ip-cref.tex revision 71e5815105fb0b86af7df9c719f7c106f05f29c0
1\documentstyle[12pt,twoside]{article} 2\def\TITLE{IP Command Reference} 3\input preamble 4\begin{center} 5\Large\bf IP Command Reference. 6\end{center} 7 8 9\begin{center} 10{ \large Alexey~N.~Kuznetsov } \\ 11\em Institute for Nuclear Research, Moscow \\ 12\verb|kuznet@ms2.inr.ac.ru| \\ 13\rm April 14, 1999 14\end{center} 15 16\vspace{5mm} 17 18\tableofcontents 19 20\newpage 21 22\section{About this document} 23 24This document presents a comprehensive description of the \verb|ip| utility 25from the \verb|iproute2| package. It is not a tutorial or user's guide. 26It is a {\em dictionary\/}, not explaining terms, 27but translating them into other terms, which may also be unknown to the reader. 28However, the document is self-contained and the reader, provided they have a 29basic networking background, will find enough information 30and examples to understand and configure Linux-2.2 IP and IPv6 31networking. 32 33This document is split into sections explaining \verb|ip| commands 34and options, decrypting \verb|ip| output and containing a few examples. 35More voluminous examples and some topics, which require more elaborate 36discussion, are in the appendix. 37 38The paragraphs beginning with NB contain side notes, warnings about 39bugs and design drawbacks. They may be skipped at the first reading. 40 41\section{{\tt ip} --- command syntax} 42 43The generic form of an \verb|ip| command is: 44\begin{verbatim} 45ip [ OPTIONS ] OBJECT [ COMMAND [ ARGUMENTS ]] 46\end{verbatim} 47where \verb|OPTIONS| is a set of optional modifiers affecting the 48general behaviour of the \verb|ip| utility or changing its output. All options 49begin with the character \verb|'-'| and may be used in either long or abbreviated 50forms. Currently, the following options are available: 51 52\begin{itemize} 53\item \verb|-V|, \verb|-Version| 54 55--- print the version of the \verb|ip| utility and exit. 56 57 58\item \verb|-s|, \verb|-stats|, \verb|-statistics| 59 60--- output more information. If the option 61appears twice or more, the amount of information increases. 62As a rule, the information is statistics or some time values. 63 64 65\item \verb|-f|, \verb|-family| followed by a protocol family 66identifier: \verb|inet|, \verb|inet6| or \verb|link|. 67 68--- enforce the protocol family to use. If the option is not present, 69the protocol family is guessed from other arguments. If the rest of the command 70line does not give enough information to guess the family, \verb|ip| falls back to the default 71one, usually \verb|inet| or \verb|any|. \verb|link| is a special family 72identifier meaning that no networking protocol is involved. 73 74\item \verb|-4| 75 76--- shortcut for \verb|-family inet|. 77 78\item \verb|-6| 79 80--- shortcut for \verb|-family inet6|. 81 82\item \verb|-0| 83 84--- shortcut for \verb|-family link|. 85 86 87\item \verb|-o|, \verb|-oneline| 88 89--- output each record on a single line, replacing line feeds 90with the \verb|'\'| character. This is convenient when you want to 91count records with \verb|wc| or to \verb|grep| the output. The trivial 92script \verb|rtpr| converts the output back into readable form. 93 94\item \verb|-r|, \verb|-resolve| 95 96--- use the system's name resolver to print DNS names instead of 97host addresses. 98 99\begin{NB} 100 Do not use this option when reporting bugs or asking for advice. 101\end{NB} 102\begin{NB} 103 \verb|ip| never uses DNS to resolve names to addresses. 104\end{NB} 105 106\end{itemize} 107 108\verb|OBJECT| is the object to manage or to get information about. 109The object types currently understood by \verb|ip| are: 110 111\begin{itemize} 112\item \verb|link| --- network device 113\item \verb|address| --- protocol (IP or IPv6) address on a device 114\item \verb|neighbour| --- ARP or NDISC cache entry 115\item \verb|route| --- routing table entry 116\item \verb|rule| --- rule in routing policy database 117\item \verb|maddress| --- multicast address 118\item \verb|mroute| --- multicast routing cache entry 119\item \verb|tunnel| --- tunnel over IP 120\end{itemize} 121 122Again, the names of all objects may be written in full or 123abbreviated form, f.e.\ \verb|address| is abbreviated as \verb|addr| 124or just \verb|a|. 125 126\verb|COMMAND| specifies the action to perform on the object. 127The set of possible actions depends on the object type. 128As a rule, it is possible to \verb|add|, \verb|delete| and 129\verb|show| (or \verb|list|) objects, but some objects 130do not allow all of these operations or have some additional commands. 131The \verb|help| command is available for all objects. It prints 132out a list of available commands and argument syntax conventions. 133 134If no command is given, some default command is assumed. 135Usually it is \verb|list| or, if the objects of this class 136cannot be listed, \verb|help|. 137 138\verb|ARGUMENTS| is a list of arguments to the command. 139The arguments depend on the command and object. There are two types of arguments: 140{\em flags\/}, consisting of a single keyword, and {\em parameters\/}, 141consisting of a keyword followed by a value. For convenience, 142each command has some {\em default parameter\/} 143which may be omitted. F.e.\ parameter \verb|dev| is the default 144for the {\tt ip link} command, so {\tt ip link ls eth0} is equivalent 145to {\tt ip link ls dev eth0}. 146In the command descriptions below such parameters 147are distinguished with the marker: ``(default)''. 148 149Almost all keywords may be abbreviated with several first (or even single) 150letters. The shortcuts are convenient when \verb|ip| is used interactively, 151but they are not recommended in scripts or when reporting bugs 152or asking for advice. ``Officially'' allowed abbreviations are listed 153in the document body. 154 155 156 157\section{{\tt ip} --- error messages} 158 159\verb|ip| may fail for one of the following reasons: 160 161\begin{itemize} 162\item 163A syntax error on the command line: an unknown keyword, incorrectly formatted 164IP address {\em et al\/}. In this case \verb|ip| prints an error message 165and exits. As a rule, the error message will contain information 166about the reason for the failure. Sometimes it also prints a help page. 167 168\item 169The arguments did not pass verification for self-consistency. 170 171\item 172\verb|ip| failed to compile a kernel request from the arguments 173because the user didn't give enough information. 174 175\item 176The kernel returned an error to some syscall. In this case \verb|ip| 177prints the error message, as it is output with \verb|perror(3)|, 178prefixed with a comment and a syscall identifier. 179 180\item 181The kernel returned an error to some RTNETLINK request. 182In this case \verb|ip| prints the error message, as it is output 183with \verb|perror(3)| prefixed with ``RTNETLINK answers:''. 184 185\end{itemize} 186 187All the operations are atomic, i.e.\ 188if the \verb|ip| utility fails, it does not change anything 189in the system. One harmful exception is \verb|ip link| command 190(Sec.\ref{IP-LINK}, p.\pageref{IP-LINK}), 191which may change only some of the device parameters given 192on command line. 193 194It is difficult to list all the error messages (especially 195syntax errors). However, as a rule, their meaning is clear 196from the context of the command. 197 198The most common mistakes are: 199 200\begin{enumerate} 201\item Netlink is not configured in the kernel. The message is: 202\begin{verbatim} 203Cannot open netlink socket: Invalid value 204\end{verbatim} 205 206\item RTNETLINK is not configured in the kernel. In this case 207one of the following messages may be printed, depending on the command: 208\begin{verbatim} 209Cannot talk to rtnetlink: Connection refused 210Cannot send dump request: Connection refused 211\end{verbatim} 212 213\item The \verb|CONFIG_IP_MULTIPLE_TABLES| option was not selected 214when configuring the kernel. In this case any attempt to use the 215\verb|ip| \verb|rule| command will fail, f.e. 216\begin{verbatim} 217kuznet@kaiser $ ip rule list 218RTNETLINK error: Invalid argument 219dump terminated 220\end{verbatim} 221 222\end{enumerate} 223 224 225\section{{\tt ip link} --- network device configuration} 226\label{IP-LINK} 227 228\paragraph{Object:} A \verb|link| is a network device and the corresponding 229commands display and change the state of devices. 230 231\paragraph{Commands:} \verb|set| and \verb|show| (or \verb|list|). 232 233\subsection{{\tt ip link set} --- change device attributes} 234 235\paragraph{Abbreviations:} \verb|set|, \verb|s|. 236 237\paragraph{Arguments:} 238 239\begin{itemize} 240\item \verb|dev NAME| (default) 241 242--- \verb|NAME| specifies the network device on which to operate. 243 244\item \verb|up| and \verb|down| 245 246--- change the state of the device to \verb|UP| or \verb|DOWN|. 247 248\item \verb|arp on| or \verb|arp off| 249 250--- change the \verb|NOARP| flag on the device. 251 252\begin{NB} 253This operation is {\em not allowed\/} if the device is in state \verb|UP|. 254Though neither the \verb|ip| utility nor the kernel check for this condition. 255You can get unpredictable results changing this flag while the 256device is running. 257\end{NB} 258 259\item \verb|multicast on| or \verb|multicast off| 260 261--- change the \verb|MULTICAST| flag on the device. 262 263\item \verb|dynamic on| or \verb|dynamic off| 264 265--- change the \verb|DYNAMIC| flag on the device. 266 267\item \verb|name NAME| 268 269--- change the name of the device. This operation is not 270recommended if the device is running or has some addresses 271already configured. 272 273\item \verb|txqueuelen NUMBER| or \verb|txqlen NUMBER| 274 275--- change the transmit queue length of the device. 276 277\item \verb|mtu NUMBER| 278 279--- change the MTU of the device. 280 281\item \verb|address LLADDRESS| 282 283--- change the station address of the interface. 284 285\item \verb|broadcast LLADDRESS|, \verb|brd LLADDRESS| or \verb|peer LLADDRESS| 286 287--- change the link layer broadcast address or the peer address when 288the interface is \verb|POINTOPOINT|. 289 290\vskip 1mm 291\begin{NB} 292For most devices (f.e.\ for Ethernet) changing the link layer 293broadcast address will break networking. 294Do not use it, if you do not understand what this operation really does. 295\end{NB} 296 297\item \verb|netns PID| 298 299--- move the device to the network namespace associated with the process PID. 300 301\end{itemize} 302 303\vskip 1mm 304\begin{NB} 305The \verb|PROMISC| and \verb|ALLMULTI| flags are considered 306obsolete and should not be changed administratively, though 307the {\tt ip} utility will allow that. 308\end{NB} 309 310\paragraph{Warning:} If multiple parameter changes are requested, 311\verb|ip| aborts immediately after any of the changes have failed. 312This is the only case when \verb|ip| can move the system to 313an unpredictable state. The solution is to avoid changing 314several parameters with one {\tt ip link set} call. 315 316\paragraph{Examples:} 317\begin{itemize} 318\item \verb|ip link set dummy address 00:00:00:00:00:01| 319 320--- change the station address of the interface \verb|dummy|. 321 322\item \verb|ip link set dummy up| 323 324--- start the interface \verb|dummy|. 325 326\end{itemize} 327 328 329\subsection{{\tt ip link show} --- display device attributes} 330\label{IP-LINK-SHOW} 331 332\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|lst|, \verb|sh|, \verb|ls|, 333\verb|l|. 334 335\paragraph{Arguments:} 336\begin{itemize} 337\item \verb|dev NAME| (default) 338 339--- \verb|NAME| specifies the network device to show. 340If this argument is omitted all devices are listed. 341 342\item \verb|up| 343 344--- only display running interfaces. 345 346\end{itemize} 347 348 349\paragraph{Output format:} 350 351\begin{verbatim} 352kuznet@alisa:~ $ ip link ls eth0 3533: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc cbq qlen 100 354 link/ether 00:a0:cc:66:18:78 brd ff:ff:ff:ff:ff:ff 355kuznet@alisa:~ $ ip link ls sit0 3565: sit0@NONE: <NOARP,UP> mtu 1480 qdisc noqueue 357 link/sit 0.0.0.0 brd 0.0.0.0 358kuznet@alisa:~ $ ip link ls dummy 3592: dummy: <BROADCAST,NOARP> mtu 1500 qdisc noop 360 link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff 361kuznet@alisa:~ $ 362\end{verbatim} 363 364 365The number before each colon is an {\em interface index\/} or {\em ifindex\/}. 366This number uniquely identifies the interface. This is followed by the {\em interface name\/} 367(\verb|eth0|, \verb|sit0| etc.). The interface name is also 368unique at every given moment. However, the interface may disappear from the 369list (f.e.\ when the corresponding driver module is unloaded) and another 370one with the same name may be created later. Besides that, 371the administrator may change the name of any device with 372\verb|ip| \verb|link| \verb|set| \verb|name| 373to make it more intelligible. 374 375The interface name may have another name or \verb|NONE| appended 376after the \verb|@| sign. This means that this device is bound to some other 377device, 378i.e.\ packets send through it are encapsulated and sent via the ``master'' 379device. If the name is \verb|NONE|, the master is unknown. 380 381Then we see the interface {\em mtu\/} (``maximal transfer unit''). This determines 382the maximal size of data which can be sent as a single packet over this interface. 383 384{\em qdisc\/} (``queuing discipline'') shows the queuing algorithm used 385on the interface. Particularly, \verb|noqueue| means that this interface 386does not queue anything and \verb|noop| means that the interface is in blackhole 387mode i.e.\ all packets sent to it are immediately discarded. 388{\em qlen\/} is the default transmit queue length of the device measured 389in packets. 390 391The interface flags are summarized in the angle brackets. 