retries1 - INTEGER |
| 134 | How many times to retry before deciding that something is wrong |
| 135 | and it is necessary to report this suspicion to network layer. |
| 136 | Minimal RFC value is 3, it is default, which corresponds |
| 137 | to ~3sec-8min depending on RTO. |
| 138 | |
| 139 | tcp_retries2 - INTEGER |
| 140 | How may times to retry before killing alive TCP connection. |
| 141 | RFC1122 says that the limit should be longer than 100 sec. |
| 142 | It is too small number. Default value 15 corresponds to ~13-30min |
| 143 | depending on RTO. |
| 144 | |
| 145 | tcp_orphan_retries - INTEGER |
| 146 | How may times to retry before killing TCP connection, closed |
| 147 | by our side. Default value 7 corresponds to ~50sec-16min |
| 148 | depending on RTO. If you machine is loaded WEB server, |
| 149 | you should think about lowering this value, such sockets |
| 150 | may consume significant resources. Cf. tcp_max_orphans. |
| 151 | |
| 152 | tcp_fin_timeout - INTEGER |
| 153 | Time to hold socket in state FIN-WAIT-2, if it was closed |
| 154 | by our side. Peer can be broken and never close its side, |
| 155 | or even died unexpectedly. Default value is 60sec. |
| 156 | Usual value used in 2.2 was 180 seconds, you may restore |
| 157 | it, but remember that if your machine is even underloaded WEB server, |
| 158 | you risk to overflow memory with kilotons of dead sockets, |
| 159 | FIN-WAIT-2 sockets are less dangerous than FIN-WAIT-1, |
| 160 | because they eat maximum 1.5K of memory, but they tend |
| 161 | to live longer. Cf. tcp_max_orphans. |
| 162 | |
| 163 | tcp_max_tw_buckets - INTEGER |
| 164 | Maximal number of timewait sockets held by system simultaneously. |
| 165 | If this number is exceeded time-wait socket is immediately destroyed |
| 166 | and warning is printed. This limit exists only to prevent |
| 167 | simple DoS attacks, you _must_ not lower the limit artificially, |
| 168 | but rather increase it (probably, after increasing installed memory), |
| 169 | if network conditions require more than default value. |
| 170 | |
| 171 | tcp_tw_recycle - BOOLEAN |
| 172 | Enable fast recycling TIME-WAIT sockets. Default value is 0. |
| 173 | It should not be changed without advice/request of technical |
| 174 | experts. |
| 175 | |
| 176 | tcp_tw_reuse - BOOLEAN |
| 177 | Allow to reuse TIME-WAIT sockets for new connections when it is |
| 178 | safe from protocol viewpoint. Default value is 0. |
| 179 | It should not be changed without advice/request of technical |
| 180 | experts. |
| 181 | |
| 182 | tcp_max_orphans - INTEGER |
| 183 | Maximal number of TCP sockets not attached to any user file handle, |
| 184 | held by system. If this number is exceeded orphaned connections are |
| 185 | reset immediately and warning is printed. This limit exists |
| 186 | only to prevent simple DoS attacks, you _must_ not rely on this |
| 187 | or lower the limit artificially, but rather increase it |
| 188 | (probably, after increasing installed memory), |
| 189 | if network conditions require more than default value, |
| 190 | and tune network services to linger and kill such states |
| 191 | more aggressively. Let me to remind again: each orphan eats |
| 192 | up to ~64K of unswappable memory. |
| 193 | |
| 194 | tcp_abort_on_overflow - BOOLEAN |
| 195 | If listening service is too slow to accept new connections, |
| 196 | reset them. Default state is FALSE. It means that if overflow |
| 197 | occurred due to a burst, connection will recover. Enable this |
| 198 | option _only_ if you are really sure that listening daemon |
| 199 | cannot be tuned to accept connections faster. Enabling this |
| 200 | option can harm clients of your server. |
| 201 | |
| 202 | tcp_syncookies - BOOLEAN |
| 203 | Only valid when the kernel was compiled with CONFIG_SYNCOOKIES |
| 204 | Send out syncookies when the syn backlog queue of a socket |
| 205 | overflows. This is to prevent against the common 'syn flood attack' |
| 206 | Default: FALSE |
| 207 | |
| 208 | Note, that syncookies is fallback facility. |
