1 # SPDX-License-Identifier: GPL-2.0-only
6 bool "IP: multicasting"
8 This is code for addressing several networked computers at once,
9 enlarging your kernel by about 2 KB. You need multicasting if you
10 intend to participate in the MBONE, a high bandwidth network on top
11 of the Internet which carries audio and video broadcasts. More
12 information about the MBONE is on the WWW at
13 <https://www.savetz.com/mbone/>. For most people, it's safe to say N.
15 config IP_ADVANCED_ROUTER
16 bool "IP: advanced router"
18 If you intend to run your Linux box mostly as a router, i.e. as a
19 computer that forwards and redistributes network packets, say Y; you
20 will then be presented with several options that allow more precise
21 control about the routing process.
23 The answer to this question won't directly affect the kernel:
24 answering N will just cause the configurator to skip all the
25 questions about advanced routing.
27 Note that your box can only act as a router if you enable IP
28 forwarding in your kernel; you can do that by saying Y to "/proc
29 file system support" and "Sysctl support" below and executing the
32 echo "1" > /proc/sys/net/ipv4/ip_forward
34 at boot time after the /proc file system has been mounted.
36 If you turn on IP forwarding, you should consider the rp_filter, which
37 automatically rejects incoming packets if the routing table entry
38 for their source address doesn't match the network interface they're
39 arriving on. This has security advantages because it prevents the
40 so-called IP spoofing, however it can pose problems if you use
41 asymmetric routing (packets from you to a host take a different path
42 than packets from that host to you) or if you operate a non-routing
43 host which has several IP addresses on different interfaces. To turn
46 echo 1 > /proc/sys/net/ipv4/conf/<device>/rp_filter
48 echo 1 > /proc/sys/net/ipv4/conf/all/rp_filter
50 Note that some distributions enable it in startup scripts.
51 For details about rp_filter strict and loose mode read
52 <file:Documentation/networking/ip-sysctl.rst>.
54 If unsure, say N here.
56 config IP_FIB_TRIE_STATS
57 bool "FIB TRIE statistics"
58 depends on IP_ADVANCED_ROUTER
60 Keep track of statistics on structure of FIB TRIE table.
61 Useful for testing and measuring TRIE performance.
63 config IP_MULTIPLE_TABLES
64 bool "IP: policy routing"
65 depends on IP_ADVANCED_ROUTER
68 Normally, a router decides what to do with a received packet based
69 solely on the packet's final destination address. If you say Y here,
70 the Linux router will also be able to take the packet's source
71 address into account. Furthermore, the TOS (Type-Of-Service) field
72 of the packet can be used for routing decisions as well.
74 If you need more information, see the Linux Advanced
75 Routing and Traffic Control documentation at
76 <https://lartc.org/howto/lartc.rpdb.html>
80 config IP_ROUTE_MULTIPATH
81 bool "IP: equal cost multipath"
82 depends on IP_ADVANCED_ROUTER
84 Normally, the routing tables specify a single action to be taken in
85 a deterministic manner for a given packet. If you say Y here
86 however, it becomes possible to attach several actions to a packet
87 pattern, in effect specifying several alternative paths to travel
88 for those packets. The router considers all these paths to be of
89 equal "cost" and chooses one of them in a non-deterministic fashion
90 if a matching packet arrives.
92 config IP_ROUTE_VERBOSE
93 bool "IP: verbose route monitoring"
94 depends on IP_ADVANCED_ROUTER
96 If you say Y here, which is recommended, then the kernel will print
97 verbose messages regarding the routing, for example warnings about
98 received packets which look strange and could be evidence of an
99 attack or a misconfigured system somewhere. The information is
100 handled by the klogd daemon which is responsible for kernel messages
103 config IP_ROUTE_CLASSID
107 bool "IP: kernel level autoconfiguration"
109 This enables automatic configuration of IP addresses of devices and
110 of the routing table during kernel boot, based on either information
111 supplied on the kernel command line or by BOOTP or RARP protocols.