392 393\begin{itemize} 394\item \verb|UP| --- the device is turned on. It is ready to accept 395packets for transmission and it may inject into the kernel packets received 396from other nodes on the network. 397 398\item \verb|LOOPBACK| --- the interface does not communicate with other 399hosts. All packets sent through it will be returned 400and nothing but bounced packets can be received. 401 402\item \verb|BROADCAST| --- the device has the facility to send packets 403to all hosts sharing the same link. A typical example is an Ethernet link. 404 405\item \verb|POINTOPOINT| --- the link has only two ends with one node 406attached to each end. All packets sent to this link will reach the peer 407and all packets received by us came from this single peer. 408 409If neither \verb|LOOPBACK| nor \verb|BROADCAST| nor \verb|POINTOPOINT| 410are set, the interface is assumed to be NMBA (Non-Broadcast Multi-Access). 411This is the most generic type of device and the most complicated one, because 412the host attached to a NBMA link has no means to send to anyone 413without additionally configured information. 414 415\item \verb|MULTICAST| --- is an advisory flag indicating that the interface 416is aware of multicasting i.e.\ sending packets to some subset of neighbouring 417nodes. Broadcasting is a particular case of multicasting, where the multicast 418group consists of all nodes on the link. It is important to emphasize 419that software {\em must not\/} interpret the absence of this flag as the inability 420to use multicasting on this interface. Any \verb|POINTOPOINT| and 421\verb|BROADCAST| link is multicasting by definition, because we have 422direct access to all the neighbours and, hence, to any part of them. 423Certainly, the use of high bandwidth multicast transfers is not recommended 424on broadcast-only links because of high expense, but it is not strictly 425prohibited. 426 427\item \verb|PROMISC| --- the device listens to and feeds to the kernel all 428traffic on the link even if it is not destined for us, not broadcasted 429and not destined for a multicast group of which we are member. Usually 430this mode exists only on broadcast links and is used by bridges and for network 431monitoring. 432 433\item \verb|ALLMULTI| --- the device receives all multicast packets 434wandering on the link. This mode is used by multicast routers. 435 436\item \verb|NOARP| --- this flag is different from the other ones. It has 437no invariant value and its interpretation depends on the network protocols 438involved. As a rule, it indicates that the device needs no address 439resolution and that the software or hardware knows how to deliver packets 440without any help from the protocol stacks. 441 442\item \verb|DYNAMIC| --- is an advisory flag indicating that the interface is 443dynamically created and destroyed. 444 445\item \verb|SLAVE| --- this interface is bonded to some other interfaces 446to share link capacities. 447 448\end{itemize} 449 450\vskip 1mm 451\begin{NB} 452There are other flags but they are either obsolete (\verb|NOTRAILERS|) 453or not implemented (\verb|DEBUG|) or specific to some devices 454(\verb|MASTER|, \verb|AUTOMEDIA| and \verb|PORTSEL|). We do not discuss 455them here. 456\end{NB} 457 458 459The second line contains information on the link layer addresses 460associated with the device. The first word (\verb|ether|, \verb|sit|) 461defines the interface hardware type. This type determines the format and semantics 462of the addresses and is logically part of the address. 463The default format of the station address and the broadcast address 464(or the peer address for pointopoint links) is a 465sequence of hexadecimal bytes separated by colons, but some link 466types may have their natural address format, f.e.\ addresses 467of tunnels over IP are printed as dotted-quad IP addresses. 468 469\vskip 1mm 470\begin{NB} 471 NBMA links have no well-defined broadcast or peer address, 472 however this field may contain useful information, f.e.\ 473 about the address of broadcast relay or about the address of the ARP server. 474\end{NB} 475\begin{NB} 476Multicast addresses are not shown by this command, see 477\verb|ip maddr ls| in~Sec.\ref{IP-MADDR} (p.\pageref{IP-MADDR} of this 478document). 479\end{NB} 480 481 482\paragraph{Statistics:} With the \verb|-statistics| option, \verb|ip| also 483prints interface statistics: 484 485\begin{verbatim} 486kuznet@alisa:~ $ ip -s link ls eth0 4873: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc cbq qlen 100 488 link/ether 00:a0:cc:66:18:78 brd ff:ff:ff:ff:ff:ff 489 RX: bytes packets errors dropped overrun mcast 490 2449949362 2786187 0 0 0 0 491 TX: bytes packets errors dropped carrier collsns 492 178558497 1783945 332 0 332 35172 493kuznet@alisa:~ $ 494\end{verbatim} 495\verb|RX:| and \verb|TX:| lines summarize receiver and transmitter 496statistics. They contain: 497\begin{itemize} 498\item \verb|bytes| --- the total number of bytes received or transmitted 499on the interface. This number wraps when the maximal length of the data type 500natural for the architecture is exceeded, so continuous monitoring requires 501a user level daemon snapping it periodically. 502\item \verb|packets| --- the total number of packets received or transmitted 503on the interface. 504\item \verb|errors| --- the total number of receiver or transmitter errors. 505\item \verb|dropped| --- the total number of packets dropped due to lack 506of resources. 507\item \verb|overrun| --- the total number of receiver overruns resulting 508in dropped packets. As a rule, if the interface is overrun, it means 509serious problems in the kernel or that your machine is too slow 510for this interface. 511\item \verb|mcast| --- the total number of received multicast packets. This option 512is only supported by a few devices. 513\item \verb|carrier| --- total number of link media failures f.e.\ because 514of lost carrier. 515\item \verb|collsns| --- the total number of collision events 516on Ethernet-like media. This number may have a different sense on other 517link types. 518\item \verb|compressed| --- the total number of compressed packets. This is 519available only for links using VJ header compression. 520\end{itemize} 521 522 523If the \verb|-s| option is entered twice or more, 524\verb|ip| prints more detailed statistics on receiver 525and transmitter errors. 526 527\begin{verbatim} 528kuznet@alisa:~ $ ip -s -s link ls eth0 5293: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc cbq qlen 100 530 link/ether 00:a0:cc:66:18:78 brd ff:ff:ff:ff:ff:ff 531 RX: bytes packets errors dropped overrun mcast 532 2449949362 2786187 0 0 0 0 533 RX errors: length crc frame fifo missed 534 0 0 0 0 0 535 TX: bytes packets errors dropped carrier collsns 536 178558497 1783945 332 0 332 35172 537 TX errors: aborted fifo window heartbeat 538 0 0 0 332 539kuznet@alisa:~ $ 540\end{verbatim} 541These error names are pure Ethernetisms. Other devices 542may have non zero values in these fields but they may be 543interpreted differently. 544 545 546\section{{\tt ip address} --- protocol address management} 547 548\paragraph{Abbreviations:} \verb|address|, \verb|addr|, \verb|a|. 549 550\paragraph{Object:} The \verb|address| is a protocol (IP or IPv6) address attached 551to a network device. Each device must have at least one address 552to use the corresponding protocol. It is possible to have several 553different addresses attached to one device. These addresses are not 554discriminated, so that the term {\em alias\/} is not quite appropriate 555for them and we do not use it in this document. 556 557The \verb|ip addr| command displays addresses and their properties, 558adds new addresses and deletes old ones. 559 560\paragraph{Commands:} \verb|add|, \verb|delete|, \verb|flush| and \verb|show| 561(or \verb|list|). 562 563 564\subsection{{\tt ip address add} --- add a new protocol address} 565\label{IP-ADDR-ADD} 566 567\paragraph{Abbreviations:} \verb|add|, \verb|a|. 568 569\paragraph{Arguments:} 570 571\begin{itemize} 572\item \verb|dev NAME| 573 574\noindent--- the name of the device to add the address to. 575 576\item \verb|local ADDRESS| (default) 577 578--- the address of the interface. The format of the address depends 579on the protocol. It is a dotted quad for IP and a sequence of hexadecimal halfwords 580separated by colons for IPv6. The \verb|ADDRESS| may be followed by 581a slash and a decimal number which encodes the network prefix length. 582 583 584\item \verb|peer ADDRESS| 585 586--- the address of the remote endpoint for pointopoint interfaces. 587Again, the \verb|ADDRESS| may be followed by a slash and a decimal number, 588encoding the network prefix length. If a peer address is specified, 589the local address {\em cannot\/} have a prefix length. The network prefix is associated 590with the peer rather than with the local address. 591 592 593\item \verb|broadcast ADDRESS| 594 595--- the broadcast address on the interface. 596 597It is possible to use the special symbols \verb|'+'| and \verb|'-'| 598instead of the broadcast address. In this case, the broadcast address 599is derived by setting/resetting the host bits of the interface prefix. 600 601\vskip 1mm 602\begin{NB} 603Unlike \verb|ifconfig|, the \verb|ip| utility {\em does not\/} set any broadcast 604address unless explicitly requested. 605\end{NB} 606 607 608\item \verb|label NAME| 609 610--- Each address may be tagged with a label string. 611In order to preserve compatibility with Linux-2.0 net aliases, 612this string must coincide with the name of the device or must be prefixed 613with the device name followed by colon. 614 615 616\item \verb|scope SCOPE_VALUE| 617 618--- the scope of the area where this address is valid. 619The available scopes are listed in file \verb|/etc/iproute2/rt_scopes|. 620Predefined scope values are: 621 622 \begin{itemize} 623 \item \verb|global| --- the address is globally valid. 624 \item \verb|site| --- (IPv6 only) the address is site local, 625 i.e.\ it is valid inside this site. 626 \item \verb|link| --- the address is link local, i.e.\ 627 it is valid only on this device. 628 \item \verb|host| --- the address is valid only inside this host. 629 \end{itemize} 630 631Appendix~\ref{ADDR-SEL} (p.\pageref{ADDR-SEL} of this document) 632contains more details on address scopes. 633 634\end{itemize} 635 636\paragraph{Examples:} 637\begin{itemize} 638\item \verb|ip addr add 127.0.0.1/8 dev lo brd + scope host| 639 640--- add the usual loopback address to the loopback device. 641 642\item \verb|ip addr add 10.0.0.1/24 brd + dev eth0 label eth0:Alias| 643 644--- add the address 10.0.0.1 with prefix length 24 (i.e.\ netmask 645\verb|255.255.255.0|), standard broadcast and label \verb|eth0:Alias| 646to the interface \verb|eth0|. 647\end{itemize} 648 649 650\subsection{{\tt ip address delete} --- delete a protocol address} 651 652\paragraph{Abbreviations:} \verb|delete|, \verb|del|, \verb|d|. 653 654\paragraph{Arguments:} coincide with the arguments of \verb|ip addr add|. 655The device name is a required argument. The rest are optional. 656If no arguments are given, the first address is deleted. 657 658\paragraph{Examples:} 659\begin{itemize} 660\item \verb|ip addr del 127.0.0.1/8 dev lo| 661 662--- deletes the loopback address from the loopback device. 663It would be best not to repeat this experiment. 664 665\item Disable IP on the interface \verb|eth0|: 666\begin{verbatim} 667 while ip -f inet addr del dev eth0; do 668 : nothing 669 done 670\end{verbatim} 671Another method to disable IP on an interface using {\tt ip addr flush} 672may be found in sec.\ref{IP-ADDR-FLUSH}, p.\pageref{IP-ADDR-FLUSH}. 673 674\end{itemize} 675 676 677\subsection{{\tt ip address show} --- display protocol addresses} 678 679\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|lst|, \verb|sh|, \verb|ls|, 680\verb|l|. 681 682\paragraph{Arguments:} 683 684\begin{itemize} 685\item \verb|dev NAME| (default) 686 687--- the name of the device. 688 689\item \verb|scope SCOPE_VAL| 690 691--- only list addresses with this scope. 692 693\item \verb|to PREFIX| 694 695--- only list addresses matching this prefix. 696 697\item \verb|label PATTERN| 698 699--- only list addresses with labels matching the \verb|PATTERN|. 700\verb|PATTERN| is a usual shell style pattern. 701 702 703\item \verb|dynamic| and \verb|permanent| 704 705--- (IPv6 only) only list addresses installed due to stateless 706address configuration or only list permanent (not dynamic) addresses. 707 708\item \verb|tentative| 709 710--- (IPv6 only) only list addresses which did not pass duplicate 711address detection. 712 713\item \verb|deprecated| 714 715--- (IPv6 only) only list deprecated addresses. 716 717 718\item \verb|primary| and \verb|secondary| 719 720--- only list primary (or secondary) addresses. 721 722\end{itemize} 723 724 725\paragraph{Output format:} 726 727\begin{verbatim} 728kuznet@alisa:~ $ ip addr ls eth0 7293: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc cbq qlen 100 730 link/ether 00:a0:cc:66:18:78 brd ff:ff:ff:ff:ff:ff 731 inet 193.233.7.90/24 brd 193.233.7.255 scope global eth0 732 inet6 3ffe:2400:0:1:2a0:ccff:fe66:1878/64 scope global dynamic 733 valid_lft forever preferred_lft 604746sec 734 inet6 fe80::2a0:ccff:fe66:1878/10 scope link 735kuznet@alisa:~ $ 736\end{verbatim} 737 738The first two lines coincide with the output of \verb|ip link ls|. 739It is natural to interpret link layer addresses 740as addresses of the protocol family \verb|AF_PACKET|. 741 742Then the list of IP and IPv6 addresses follows, accompanied by 743additional address attributes: scope value (see Sec.\ref{IP-ADDR-ADD}, 744p.\pageref{IP-ADDR-ADD} above), flags and the address label. 745 746Address flags are set by the kernel and cannot be changed 747administratively. Currently, the following flags are defined: 748 749\begin{enumerate} 750\item \verb|secondary| 751 752--- the address is not used when selecting the default source address 753of outgoing packets (Cf.\ Appendix~\ref{ADDR-SEL}, p.