| 209 | It MUST NOT be used to help highly loaded servers to stand |
| 210 | against legal connection rate. If you see synflood warnings |
| 211 | in your logs, but investigation shows that they occur |
| 212 | because of overload with legal connections, you should tune |
| 213 | another parameters until this warning disappear. |
| 214 | See: tcp_max_syn_backlog, tcp_synack_retries, tcp_abort_on_overflow. |
| 215 | |
| 216 | syncookies seriously violate TCP protocol, do not allow |
| 217 | to use TCP extensions, can result in serious degradation |
| 218 | of some services (f.e. SMTP relaying), visible not by you, |
| 219 | but your clients and relays, contacting you. While you see |
| 220 | synflood warnings in logs not being really flooded, your server |
| 221 | is seriously misconfigured. |
| 222 | |
| 223 | tcp_stdurg - BOOLEAN |
| 224 | Use the Host requirements interpretation of the TCP urg pointer field. |
| 225 | Most hosts use the older BSD interpretation, so if you turn this on |
| 226 | Linux might not communicate correctly with them. |
| 227 | Default: FALSE |
| 228 | |
| 229 | tcp_max_syn_backlog - INTEGER |
| 230 | Maximal number of remembered connection requests, which are |
| 231 | still did not receive an acknowledgment from connecting client. |
| 232 | Default value is 1024 for systems with more than 128Mb of memory, |
| 233 | and 128 for low memory machines. If server suffers of overload, |
| 234 | try to increase this number. |
| 235 | |
| 236 | tcp_window_scaling - BOOLEAN |
| 237 | Enable window scaling as defined in RFC1323. |
| 238 | |
| 239 | tcp_timestamps - BOOLEAN |
| 240 | Enable timestamps as defined in RFC1323. |
| 241 | |
| 242 | tcp_sack - BOOLEAN |
| 243 | Enable select acknowledgments (SACKS). |
| 244 | |
| 245 | tcp_fack - BOOLEAN |
| 246 | Enable FACK congestion avoidance and fast retransmission. |
| 247 | The value is not used, if tcp_sack is not enabled. |
| 248 | |
| 249 | tcp_dsack - BOOLEAN |
| 250 | Allows TCP to send "duplicate" SACKs. |
| 251 | |
| 252 | tcp_ecn - BOOLEAN |
| 253 | Enable Explicit Congestion Notification in TCP. |
| 254 | |
| 255 | tcp_reordering - INTEGER |
| 256 | Maximal reordering of packets in a TCP stream. |
| 257 | Default: 3 |
| 258 | |
| 259 | tcp_retrans_collapse - BOOLEAN |
| 260 | Bug-to-bug compatibility with some broken printers. |
| 261 | On retransmit try to send bigger packets to work around bugs in |
| 262 | certain TCP stacks. |
| 263 | |
| 264 | tcp_wmem - vector of 3 INTEGERs: min, default, max |
| 265 | min: Amount of memory reserved for send buffers for TCP socket. |
| 266 | Each TCP socket has rights to use it due to fact of its birth. |
| 267 | Default: 4K |
| 268 | |
| 269 | default: Amount of memory allowed for send buffers for TCP socket |
| 270 | by default. This value overrides net.core.wmem_default used |
| 271 | by other protocols, it is usually lower than net.core.wmem_default. |
| 272 | Default: 16K |
| 273 | |
| 274 | max: Maximal amount of memory allowed for automatically selected |
| 275 | send buffers for TCP socket. This value does not override |
| 276 | net.core.wmem_max, "static" selection via SO_SNDBUF does not use this. |
| 277 | Default: 128K |
| 278 | |
| 279 | tcp_rmem - vector of 3 INTEGERs: min, default, max |
| 280 | min: Minimal size of receive buffer used by TCP sockets. |
| 281 | It is guaranteed to each TCP socket, even under moderate memory |
| 282 | pressure. |
| 283 | Default: 8K |
| 284 | |
| 285 | default: default size of receive buffer used by TCP sockets. |
| 286 | This value overrides net.core.rmem_default used by other protocols. |
| 287 | Default: 87380 bytes. This value results in window of 65535 with |
| 288 | default setting of tcp_adv_win_scale and tcp_app_win:0 and a bit |
| 289 | less for default tcp_app_win. See below about these variables. |
| 290 | |
| 291 | max: maximal size of receive buffer allowed for automatically |
| 292 | selected receiver buffers for TCP socket. This value does not override |
| 293 | net.core.rmem_max, "static" selection via SO_RCVBUF does not use this. |