112 You need to say Y only for diskless machines requiring network
113 access to boot (in which case you want to say Y to "Root file system
114 on NFS" as well), because all other machines configure the network
115 in their startup scripts.
118 bool "IP: DHCP support"
121 If you want your Linux box to mount its whole root file system (the
122 one containing the directory /) from some other computer over the
123 net via NFS and you want the IP address of your computer to be
124 discovered automatically at boot time using the DHCP protocol (a
125 special protocol designed for doing this job), say Y here. In case
126 the boot ROM of your network card was designed for booting Linux and
127 does DHCP itself, providing all necessary information on the kernel
128 command line, you can say N here.
130 If unsure, say Y. Note that if you want to use DHCP, a DHCP server
131 must be operating on your network. Read
132 <file:Documentation/admin-guide/nfs/nfsroot.rst> for details.
135 bool "IP: BOOTP support"
138 If you want your Linux box to mount its whole root file system (the
139 one containing the directory /) from some other computer over the
140 net via NFS and you want the IP address of your computer to be
141 discovered automatically at boot time using the BOOTP protocol (a
142 special protocol designed for doing this job), say Y here. In case
143 the boot ROM of your network card was designed for booting Linux and
144 does BOOTP itself, providing all necessary information on the kernel
145 command line, you can say N here. If unsure, say Y. Note that if you
146 want to use BOOTP, a BOOTP server must be operating on your network.
147 Read <file:Documentation/admin-guide/nfs/nfsroot.rst> for details.
150 bool "IP: RARP support"
153 If you want your Linux box to mount its whole root file system (the
154 one containing the directory /) from some other computer over the
155 net via NFS and you want the IP address of your computer to be
156 discovered automatically at boot time using the RARP protocol (an
157 older protocol which is being obsoleted by BOOTP and DHCP), say Y
158 here. Note that if you want to use RARP, a RARP server must be
159 operating on your network. Read
160 <file:Documentation/admin-guide/nfs/nfsroot.rst> for details.
163 tristate "IP: tunneling"
167 Tunneling means encapsulating data of one protocol type within
168 another protocol and sending it over a channel that understands the
169 encapsulating protocol. This particular tunneling driver implements
170 encapsulation of IP within IP, which sounds kind of pointless, but
171 can be useful if you want to make your (or some other) machine
172 appear on a different network than it physically is, or to use
173 mobile-IP facilities (allowing laptops to seamlessly move between
174 networks without changing their IP addresses).
176 Saying Y to this option will produce two modules ( = code which can
177 be inserted in and removed from the running kernel whenever you
178 want). Most people won't need this and can say N.
180 config NET_IPGRE_DEMUX
181 tristate "IP: GRE demultiplexer"
183 This is helper module to demultiplex GRE packets on GRE version field criteria.
184 Required by ip_gre and pptp modules.
193 tristate "IP: GRE tunnels over IP"
194 depends on (IPV6 || IPV6=n) && NET_IPGRE_DEMUX
197 Tunneling means encapsulating data of one protocol type within
198 another protocol and sending it over a channel that understands the
199 encapsulating protocol. This particular tunneling driver implements
200 GRE (Generic Routing Encapsulation) and at this time allows
201 encapsulating of IPv4 or IPv6 over existing IPv4 infrastructure.
202 This driver is useful if the other endpoint is a Cisco router: Cisco
203 likes GRE much better than the other Linux tunneling driver ("IP
204 tunneling" above). In addition, GRE allows multicast redistribution
207 config NET_IPGRE_BROADCAST
208 bool "IP: broadcast GRE over IP"
209 depends on IP_MULTICAST && NET_IPGRE
211 One application of GRE/IP is to construct a broadcast WAN (Wide Area
212 Network), which looks like a normal Ethernet LAN (Local Area
213 Network), but can be distributed all over the Internet. If you want
214 to do that, say Y here and to "IP multicast routing" below.