\pageref{ADDR-SEL}.). 754An IP address becomes secondary if another address with the same 755prefix bits already exists. The first address is primary. 756It is the leader of the group of all secondary addresses. When the leader 757is deleted, all secondaries are purged too. 758There is a tweak in \verb|/proc/sys/net/ipv4/conf/<dev>/promote_secondaries| 759which activate secondaries promotion when a primary is deleted. 760To permanently enable this feature on all devices add 761\verb|net.ipv4.conf.all.promote_secondaries=1| to \verb|/etc/sysctl.conf|. 762This tweak is available in linux 2.6.15 and later. 763 764 765\item \verb|dynamic| 766 767--- the address was created due to stateless autoconfiguration~\cite{RFC-ADDRCONF}. 768In this case the output also contains information on times, when 769the address is still valid. After \verb|preferred_lft| expires the address is 770moved to the deprecated state. After \verb|valid_lft| expires the address 771is finally invalidated. 772 773\item \verb|deprecated| 774 775--- the address is deprecated, i.e.\ it is still valid, but cannot 776be used by newly created connections. 777 778\item \verb|tentative| 779 780--- the address is not used because duplicate address detection~\cite{RFC-ADDRCONF} 781is still not complete or failed. 782 783\end{enumerate} 784 785 786\subsection{{\tt ip address flush} --- flush protocol addresses} 787\label{IP-ADDR-FLUSH} 788 789\paragraph{Abbreviations:} \verb|flush|, \verb|f|. 790 791\paragraph{Description:}This command flushes the protocol addresses 792selected by some criteria. 793 794\paragraph{Arguments:} This command has the same arguments as \verb|show|. 795The difference is that it does not run when no arguments are given. 796 797\paragraph{Warning:} This command (and other \verb|flush| commands 798described below) is pretty dangerous. If you make a mistake, it will 799not forgive it, but will cruelly purge all the addresses. 800 801\paragraph{Statistics:} With the \verb|-statistics| option, the command 802becomes verbose. It prints out the number of deleted addresses and the number 803of rounds made to flush the address list. If this option is given 804twice, \verb|ip addr flush| also dumps all the deleted addresses 805in the format described in the previous subsection. 806 807\paragraph{Example:} Delete all the addresses from the private network 80810.0.0.0/8: 809\begin{verbatim} 810netadm@amber:~ # ip -s -s a f to 10/8 8112: dummy inet 10.7.7.7/16 brd 10.7.255.255 scope global dummy 8123: eth0 inet 10.10.7.7/16 brd 10.10.255.255 scope global eth0 8134: eth1 inet 10.8.7.7/16 brd 10.8.255.255 scope global eth1 814 815*** Round 1, deleting 3 addresses *** 816*** Flush is complete after 1 round *** 817netadm@amber:~ # 818\end{verbatim} 819Another instructive example is disabling IP on all the Ethernets: 820\begin{verbatim} 821netadm@amber:~ # ip -4 addr flush label "eth*" 822\end{verbatim} 823And the last example shows how to flush all the IPv6 addresses 824acquired by the host from stateless address autoconfiguration 825after you enabled forwarding or disabled autoconfiguration. 826\begin{verbatim} 827netadm@amber:~ # ip -6 addr flush dynamic 828\end{verbatim} 829 830 831 832\section{{\tt ip neighbour} --- neighbour/arp tables management} 833 834\paragraph{Abbreviations:} \verb|neighbour|, \verb|neighbor|, \verb|neigh|, 835\verb|n|. 836 837\paragraph{Object:} \verb|neighbour| objects establish bindings between protocol 838addresses and link layer addresses for hosts sharing the same link. 839Neighbour entries are organized into tables. The IPv4 neighbour table 840is known by another name --- the ARP table. 841 842The corresponding commands display neighbour bindings 843and their properties, add new neighbour entries and delete old ones. 844 845\paragraph{Commands:} \verb|add|, \verb|change|, \verb|replace|, 846\verb|delete|, \verb|flush| and \verb|show| (or \verb|list|). 847 848\paragraph{See also:} Appendix~\ref{PROXY-NEIGH}, p.\pageref{PROXY-NEIGH} 849describes how to manage proxy ARP/NDISC with the \verb|ip| utility. 850 851 852\subsection{{\tt ip neighbour add} --- add a new neighbour entry\\ 853 {\tt ip neighbour change} --- change an existing entry\\ 854 {\tt ip neighbour replace} --- add a new entry or change an existing one} 855 856\paragraph{Abbreviations:} \verb|add|, \verb|a|; \verb|change|, \verb|chg|; 857\verb|replace|, \verb|repl|. 858 859\paragraph{Description:} These commands create new neighbour records 860or update existing ones. 861 862\paragraph{Arguments:} 863 864\begin{itemize} 865\item \verb|to ADDRESS| (default) 866 867--- the protocol address of the neighbour. It is either an IPv4 or IPv6 address. 868 869\item \verb|dev NAME| 870 871--- the interface to which this neighbour is attached. 872 873 874\item \verb|lladdr LLADDRESS| 875 876--- the link layer address of the neighbour. \verb|LLADDRESS| can also be 877\verb|null|. 878 879\item \verb|nud NUD_STATE| 880 881--- the state of the neighbour entry. \verb|nud| is an abbreviation for ``Neighbour 882Unreachability Detection''. The state can take one of the following values: 883 884\begin{enumerate} 885\item \verb|permanent| --- the neighbour entry is valid forever and can be only be removed 886administratively. 887\item \verb|noarp| --- the neighbour entry is valid. No attempts to validate 888this entry will be made but it can be removed when its lifetime expires. 889\item \verb|reachable| --- the neighbour entry is valid until the reachability 890timeout expires. 891\item \verb|stale| --- the neighbour entry is valid but suspicious. 892This option to \verb|ip neigh| does not change the neighbour state if 893it was valid and the address is not changed by this command. 894\end{enumerate} 895 896\end{itemize} 897 898\paragraph{Examples:} 899\begin{itemize} 900\item \verb|ip neigh add 10.0.0.3 lladdr 0:0:0:0:0:1 dev eth0 nud perm| 901 902--- add a permanent ARP entry for the neighbour 10.0.0.3 on the device \verb|eth0|. 903 904\item \verb|ip neigh chg 10.0.0.3 dev eth0 nud reachable| 905 906--- change its state to \verb|reachable|. 907\end{itemize} 908 909 910\subsection{{\tt ip neighbour delete} --- delete a neighbour entry} 911 912\paragraph{Abbreviations:} \verb|delete|, \verb|del|, \verb|d|. 913 914\paragraph{Description:} This command invalidates a neighbour entry. 915 916\paragraph{Arguments:} The arguments are the same as with \verb|ip neigh add|, 917except that \verb|lladdr| and \verb|nud| are ignored. 918 919 920\paragraph{Example:} 921\begin{itemize} 922\item \verb|ip neigh del 10.0.0.3 dev eth0| 923 924--- invalidate an ARP entry for the neighbour 10.0.0.3 on the device \verb|eth0|. 925 926\end{itemize} 927 928\begin{NB} 929 The deleted neighbour entry will not disappear from the tables 930 immediately. If it is in use it cannot be deleted until the last 931 client releases it. Otherwise it will be destroyed during 932 the next garbage collection. 933\end{NB} 934 935 936\paragraph{Warning:} Attempts to delete or manually change 937a \verb|noarp| entry created by the kernel may result in unpredictable behaviour. 938Particularly, the kernel may try to resolve this address even 939on a \verb|NOARP| interface or if the address is multicast or broadcast. 940 941 942\subsection{{\tt ip neighbour show} --- list neighbour entries} 943 944\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|sh|, \verb|ls|. 945 946\paragraph{Description:}This commands displays neighbour tables. 947 948\paragraph{Arguments:} 949 950\begin{itemize} 951 952\item \verb|to ADDRESS| (default) 953 954--- the prefix selecting the neighbours to list. 955 956\item \verb|dev NAME| 957 958--- only list the neighbours attached to this device. 959 960\item \verb|unused| 961 962--- only list neighbours which are not currently in use. 963 964\item \verb|nud NUD_STATE| 965 966--- only list neighbour entries in this state. \verb|NUD_STATE| takes 967values listed below or the special value \verb|all| which means all states. 968This option may occur more than once. If this option is absent, \verb|ip| 969lists all entries except for \verb|none| and \verb|noarp|. 970 971\end{itemize} 972 973 974\paragraph{Output format:} 975 976\begin{verbatim} 977kuznet@alisa:~ $ ip neigh ls 978:: dev lo lladdr 00:00:00:00:00:00 nud noarp 979fe80::200:cff:fe76:3f85 dev eth0 lladdr 00:00:0c:76:3f:85 router \ 980 nud stale 9810.0.0.0 dev lo lladdr 00:00:00:00:00:00 nud noarp 982193.233.7.254 dev eth0 lladdr 00:00:0c:76:3f:85 nud reachable 983193.233.7.85 dev eth0 lladdr 00:e0:1e:63:39:00 nud stale 984kuznet@alisa:~ $ 985\end{verbatim} 986 987The first word of each line is the protocol address of the neighbour. 988Then the device name follows. The rest of the line describes the contents of 989the neighbour entry identified by the pair (device, address). 990 991\verb|lladdr| is the link layer address of the neighbour. 992 993\verb|nud| is the state of the ``neighbour unreachability detection'' machine 994for this entry. The detailed description of the neighbour 995state machine can be found in~\cite{RFC-NDISC}. Here is the full list 996of the states with short descriptions: 997 998\begin{enumerate} 999\item\verb|none| --- the state of the neighbour is void. 1000\item\verb|incomplete| --- the neighbour is in the process of resolution. 1001\item\verb|reachable| --- the neighbour is valid and apparently reachable. 1002\item\verb|stale| --- the neighbour is valid, but is probably already 1003unreachable, so the kernel will try to check it at the first transmission. 1004\item\verb|delay| --- a packet has been sent to the stale neighbour and the kernel is waiting 1005for confirmation. 1006\item\verb|probe| --- the delay timer expired but no confirmation was received. 1007The kernel has started to probe the neighbour with ARP/NDISC messages. 1008\item\verb|failed| --- resolution has failed. 1009\item\verb|noarp| --- the neighbour is valid. No attempts to check the entry 1010will be made. 1011\item\verb|permanent| --- it is a \verb|noarp| entry, but only the administrator 1012may remove the entry from the neighbour table. 1013\end{enumerate} 1014 1015The link layer address is valid in all states except for \verb|none|, 1016\verb|failed| and \verb|incomplete|. 1017 1018IPv6 neighbours can be marked with the additional flag \verb|router| 1019which means that the neighbour introduced itself as an IPv6 router~\cite{RFC-NDISC}. 1020 1021\paragraph{Statistics:} The \verb|-statistics| option displays some usage 1022statistics, f.e.\ 1023 1024\begin{verbatim} 1025kuznet@alisa:~ $ ip -s n ls 193.233.7.254 1026193.233.7.254 dev eth0 lladdr 00:00:0c:76:3f:85 ref 5 used 12/13/20 \ 1027 nud reachable 1028kuznet@alisa:~ $ 1029\end{verbatim} 1030 1031Here \verb|ref| is the number of users of this entry 1032and \verb|used| is a triplet of time intervals in seconds 1033separated by slashes. In this case they show that: 1034 1035\begin{enumerate} 1036\item the entry was used 12 seconds ago. 1037\item the entry was confirmed 13 seconds ago. 1038\item the entry was updated 20 seconds ago. 1039\end{enumerate} 1040 1041\subsection{{\tt ip neighbour flush} --- flush neighbour entries} 1042 1043\paragraph{Abbreviations:} \verb|flush|, \verb|f|. 1044 1045\paragraph{Description:}This command flushes neighbour tables, selecting 1046entries to flush by some criteria. 1047 1048\paragraph{Arguments:} This command has the same arguments as \verb|show|. 1049The differences are that it does not run when no arguments are given, 1050and that the default neighbour states to be flushed do not include 1051\verb|permanent| and \verb|noarp|. 1052 1053 1054\paragraph{Statistics:} With the \verb|-statistics| option, the command 1055becomes verbose. It prints out the number of deleted neighbours and the number 1056of rounds made to flush the neighbour table. If the option is given 1057twice, \verb|ip neigh flush| also dumps all the deleted neighbours 1058in the format described in the previous subsection. 1059 1060\paragraph{Example:} 1061\begin{verbatim} 1062netadm@alisa:~ # ip -s -s n f 193.233.7.254 1063193.233.7.254 dev eth0 lladdr 00:00:0c:76:3f:85 ref 5 used 12/13/20 \ 1064 nud reachable 1065 1066*** Round 1, deleting 1 entries *** 1067*** Flush is complete after 1 round *** 1068netadm@alisa:~ # 1069\end{verbatim} 1070 1071 1072\section{{\tt ip route} --- routing table management} 1073\label{IP-ROUTE} 1074 1075\paragraph{Abbreviations:} \verb|route|, \verb|ro|, \verb|r|. 1076 1077\paragraph{Object:} \verb|route| entries in the kernel routing tables keep 1078information about paths to other networked nodes. 1079 1080Each route entry has a {\em key\/} consisting of a {\em prefix\/} 1081(i.e.\ a pair containing a network address and the length of its mask) and, 1082optionally, the TOS value. An IP packet matches the route if the highest 1083bits of its destination address are equal to the route prefix at least 1084up to the prefix length and if the TOS of the route is zero or equal to 1085the TOS of the packet. 1086 1087If several routes match the packet, the following pruning rules 1088are used to select the best one (see~\cite{RFC1812}): 1089\begin{enumerate} 1090\item The longest matching prefix is selected. All shorter ones 1091are dropped. 1092 1093\item If the TOS of some route with the longest prefix is equal to the TOS 1094of the packet, the routes with different TOS are dropped. 1095 1096If no exact TOS match was found and routes with TOS=0 exist, 1097the rest of routes are pruned. 1098 1099Otherwise, the route lookup fails. 1100 1101\item If several routes remain after the previous steps, then 1102the routes with the best preference values are selected. 1103 1104\item If we still have several routes, then the {\em first\/} of them 1105is selected. 1106 1107\begin{NB} 1108 Note the ambiguity of the last step. Unfortunately, Linux 1109 historically allows such a bizarre situation. The sense of the 1110word ``first'' depends on the order of route additions and it is practically 1111impossible to maintain a bundle of such routes in this order. 