| 294 | Default: 87380*2 bytes. |
| 295 | |
| 296 | tcp_mem - vector of 3 INTEGERs: min, pressure, max |
| 297 | low: below this number of pages TCP is not bothered about its |
| 298 | memory appetite. |
| 299 | |
| 300 | pressure: when amount of memory allocated by TCP exceeds this number |
| 301 | of pages, TCP moderates its memory consumption and enters memory |
| 302 | pressure mode, which is exited when memory consumption falls |
| 303 | under "low". |
| 304 | |
| 305 | high: number of pages allowed for queueing by all TCP sockets. |
| 306 | |
| 307 | Defaults are calculated at boot time from amount of available |
| 308 | memory. |
| 309 | |
| 310 | tcp_app_win - INTEGER |
| 311 | Reserve max(window/2^tcp_app_win, mss) of window for application |
| 312 | buffer. Value 0 is special, it means that nothing is reserved. |
| 313 | Default: 31 |
| 314 | |
| 315 | tcp_adv_win_scale - INTEGER |
| 316 | Count buffering overhead as bytes/2^tcp_adv_win_scale |
| 317 | (if tcp_adv_win_scale > 0) or bytes-bytes/2^(-tcp_adv_win_scale), |
| 318 | if it is <= 0. |
| 319 | Default: 2 |
| 320 | |
| 321 | tcp_rfc1337 - BOOLEAN |
| 322 | If set, the TCP stack behaves conforming to RFC1337. If unset, |
| 323 | we are not conforming to RFC, but prevent TCP TIME_WAIT |
| 324 | assassination. |
| 325 | Default: 0 |
| 326 | |
| 327 | tcp_low_latency - BOOLEAN |
| 328 | If set, the TCP stack makes decisions that prefer lower |
| 329 | latency as opposed to higher throughput. By default, this |
| 330 | option is not set meaning that higher throughput is preferred. |
| 331 | An example of an application where this default should be |
| 332 | changed would be a Beowulf compute cluster. |
| 333 | Default: 0 |
| 334 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 335 | tcp_tso_win_divisor - INTEGER |
| 336 | This allows control over what percentage of the congestion window |
| 337 | can be consumed by a single TSO frame. |
| 338 | The setting of this parameter is a choice between burstiness and |
| 339 | building larger TSO frames. |
Akinobu Mita | e83b860 | 2005-10-21 18:00:48 +0900 | [diff] [blame] | 340 | Default: 3 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 341 | |
| 342 | tcp_frto - BOOLEAN |
| 343 | Enables F-RTO, an enhanced recovery algorithm for TCP retransmission |
| 344 | timeouts. It is particularly beneficial in wireless environments |
| 345 | where packet loss is typically due to random radio interference |
| 346 | rather than intermediate router congestion. |
| 347 | |
Stephen Hemminger | 9d7bcfc | 2005-06-23 12:22:36 -0700 | [diff] [blame] | 348 | tcp_congestion_control - STRING |
| 349 | Set the congestion control algorithm to be used for new |
| 350 | connections. The algorithm "reno" is always available, but |
| 351 | additional choices may be available based on kernel configuration. |
| 352 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 353 | somaxconn - INTEGER |
| 354 | Limit of socket listen() backlog, known in userspace as SOMAXCONN. |
| 355 | Defaults to 128. See also tcp_max_syn_backlog for additional tuning |
| 356 | for TCP sockets. |
| 357 | |
| 358 | IP Variables: |
| 359 | |
| 360 | ip_local_port_range - 2 INTEGERS |
| 361 | Defines the local port range that is used by TCP and UDP to |
| 362 | choose the local port. The first number is the first, the |
| 363 | second the last local port number. Default value depends on |
| 364 | amount of memory available on the system: |
| 365 | > 128Mb 32768-61000 |
| 366 | < 128Mb 1024-4999 or even less. |
| 367 | This number defines number of active connections, which this |
| 368 | system can issue simultaneously to systems not supporting |
| 369 | TCP extensions (timestamps). With tcp_tw_recycle enabled |
| 370 | (i.e. by default) range 1024-4999 is enough to issue up to |
| 371 | 2000 connections per second to systems supporting timestamps. |
| 372 | |
| 373 | ip_nonlocal_bind - BOOLEAN |
| 374 | If set, allows processes to bind() to non-local IP addresses, |