216 config IP_MROUTE_COMMON
218 depends on IP_MROUTE || IPV6_MROUTE
221 bool "IP: multicast routing"
222 depends on IP_MULTICAST
223 select IP_MROUTE_COMMON
225 This is used if you want your machine to act as a router for IP
226 packets that have several destination addresses. It is needed on the
227 MBONE, a high bandwidth network on top of the Internet which carries
228 audio and video broadcasts. In order to do that, you would most
229 likely run the program mrouted. If you haven't heard about it, you
232 config IP_MROUTE_MULTIPLE_TABLES
233 bool "IP: multicast policy routing"
234 depends on IP_MROUTE && IP_ADVANCED_ROUTER
237 Normally, a multicast router runs a userspace daemon and decides
238 what to do with a multicast packet based on the source and
239 destination addresses. If you say Y here, the multicast router
240 will also be able to take interfaces and packet marks into
241 account and run multiple instances of userspace daemons
242 simultaneously, each one handling a single table.
247 bool "IP: PIM-SM version 1 support"
250 Kernel side support for Sparse Mode PIM (Protocol Independent
251 Multicast) version 1. This multicast routing protocol is used widely
252 because Cisco supports it. You need special software to use it
253 (pimd-v1). Please see <http://netweb.usc.edu/pim/> for more
254 information about PIM.
256 Say Y if you want to use PIM-SM v1. Note that you can say N here if
257 you just want to use Dense Mode PIM.
260 bool "IP: PIM-SM version 2 support"
263 Kernel side support for Sparse Mode PIM version 2. In order to use
264 this, you need an experimental routing daemon supporting it (pimd or
265 gated-5). This routing protocol is not used widely, so say N unless
266 you want to play with it.
269 bool "IP: TCP syncookie support"
271 Normal TCP/IP networking is open to an attack known as "SYN
272 flooding". This denial-of-service attack prevents legitimate remote
273 users from being able to connect to your computer during an ongoing
274 attack and requires very little work from the attacker, who can
275 operate from anywhere on the Internet.
277 SYN cookies provide protection against this type of attack. If you
278 say Y here, the TCP/IP stack will use a cryptographic challenge
279 protocol known as "SYN cookies" to enable legitimate users to
280 continue to connect, even when your machine is under attack. There
281 is no need for the legitimate users to change their TCP/IP software;
282 SYN cookies work transparently to them. For technical information
283 about SYN cookies, check out <https://cr.yp.to/syncookies.html>.
285 If you are SYN flooded, the source address reported by the kernel is
286 likely to have been forged by the attacker; it is only reported as
287 an aid in tracing the packets to their actual source and should not
288 be taken as absolute truth.
290 SYN cookies may prevent correct error reporting on clients when the
291 server is really overloaded. If this happens frequently better turn
294 If you say Y here, you can disable SYN cookies at run time by
295 saying Y to "/proc file system support" and
296 "Sysctl support" below and executing the command
298 echo 0 > /proc/sys/net/ipv4/tcp_syncookies
300 after the /proc file system has been mounted.
305 tristate "Virtual (secure) IP: tunneling"
306 depends on IPV6 || IPV6=n
311 Tunneling means encapsulating data of one protocol type within
312 another protocol and sending it over a channel that understands the
313 encapsulating protocol. This can be used with xfrm mode tunnel to give
314 the notion of a secure tunnel for IPSEC and then use routing protocol
317 config NET_UDP_TUNNEL
323 tristate "IP: Foo (IP protocols) over UDP"
324 select NET_UDP_TUNNEL
326 Foo over UDP allows any IP protocol to be directly encapsulated
327 over UDP include tunnels (IPIP, GRE, SIT). By encapsulating in UDP
328 network mechanisms and optimizations for UDP (such as ECMP
329 and RSS) can be leveraged to provide better service.
331 config NET_FOU_IP_TUNNELS
332 bool "IP: FOU encapsulation of IP tunnels"
333 depends on NET_IPIP || NET_IPGRE || IPV6_SIT
336 Allow configuration of FOU or GUE encapsulation for IP tunnels.
337 When this option is enabled IP tunnels can be configured to use
338 FOU or GUE encapsulation.
341 tristate "IP: AH transformation"
344 Support for IPsec AH (Authentication Header).
346 AH can be used with various authentication algorithms. Besides
347 enabling AH support itself, this option enables the generic
348 implementations of the algorithms that RFC 8221 lists as MUST be
349 implemented. If you need any other algorithms, you'll need to enable
350 them in the crypto API. You should also enable accelerated
351 implementations of any needed algorithms when available.