1112\end{NB} 1113 1114For simplicity we will limit ourselves to the case where such a situation 1115is impossible and routes are uniquely identified by the triplet 1116\{prefix, tos, preference\}. Actually, it is impossible to create 1117non-unique routes with \verb|ip| commands described in this section. 1118 1119One useful exception to this rule is the default route on non-forwarding 1120hosts. It is ``officially'' allowed to have several fallback routes 1121when several routers are present on directly connected networks. 1122In this case, Linux-2.2 makes ``dead gateway detection''~\cite{RFC1122} 1123controlled by neighbour unreachability detection and by advice 1124from transport protocols to select a working router, so the order 1125of the routes is not essential. However, in this case, 1126fiddling with default routes manually is not recommended. Use the Router Discovery 1127protocol (see Appendix~\ref{EXAMPLE-SETUP}, p.\pageref{EXAMPLE-SETUP}) 1128instead. Actually, Linux-2.2 IPv6 does not give user level applications 1129any access to default routes. 1130\end{enumerate} 1131 1132Certainly, the steps above are not performed exactly 1133in this sequence. Instead, the routing table in the kernel is kept 1134in some data structure to achieve the final result 1135with minimal cost. However, not depending on a particular 1136routing algorithm implemented in the kernel, we can summarize 1137the statements above as: a route is identified by the triplet 1138\{prefix, tos, preference\}. This {\em key\/} lets us locate 1139the route in the routing table. 1140 1141\paragraph{Route attributes:} Each route key refers to a routing 1142information record containing 1143the data required to deliver IP packets (f.e.\ output device and 1144next hop router) and some optional attributes (f.e. the path MTU or 1145the preferred source address when communicating with this destination). 1146These attributes are described in the following subsection. 1147 1148\paragraph{Route types:} \label{IP-ROUTE-TYPES} 1149It is important that the set 1150of required and optional attributes depend on the route {\em type\/}. 1151The most important route type 1152is \verb|unicast|. It describes real paths to other hosts. 1153As a rule, common routing tables contain only such routes. However, 1154there are other types of routes with different semantics. The 1155full list of types understood by Linux-2.2 is: 1156\begin{itemize} 1157\item \verb|unicast| --- the route entry describes real paths to the 1158destinations covered by the route prefix. 1159\item \verb|unreachable| --- these destinations are unreachable. Packets 1160are discarded and the ICMP message {\em host unreachable\/} is generated. 1161The local senders get an \verb|EHOSTUNREACH| error. 1162\item \verb|blackhole| --- these destinations are unreachable. Packets 1163are discarded silently. The local senders get an \verb|EINVAL| error. 1164\item \verb|prohibit| --- these destinations are unreachable. Packets 1165are discarded and the ICMP message {\em communication administratively 1166prohibited\/} is generated. The local senders get an \verb|EACCES| error. 1167\item \verb|local| --- the destinations are assigned to this 1168host. The packets are looped back and delivered locally. 1169\item \verb|broadcast| --- the destinations are broadcast addresses. 1170The packets are sent as link broadcasts. 1171\item \verb|throw| --- a special control route used together with policy 1172rules (see sec.\ref{IP-RULE}, p.\pageref{IP-RULE}). If such a route is selected, lookup 1173in this table is terminated pretending that no route was found. 1174Without policy routing it is equivalent to the absence of the route in the routing 1175table. The packets are dropped and the ICMP message {\em net unreachable\/} 1176is generated. The local senders get an \verb|ENETUNREACH| error. 1177\item \verb|nat| --- a special NAT route. Destinations covered by the prefix 1178are considered to be dummy (or external) addresses which require translation 1179to real (or internal) ones before forwarding. The addresses to translate to 1180are selected with the attribute \verb|via|. More about NAT is 1181in Appendix~\ref{ROUTE-NAT}, p.\pageref{ROUTE-NAT}. 1182\item \verb|anycast| --- ({\em not implemented\/}) the destinations are 1183{\em anycast\/} addresses assigned to this host. They are mainly equivalent 1184to \verb|local| with one difference: such addresses are invalid when used 1185as the source address of any packet. 1186\item \verb|multicast| --- a special type used for multicast routing. 1187It is not present in normal routing tables. 1188\end{itemize} 1189 1190\paragraph{Route tables:} Linux-2.2 can pack routes into several routing 1191tables identified by a number in the range from 1 to 255 or by 1192name from the file \verb|/etc/iproute2/rt_tables|. By default all normal 1193routes are inserted into the \verb|main| table (ID 254) and the kernel only uses 1194this table when calculating routes. 1195 1196Actually, one other table always exists, which is invisible but 1197even more important. It is the \verb|local| table (ID 255). This table 1198consists of routes for local and broadcast addresses. The kernel maintains 1199this table automatically and the administrator usually need not modify it 1200or even look at it. 1201 1202The multiple routing tables enter the game when {\em policy routing\/} 1203is used. See sec.\ref{IP-RULE}, p.\pageref{IP-RULE}. 1204In this case, the table identifier effectively becomes 1205one more parameter, which should be added to the triplet 1206\{prefix, tos, preference\} to uniquely identify the route. 1207 1208 1209\subsection{{\tt ip route add} --- add a new route\\ 1210 {\tt ip route change} --- change a route\\ 1211 {\tt ip route replace} --- change a route or add a new one} 1212\label{IP-ROUTE-ADD} 1213 1214\paragraph{Abbreviations:} \verb|add|, \verb|a|; \verb|change|, \verb|chg|; 1215 \verb|replace|, \verb|repl|. 1216 1217 1218\paragraph{Arguments:} 1219\begin{itemize} 1220\item \verb|to PREFIX| or \verb|to TYPE PREFIX| (default) 1221 1222--- the destination prefix of the route. If \verb|TYPE| is omitted, 1223\verb|ip| assumes type \verb|unicast|. Other values of \verb|TYPE| 1224are listed above. \verb|PREFIX| is an IP or IPv6 address optionally followed 1225by a slash and the prefix length. If the length of the prefix is missing, 1226\verb|ip| assumes a full-length host route. There is also a special 1227\verb|PREFIX| --- \verb|default| --- which is equivalent to IP \verb|0/0| or 1228to IPv6 \verb|::/0|. 1229 1230\item \verb|tos TOS| or \verb|dsfield TOS| 1231 1232--- the Type Of Service (TOS) key. This key has no associated mask and 1233the longest match is understood as: First, compare the TOS 1234of the route and of the packet. If they are not equal, then the packet 1235may still match a route with a zero TOS. \verb|TOS| is either an 8 bit hexadecimal 1236number or an identifier from {\tt /etc/iproute2/rt\_dsfield}. 1237 1238 1239\item \verb|metric NUMBER| or \verb|preference NUMBER| 1240 1241--- the preference value of the route. \verb|NUMBER| is an arbitrary 32bit number. 1242 1243\item \verb|table TABLEID| 1244 1245--- the table to add this route to. 1246\verb|TABLEID| may be a number or a string from the file 1247\verb|/etc/iproute2/rt_tables|. If this parameter is omitted, 1248\verb|ip| assumes the \verb|main| table, with the exception of 1249\verb|local|, \verb|broadcast| and \verb|nat| routes, which are 1250put into the \verb|local| table by default. 1251 1252\item \verb|dev NAME| 1253 1254--- the output device name. 1255 1256\item \verb|via ADDRESS| 1257 1258--- the address of the nexthop router. Actually, the sense of this field depends 1259on the route type. For normal \verb|unicast| routes it is either the true nexthop 1260router or, if it is a direct route installed in BSD compatibility mode, 1261it can be a local address of the interface. 1262For NAT routes it is the first address of the block of translated IP destinations. 1263 1264\item \verb|src ADDRESS| 1265 1266--- the source address to prefer when sending to the destinations 1267covered by the route prefix. 1268 1269\item \verb|realm REALMID| 1270 1271--- the realm to which this route is assigned. 1272\verb|REALMID| may be a number or a string from the file 1273\verb|/etc/iproute2/rt_realms|. Sec.\ref{RT-REALMS} (p.\pageref{RT-REALMS}) 1274contains more information on realms. 1275 1276\item \verb|mtu MTU| or \verb|mtu lock MTU| 1277 1278--- the MTU along the path to the destination. If the modifier \verb|lock| is 1279not used, the MTU may be updated by the kernel due to Path MTU Discovery. 1280If the modifier \verb|lock| is used, no path MTU discovery will be tried, 1281all packets will be sent without the DF bit in IPv4 case 1282or fragmented to MTU for IPv6. 1283 1284\item \verb|window NUMBER| 1285 1286--- the maximal window for TCP to advertise to these destinations, 1287measured in bytes. It limits maximal data bursts that our TCP 1288peers are allowed to send to us. 1289 1290\item \verb|rtt NUMBER| 1291 1292--- the initial RTT (``Round Trip Time'') estimate. 1293 1294 1295\item \verb|rttvar NUMBER| 1296 1297--- \threeonly the initial RTT variance estimate. 1298 1299 1300\item \verb|ssthresh NUMBER| 1301 1302--- \threeonly an estimate for the initial slow start threshold. 1303 1304 1305\item \verb|cwnd NUMBER| 1306 1307--- \threeonly the clamp for congestion window. It is ignored if the \verb|lock| 1308 flag is not used. 1309 1310 1311\item \verb|advmss NUMBER| 1312 1313--- \threeonly the MSS (``Maximal Segment Size'') to advertise to these 1314 destinations when establishing TCP connections. If it is not given, 1315 Linux uses a default value calculated from the first hop device MTU. 1316 1317\begin{NB} 1318 If the path to these destination is asymmetric, this guess may be wrong. 1319\end{NB} 1320 1321\item \verb|reordering NUMBER| 1322 1323--- \threeonly Maximal reordering on the path to this destination. 1324 If it is not given, Linux uses the value selected with \verb|sysctl| 1325 variable \verb|net/ipv4/tcp_reordering|. 1326 1327\item \verb|hoplimit NUMBER| 1328 1329--- [2.5.74+ only] Maximum number of hops on the path to this destination. 1330 The default is the value selected with the \verb|sysctl| variable 1331 \verb|net/ipv4/ip_default_ttl|. 1332 1333\item \verb|initcwnd NUMBER| 1334--- [2.5.70+ only] Initial congestion window size for connections to 1335 this destination. Actual window size is this value multiplied by the 1336 MSS (``Maximal Segment Size'') for same connection. The default is 1337 zero, meaning to use the values specified in~\cite{RFC2414}. 1338 1339\item \verb|nexthop NEXTHOP| 1340 1341--- the nexthop of a multipath route. \verb|NEXTHOP| is a complex value 1342with its own syntax similar to the top level argument lists: 1343\begin{itemize} 1344\item \verb|via ADDRESS| is the nexthop router. 1345\item \verb|dev NAME| is the output device. 1346\item \verb|weight NUMBER| is a weight for this element of a multipath 1347route reflecting its relative bandwidth or quality. 1348\end{itemize} 1349 1350\item \verb|scope SCOPE_VAL| 1351 1352--- the scope of the destinations covered by the route prefix. 1353\verb|SCOPE_VAL| may be a number or a string from the file 1354\verb|/etc/iproute2/rt_scopes|. 1355If this parameter is omitted, 1356\verb|ip| assumes scope \verb|global| for all gatewayed \verb|unicast| 1357routes, scope \verb|link| for direct \verb|unicast| and \verb|broadcast| routes 1358and scope \verb|host| for \verb|local| routes. 1359 1360\item \verb|protocol RTPROTO| 1361 1362--- the routing protocol identifier of this route. 1363\verb|RTPROTO| may be a number or a string from the file 1364\verb|/etc/iproute2/rt_protos|. If the routing protocol ID is 1365not given, \verb|ip| assumes protocol \verb|boot| (i.e.\ 1366it assumes the route was added by someone who doesn't 1367understand what they are doing). Several protocol values have a fixed interpretation. 1368Namely: 1369\begin{itemize} 1370\item \verb|redirect| --- the route was installed due to an ICMP redirect. 1371\item \verb|kernel| --- the route was installed by the kernel during 1372autoconfiguration. 1373\item \verb|boot| --- the route was installed during the bootup sequence. 1374If a routing daemon starts, it will purge all of them. 1375\item \verb|static| --- the route was installed by the administrator 1376to override dynamic routing. Routing daemon will respect them 1377and, probably, even advertise them to its peers. 1378\item \verb|ra| --- the route was installed by Router Discovery protocol. 1379\end{itemize} 1380The rest of the values are not reserved and the administrator is free 1381to assign (or not to assign) protocol tags. At least, routing 1382daemons should take care of setting some unique protocol values, 1383f.e.\ as they are assigned in \verb|rtnetlink.h| or in \verb|rt_protos| 1384database. 1385 1386 1387\item \verb|onlink| 1388 1389--- pretend that the nexthop is directly attached to this link, 1390even if it does not match any interface prefix. One application of this 1391option may be found in~\cite{IP-TUNNELS}. 1392 1393\item \verb|equalize| 1394 1395--- allow packet by packet randomization on multipath routes. 1396Without this modifier, the route will be frozen to one selected 1397nexthop, so that load splitting will only occur on per-flow base. 1398\verb|equalize| only works if the kernel is patched. 1399 1400 1401\end{itemize} 1402 1403 1404\begin{NB} 1405 Actually there are more commands: \verb|prepend| does the same 1406 thing as classic \verb|route add|, i.e.\ adds a route, even if another 1407 route to the same destination exists. Its opposite case is \verb|append|, 1408 which adds the route to the end of the list. Avoid these 1409 features. 1410\end{NB} 1411\begin{NB} 1412 More sad news, IPv6 only understands the \verb|append| command correctly. 