| 375 | which can be quite useful - but may break some applications. |
| 376 | Default: 0 |
| 377 | |
| 378 | ip_dynaddr - BOOLEAN |
| 379 | If set non-zero, enables support for dynamic addresses. |
| 380 | If set to a non-zero value larger than 1, a kernel log |
| 381 | message will be printed when dynamic address rewriting |
| 382 | occurs. |
| 383 | Default: 0 |
| 384 | |
| 385 | icmp_echo_ignore_all - BOOLEAN |
David S. Miller | 7ce31246 | 2005-10-03 16:07:30 -0700 | [diff] [blame] | 386 | If set non-zero, then the kernel will ignore all ICMP ECHO |
| 387 | requests sent to it. |
| 388 | Default: 0 |
| 389 | |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 390 | icmp_echo_ignore_broadcasts - BOOLEAN |
David S. Miller | 7ce31246 | 2005-10-03 16:07:30 -0700 | [diff] [blame] | 391 | If set non-zero, then the kernel will ignore all ICMP ECHO and |
| 392 | TIMESTAMP requests sent to it via broadcast/multicast. |
| 393 | Default: 1 |
Linus Torvalds | 1da177e | 2005-04-16 15:20:36 -0700 | [diff] [blame] | 394 | |
| 395 | icmp_ratelimit - INTEGER |
| 396 | Limit the maximal rates for sending ICMP packets whose type matches |
| 397 | icmp_ratemask (see below) to specific targets. |
| 398 | 0 to disable any limiting, otherwise the maximal rate in jiffies(1) |
| 399 | Default: 100 |
| 400 | |
| 401 | icmp_ratemask - INTEGER |
| 402 | Mask made of ICMP types for which rates are being limited. |
| 403 | Significant bits: IHGFEDCBA9876543210 |
| 404 | Default mask: 0000001100000011000 (6168) |
| 405 | |
| 406 | Bit definitions (see include/linux/icmp.h): |
| 407 | 0 Echo Reply |
| 408 | 3 Destination Unreachable * |
| 409 | 4 Source Quench * |
| 410 | 5 Redirect |
| 411 | 8 Echo Request |
| 412 | B Time Exceeded * |
| 413 | C Parameter Problem * |
| 414 | D Timestamp Request |
| 415 | E Timestamp Reply |
| 416 | F Info Request |
| 417 | G Info Reply |
| 418 | H Address Mask Request |
| 419 | I Address Mask Reply |
| 420 | |
| 421 | * These are rate limited by default (see default mask above) |
| 422 | |
| 423 | icmp_ignore_bogus_error_responses - BOOLEAN |
| 424 | Some routers violate RFC1122 by sending bogus responses to broadcast |
| 425 | frames. Such violations are normally logged via a kernel warning. |
| 426 | If this is set to TRUE, the kernel will not give such warnings, which |
| 427 | will avoid log file clutter. |
| 428 | Default: FALSE |
| 429 | |
| 430 | igmp_max_memberships - INTEGER |
| 431 | Change the maximum number of multicast groups we can subscribe to. |
| 432 | Default: 20 |
| 433 | |
| 434 | conf/interface/* changes special settings per interface (where "interface" is |
| 435 | the name of your network interface) |
| 436 | conf/all/* is special, changes the settings for all interfaces |
| 437 | |
| 438 | |
| 439 | log_martians - BOOLEAN |
| 440 | Log packets with impossible addresses to kernel log. |
| 441 | log_martians for the interface will be enabled if at least one of |
| 442 | conf/{all,interface}/log_martians is set to TRUE, |
| 443 | it will be disabled otherwise |
| 444 | |
| 445 | accept_redirects - BOOLEAN |
| 446 | Accept ICMP redirect messages. |
| 447 | accept_redirects for the interface will be enabled if: |
| 448 | - both conf/{all,interface}/accept_redirects are TRUE in the case forwarding |
| 449 | for the interface is enabled |
| 450 | or |
| 451 | - at least one of conf/{all,interface}/accept_redirects is TRUE in the case |
| 452 | forwarding for the interface is disabled |
| 453 | accept_redirects for the interface will be disabled otherwise |
| 454 | default TRUE (host) |
| 455 | FALSE (router) |
| 456 | |
| 457 | forwarding - BOOLEAN |
| 458 | Enable IP forwarding on this interface. |
| 459 | |
| 460 | mc_forwarding - BOOLEAN |
| 461 | Do multicast routing. The kernel needs to be compiled with CONFIG_MROUTE |
| 462 | and a multicast routing daemon is required. |
| 463 | conf/all/mc_forwarding must also be set to TRUE to enable multicast routing |
| 464 | for the interface |
| 465 | |
| 466 | medium_id - INTEGER |
| 467 | Integer value used to differentiate the devices by the medium they |