356 tristate "IP: ESP transformation"
359 Support for IPsec ESP (Encapsulating Security Payload).
361 ESP can be used with various encryption and authentication algorithms.
362 Besides enabling ESP support itself, this option enables the generic
363 implementations of the algorithms that RFC 8221 lists as MUST be
364 implemented. If you need any other algorithms, you'll need to enable
365 them in the crypto API. You should also enable accelerated
366 implementations of any needed algorithms when available.
370 config INET_ESP_OFFLOAD
371 tristate "IP: ESP transformation offload"
376 Support for ESP transformation offload. This makes sense
377 only if this system really does IPsec and want to do it
378 with high throughput. A typical desktop system does not
379 need it, even if it does IPsec.
384 bool "IP: ESP in TCP encapsulation (RFC 8229)"
385 depends on XFRM && INET_ESP
390 Support for RFC 8229 encapsulation of ESP and IKE over
396 tristate "IP: IPComp transformation"
397 select INET_XFRM_TUNNEL
400 Support for IP Payload Compression Protocol (IPComp) (RFC3173),
401 typically needed for IPsec.
405 config INET_TABLE_PERTURB_ORDER
406 int "INET: Source port perturbation table size (as power of 2)" if EXPERT
409 Source port perturbation table size (as power of 2) for
410 RFC 6056 3.3.4. Algorithm 4: Double-Hash Port Selection Algorithm.
412 The default is almost always what you want.
413 Only change this if you know what you are doing.
415 config INET_XFRM_TUNNEL
425 tristate "INET: socket monitoring interface"
428 Support for INET (TCP, DCCP, etc) socket monitoring interface used by
429 native Linux tools such as ss. ss is included in iproute2, currently
432 http://www.linuxfoundation.org/collaborate/workgroups/networking/iproute2
438 def_tristate INET_DIAG
441 tristate "UDP: socket monitoring interface"
442 depends on INET_DIAG && (IPV6 || IPV6=n)
445 Support for UDP socket monitoring interface used by the ss tool.
449 tristate "RAW: socket monitoring interface"
450 depends on INET_DIAG && (IPV6 || IPV6=n)
453 Support for RAW socket monitoring interface used by the ss tool.
456 config INET_DIAG_DESTROY
457 bool "INET: allow privileged process to administratively close sockets"
461 Provides a SOCK_DESTROY operation that allows privileged processes
462 (e.g., a connection manager or a network administration tool such as
463 ss) to close sockets opened by other processes. Closing a socket in
464 this way interrupts any blocking read/write/connect operations on
465 the socket and causes future socket calls to behave as if the socket
466 had been disconnected.
469 menuconfig TCP_CONG_ADVANCED
470 bool "TCP: advanced congestion control"
472 Support for selection of various TCP congestion control
475 Nearly all users can safely say no here, and a safe default
476 selection will be made (CUBIC with new Reno as a fallback).
483 tristate "Binary Increase Congestion (BIC) control"
486 BIC-TCP is a sender-side only change that ensures a linear RTT
487 fairness under large windows while offering both scalability and
488 bounded TCP-friendliness. The protocol combines two schemes
489 called additive increase and binary search increase. When the
490 congestion window is large, additive increase with a large
491 increment ensures linear RTT fairness as well as good
492 scalability. Under small congestion windows, binary search
493 increase provides TCP friendliness.
494 See http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/
496 config TCP_CONG_CUBIC
500 This is version 2.0 of BIC-TCP which uses a cubic growth function
501 among other techniques.