1413 All the others are translated into \verb|append| commands. Certainly, 1414 this will change in the future. 1415\end{NB} 1416 1417\paragraph{Examples:} 1418\begin{itemize} 1419\item add a plain route to network 10.0.0/24 via gateway 193.233.7.65 1420\begin{verbatim} 1421 ip route add 10.0.0/24 via 193.233.7.65 1422\end{verbatim} 1423\item change it to a direct route via the \verb|dummy| device 1424\begin{verbatim} 1425 ip ro chg 10.0.0/24 dev dummy 1426\end{verbatim} 1427\item add a default multipath route splitting the load between \verb|ppp0| 1428and \verb|ppp1| 1429\begin{verbatim} 1430 ip route add default scope global nexthop dev ppp0 \ 1431 nexthop dev ppp1 1432\end{verbatim} 1433Note the scope value. It is not necessary but it informs the kernel 1434that this route is gatewayed rather than direct. Actually, if you 1435know the addresses of remote endpoints it would be better to use the 1436\verb|via| parameter. 1437\item announce that the address 192.203.80.144 is not a real one, but 1438should be translated to 193.233.7.83 before forwarding 1439\begin{verbatim} 1440 ip route add nat 192.203.80.144 via 193.233.7.83 1441\end{verbatim} 1442Backward translation is setup with policy rules described 1443in the following section (sec.\ref{IP-RULE}, p.\pageref{IP-RULE}). 1444\end{itemize} 1445 1446\subsection{{\tt ip route delete} --- delete a route} 1447 1448\paragraph{Abbreviations:} \verb|delete|, \verb|del|, \verb|d|. 1449 1450\paragraph{Arguments:} \verb|ip route del| has the same arguments as 1451\verb|ip route add|, but their semantics are a bit different. 1452 1453Key values (\verb|to|, \verb|tos|, \verb|preference| and \verb|table|) 1454select the route to delete. If optional attributes are present, \verb|ip| 1455verifies that they coincide with the attributes of the route to delete. 1456If no route with the given key and attributes was found, \verb|ip route del| 1457fails. 1458\begin{NB} 1459Linux-2.0 had the option to delete a route selected only by prefix address, 1460ignoring its length (i.e.\ netmask). This option no longer exists 1461because it was ambiguous. However, look at {\tt ip route flush} 1462(sec.\ref{IP-ROUTE-FLUSH}, p.\pageref{IP-ROUTE-FLUSH}) which 1463provides similar and even richer functionality. 1464\end{NB} 1465 1466\paragraph{Example:} 1467\begin{itemize} 1468\item delete the multipath route created by the command in previous subsection 1469\begin{verbatim} 1470 ip route del default scope global nexthop dev ppp0 \ 1471 nexthop dev ppp1 1472\end{verbatim} 1473\end{itemize} 1474 1475 1476 1477\subsection{{\tt ip route show} --- list routes} 1478 1479\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|sh|, \verb|ls|, \verb|l|. 1480 1481\paragraph{Description:} the command displays the contents of the routing tables 1482or the route(s) selected by some criteria. 1483 1484 1485\paragraph{Arguments:} 1486\begin{itemize} 1487\item \verb|to SELECTOR| (default) 1488 1489--- only select routes from the given range of destinations. \verb|SELECTOR| 1490consists of an optional modifier (\verb|root|, \verb|match| or \verb|exact|) 1491and a prefix. \verb|root PREFIX| selects routes with prefixes not shorter 1492than \verb|PREFIX|. F.e.\ \verb|root 0/0| selects the entire routing table. 1493\verb|match PREFIX| selects routes with prefixes not longer than 1494\verb|PREFIX|. F.e.\ \verb|match 10.0/16| selects \verb|10.0/16|, 1495\verb|10/8| and \verb|0/0|, but it does not select \verb|10.1/16| and 1496\verb|10.0.0/24|. And \verb|exact PREFIX| (or just \verb|PREFIX|) 1497selects routes with this exact prefix. If neither of these options 1498are present, \verb|ip| assumes \verb|root 0/0| i.e.\ it lists the entire table. 1499 1500 1501\item \verb|tos TOS| or \verb|dsfield TOS| 1502 1503 --- only select routes with the given TOS. 1504 1505 1506\item \verb|table TABLEID| 1507 1508 --- show the routes from this table(s). The default setting is to show 1509\verb|table| \verb|main|. \verb|TABLEID| may either be the ID of a real table 1510or one of the special values: 1511 \begin{itemize} 1512 \item \verb|all| --- list all of the tables. 1513 \item \verb|cache| --- dump the routing cache. 1514 \end{itemize} 1515\begin{NB} 1516 IPv6 has a single table. However, splitting it into \verb|main|, \verb|local| 1517 and \verb|cache| is emulated by the \verb|ip| utility. 1518\end{NB} 1519 1520\item \verb|cloned| or \verb|cached| 1521 1522--- list cloned routes i.e.\ routes which were dynamically forked from 1523other routes because some route attribute (f.e.\ MTU) was updated. 1524Actually, it is equivalent to \verb|table cache|. 1525 1526\item \verb|from SELECTOR| 1527 1528--- the same syntax as for \verb|to|, but it binds the source address range 1529rather than destinations. Note that the \verb|from| option only works with 1530cloned routes. 1531 1532\item \verb|protocol RTPROTO| 1533 1534--- only list routes of this protocol. 1535 1536 1537\item \verb|scope SCOPE_VAL| 1538 1539--- only list routes with this scope. 1540 1541\item \verb|type TYPE| 1542 1543--- only list routes of this type. 1544 1545\item \verb|dev NAME| 1546 1547--- only list routes going via this device. 1548 1549\item \verb|via PREFIX| 1550 1551--- only list routes going via the nexthop routers selected by \verb|PREFIX|. 1552 1553\item \verb|src PREFIX| 1554 1555--- only list routes with preferred source addresses selected 1556by \verb|PREFIX|. 1557 1558\item \verb|realm REALMID| or \verb|realms FROMREALM/TOREALM| 1559 1560--- only list routes with these realms. 1561 1562\end{itemize} 1563 1564\paragraph{Examples:} Let us count routes of protocol \verb|gated/bgp| 1565on a router: 1566\begin{verbatim} 1567kuznet@amber:~ $ ip ro ls proto gated/bgp | wc 1568 1413 9891 79010 1569kuznet@amber:~ $ 1570\end{verbatim} 1571To count the size of the routing cache, we have to use the \verb|-o| option 1572because cached attributes can take more than one line of output: 1573\begin{verbatim} 1574kuznet@amber:~ $ ip -o ro ls cloned | wc 1575 159 2543 18707 1576kuznet@amber:~ $ 1577\end{verbatim} 1578 1579 1580\paragraph{Output format:} The output of this command consists 1581of per route records separated by line feeds. 1582However, some records may consist 1583of more than one line: particularly, this is the case when the route 1584is cloned or you requested additional statistics. If the 1585\verb|-o| option was given, then line feeds separating lines inside 1586records are replaced with the backslash sign. 1587 1588The output has the same syntax as arguments given to {\tt ip route add}, 1589so that it can be understood easily. F.e.\ 1590\begin{verbatim} 1591kuznet@amber:~ $ ip ro ls 193.233.7/24 1592193.233.7.0/24 dev eth0 proto gated/conn scope link \ 1593 src 193.233.7.65 realms inr.ac 1594kuznet@amber:~ $ 1595\end{verbatim} 1596 1597If you list cloned entries, the output contains other attributes which 1598are evaluated during route calculation and updated during route 1599lifetime. An example of the output is: 1600\begin{verbatim} 1601kuznet@amber:~ $ ip ro ls 193.233.7.82 tab cache 1602193.233.7.82 from 193.233.7.82 dev eth0 src 193.233.7.65 \ 1603 realms inr.ac/inr.ac 1604 cache <src-direct,redirect> mtu 1500 rtt 300 iif eth0 1605193.233.7.82 dev eth0 src 193.233.7.65 realms inr.ac 1606 cache mtu 1500 rtt 300 1607kuznet@amber:~ $ 1608\end{verbatim} 1609\begin{NB} 1610 \label{NB-strange-route} 1611 The route looks a bit strange, doesn't it? Did you notice that 1612 it is a path from 193.233.7.82 back to 193.233.82? Well, you will 1613 see in the section on \verb|ip route get| (p.\pageref{NB-nature-of-strangeness}) 1614 how it appeared. 1615\end{NB} 1616The second line, starting with the word \verb|cache|, shows 1617additional attributes which normal routes do not possess. 1618Cached flags are summarized in angle brackets: 1619\begin{itemize} 1620\item \verb|local| --- packets are delivered locally. 1621It stands for loopback unicast routes, for broadcast routes 1622and for multicast routes, if this host is a member of the corresponding 1623group. 1624 1625\item \verb|reject| --- the path is bad. Any attempt to use it results 1626in an error. See attribute \verb|error| below (p.\pageref{IP-ROUTE-GET-error}). 1627 1628\item \verb|mc| --- the destination is multicast. 1629 1630\item \verb|brd| --- the destination is broadcast. 1631 1632\item \verb|src-direct| --- the source is on a directly connected 1633interface. 1634 1635\item \verb|redirected| --- the route was created by an ICMP Redirect. 1636 1637\item \verb|redirect| --- packets going via this route will 1638trigger an ICMP redirect. 1639 1640\item \verb|fastroute| --- the route is eligible to be used for fastroute. 1641 1642\item \verb|equalize| --- make packet by packet randomization 1643along this path. 1644 1645\item \verb|dst-nat| --- the destination address requires translation. 1646 1647\item \verb|src-nat| --- the source address requires translation. 1648 1649\item \verb|masq| --- the source address requires masquerading. 1650This feature disappeared in linux-2.4. 1651 1652\item \verb|notify| --- ({\em not implemented}) change/deletion 1653of this route will trigger RTNETLINK notification. 1654\end{itemize} 1655 1656Then some optional attributes follow: 1657\begin{itemize} 1658\item \verb|error| --- on \verb|reject| routes it is error code 1659returned to local senders when they try to use this route. 1660These error codes are translated into ICMP error codes, sent to remote 1661senders, according to the rules described above in the subsection 1662devoted to route types (p.\pageref{IP-ROUTE-TYPES}). 1663\label{IP-ROUTE-GET-error} 1664 1665\item \verb|expires| --- this entry will expire after this timeout. 1666 1667\item \verb|iif| --- the packets for this path are expected to arrive 1668on this interface. 1669\end{itemize} 1670 1671\paragraph{Statistics:} With the \verb|-statistics| option, more 1672information about this route is shown: 1673\begin{itemize} 1674\item \verb|users| --- the number of users of this entry. 1675\item \verb|age| --- shows when this route was last used. 1676\item \verb|used| --- the number of lookups of this route since its creation. 1677\end{itemize} 1678 1679 1680\subsection{{\tt ip route flush} --- flush routing tables} 1681\label{IP-ROUTE-FLUSH} 1682 1683\paragraph{Abbreviations:} \verb|flush|, \verb|f|. 1684 1685\paragraph{Description:} this command flushes routes selected 1686by some criteria. 1687 1688\paragraph{Arguments:} the arguments have the same syntax and semantics 1689as the arguments of \verb|ip route show|, but routing tables are not 1690listed but purged. The only difference is the default action: \verb|show| 1691dumps all the IP main routing table but \verb|flush| prints the helper page. 1692The reason for this difference does not require any explanation, does it? 1693 1694 1695\paragraph{Statistics:} With the \verb|-statistics| option, the command 1696becomes verbose. It prints out the number of deleted routes and the number 1697of rounds made to flush the routing table. If the option is given 1698twice, \verb|ip route flush| also dumps all the deleted routes 1699in the format described in the previous subsection. 1700 1701\paragraph{Examples:} The first example flushes all the 1702gatewayed routes from the main table (f.e.\ after a routing daemon crash). 1703\begin{verbatim} 1704netadm@amber:~ # ip -4 ro flush scope global type unicast 1705\end{verbatim} 1706This option deserves to be put into a scriptlet \verb|routef|. 1707\begin{NB} 1708This option was described in the \verb|route(8)| man page borrowed 1709from BSD, but was never implemented in Linux. 1710\end{NB} 1711 1712The second example flushes all IPv6 cloned routes: 1713\begin{verbatim} 1714netadm@amber:~ # ip -6 -s -s ro flush cache 17153ffe:2400::220:afff:fef4:c5d1 via 3ffe:2400::220:afff:fef4:c5d1 \ 1716 dev eth0 metric 0 1717 cache used 2 age 12sec mtu 1500 rtt 300 17183ffe:2400::280:adff:feb7:8034 via 3ffe:2400::280:adff:feb7:8034 \ 1719 dev eth0 metric 0 1720 cache used 2 age 15sec mtu 1500 rtt 300 17213ffe:2400::280:c8ff:fe59:5bcc via 3ffe:2400::280:c8ff:fe59:5bcc \ 1722 dev eth0 metric 0 1723 cache users 1 used 1 age 23sec mtu 1500 rtt 300 17243ffe:2400:0:1:2a0:ccff:fe66:1878 via 3ffe:2400:0:1:2a0:ccff:fe66:1878 \ 1725 dev eth1 metric 0 1726 cache used 2 age 20sec mtu 1500 rtt 300 17273ffe:2400:0:1:a00:20ff:fe71:fb30 via 3ffe:2400:0:1:a00:20ff:fe71:fb30 \ 1728 dev eth1 metric 0 1729 cache used 2 age 33sec mtu 1500 rtt 300 1730ff02::1 via ff02::1 dev eth1 metric 0 1731 cache users 1 used 1 age 45sec mtu 1500 rtt 300 1732 1733*** Round 1, deleting 6 entries *** 1734*** Flush is complete after 1 round *** 1735netadm@amber:~ # ip -6 -s -s ro flush cache 1736Nothing to flush. 1737netadm@amber:~ # 1738\end{verbatim} 1739 1740The third example flushes BGP routing tables after a \verb|gated| 1741death. 1742\begin{verbatim} 1743netadm@amber:~ # ip ro ls proto gated/bgp | wc 1744 1408 9856 78730 1745netadm@amber:~ # ip -s ro f proto gated/bgp 1746 1747*** Round 1, deleting 1408 entries *** 1748*** Flush is complete after 1 round *** 1749netadm@amber:~ # ip ro f proto gated/bgp 1750Nothing to flush. 1751netadm@amber:~ # ip ro ls proto gated/bgp 1752netadm@amber:~ # 1753\end{verbatim} 1754 1755 1756\subsection{{\tt ip route get} --- get a single route} 1757\label{IP-ROUTE-GET} 1758 1759\paragraph{Abbreviations:} \verb|get|, \verb|g|. 1760 1761\paragraph{Description:} this command gets a single route to a destination 1762and prints its contents exactly as the kernel sees it. 1763 1764\paragraph{Arguments:} 1765\begin{itemize} 1766\item \verb|to ADDRESS| (default) 1767 1768--- the destination address. 1769 1770\item \verb|from ADDRESS| 1771 1772--- the source address. 1773 1774\item \verb|tos TOS| or \verb|dsfield TOS| 1775 1776--- the Type Of Service. 1777 1778\item \verb|iif NAME| 1779 1780--- the device from which this packet is expected to arrive. 1781 1782\item \verb|oif NAME| 1783 1784--- force the output device on which this packet will be routed. 