5 bool "IP: multicasting"
7 This is code for addressing several networked computers at once,
8 enlarging your kernel by about 2 KB. You need multicasting if you
9 intend to participate in the MBONE, a high bandwidth network on top
10 of the Internet which carries audio and video broadcasts. More
11 information about the MBONE is on the WWW at
12 <http://www-itg.lbl.gov/mbone/>. Information about the multicast
13 capabilities of the various network cards is contained in
14 <file:Documentation/networking/multicast.txt>. For most people, it's
17 config IP_ADVANCED_ROUTER
18 bool "IP: advanced router"
20 If you intend to run your Linux box mostly as a router, i.e. as a
21 computer that forwards and redistributes network packets, say Y; you
22 will then be presented with several options that allow more precise
23 control about the routing process.
25 The answer to this question won't directly affect the kernel:
26 answering N will just cause the configurator to skip all the
27 questions about advanced routing.
29 Note that your box can only act as a router if you enable IP
30 forwarding in your kernel; you can do that by saying Y to "/proc
31 file system support" and "Sysctl support" below and executing the
34 echo "1" > /proc/sys/net/ipv4/ip_forward
36 at boot time after the /proc file system has been mounted.
38 If you turn on IP forwarding, you will also get the rp_filter, which
39 automatically rejects incoming packets if the routing table entry
40 for their source address doesn't match the network interface they're
41 arriving on. This has security advantages because it prevents the
42 so-called IP spoofing, however it can pose problems if you use
43 asymmetric routing (packets from you to a host take a different path
44 than packets from that host to you) or if you operate a non-routing
45 host which has several IP addresses on different interfaces. To turn
48 echo 0 > /proc/sys/net/ipv4/conf/<device>/rp_filter
50 echo 0 > /proc/sys/net/ipv4/conf/all/rp_filter
52 If unsure, say N here.
55 prompt "Choose IP: FIB lookup algorithm (choose FIB_HASH if unsure)"
56 depends on IP_ADVANCED_ROUTER
57 default ASK_IP_FIB_HASH
59 config ASK_IP_FIB_HASH
62 Current FIB is very proven and good enough for most users.
67 Use new experimental LC-trie as FIB lookup algoritm.
68 This improves lookup performance if you have a large
71 LC-trie is a longest matching prefix lookup algorithm which
72 performs better than FIB_HASH for large routing tables.
73 But, it consumes more memory and is more complex.
75 LC-trie is described in:
77 IP-address lookup using LC-tries. Stefan Nilsson and Gunnar Karlsson
78 IEEE Journal on Selected Areas in Communications, 17(6):1083-1092, June 1999
79 An experimental study of compression methods for dynamic tries
80 Stefan Nilsson and Matti Tikkanen. Algorithmica, 33(1):19-33, 2002.
81 http://www.nada.kth.se/~snilsson/public/papers/dyntrie2/
86 def_bool ASK_IP_FIB_HASH || !IP_ADVANCED_ROUTER
88 config IP_MULTIPLE_TABLES
89 bool "IP: policy routing"
90 depends on IP_ADVANCED_ROUTER
92 Normally, a router decides what to do with a received packet based
93 solely on the packet's final destination address. If you say Y here,
94 the Linux router will also be able to take the packet's source
95 address into account. Furthermore, the TOS (Type-Of-Service) field
96 of the packet can be used for routing decisions as well.
98 If you are interested in this, please see the preliminary
99 documentation at <http://www.compendium.com.ar/policy-routing.txt>
100 and <ftp://post.tepkom.ru/pub/vol2/Linux/docs/advanced-routing.tex>.
101 You will need supporting software from
102 <ftp://ftp.tux.org/pub/net/ip-routing/>.
106 config IP_ROUTE_FWMARK
107 bool "IP: use netfilter MARK value as routing key"
108 depends on IP_MULTIPLE_TABLES && NETFILTER
110 If you say Y here, you will be able to specify different routes for
111 packets with different mark values (see iptables(8), MARK target).
113 config IP_ROUTE_MULTIPATH
114 bool "IP: equal cost multipath"
115 depends on IP_ADVANCED_ROUTER
117 Normally, the routing tables specify a single action to be taken in
118 a deterministic manner for a given packet. If you say Y here
119 however, it becomes possible to attach several actions to a packet
120 pattern, in effect specifying several alternative paths to travel
121 for those packets. The router considers all these paths to be of
122 equal "cost" and chooses one of them in a non-deterministic fashion
123 if a matching packet arrives.