| 468 | are attached to. Two devices can have different id values when |
| 469 | the broadcast packets are received only on one of them. |
| 470 | The default value 0 means that the device is the only interface |
| 471 | to its medium, value of -1 means that medium is not known. |
| 472 | |
| 473 | Currently, it is used to change the proxy_arp behavior: |
| 474 | the proxy_arp feature is enabled for packets forwarded between |
| 475 | two devices attached to different media. |
| 476 | |
| 477 | proxy_arp - BOOLEAN |
| 478 | Do proxy arp. |
| 479 | proxy_arp for the interface will be enabled if at least one of |
| 480 | conf/{all,interface}/proxy_arp is set to TRUE, |
| 481 | it will be disabled otherwise |
| 482 | |
| 483 | shared_media - BOOLEAN |
| 484 | Send(router) or accept(host) RFC1620 shared media redirects. |
| 485 | Overrides ip_secure_redirects. |
| 486 | shared_media for the interface will be enabled if at least one of |
| 487 | conf/{all,interface}/shared_media is set to TRUE, |
| 488 | it will be disabled otherwise |
| 489 | default TRUE |
| 490 | |
| 491 | secure_redirects - BOOLEAN |
| 492 | Accept ICMP redirect messages only for gateways, |
| 493 | listed in default gateway list. |
| 494 | secure_redirects for the interface will be enabled if at least one of |
| 495 | conf/{all,interface}/secure_redirects is set to TRUE, |
| 496 | it will be disabled otherwise |
| 497 | default TRUE |
| 498 | |
| 499 | send_redirects - BOOLEAN |
| 500 | Send redirects, if router. |
| 501 | send_redirects for the interface will be enabled if at least one of |
| 502 | conf/{all,interface}/send_redirects is set to TRUE, |
| 503 | it will be disabled otherwise |
| 504 | Default: TRUE |
| 505 | |
| 506 | bootp_relay - BOOLEAN |
| 507 | Accept packets with source address 0.b.c.d destined |
| 508 | not to this host as local ones. It is supposed, that |
| 509 | BOOTP relay daemon will catch and forward such packets. |
| 510 | conf/all/bootp_relay must also be set to TRUE to enable BOOTP relay |
| 511 | for the interface |
| 512 | default FALSE |
| 513 | Not Implemented Yet. |
| 514 | |
| 515 | accept_source_route - BOOLEAN |
| 516 | Accept packets with SRR option. |
| 517 | conf/all/accept_source_route must also be set to TRUE to accept packets |
| 518 | with SRR option on the interface |
| 519 | default TRUE (router) |
| 520 | FALSE (host) |
| 521 | |
| 522 | rp_filter - BOOLEAN |
| 523 | 1 - do source validation by reversed path, as specified in RFC1812 |
| 524 | Recommended option for single homed hosts and stub network |
| 525 | routers. Could cause troubles for complicated (not loop free) |
| 526 | networks running a slow unreliable protocol (sort of RIP), |
| 527 | or using static routes. |
| 528 | |
| 529 | 0 - No source validation. |
| 530 | |
| 531 | conf/all/rp_filter must also be set to TRUE to do source validation |
| 532 | on the interface |
| 533 | |
| 534 | Default value is 0. Note that some distributions enable it |
| 535 | in startup scripts. |
| 536 | |
| 537 | arp_filter - BOOLEAN |
| 538 | 1 - Allows you to have multiple network interfaces on the same |
| 539 | subnet, and have the ARPs for each interface be answered |
| 540 | based on whether or not the kernel would route a packet from |
| 541 | the ARP'd IP out that interface (therefore you must use source |
| 542 | based routing for this to work). In other words it allows control |
| 543 | of which cards (usually 1) will respond to an arp request. |
| 544 | |
| 545 | 0 - (default) The kernel can respond to arp requests with addresses |
| 546 | from other interfaces. This may seem wrong but it usually makes |
| 547 | sense, because it increases the chance of successful communication. |
| 548 | IP addresses are owned by the complete host on Linux, not by |
| 549 | particular interfaces. Only for more complex setups like load- |
| 550 | balancing, does this behaviour cause problems. |
| 551 | |
| 552 | arp_filter for the interface will be enabled if at least one of |
| 553 | conf/{all,interface}/arp_filter is set to TRUE, |