1785 1786\item \verb|connected| 1787 1788--- if no source address (option \verb|from|) was given, relookup 1789the route with the source set to the preferred address received from the first lookup. 1790If policy routing is used, it may be a different route. 1791 1792\end{itemize} 1793 1794Note that this operation is not equivalent to \verb|ip route show|. 1795\verb|show| shows existing routes. \verb|get| resolves them and 1796creates new clones if necessary. Essentially, \verb|get| 1797is equivalent to sending a packet along this path. 1798If the \verb|iif| argument is not given, the kernel creates a route 1799to output packets towards the requested destination. 1800This is equivalent to pinging the destination 1801with a subsequent {\tt ip route ls cache}, however, no packets are 1802actually sent. With the \verb|iif| argument, the kernel pretends 1803that a packet arrived from this interface and searches for 1804a path to forward the packet. 1805 1806\paragraph{Output format:} This command outputs routes in the same 1807format as \verb|ip route ls|. 1808 1809\paragraph{Examples:} 1810\begin{itemize} 1811\item Find a route to output packets to 193.233.7.82: 1812\begin{verbatim} 1813kuznet@amber:~ $ ip route get 193.233.7.82 1814193.233.7.82 dev eth0 src 193.233.7.65 realms inr.ac 1815 cache mtu 1500 rtt 300 1816kuznet@amber:~ $ 1817\end{verbatim} 1818 1819\item Find a route to forward packets arriving on \verb|eth0| 1820from 193.233.7.82 and destined for 193.233.7.82: 1821\begin{verbatim} 1822kuznet@amber:~ $ ip r g 193.233.7.82 from 193.233.7.82 iif eth0 1823193.233.7.82 from 193.233.7.82 dev eth0 src 193.233.7.65 \ 1824 realms inr.ac/inr.ac 1825 cache <src-direct,redirect> mtu 1500 rtt 300 iif eth0 1826kuznet@amber:~ $ 1827\end{verbatim} 1828\begin{NB} 1829 \label{NB-nature-of-strangeness} 1830 This is the command that created the funny route from 193.233.7.82 1831 looped back to 193.233.7.82 (cf.\ NB on~p.\pageref{NB-strange-route}). 1832 Note the \verb|redirect| flag on it. 1833\end{NB} 1834 1835\item Find a multicast route for packets arriving on \verb|eth0| 1836from host 193.233.7.82 and destined for multicast group 224.2.127.254 1837(it is assumed that a multicast routing daemon is running. 1838In this case, it is \verb|pimd|) 1839\begin{verbatim} 1840kuznet@amber:~ $ ip r g 224.2.127.254 from 193.233.7.82 iif eth0 1841multicast 224.2.127.254 from 193.233.7.82 dev lo \ 1842 src 193.233.7.65 realms inr.ac/cosmos 1843 cache <mc> iif eth0 Oifs: eth1 pimreg 1844kuznet@amber:~ $ 1845\end{verbatim} 1846This route differs from the ones seen before. It contains a ``normal'' part 1847and a ``multicast'' part. The normal part is used to deliver (or not to 1848deliver) the packet to local IP listeners. In this case the router 1849is not a member 1850of this group, so that route has no \verb|local| flag and only 1851forwards packets. The output device for such entries is always loopback. 1852The multicast part consists of an additional \verb|Oifs:| list showing 1853the output interfaces. 1854\end{itemize} 1855 1856 1857It is time for a more complicated example. Let us add an invalid 1858gatewayed route for a destination which is really directly connected: 1859\begin{verbatim} 1860netadm@alisa:~ # ip route add 193.233.7.98 via 193.233.7.254 1861netadm@alisa:~ # ip route get 193.233.7.98 1862193.233.7.98 via 193.233.7.254 dev eth0 src 193.233.7.90 1863 cache mtu 1500 rtt 3072 1864netadm@alisa:~ # 1865\end{verbatim} 1866and probe it with ping: 1867\begin{verbatim} 1868netadm@alisa:~ # ping -n 193.233.7.98 1869PING 193.233.7.98 (193.233.7.98) from 193.233.7.90 : 56 data bytes 1870From 193.233.7.254: Redirect Host(New nexthop: 193.233.7.98) 187164 bytes from 193.233.7.98: icmp_seq=0 ttl=255 time=3.5 ms 1872From 193.233.7.254: Redirect Host(New nexthop: 193.233.7.98) 187364 bytes from 193.233.7.98: icmp_seq=1 ttl=255 time=2.2 ms 187464 bytes from 193.233.7.98: icmp_seq=2 ttl=255 time=0.4 ms 187564 bytes from 193.233.7.98: icmp_seq=3 ttl=255 time=0.4 ms 187664 bytes from 193.233.7.98: icmp_seq=4 ttl=255 time=0.4 ms 1877^C 1878--- 193.233.7.98 ping statistics --- 18795 packets transmitted, 5 packets received, 0% packet loss 1880round-trip min/avg/max = 0.4/1.3/3.5 ms 1881netadm@alisa:~ # 1882\end{verbatim} 1883What happened? Router 193.233.7.254 understood that we have a much 1884better path to the destination and sent us an ICMP redirect message. 1885We may retry \verb|ip route get| to see what we have in the routing 1886tables now: 1887\begin{verbatim} 1888netadm@alisa:~ # ip route get 193.233.7.98 1889193.233.7.98 dev eth0 src 193.233.7.90 1890 cache <redirected> mtu 1500 rtt 3072 1891netadm@alisa:~ # 1892\end{verbatim} 1893 1894 1895 1896\section{{\tt ip rule} --- routing policy database management} 1897\label{IP-RULE} 1898 1899\paragraph{Abbreviations:} \verb|rule|, \verb|ru|. 1900 1901\paragraph{Object:} \verb|rule|s in the routing policy database control 1902the route selection algorithm. 1903 1904Classic routing algorithms used in the Internet make routing decisions 1905based only on the destination address of packets (and in theory, 1906but not in practice, on the TOS field). The seminal review of classic 1907routing algorithms and their modifications can be found in~\cite{RFC1812}. 1908 1909In some circumstances we want to route packets differently depending not only 1910on destination addresses, but also on other packet fields: source address, 1911IP protocol, transport protocol ports or even packet payload. 1912This task is called ``policy routing''. 1913 1914\begin{NB} 1915 ``policy routing'' $\neq$ ``routing policy''. 1916 1917\noindent ``policy routing'' $=$ ``cunning routing''. 1918 1919\noindent ``routing policy'' $=$ ``routing tactics'' or ``routing plan''. 1920\end{NB} 1921 1922To solve this task, the conventional destination based routing table, ordered 1923according to the longest match rule, is replaced with a ``routing policy 1924database'' (or RPDB), which selects routes 1925by executing some set of rules. The rules may have lots of keys of different 1926natures and therefore they have no natural ordering, but one imposed 1927by the administrator. Linux-2.2 RPDB is a linear list of rules 1928ordered by numeric priority value. 1929RPDB explicitly allows matching a few packet fields: 1930 1931\begin{itemize} 1932\item packet source address. 1933\item packet destination address. 1934\item TOS. 1935\item incoming interface (which is packet metadata, rather than a packet field). 1936\end{itemize} 1937 1938Matching IP protocols and transport ports is also possible, 1939indirectly, via \verb|ipchains|, by exploiting their ability 1940to mark some classes of packets with \verb|fwmark|. Therefore, 1941\verb|fwmark| is also included in the set of keys checked by rules. 1942 1943Each policy routing rule consists of a {\em selector\/} and an {\em action\/} 1944predicate. The RPDB is scanned in the order of increasing priority. The selector 1945of each rule is applied to \{source address, destination address, incoming 1946interface, tos, fwmark\} and, if the selector matches the packet, 1947the action is performed. The action predicate may return with success. 1948In this case, it will either give a route or failure indication 1949and the RPDB lookup is terminated. Otherwise, the RPDB program 1950continues on the next rule. 1951 1952What is the action, semantically? The natural action is to select the 1953nexthop and the output device. This is what 1954Cisco IOS~\cite{IOS} does. Let us call it ``match \& set''. 1955The Linux-2.2 approach is more flexible. The action includes 1956lookups in destination-based routing tables and selecting 1957a route from these tables according to the classic longest match algorithm. 1958The ``match \& set'' approach is the simplest case of the Linux one. It is realized 1959when a second level routing table contains a single default route. 1960Recall that Linux-2.2 supports multiple tables 1961managed with the \verb|ip route| command, described in the previous section. 1962 1963At startup time the kernel configures the default RPDB consisting of three 1964rules: 1965 1966\begin{enumerate} 1967\item Priority: 0, Selector: match anything, Action: lookup routing 1968table \verb|local| (ID 255). 1969The \verb|local| table is a special routing table containing 1970high priority control routes for local and broadcast addresses. 1971 1972Rule 0 is special. It cannot be deleted or overridden. 1973 1974 1975\item Priority: 32766, Selector: match anything, Action: lookup routing 1976table \verb|main| (ID 254). 1977The \verb|main| table is the normal routing table containing all non-policy 1978routes. This rule may be deleted and/or overridden with other 1979ones by the administrator. 1980 1981\item Priority: 32767, Selector: match anything, Action: lookup routing 1982table \verb|default| (ID 253). 1983The \verb|default| table is empty. It is reserved for some 1984post-processing if no previous default rules selected the packet. 1985This rule may also be deleted. 1986 1987\end{enumerate} 1988 1989Do not confuse routing tables with rules: rules point to routing tables, 1990several rules may refer to one routing table and some routing tables 1991may have no rules pointing to them. If the administrator deletes all the rules 1992referring to a table, the table is not used, but it still exists 1993and will disappear only after all the routes contained in it are deleted. 1994 1995 1996\paragraph{Rule attributes:} Each RPDB entry has additional 1997attributes. F.e.\ each rule has a pointer to some routing 1998table. NAT and masquerading rules have an attribute to select new IP 1999address to translate/masquerade. Besides that, rules have some 2000optional attributes, which routes have, namely \verb|realms|. 2001These values do not override those contained in the routing tables. They 2002are only used if the route did not select any attributes. 2003 2004 2005\paragraph{Rule types:} The RPDB may contain rules of the following 2006types: 2007\begin{itemize} 2008\item \verb|unicast| --- the rule prescribes to return the route found 2009in the routing table referenced by the rule. 2010\item \verb|blackhole| --- the rule prescribes to silently drop the packet. 2011\item \verb|unreachable| --- the rule prescribes to generate a ``Network 2012is unreachable'' error. 2013\item \verb|prohibit| --- the rule prescribes to generate 2014``Communication is administratively prohibited'' error. 2015\item \verb|nat| --- the rule prescribes to translate the source address 2016of the IP packet into some other value. More about NAT is 2017in Appendix~\ref{ROUTE-NAT}, p.\pageref{ROUTE-NAT}. 2018\end{itemize} 2019 2020 2021\paragraph{Commands:} \verb|add|, \verb|delete| and \verb|show| 2022(or \verb|list|). 2023 2024\subsection{{\tt ip rule add} --- insert a new rule\\ 2025 {\tt ip rule delete} --- delete a rule} 2026\label{IP-RULE-ADD} 2027 2028\paragraph{Abbreviations:} \verb|add|, \verb|a|; \verb|delete|, \verb|del|, 2029 \verb|d|. 2030 2031\paragraph{Arguments:} 2032 2033\begin{itemize} 2034\item \verb|type TYPE| (default) 2035 2036--- the type of this rule. The list of valid types was given in the previous 2037subsection. 2038 2039\item \verb|from PREFIX| 2040 2041--- select the source prefix to match. 2042 2043\item \verb|to PREFIX| 2044 2045--- select the destination prefix to match. 2046 2047\item \verb|iif NAME| 2048 2049--- select the incoming device to match. If the interface is loopback, 2050the rule only matches packets originating from this host. This means that you 2051may create separate routing tables for forwarded and local packets and, 2052hence, completely segregate them. 2053 2054\item \verb|tos TOS| or \verb|dsfield TOS| 2055 2056--- select the TOS value to match. 2057 2058\item \verb|fwmark MARK| 2059 2060--- select the \verb|fwmark| value to match. 2061 2062\item \verb|priority PREFERENCE| 2063 2064--- the priority of this rule. Each rule should have an explicitly 2065set {\em unique\/} priority value. 2066\begin{NB} 2067 Really, for historical reasons \verb|ip rule add| does not require a 2068 priority value and allows them to be non-unique. 2069 If the user does not supplied a priority, it is selected by the kernel. 2070 If the user creates a rule with a priority value that 2071 already exists, the kernel does not reject the request. It adds 2072 the new rule before all old rules of the same priority. 2073 2074 It is mistake in design, no more. And it will be fixed one day, 2075 so do not rely on this feature. Use explicit priorities. 2076\end{NB} 2077 2078 2079\item \verb|table TABLEID| 2080 2081--- the routing table identifier to lookup if the rule selector matches. 2082 2083\item \verb|realms FROM/TO| 2084 2085--- Realms to select if the rule matched and the routing table lookup 2086succeeded. Realm \verb|TO| is only used if the route did not select 2087any realm. 2088 2089\item \verb|nat ADDRESS| 2090 2091--- The base of the IP address block to translate (for source addresses). 2092The \verb|ADDRESS| may be either the start of the block of NAT addresses 2093(selected by NAT routes) or in linux-2.2 a local host address (or even zero). 2094In the last case the router does not translate the packets, 2095but masquerades them to this address; this feature disappered in 2.4. 2096More about NAT is in Appendix~\ref{ROUTE-NAT}, 2097p.\pageref{ROUTE-NAT}. 2098 2099\end{itemize} 2100 2101\paragraph{Warning:} Changes to the RPDB made with these commands 2102do not become active immediately. It is assumed that after 2103a script finishes a batch of updates, it flushes the routing cache 2104with \verb|ip route flush cache|. 