125 config IP_ROUTE_MULTIPATH_CACHED
126 bool "IP: equal cost multipath with caching support (EXPERIMENTAL)"
127 depends on IP_ROUTE_MULTIPATH
129 Normally, equal cost multipath routing is not supported by the
130 routing cache. If you say Y here, alternative routes are cached
131 and on cache lookup a route is chosen in a configurable fashion.
135 config IP_ROUTE_MULTIPATH_RR
136 tristate "MULTIPATH: round robin algorithm"
137 depends on IP_ROUTE_MULTIPATH_CACHED
139 Mulitpath routes are chosen according to Round Robin
141 config IP_ROUTE_MULTIPATH_RANDOM
142 tristate "MULTIPATH: random algorithm"
143 depends on IP_ROUTE_MULTIPATH_CACHED
145 Multipath routes are chosen in a random fashion. Actually,
146 there is no weight for a route. The advantage of this policy
147 is that it is implemented stateless and therefore introduces only
150 config IP_ROUTE_MULTIPATH_WRANDOM
151 tristate "MULTIPATH: weighted random algorithm"
152 depends on IP_ROUTE_MULTIPATH_CACHED
154 Multipath routes are chosen in a weighted random fashion.
155 The per route weights are the weights visible via ip route 2. As the
156 corresponding state management introduces some overhead routing delay
159 config IP_ROUTE_MULTIPATH_DRR
160 tristate "MULTIPATH: interface round robin algorithm"
161 depends on IP_ROUTE_MULTIPATH_CACHED
163 Connections are distributed in a round robin fashion over the
164 available interfaces. This policy makes sense if the connections
165 should be primarily distributed on interfaces and not on routes.
167 config IP_ROUTE_VERBOSE
168 bool "IP: verbose route monitoring"
169 depends on IP_ADVANCED_ROUTER
171 If you say Y here, which is recommended, then the kernel will print
172 verbose messages regarding the routing, for example warnings about
173 received packets which look strange and could be evidence of an
174 attack or a misconfigured system somewhere. The information is
175 handled by the klogd daemon which is responsible for kernel messages
179 bool "IP: kernel level autoconfiguration"
181 This enables automatic configuration of IP addresses of devices and
182 of the routing table during kernel boot, based on either information
183 supplied on the kernel command line or by BOOTP or RARP protocols.
184 You need to say Y only for diskless machines requiring network
185 access to boot (in which case you want to say Y to "Root file system
186 on NFS" as well), because all other machines configure the network
187 in their startup scripts.
190 bool "IP: DHCP support"
193 If you want your Linux box to mount its whole root file system (the
194 one containing the directory /) from some other computer over the
195 net via NFS and you want the IP address of your computer to be
196 discovered automatically at boot time using the DHCP protocol (a
197 special protocol designed for doing this job), say Y here. In case
198 the boot ROM of your network card was designed for booting Linux and
199 does DHCP itself, providing all necessary information on the kernel
200 command line, you can say N here.
202 If unsure, say Y. Note that if you want to use DHCP, a DHCP server
203 must be operating on your network. Read
204 <file:Documentation/nfsroot.txt> for details.
207 bool "IP: BOOTP support"
210 If you want your Linux box to mount its whole root file system (the
211 one containing the directory /) from some other computer over the
212 net via NFS and you want the IP address of your computer to be
213 discovered automatically at boot time using the BOOTP protocol (a
214 special protocol designed for doing this job), say Y here. In case
215 the boot ROM of your network card was designed for booting Linux and
216 does BOOTP itself, providing all necessary information on the kernel
217 command line, you can say N here. If unsure, say Y. Note that if you
218 want to use BOOTP, a BOOTP server must be operating on your network.
219 Read <file:Documentation/nfsroot.txt> for details.
222 bool "IP: RARP support"
225 If you want your Linux box to mount its whole root file system (the
226 one containing the directory /) from some other computer over the
227 net via NFS and you want the IP address of your computer to be
228 discovered automatically at boot time using the RARP protocol (an
229 older protocol which is being obsoleted by BOOTP and DHCP), say Y
230 here. Note that if you want to use RARP, a RARP server must be
231 operating on your network. Read <file:Documentation/nfsroot.txt> for
235 # bool ' IP: ARP support' CONFIG_IP_PNP_ARP
237 tristate "IP: tunneling"
239 Tunneling means encapsulating data of one protocol type within
240 another protocol and sending it over a channel that understands the
241 encapsulating protocol. This particular tunneling driver implements
242 encapsulation of IP within IP, which sounds kind of pointless, but
243 can be useful if you want to make your (or some other) machine
244 appear on a different network than it physically is, or to use
245 mobile-IP facilities (allowing laptops to seamlessly move between
246 networks without changing their IP addresses).