| 554 | it will be disabled otherwise |
| 555 | |
| 556 | arp_announce - INTEGER |
| 557 | Define different restriction levels for announcing the local |
| 558 | source IP address from IP packets in ARP requests sent on |
| 559 | interface: |
| 560 | 0 - (default) Use any local address, configured on any interface |
| 561 | 1 - Try to avoid local addresses that are not in the target's |
| 562 | subnet for this interface. This mode is useful when target |
| 563 | hosts reachable via this interface require the source IP |
| 564 | address in ARP requests to be part of their logical network |
| 565 | configured on the receiving interface. When we generate the |
| 566 | request we will check all our subnets that include the |
| 567 | target IP and will preserve the source address if it is from |
| 568 | such subnet. If there is no such subnet we select source |
| 569 | address according to the rules for level 2. |
| 570 | 2 - Always use the best local address for this target. |
| 571 | In this mode we ignore the source address in the IP packet |
| 572 | and try to select local address that we prefer for talks with |
| 573 | the target host. Such local address is selected by looking |
| 574 | for primary IP addresses on all our subnets on the outgoing |
| 575 | interface that include the target IP address. If no suitable |
| 576 | local address is found we select the first local address |
| 577 | we have on the outgoing interface or on all other interfaces, |
| 578 | with the hope we will receive reply for our request and |
| 579 | even sometimes no matter the source IP address we announce. |
| 580 | |
| 581 | The max value from conf/{all,interface}/arp_announce is used. |
| 582 | |
| 583 | Increasing the restriction level gives more chance for |
| 584 | receiving answer from the resolved target while decreasing |
| 585 | the level announces more valid sender's information. |
| 586 | |
| 587 | arp_ignore - INTEGER |
| 588 | Define different modes for sending replies in response to |
| 589 | received ARP requests that resolve local target IP addresses: |
| 590 | 0 - (default): reply for any local target IP address, configured |
| 591 | on any interface |
| 592 | 1 - reply only if the target IP address is local address |
| 593 | configured on the incoming interface |
| 594 | 2 - reply only if the target IP address is local address |
| 595 | configured on the incoming interface and both with the |
| 596 | sender's IP address are part from same subnet on this interface |
| 597 | 3 - do not reply for local addresses configured with scope host, |
| 598 | only resolutions for global and link addresses are replied |
| 599 | 4-7 - reserved |
| 600 | 8 - do not reply for all local addresses |
| 601 | |
| 602 | The max value from conf/{all,interface}/arp_ignore is used |
| 603 | when ARP request is received on the {interface} |
| 604 | |
| 605 | app_solicit - INTEGER |
| 606 | The maximum number of probes to send to the user space ARP daemon |
| 607 | via netlink before dropping back to multicast probes (see |
| 608 | mcast_solicit). Defaults to 0. |
| 609 | |
| 610 | disable_policy - BOOLEAN |
| 611 | Disable IPSEC policy (SPD) for this interface |
| 612 | |
| 613 | disable_xfrm - BOOLEAN |
| 614 | Disable IPSEC encryption on this interface, whatever the policy |
| 615 | |
| 616 | |
| 617 | |
| 618 | tag - INTEGER |
| 619 | Allows you to write a number, which can be used as required. |
| 620 | Default value is 0. |
| 621 | |
| 622 | (1) Jiffie: internal timeunit for the kernel. On the i386 1/100s, on the |
| 623 | Alpha 1/1024s. See the HZ define in /usr/include/asm/param.h for the exact |
| 624 | value on your system. |
| 625 | |
| 626 | Alexey Kuznetsov. |
| 627 | kuznet@ms2.inr.ac.ru |
| 628 | |
| 629 | Updated by: |
| 630 | Andi Kleen |
| 631 | ak@muc.de |
| 632 | Nicolas Delon |
| 633 | delon.nicolas@wanadoo.fr |
| 634 | |
| 635 | |
| 636 | |
| 637 | |
| 638 | /proc/sys/net/ipv6/* Variables: |
| 639 | |
| 640 | IPv6 has no global variables such as tcp_*. tcp_* settings under ipv4/ also |