2105 2106\paragraph{Examples:} 2107\begin{itemize} 2108\item Route packets with source addresses from 192.203.80/24 2109according to routing table \verb|inr.ruhep|: 2110\begin{verbatim} 2111ip ru add from 192.203.80.0/24 table inr.ruhep prio 220 2112\end{verbatim} 2113 2114\item Translate packet source address 193.233.7.83 into 192.203.80.144 2115and route it according to table \#1 (actually, it is \verb|inr.ruhep|): 2116\begin{verbatim} 2117ip ru add from 193.233.7.83 nat 192.203.80.144 table 1 prio 320 2118\end{verbatim} 2119 2120\item Delete the unused default rule: 2121\begin{verbatim} 2122ip ru del prio 32767 2123\end{verbatim} 2124 2125\end{itemize} 2126 2127 2128 2129\subsection{{\tt ip rule show} --- list rules} 2130\label{IP-RULE-SHOW} 2131 2132\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|sh|, \verb|ls|, \verb|l|. 2133 2134 2135\paragraph{Arguments:} Good news, this is one command that has no arguments. 2136 2137\paragraph{Output format:} 2138 2139\begin{verbatim} 2140kuznet@amber:~ $ ip ru ls 21410: from all lookup local 2142200: from 192.203.80.0/24 to 193.233.7.0/24 lookup main 2143210: from 192.203.80.0/24 to 192.203.80.0/24 lookup main 2144220: from 192.203.80.0/24 lookup inr.ruhep realms inr.ruhep/radio-msu 2145300: from 193.233.7.83 to 193.233.7.0/24 lookup main 2146310: from 193.233.7.83 to 192.203.80.0/24 lookup main 2147320: from 193.233.7.83 lookup inr.ruhep map-to 192.203.80.144 214832766: from all lookup main 2149kuznet@amber:~ $ 2150\end{verbatim} 2151 2152In the first column is the rule priority value followed 2153by a colon. Then the selectors follow. Each key is prefixed 2154with the same keyword that was used to create the rule. 2155 2156The keyword \verb|lookup| is followed by a routing table identifier, 2157as it is recorded in the file \verb|/etc/iproute2/rt_tables|. 2158 2159If the rule does NAT (f.e.\ rule \#320), it is shown by the keyword 2160\verb|map-to| followed by the start of the block of addresses to map. 2161 2162The sense of this example is pretty simple. The prefixes 2163192.203.80.0/24 and 193.233.7.0/24 form the internal network, but 2164they are routed differently when the packets leave it. 2165Besides that, the host 193.233.7.83 is translated into 2166another prefix to look like 192.203.80.144 when talking 2167to the outer world. 2168 2169 2170 2171\section{{\tt ip maddress} --- multicast addresses management} 2172\label{IP-MADDR} 2173 2174\paragraph{Object:} \verb|maddress| objects are multicast addresses. 2175 2176\paragraph{Commands:} \verb|add|, \verb|delete|, \verb|show| (or \verb|list|). 2177 2178\subsection{{\tt ip maddress show} --- list multicast addresses} 2179 2180\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|sh|, \verb|ls|, \verb|l|. 2181 2182\paragraph{Arguments:} 2183 2184\begin{itemize} 2185 2186\item \verb|dev NAME| (default) 2187 2188--- the device name. 2189 2190\end{itemize} 2191 2192\paragraph{Output format:} 2193 2194\begin{verbatim} 2195kuznet@alisa:~ $ ip maddr ls dummy 21962: dummy 2197 link 33:33:00:00:00:01 2198 link 01:00:5e:00:00:01 2199 inet 224.0.0.1 users 2 2200 inet6 ff02::1 2201kuznet@alisa:~ $ 2202\end{verbatim} 2203 2204The first line of the output shows the interface index and its name. 2205Then the multicast address list follows. Each line starts with the 2206protocol identifier. The word \verb|link| denotes a link layer 2207multicast addresses. 2208 2209If a multicast address has more than one user, the number 2210of users is shown after the \verb|users| keyword. 2211 2212One additional feature not present in the example above 2213is the \verb|static| flag, which indicates that the address was joined 2214with \verb|ip maddr add|. See the following subsection. 2215 2216 2217 2218\subsection{{\tt ip maddress add} --- add a multicast address\\ 2219 {\tt ip maddress delete} --- delete a multicast address} 2220 2221\paragraph{Abbreviations:} \verb|add|, \verb|a|; \verb|delete|, \verb|del|, \verb|d|. 2222 2223\paragraph{Description:} these commands attach/detach 2224a static link layer multicast address to listen on the interface. 2225Note that it is impossible to join protocol multicast groups 2226statically. This command only manages link layer addresses. 2227 2228 2229\paragraph{Arguments:} 2230 2231\begin{itemize} 2232\item \verb|address LLADDRESS| (default) 2233 2234--- the link layer multicast address. 2235 2236\item \verb|dev NAME| 2237 2238--- the device to join/leave this multicast address. 2239 2240\end{itemize} 2241 2242 2243\paragraph{Example:} Let us continue with the example from the previous subsection. 2244 2245\begin{verbatim} 2246netadm@alisa:~ # ip maddr add 33:33:00:00:00:01 dev dummy 2247netadm@alisa:~ # ip -0 maddr ls dummy 22482: dummy 2249 link 33:33:00:00:00:01 users 2 static 2250 link 01:00:5e:00:00:01 2251netadm@alisa:~ # ip maddr del 33:33:00:00:00:01 dev dummy 2252\end{verbatim} 2253 2254\begin{NB} 2255 Neither \verb|ip| nor the kernel check for multicast address validity. 2256 Particularly, this means that you can try to load a unicast address 2257 instead of a multicast address. Most drivers will ignore such addresses, 2258 but several (f.e.\ Tulip) will intern it to their on-board filter. 2259 The effects may be strange. Namely, the addresses become additional 2260 local link addresses and, if you loaded the address of another host 2261 to the router, wait for duplicated packets on the wire. 2262 It is not a bug, but rather a hole in the API and intra-kernel interfaces. 2263 This feature is really more useful for traffic monitoring, but using it 2264 with Linux-2.2 you {\em have to\/} be sure that the host is not 2265 a router and, especially, that it is not a transparent proxy or masquerading 2266 agent. 2267\end{NB} 2268 2269 2270 2271\section{{\tt ip mroute} --- multicast routing cache management} 2272\label{IP-MROUTE} 2273 2274\paragraph{Abbreviations:} \verb|mroute|, \verb|mr|. 2275 2276\paragraph{Object:} \verb|mroute| objects are multicast routing cache 2277entries created by a user level mrouting daemon 2278(f.e.\ \verb|pimd| or \verb|mrouted|). 2279 2280Due to the limitations of the current interface to the multicast routing 2281engine, it is impossible to change \verb|mroute| objects administratively, 2282so we may only display them. This limitation will be removed 2283in the future. 2284 2285\paragraph{Commands:} \verb|show| (or \verb|list|). 2286 2287 2288\subsection{{\tt ip mroute show} --- list mroute cache entries} 2289 2290\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|sh|, \verb|ls|, \verb|l|. 2291 2292\paragraph{Arguments:} 2293 2294\begin{itemize} 2295\item \verb|to PREFIX| (default) 2296 2297--- the prefix selecting the destination multicast addresses to list. 2298 2299 2300\item \verb|iif NAME| 2301 2302--- the interface on which multicast packets are received. 2303 2304 2305\item \verb|from PREFIX| 2306 2307--- the prefix selecting the IP source addresses of the multicast route. 2308 2309 2310\end{itemize} 2311 2312\paragraph{Output format:} 2313 2314\begin{verbatim} 2315kuznet@amber:~ $ ip mroute ls 2316(193.232.127.6, 224.0.1.39) Iif: unresolved 2317(193.232.244.34, 224.0.1.40) Iif: unresolved 2318(193.233.7.65, 224.66.66.66) Iif: eth0 Oifs: pimreg 2319kuznet@amber:~ $ 2320\end{verbatim} 2321 2322Each line shows one (S,G) entry in the multicast routing cache, 2323where S is the source address and G is the multicast group. \verb|Iif| is 2324the interface on which multicast packets are expected to arrive. 2325If the word \verb|unresolved| is there instead of the interface name, 2326it means that the routing daemon still hasn't resolved this entry. 2327The keyword \verb|oifs| is followed by a list of output interfaces, separated 2328by spaces. If a multicast routing entry is created with non-trivial 2329TTL scope, administrative distances are appended to the device names 2330in the \verb|oifs| list. 2331 2332\paragraph{Statistics:} The \verb|-statistics| option also prints the 2333number of packets and bytes forwarded along this route and 2334the number of packets that arrived on the wrong interface, if this number is not zero. 2335 2336\begin{verbatim} 2337kuznet@amber:~ $ ip -s mr ls 224.66/16 2338(193.233.7.65, 224.66.66.66) Iif: eth0 Oifs: pimreg 2339 9383 packets, 300256 bytes 2340kuznet@amber:~ $ 2341\end{verbatim} 2342 2343 2344\section{{\tt ip tunnel} --- tunnel configuration} 2345\label{IP-TUNNEL} 2346 2347\paragraph{Abbreviations:} \verb|tunnel|, \verb|tunl|. 2348 2349\paragraph{Object:} \verb|tunnel| objects are tunnels, encapsulating 2350packets in IPv4 packets and then sending them over the IP infrastructure. 2351 2352\paragraph{Commands:} \verb|add|, \verb|delete|, \verb|change|, \verb|show| 2353(or \verb|list|). 2354 2355\paragraph{See also:} A more informal discussion of tunneling 2356over IP and the \verb|ip tunnel| command can be found in~\cite{IP-TUNNELS}. 2357 2358\subsection{{\tt ip tunnel add} --- add a new tunnel\\ 2359 {\tt ip tunnel change} --- change an existing tunnel\\ 2360 {\tt ip tunnel delete} --- destroy a tunnel} 2361 2362\paragraph{Abbreviations:} \verb|add|, \verb|a|; \verb|change|, \verb|chg|; 2363\verb|delete|, \verb|del|, \verb|d|. 2364 2365 2366\paragraph{Arguments:} 2367 2368\begin{itemize} 2369 2370\item \verb|name NAME| (default) 2371 2372--- select the tunnel device name. 2373 2374\item \verb|mode MODE| 2375 2376--- set the tunnel mode. Three modes are currently available: 2377 \verb|ipip|, \verb|sit| and \verb|gre|. 2378 2379\item \verb|remote ADDRESS| 2380 2381--- set the remote endpoint of the tunnel. 2382 2383\item \verb|local ADDRESS| 2384 2385--- set the fixed local address for tunneled packets. 2386It must be an address on another interface of this host. 2387 2388\item \verb|ttl N| 2389 2390--- set a fixed TTL \verb|N| on tunneled packets. 2391 \verb|N| is a number in the range 1--255. 0 is a special value 2392 meaning that packets inherit the TTL value. 2393 The default value is: \verb|inherit|. 2394 2395\item \verb|tos T| or \verb|dsfield T| 2396 2397--- set a fixed TOS \verb|T| on tunneled packets. 2398 The default value is: \verb|inherit|. 2399 2400 2401 2402\item \verb|dev NAME| 2403 2404--- bind the tunnel to the device \verb|NAME| so that 2405 tunneled packets will only be routed via this device and will 2406 not be able to escape to another device when the route to endpoint changes. 2407 2408\item \verb|nopmtudisc| 2409 2410--- disable Path MTU Discovery on this tunnel. 2411 It is enabled by default. Note that a fixed ttl is incompatible 2412 with this option: tunnelling with a fixed ttl always makes pmtu discovery. 2413 2414\item \verb|key K|, \verb|ikey K|, \verb|okey K| 2415 2416--- (only GRE tunnels) use keyed GRE with key \verb|K|. \verb|K| is 2417 either a number or an IP address-like dotted quad. 2418 The \verb|key| parameter sets the key to use in both directions. 2419 The \verb|ikey| and \verb|okey| parameters set different keys for input and output. 2420 2421 2422\item \verb|csum|, \verb|icsum|, \verb|ocsum| 2423 2424--- (only GRE tunnels) generate/require checksums for tunneled packets. 2425 The \verb|ocsum| flag calculates checksums for outgoing packets. 2426 The \verb|icsum| flag requires that all input packets have the correct 2427 checksum. The \verb|csum| flag is equivalent to the combination 2428 ``\verb|icsum| \verb|ocsum|''. 2429 2430\item \verb|seq|, \verb|iseq|, \verb|oseq| 2431 2432--- (only GRE tunnels) serialize packets. 2433 The \verb|oseq| flag enables sequencing of outgoing packets. 2434 The \verb|iseq| flag requires that all input packets are serialized. 2435 The \verb|seq| flag is equivalent to the combination ``\verb|iseq| \verb|oseq|''. 2436 2437\begin{NB} 2438 I think this option does not 2439 work. At least, I did not test it, did not debug it and 2440 do not even understand how it is supposed to work or for what 2441 purpose Cisco planned to use it. Do not use it. 2442\end{NB} 2443 2444 2445\end{itemize} 2446 2447\paragraph{Example:} Create a pointopoint IPv6 tunnel with maximal TTL of 32. 2448\begin{verbatim} 2449netadm@amber:~ # ip tunl add Cisco mode sit remote 192.31.7.104 \ 2450 local 192.203.80.142 ttl 32 2451\end{verbatim} 2452 2453\subsection{{\tt ip tunnel show} --- list tunnels} 2454 2455\paragraph{Abbreviations:} \verb|show|, \verb|list|, \verb|sh|, \verb|ls|, \verb|l|. 2456 2457 2458\paragraph{Arguments:} None. 2459 2460\paragraph{Output format:} 2461\begin{verbatim} 2462kuznet@amber:~ $ ip tunl ls Cisco 2463Cisco: ipv6/ip remote 192.31.7.104 local 192.203.80.142 ttl 32 2464kuznet@amber:~ $ 2465\end{verbatim} 2466The line starts with the tunnel device name followed by a colon. 2467Then the tunnel mode follows. The parameters of the tunnel are listed 2468with the same keywords that were used when creating the tunnel. 2469 2470\paragraph{Statistics:} 2471 2472\begin{verbatim} 2473kuznet@amber:~ $ ip -s tunl ls Cisco 2474Cisco: ipv6/ip remote 192.31.7.104 local 192.203.80.142 ttl 32 2475RX: Packets Bytes Errors CsumErrs OutOfSeq Mcasts 2476 12566 1707516 0 0 0 0 2477TX: Packets Bytes Errors DeadLoop NoRoute NoBufs 2478 13445 1879677 0 0 0 0 2479kuznet@amber:~ $ 2480\end{verbatim} 2481Essentially, these numbers are the same as the numbers 2482printed with {\tt ip -s link show} 2483(sec.\ref{IP-LINK-SHOW}, p.\pageref{IP-LINK-SHOW}) but the tags are different 2484to reflect that they are tunnel specific. 2485\begin{itemize} 2486\item \verb|CsumErrs| --- the total number of packets dropped 2487because of checksum failures for a GRE tunnel with checksumming enabled. 2488\item \verb|OutOfSeq| --- the total number of packets dropped 2489because they arrived out of sequence for a GRE tunnel with 2490serialization enabled. 2491\item \verb|Mcasts| --- the total number of multicast packets 2492received on a broadcast GRE tunnel. 2493\item \verb|DeadLoop| --- the total number of packets which were not 2494transmitted because the tunnel is looped back to itself. 2495\item \verb|NoRoute| --- the total number of packets which were not 2496transmitted because there is no IP route to the remote endpoint. 2497\item \verb|NoBufs| --- the total number of packets which were not 2498transmitted because the kernel failed to allocate a buffer. 