502 See http://www.csc.ncsu.edu/faculty/rhee/export/bitcp/cubic-paper.pdf
504 config TCP_CONG_WESTWOOD
505 tristate "TCP Westwood+"
508 TCP Westwood+ is a sender-side only modification of the TCP Reno
509 protocol stack that optimizes the performance of TCP congestion
510 control. It is based on end-to-end bandwidth estimation to set
511 congestion window and slow start threshold after a congestion
512 episode. Using this estimation, TCP Westwood+ adaptively sets a
513 slow start threshold and a congestion window which takes into
514 account the bandwidth used at the time congestion is experienced.
515 TCP Westwood+ significantly increases fairness wrt TCP Reno in
516 wired networks and throughput over wireless links.
522 H-TCP is a send-side only modifications of the TCP Reno
523 protocol stack that optimizes the performance of TCP
524 congestion control for high speed network links. It uses a
525 modeswitch to change the alpha and beta parameters of TCP Reno
526 based on network conditions and in a way so as to be fair with
527 other Reno and H-TCP flows.
529 config TCP_CONG_HSTCP
530 tristate "High Speed TCP"
533 Sally Floyd's High Speed TCP (RFC 3649) congestion control.
534 A modification to TCP's congestion control mechanism for use
535 with large congestion windows. A table indicates how much to
536 increase the congestion window by when an ACK is received.
537 For more detail see https://www.icir.org/floyd/hstcp.html
539 config TCP_CONG_HYBLA
540 tristate "TCP-Hybla congestion control algorithm"
543 TCP-Hybla is a sender-side only change that eliminates penalization of
544 long-RTT, large-bandwidth connections, like when satellite legs are
545 involved, especially when sharing a common bottleneck with normal
546 terrestrial connections.
548 config TCP_CONG_VEGAS
552 TCP Vegas is a sender-side only change to TCP that anticipates
553 the onset of congestion by estimating the bandwidth. TCP Vegas
554 adjusts the sending rate by modifying the congestion
555 window. TCP Vegas should provide less packet loss, but it is
556 not as aggressive as TCP Reno.
562 TCP NV is a follow up to TCP Vegas. It has been modified to deal with
563 10G networks, measurement noise introduced by LRO, GRO and interrupt
564 coalescence. In addition, it will decrease its cwnd multiplicatively
567 Note that in general congestion avoidance (cwnd decreased when # packets
568 queued grows) cannot coexist with congestion control (cwnd decreased only
569 when there is packet loss) due to fairness issues. One scenario when they
570 can coexist safely is when the CA flows have RTTs << CC flows RTTs.
572 For further details see http://www.brakmo.org/networking/tcp-nv/
574 config TCP_CONG_SCALABLE
575 tristate "Scalable TCP"
578 Scalable TCP is a sender-side only change to TCP which uses a
579 MIMD congestion control algorithm which has some nice scaling
580 properties, though is known to have fairness issues.
581 See http://www.deneholme.net/tom/scalable/
584 tristate "TCP Low Priority"
587 TCP Low Priority (TCP-LP), a distributed algorithm whose goal is
588 to utilize only the excess network bandwidth as compared to the
589 ``fair share`` of bandwidth as targeted by TCP.
590 See http://www-ece.rice.edu/networks/TCP-LP/
596 TCP Veno is a sender-side only enhancement of TCP to obtain better
597 throughput over wireless networks. TCP Veno makes use of state
598 distinguishing to circumvent the difficult judgment of the packet loss
599 type. TCP Veno cuts down less congestion window in response to random
601 See <http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=1177186>
605 select TCP_CONG_VEGAS
608 YeAH-TCP is a sender-side high-speed enabled TCP congestion control
609 algorithm, which uses a mixed loss/delay approach to compute the
610 congestion window. It's design goals target high efficiency,
611 internal, RTT and Reno fairness, resilience to link loss while
612 keeping network elements load as low as possible.
614 For further details look here:
615 http://wil.cs.caltech.edu/pfldnet2007/paper/YeAH_TCP.pdf
617 config TCP_CONG_ILLINOIS
618 tristate "TCP Illinois"
621 TCP-Illinois is a sender-side modification of TCP Reno for
622 high speed long delay links. It uses round-trip-time to
623 adjust the alpha and beta parameters to achieve a higher average
624 throughput and maintain fairness.