| 641 | apply to IPv6 [XXX?]. |
| 642 | |
| 643 | bindv6only - BOOLEAN |
| 644 | Default value for IPV6_V6ONLY socket option, |
| 645 | which restricts use of the IPv6 socket to IPv6 communication |
| 646 | only. |
| 647 | TRUE: disable IPv4-mapped address feature |
| 648 | FALSE: enable IPv4-mapped address feature |
| 649 | |
| 650 | Default: FALSE (as specified in RFC2553bis) |
| 651 | |
| 652 | IPv6 Fragmentation: |
| 653 | |
| 654 | ip6frag_high_thresh - INTEGER |
| 655 | Maximum memory used to reassemble IPv6 fragments. When |
| 656 | ip6frag_high_thresh bytes of memory is allocated for this purpose, |
| 657 | the fragment handler will toss packets until ip6frag_low_thresh |
| 658 | is reached. |
| 659 | |
| 660 | ip6frag_low_thresh - INTEGER |
| 661 | See ip6frag_high_thresh |
| 662 | |
| 663 | ip6frag_time - INTEGER |
| 664 | Time in seconds to keep an IPv6 fragment in memory. |
| 665 | |
| 666 | ip6frag_secret_interval - INTEGER |
| 667 | Regeneration interval (in seconds) of the hash secret (or lifetime |
| 668 | for the hash secret) for IPv6 fragments. |
| 669 | Default: 600 |
| 670 | |
| 671 | conf/default/*: |
| 672 | Change the interface-specific default settings. |
| 673 | |
| 674 | |
| 675 | conf/all/*: |
| 676 | Change all the interface-specific settings. |
| 677 | |
| 678 | [XXX: Other special features than forwarding?] |
| 679 | |
| 680 | conf/all/forwarding - BOOLEAN |
| 681 | Enable global IPv6 forwarding between all interfaces. |
| 682 | |
| 683 | IPv4 and IPv6 work differently here; e.g. netfilter must be used |
| 684 | to control which interfaces may forward packets and which not. |
| 685 | |
| 686 | This also sets all interfaces' Host/Router setting |
| 687 | 'forwarding' to the specified value. See below for details. |
| 688 | |
| 689 | This referred to as global forwarding. |
| 690 | |
| 691 | conf/interface/*: |
| 692 | Change special settings per interface. |
| 693 | |
| 694 | The functional behaviour for certain settings is different |
| 695 | depending on whether local forwarding is enabled or not. |
| 696 | |
| 697 | accept_ra - BOOLEAN |
| 698 | Accept Router Advertisements; autoconfigure using them. |
| 699 | |
| 700 | Functional default: enabled if local forwarding is disabled. |
| 701 | disabled if local forwarding is enabled. |
| 702 | |
| 703 | accept_redirects - BOOLEAN |
| 704 | Accept Redirects. |
| 705 | |
| 706 | Functional default: enabled if local forwarding is disabled. |
| 707 | disabled if local forwarding is enabled. |
| 708 | |
| 709 | autoconf - BOOLEAN |
| 710 | Autoconfigure addresses using Prefix Information in Router |
| 711 | Advertisements. |
| 712 | |
| 713 | Functional default: enabled if accept_ra is enabled. |
| 714 | disabled if accept_ra is disabled. |
| 715 | |
| 716 | dad_transmits - INTEGER |
| 717 | The amount of Duplicate Address Detection probes to send. |
| 718 | Default: 1 |
| 719 | |
| 720 | forwarding - BOOLEAN |
| 721 | Configure interface-specific Host/Router behaviour. |
| 722 | |
| 723 | Note: It is recommended to have the same setting on all |
| 724 | interfaces; mixed router/host scenarios are rather uncommon. |
| 725 | |
| 726 | FALSE: |
| 727 | |
| 728 | By default, Host behaviour is assumed. This means: |
| 729 | |
| 730 | 1. IsRouter flag is not set in Neighbour Advertisements. |
| 731 | 2. Router Solicitations are being sent when necessary. |
| 732 | 3. If accept_ra is TRUE (default), accept Router |
| 733 | Advertisements (and do autoconfiguration). |
| 734 | 4. If accept_redirects is TRUE (default), accept Redirects. |
| 735 | |
| 736 | TRUE: |
| 737 | |
| 738 | If local forwarding is enabled, Router behaviour is assumed. |
| 739 | This means exactly the reverse from the above: |
| 740 | |
| 741 | 1. IsRouter flag is set in Neighbour Advertisements. |
| 742 | 2. Router Solicitations are not sent. |
| 743 | 3. Router Advertisements are ignored. |
| 744 | 4. Redirects are ignored. |
| 745 | |
| 746 | Default: FALSE if global forwarding is disabled (default), |
| 747 | otherwise TRUE. |
| 748 | |
| 749 | hop_limit - INTEGER |