248 Saying Y to this option will produce two modules ( = code which can
249 be inserted in and removed from the running kernel whenever you
250 want). Most people won't need this and can say N.
253 tristate "IP: GRE tunnels over IP"
255 Tunneling means encapsulating data of one protocol type within
256 another protocol and sending it over a channel that understands the
257 encapsulating protocol. This particular tunneling driver implements
258 GRE (Generic Routing Encapsulation) and at this time allows
259 encapsulating of IPv4 or IPv6 over existing IPv4 infrastructure.
260 This driver is useful if the other endpoint is a Cisco router: Cisco
261 likes GRE much better than the other Linux tunneling driver ("IP
262 tunneling" above). In addition, GRE allows multicast redistribution
265 config NET_IPGRE_BROADCAST
266 bool "IP: broadcast GRE over IP"
267 depends on IP_MULTICAST && NET_IPGRE
269 One application of GRE/IP is to construct a broadcast WAN (Wide Area
270 Network), which looks like a normal Ethernet LAN (Local Area
271 Network), but can be distributed all over the Internet. If you want
272 to do that, say Y here and to "IP multicast routing" below.
275 bool "IP: multicast routing"
276 depends on IP_MULTICAST
278 This is used if you want your machine to act as a router for IP
279 packets that have several destination addresses. It is needed on the
280 MBONE, a high bandwidth network on top of the Internet which carries
281 audio and video broadcasts. In order to do that, you would most
282 likely run the program mrouted. Information about the multicast
283 capabilities of the various network cards is contained in
284 <file:Documentation/networking/multicast.txt>. If you haven't heard
285 about it, you don't need it.
288 bool "IP: PIM-SM version 1 support"
291 Kernel side support for Sparse Mode PIM (Protocol Independent
292 Multicast) version 1. This multicast routing protocol is used widely
293 because Cisco supports it. You need special software to use it
294 (pimd-v1). Please see <http://netweb.usc.edu/pim/> for more
295 information about PIM.
297 Say Y if you want to use PIM-SM v1. Note that you can say N here if
298 you just want to use Dense Mode PIM.
301 bool "IP: PIM-SM version 2 support"
304 Kernel side support for Sparse Mode PIM version 2. In order to use
305 this, you need an experimental routing daemon supporting it (pimd or
306 gated-5). This routing protocol is not used widely, so say N unless
307 you want to play with it.
310 bool "IP: ARP daemon support (EXPERIMENTAL)"
311 depends on EXPERIMENTAL
313 Normally, the kernel maintains an internal cache which maps IP
314 addresses to hardware addresses on the local network, so that
315 Ethernet/Token Ring/ etc. frames are sent to the proper address on
316 the physical networking layer. For small networks having a few
317 hundred directly connected hosts or less, keeping this address
318 resolution (ARP) cache inside the kernel works well. However,
319 maintaining an internal ARP cache does not work well for very large
320 switched networks, and will use a lot of kernel memory if TCP/IP
321 connections are made to many machines on the network.
323 If you say Y here, the kernel's internal ARP cache will never grow
324 to more than 256 entries (the oldest entries are expired in a LIFO
325 manner) and communication will be attempted with the user space ARP
326 daemon arpd. Arpd then answers the address resolution request either
327 from its own cache or by asking the net.
329 This code is experimental and also obsolete. If you want to use it,
330 you need to find a version of the daemon arpd on the net somewhere,
331 and you should also say Y to "Kernel/User network link driver",
332 below. If unsure, say N.
335 bool "IP: TCP syncookie support (disabled per default)"
337 Normal TCP/IP networking is open to an attack known as "SYN
338 flooding". This denial-of-service attack prevents legitimate remote
339 users from being able to connect to your computer during an ongoing
340 attack and requires very little work from the attacker, who can
341 operate from anywhere on the Internet.
343 SYN cookies provide protection against this type of attack. If you
344 say Y here, the TCP/IP stack will use a cryptographic challenge
345 protocol known as "SYN cookies" to enable legitimate users to
346 continue to connect, even when your machine is under attack. There
347 is no need for the legitimate users to change their TCP/IP software;
348 SYN cookies work transparently to them. For technical information
349 about SYN cookies, check out <http://cr.yp.to/syncookies.html>.
351 If you are SYN flooded, the source address reported by the kernel is
352 likely to have been forged by the attacker; it is only reported as
353 an aid in tracing the packets to their actual source and should not
354 be taken as absolute truth.