626 For further details see:
627 http://www.ews.uiuc.edu/~shaoliu/tcpillinois/index.html
629 config TCP_CONG_DCTCP
630 tristate "DataCenter TCP (DCTCP)"
633 DCTCP leverages Explicit Congestion Notification (ECN) in the network to
634 provide multi-bit feedback to the end hosts. It is designed to provide:
636 - High burst tolerance (incast due to partition/aggregate),
637 - Low latency (short flows, queries),
638 - High throughput (continuous data updates, large file transfers) with
639 commodity, shallow-buffered switches.
641 All switches in the data center network running DCTCP must support
642 ECN marking and be configured for marking when reaching defined switch
643 buffer thresholds. The default ECN marking threshold heuristic for
644 DCTCP on switches is 20 packets (30KB) at 1Gbps, and 65 packets
645 (~100KB) at 10Gbps, but might need further careful tweaking.
647 For further details see:
648 http://simula.stanford.edu/~alizade/Site/DCTCP_files/dctcp-final.pdf
651 tristate "CAIA Delay-Gradient (CDG)"
654 CAIA Delay-Gradient (CDG) is a TCP congestion control that modifies
655 the TCP sender in order to:
657 o Use the delay gradient as a congestion signal.
658 o Back off with an average probability that is independent of the RTT.
659 o Coexist with flows that use loss-based congestion control.
660 o Tolerate packet loss unrelated to congestion.
662 For further details see:
663 D.A. Hayes and G. Armitage. "Revisiting TCP congestion control using
664 delay gradients." In Networking 2011. Preprint: http://goo.gl/No3vdg
671 BBR (Bottleneck Bandwidth and RTT) TCP congestion control aims to
672 maximize network utilization and minimize queues. It builds an explicit
673 model of the bottleneck delivery rate and path round-trip propagation
674 delay. It tolerates packet loss and delay unrelated to congestion. It
675 can operate over LAN, WAN, cellular, wifi, or cable modem links. It can
676 coexist with flows that use loss-based congestion control, and can
677 operate with shallow buffers, deep buffers, bufferbloat, policers, or
678 AQM schemes that do not provide a delay signal. It requires the fq
679 ("Fair Queue") pacing packet scheduler.
682 prompt "Default TCP congestion control"
683 default DEFAULT_CUBIC
685 Select the TCP congestion control that will be used by default
689 bool "Bic" if TCP_CONG_BIC=y
692 bool "Cubic" if TCP_CONG_CUBIC=y
695 bool "Htcp" if TCP_CONG_HTCP=y
698 bool "Hybla" if TCP_CONG_HYBLA=y
701 bool "Vegas" if TCP_CONG_VEGAS=y
704 bool "Veno" if TCP_CONG_VENO=y
706 config DEFAULT_WESTWOOD
707 bool "Westwood" if TCP_CONG_WESTWOOD=y
710 bool "DCTCP" if TCP_CONG_DCTCP=y
713 bool "CDG" if TCP_CONG_CDG=y
716 bool "BBR" if TCP_CONG_BBR=y
724 config TCP_CONG_CUBIC
726 depends on !TCP_CONG_ADVANCED
729 config DEFAULT_TCP_CONG
731 default "bic" if DEFAULT_BIC
732 default "cubic" if DEFAULT_CUBIC
733 default "htcp" if DEFAULT_HTCP
734 default "hybla" if DEFAULT_HYBLA
735 default "vegas" if DEFAULT_VEGAS
736 default "westwood" if DEFAULT_WESTWOOD
737 default "veno" if DEFAULT_VENO
738 default "reno" if DEFAULT_RENO
739 default "dctcp" if DEFAULT_DCTCP
740 default "cdg" if DEFAULT_CDG
741 default "bbr" if DEFAULT_BBR
745 bool "TCP: MD5 Signature Option support (RFC2385)"
749 RFC2385 specifies a method of giving MD5 protection to TCP sessions.
750 Its main (only?) use is to protect BGP sessions between core routers