| 750 | Default Hop Limit to set. |
| 751 | Default: 64 |
| 752 | |
| 753 | mtu - INTEGER |
| 754 | Default Maximum Transfer Unit |
| 755 | Default: 1280 (IPv6 required minimum) |
| 756 | |
| 757 | router_solicitation_delay - INTEGER |
| 758 | Number of seconds to wait after interface is brought up |
| 759 | before sending Router Solicitations. |
| 760 | Default: 1 |
| 761 | |
| 762 | router_solicitation_interval - INTEGER |
| 763 | Number of seconds to wait between Router Solicitations. |
| 764 | Default: 4 |
| 765 | |
| 766 | router_solicitations - INTEGER |
| 767 | Number of Router Solicitations to send until assuming no |
| 768 | routers are present. |
| 769 | Default: 3 |
| 770 | |
| 771 | use_tempaddr - INTEGER |
| 772 | Preference for Privacy Extensions (RFC3041). |
| 773 | <= 0 : disable Privacy Extensions |
| 774 | == 1 : enable Privacy Extensions, but prefer public |
| 775 | addresses over temporary addresses. |
| 776 | > 1 : enable Privacy Extensions and prefer temporary |
| 777 | addresses over public addresses. |
| 778 | Default: 0 (for most devices) |
| 779 | -1 (for point-to-point devices and loopback devices) |
| 780 | |
| 781 | temp_valid_lft - INTEGER |
| 782 | valid lifetime (in seconds) for temporary addresses. |
| 783 | Default: 604800 (7 days) |
| 784 | |
| 785 | temp_prefered_lft - INTEGER |
| 786 | Preferred lifetime (in seconds) for temporary addresses. |
| 787 | Default: 86400 (1 day) |
| 788 | |
| 789 | max_desync_factor - INTEGER |
| 790 | Maximum value for DESYNC_FACTOR, which is a random value |
| 791 | that ensures that clients don't synchronize with each |
| 792 | other and generate new addresses at exactly the same time. |
| 793 | value is in seconds. |
| 794 | Default: 600 |
| 795 | |
| 796 | regen_max_retry - INTEGER |
| 797 | Number of attempts before give up attempting to generate |
| 798 | valid temporary addresses. |
| 799 | Default: 5 |
| 800 | |
| 801 | max_addresses - INTEGER |
| 802 | Number of maximum addresses per interface. 0 disables limitation. |
| 803 | It is recommended not set too large value (or 0) because it would |
| 804 | be too easy way to crash kernel to allow to create too much of |
| 805 | autoconfigured addresses. |
| 806 | Default: 16 |
| 807 | |
| 808 | icmp/*: |
| 809 | ratelimit - INTEGER |
| 810 | Limit the maximal rates for sending ICMPv6 packets. |
| 811 | 0 to disable any limiting, otherwise the maximal rate in jiffies(1) |
| 812 | Default: 100 |
| 813 | |
| 814 | |
| 815 | IPv6 Update by: |
| 816 | Pekka Savola <pekkas@netcore.fi> |
| 817 | YOSHIFUJI Hideaki / USAGI Project <yoshfuji@linux-ipv6.org> |
| 818 | |
| 819 | |
| 820 | /proc/sys/net/bridge/* Variables: |
| 821 | |
| 822 | bridge-nf-call-arptables - BOOLEAN |
| 823 | 1 : pass bridged ARP traffic to arptables' FORWARD chain. |
| 824 | 0 : disable this. |
| 825 | Default: 1 |
| 826 | |
| 827 | bridge-nf-call-iptables - BOOLEAN |
| 828 | 1 : pass bridged IPv4 traffic to iptables' chains. |
| 829 | 0 : disable this. |
| 830 | Default: 1 |
| 831 | |
| 832 | bridge-nf-call-ip6tables - BOOLEAN |
| 833 | 1 : pass bridged IPv6 traffic to ip6tables' chains. |
| 834 | 0 : disable this. |
| 835 | Default: 1 |
| 836 | |
| 837 | bridge-nf-filter-vlan-tagged - BOOLEAN |
| 838 | 1 : pass bridged vlan-tagged ARP/IP traffic to arptables/iptables. |
| 839 | 0 : disable this. |
| 840 | Default: 1 |
| 841 | |
| 842 | |
| 843 | UNDOCUMENTED: |
| 844 | |
| 845 | dev_weight FIXME |
| 846 | discovery_slots FIXME |
| 847 | discovery_timeout FIXME |
| 848 | fast_poll_increase FIXME |
| 849 | ip6_queue_maxlen FIXME |
| 850 | lap_keepalive_time FIXME |
| 851 | lo_cong FIXME |
| 852 | max_baud_rate FIXME |
| 853 | max_dgram_qlen FIXME |
| 854 | max_noreply_time FIXME |
| 855 | max_tx_data_size FIXME |
| 856 | max_tx_window FIXME |
| 857 | min_tx_turn_time FIXME |
| 858 | mod_cong FIXME |
| 859 | no_cong FIXME |
| 860 | no_cong_thresh FIXME |
| 861 | slot_timeout FIXME |
| 862 | warn_noreply_time FIXME |
| 863 | |
| 864 | $Id: ip-sysctl.txt,v 1.20 2001/12/13 09:00:18 davem Exp $ |