356 SYN cookies may prevent correct error reporting on clients when the
357 server is really overloaded. If this happens frequently better turn
360 If you say Y here, note that SYN cookies aren't enabled by default;
361 you can enable them by saying Y to "/proc file system support" and
362 "Sysctl support" below and executing the command
364 echo 1 >/proc/sys/net/ipv4/tcp_syncookies
366 at boot time after the /proc file system has been mounted.
371 tristate "IP: AH transformation"
378 Support for IPsec AH.
383 tristate "IP: ESP transformation"
391 Support for IPsec ESP.
396 tristate "IP: IPComp transformation"
400 select CRYPTO_DEFLATE
402 Support for IP Payload Compression Protocol (IPComp) (RFC3173),
403 typically needed for IPsec.
408 tristate "IP: tunnel transformation"
411 Support for generic IP tunnel transformation, which is required by
412 the IP tunneling module as well as tunnel mode IPComp.
417 tristate "INET: socket monitoring interface"
420 Support for INET (TCP, DCCP, etc) socket monitoring interface used by
421 native Linux tools such as ss. ss is included in iproute2, currently
422 downloadable at <http://developer.osdl.org/dev/iproute2>.
428 def_tristate INET_DIAG
430 config TCP_CONG_ADVANCED
431 bool "TCP: advanced congestion control"
433 Support for selection of various TCP congestion control
436 Nearly all users can safely say no here, and a safe default
437 selection will be made (BIC-TCP with new Reno as a fallback).
441 # TCP Reno is builtin (required as fallback)
442 menu "TCP congestion control"
443 depends on TCP_CONG_ADVANCED
446 tristate "Binary Increase Congestion (BIC) control"
449 BIC-TCP is a sender-side only change that ensures a linear RTT
450 fairness under large windows while offering both scalability and
451 bounded TCP-friendliness. The protocol combines two schemes
452 called additive increase and binary search increase. When the
453 congestion window is large, additive increase with a large
454 increment ensures linear RTT fairness as well as good
455 scalability. Under small congestion windows, binary search
456 increase provides TCP friendliness.
457 See http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/
459 config TCP_CONG_WESTWOOD
460 tristate "TCP Westwood+"
463 TCP Westwood+ is a sender-side only modification of the TCP Reno
464 protocol stack that optimizes the performance of TCP congestion
465 control. It is based on end-to-end bandwidth estimation to set
466 congestion window and slow start threshold after a congestion
467 episode. Using this estimation, TCP Westwood+ adaptively sets a
468 slow start threshold and a congestion window which takes into
469 account the bandwidth used at the time congestion is experienced.
470 TCP Westwood+ significantly increases fairness wrt TCP Reno in
471 wired networks and throughput over wireless links.
477 H-TCP is a send-side only modifications of the TCP Reno
478 protocol stack that optimizes the performance of TCP
479 congestion control for high speed network links. It uses a
480 modeswitch to change the alpha and beta parameters of TCP Reno
481 based on network conditions and in a way so as to be fair with
482 other Reno and H-TCP flows.
484 config TCP_CONG_HSTCP
485 tristate "High Speed TCP"
486 depends on EXPERIMENTAL
489 Sally Floyd's High Speed TCP (RFC 3649) congestion control.
490 A modification to TCP's congestion control mechanism for use
491 with large congestion windows. A table indicates how much to
492 increase the congestion window by when an ACK is received.
493 For more detail see http://www.icir.org/floyd/hstcp.html
495 config TCP_CONG_HYBLA
496 tristate "TCP-Hybla congestion control algorithm"
497 depends on EXPERIMENTAL
500 TCP-Hybla is a sender-side only change that eliminates penalization of
501 long-RTT, large-bandwidth connections, like when satellite legs are
502 involved, expecially when sharing a common bottleneck with normal
503 terrestrial connections.
505 config TCP_CONG_VEGAS
507 depends on EXPERIMENTAL
510 TCP Vegas is a sender-side only change to TCP that anticipates
511 the onset of congestion by estimating the bandwidth. TCP Vegas
512 adjusts the sending rate by modifying the congestion
513 window. TCP Vegas should provide less packet loss, but it is
514 not as aggressive as TCP Reno.
516 config TCP_CONG_SCALABLE
517 tristate "Scalable TCP"
518 depends on EXPERIMENTAL
521 Scalable TCP is a sender-side only change to TCP which uses a
522 MIMD congestion control algorithm which has some nice scaling
523 properties, though is known to have fairness issues.
524 See http://www-lce.eng.cam.ac.uk/~ctk21/scalable/
530 depends on !TCP_CONG_ADVANCED
533 source "net/ipv4/ipvs/Kconfig"