2499\end{itemize} 2500 2501 2502\section{{\tt ip monitor} and {\tt rtmon} --- state monitoring} 2503\label{IP-MONITOR} 2504 2505The \verb|ip| utility can monitor the state of devices, addresses 2506and routes continuously. This option has a slightly different format. 2507Namely, 2508the \verb|monitor| command is the first in the command line and then 2509the object list follows: 2510\begin{verbatim} 2511 ip monitor [ file FILE ] [ all | OBJECT-LIST ] 2512\end{verbatim} 2513\verb|OBJECT-LIST| is the list of object types that we want to monitor. 2514It may contain \verb|link|, \verb|address| and \verb|route|. 2515If no \verb|file| argument is given, \verb|ip| opens RTNETLINK, 2516listens on it and dumps state changes in the format described 2517in previous sections. 2518 2519If a file name is given, it does not listen on RTNETLINK, 2520but opens the file containing RTNETLINK messages saved in binary format 2521and dumps them. Such a history file can be generated with the 2522\verb|rtmon| utility. This utility has a command line syntax similar to 2523\verb|ip monitor|. 2524Ideally, \verb|rtmon| should be started before 2525the first network configuration command is issued. F.e.\ if 2526you insert: 2527\begin{verbatim} 2528 rtmon file /var/log/rtmon.log 2529\end{verbatim} 2530in a startup script, you will be able to view the full history 2531later. 2532 2533Certainly, it is possible to start \verb|rtmon| at any time. 2534It prepends the history with the state snapshot dumped at the moment 2535of starting. 2536 2537 2538\section{Route realms and policy propagation, {\tt rtacct}} 2539\label{RT-REALMS} 2540 2541On routers using OSPF ASE or, especially, the BGP protocol, routing 2542tables may be huge. If we want to classify or to account for the packets 2543per route, we will have to keep lots of information. Even worse, if we 2544want to distinguish the packets not only by their destination, but 2545also by their source, the task gets quadratic complexity and its solution 2546is physically impossible. 2547 2548One approach to propagating the policy from routing protocols 2549to the forwarding engine has been proposed in~\cite{IOS-BGP-PP}. 2550Essentially, Cisco Policy Propagation via BGP is based on the fact 2551that dedicated routers all have the RIB (Routing Information Base) 2552close to the forwarding engine, so policy routing rules can 2553check all the route attributes, including ASPATH information 2554and community strings. 2555 2556The Linux architecture, splitting the RIB (maintained by a user level 2557daemon) and the kernel based FIB (Forwarding Information Base), 2558does not allow such a simple approach. 2559 2560It is to our fortune because there is another solution 2561which allows even more flexible policy and richer semantics. 2562 2563Namely, routes can be clustered together in user space, based on their 2564attributes. F.e.\ a BGP router knows route ASPATH, its community; 2565an OSPF router knows the route tag or its area. The administrator, when adding 2566routes manually, also knows their nature. Providing that the number of such 2567aggregates (we call them {\em realms\/}) is low, the task of full 2568classification both by source and destination becomes quite manageable. 2569 2570So each route may be assigned to a realm. It is assumed that 2571this identification is made by a routing daemon, but static routes 2572can also be handled manually with \verb|ip route| (see sec.\ref{IP-ROUTE}, 2573p.\pageref{IP-ROUTE}). 2574\begin{NB} 2575 There is a patch to \verb|gated|, allowing classification of routes 2576 to realms with all the set of policy rules implemented in \verb|gated|: 2577 by prefix, by ASPATH, by origin, by tag etc. 2578\end{NB} 2579 2580To facilitate the construction (f.e.\ in case the routing 2581daemon is not aware of realms), missing realms may be completed 2582with routing policy rules, see sec.~\ref{IP-RULE}, p.\pageref{IP-RULE}. 2583 2584For each packet the kernel calculates a tuple of realms: source realm 2585and destination realm, using the following algorithm: 2586 2587\begin{enumerate} 2588\item If the route has a realm, the destination realm of the packet is set to it. 2589\item If the rule has a source realm, the source realm of the packet is set to it. 2590If the destination realm was not inherited from the route and the rule has a destination realm, 2591it is also set. 2592\item If at least one of the realms is still unknown, the kernel finds 2593the reversed route to the source of the packet. 2594\item If the source realm is still unknown, get it from the reversed route. 2595\item If one of the realms is still unknown, swap the realms of reversed 2596routes and apply step 2 again. 2597\end{enumerate} 2598 2599After this procedure is completed we know what realm the packet 2600arrived from and the realm where it is going to propagate to. 2601If some of the realms are unknown, they are initialized to zero 2602(or realm \verb|unknown|). 2603 2604The main application of realms is the TC \verb|route| classifier~\cite{TC-CREF}, 2605where they are used to help assign packets to traffic classes, 2606to account, police and schedule them according to this 2607classification. 2608 2609A much simpler but still very useful application is incoming packet 2610accounting by realms. The kernel gathers a packet statistics summary 2611which can be viewed with the \verb|rtacct| utility. 2612\begin{verbatim} 2613kuznet@amber:~ $ rtacct russia 2614Realm BytesTo PktsTo BytesFrom PktsFrom 2615russia 20576778 169176 47080168 153805 2616kuznet@amber:~ $ 2617\end{verbatim} 2618This shows that this router received 153805 packets from 2619the realm \verb|russia| and forwarded 169176 packets to \verb|russia|. 2620The realm \verb|russia| consists of routes with ASPATHs not leaving 2621Russia. 2622 2623Note that locally originating packets are not accounted here, 2624\verb|rtacct| shows incoming packets only. Using the \verb|route| 2625classifier (see~\cite{TC-CREF}) you can get even more detailed 2626accounting information about outgoing packets, optionally 2627summarizing traffic not only by source or destination, but 2628by any pair of source and destination realms. 2629 2630 2631\begin{thebibliography}{99} 2632\addcontentsline{toc}{section}{References} 2633\bibitem{RFC-NDISC} T.~Narten, E.~Nordmark, W.~Simpson. 2634``Neighbor Discovery for IP Version 6 (IPv6)'', RFC-2461. 2635 2636\bibitem{RFC-ADDRCONF} S.~Thomson, T.~Narten. 2637``IPv6 Stateless Address Autoconfiguration'', RFC-2462. 2638 2639\bibitem{RFC1812} F.~Baker. 2640``Requirements for IP Version 4 Routers'', RFC-1812. 2641 2642\bibitem{RFC1122} R.~T.~Braden. 2643``Requirements for Internet hosts --- communication layers'', RFC-1122. 2644 2645\bibitem{IOS} ``Cisco IOS Release 12.0 Network Protocols 2646Command Reference, Part 1'' and 2647``Cisco IOS Release 12.0 Quality of Service Solutions 2648Configuration Guide: Configuring Policy-Based Routing'',\\ 2649http://www.cisco.com/univercd/cc/td/doc/product/software/ios120. 2650 2651\bibitem{IP-TUNNELS} A.~N.~Kuznetsov. 2652``Tunnels over IP in Linux-2.2'', \\ 2653In: {\tt ftp://ftp.inr.ac.ru/ip-routing/iproute2-current.tar.gz}. 2654 2655\bibitem{TC-CREF} A.~N.~Kuznetsov. ``TC Command Reference'',\\ 2656In: {\tt ftp://ftp.inr.ac.ru/ip-routing/iproute2-current.tar.gz}. 2657 2658\bibitem{IOS-BGP-PP} ``Cisco IOS Release 12.0 Quality of Service Solutions 2659Configuration Guide: Configuring QoS Policy Propagation via 2660Border Gateway Protocol'',\\ 2661http://www.cisco.com/univercd/cc/td/doc/product/software/ios120. 2662 2663\bibitem{RFC-DHCP} R.~Droms. 2664``Dynamic Host Configuration Protocol.'', RFC-2131 2665 2666\bibitem{RFC2414} M.~Allman, S.~Floyd, C.~Partridge. 2667``Increasing TCP's Initial Window'', RFC-2414. 2668 2669\end{thebibliography} 2670 2671 2672 2673 2674\appendix 2675\addcontentsline{toc}{section}{Appendix} 2676 2677\section{Source address selection} 2678\label{ADDR-SEL} 2679 2680When a host creates an IP packet, it must select some source 2681address. Correct source address selection is a critical procedure, 2682because it gives the receiver the information needed to deliver a 2683reply. If the source is selected incorrectly, in the best case, 2684the backward path may appear different to the forward one which 2685is harmful for performance. In the worst case, when the addresses 2686are administratively scoped, the reply may be lost entirely. 2687 2688Linux-2.2 selects source addresses using the following algorithm: 2689 2690\begin{itemize} 2691\item 2692The application may select a source address explicitly with \verb|bind(2)| 2693syscall or supplying it to \verb|sendmsg(2)| via the ancillary data object 2694\verb|IP_PKTINFO|. In this case the kernel only checks the validity 2695of the address and never tries to ``improve'' an incorrect user choice, 2696generating an error instead. 2697\begin{NB} 2698 Never say ``Never''. The sysctl option \verb|ip_dynaddr| breaks 2699 this axiom. It has been made deliberately with the purpose 2700 of automatically reselecting the address on hosts with dynamic dial-out interfaces. 2701 However, this hack {\em must not\/} be used on multihomed hosts 2702 and especially on routers: it would break them. 2703\end{NB} 2704 2705 2706\item Otherwise, IP routing tables can contain an explicit source 2707address hint for this destination. The hint is set with the \verb|src| parameter 2708to the \verb|ip route| command, sec.\ref{IP-ROUTE}, p.\pageref{IP-ROUTE}. 2709 2710 2711\item Otherwise, the kernel searches through the list of addresses 2712attached to the interface through which the packets will be routed. 2713The search strategies are different for IP and IPv6. Namely: 2714 2715\begin{itemize} 2716\item IPv6 searches for the first valid, not deprecated address 2717with the same scope as the destination. 2718 2719\item IP searches for the first valid address with a scope wider 2720than the scope of the destination but it prefers addresses 2721which fall to the same subnet as the nexthop of the route 2722to the destination. Unlike IPv6, the scopes of IPv4 destinations 2723are not encoded in their addresses but are supplied 2724in routing tables instead (the \verb|scope| parameter to the \verb|ip route| command, 2725sec.\ref{IP-ROUTE}, p.\pageref{IP-ROUTE}). 2726 2727\end{itemize} 2728 2729 2730\item Otherwise, if the scope of the destination is \verb|link| or \verb|host|, 2731the algorithm fails and returns a zero source address. 2732 2733\item Otherwise, all interfaces are scanned to search for an address 2734with an appropriate scope. The loopback device \verb|lo| is always the first 2735in the search list, so that if an address with global scope (not 127.0.0.1!) 2736is configured on loopback, it is always preferred. 2737 2738\end{itemize} 2739 2740 2741\section{Proxy ARP/NDISC} 2742\label{PROXY-NEIGH} 2743 2744Routers may answer ARP/NDISC solicitations on behalf of other hosts. 2745In Linux-2.2 proxy ARP on an interface may be enabled 2746by setting the kernel \verb|sysctl| variable 2747\verb|/proc/sys/net/ipv4/conf/<dev>/proxy_arp| to 1. After this, the router 2748starts to answer ARP requests on the interface \verb|<dev>|, provided 2749the route to the requested destination does {\em not\/} go back via the same 2750device. 2751 2752The variable \verb|/proc/sys/net/ipv4/conf/all/proxy_arp| enables proxy 2753ARP on all the IP devices. 2754 2755However, this approach fails in the case of IPv6 because the router 2756must join the solicited node multicast address to listen for the corresponding 2757NDISC queries. It means that proxy NDISC is possible only on a per destination 2758basis. 2759 2760Logically, proxy ARP/NDISC is not a kernel task. It can easily be implemented 2761in user space. However, similar functionality was present in BSD kernels 2762and in Linux-2.0, so we have to preserve it at least to the extent that 2763is standardized in BSD. 2764\begin{NB} 2765 Linux-2.0 ARP had a feature called {\em subnet\/} proxy ARP. 2766 It is replaced with the sysctl flag in Linux-2.2. 2767\end{NB} 2768 2769 2770The \verb|ip| utility provides a way to manage proxy ARP/NDISC 2771with the \verb|ip neigh| command, namely: 2772\begin{verbatim} 2773 ip neigh add proxy ADDRESS [ dev NAME ] 2774\end{verbatim} 2775adds a new proxy ARP/NDISC record and 2776\begin{verbatim} 2777 ip neigh del proxy ADDRESS [ dev NAME ] 2778\end{verbatim} 2779deletes it. 2780 2781If the name of the device is not given, the router will answer solicitations 2782for address \verb|ADDRESS| on all devices, otherwise it will only serve 2783the device \verb|NAME|. Even if the proxy entry is created with 2784\verb|ip neigh|, the router {\em will not\/} answer a query if the route 2785to the destination goes back via the interface from which the solicitation 2786was received. 2787 2788It is important to emphasize that proxy entries have {\em no\/} 2789parameters other than these (IP/IPv6 address and optional device). 2790Particularly, the entry does not store any link layer address. 2791It always advertises the station address of the interface 2792on which it sends advertisements (i.e. it's own station address). 2793 2794\section{Route NAT status} 2795\label{ROUTE-NAT} 2796 2797NAT (or ``Network Address Translation'') remaps some parts 2798of the IP address space into other ones. Linux-2.2 route NAT is supposed 2799to be used to facilitate policy routing by rewriting addresses 2800to other routing domains or to help while renumbering sites 2801to another prefix. 2802 2803\paragraph{What it is not:} 2804It is necessary to emphasize that {\em it is not supposed\/} 2805to be used to compress address space or to split load. 2806This is not missing functionality but a design principle. 2807Route NAT is {\em stateless\/}. It does not hold any state 2808about translated sessions. This means that it handles any number 2809of sessions flawlessly. But it also means that it is {\em static\/}. 2810It cannot detect the moment when the last TCP client stops 2811using an address. For the same reason, it will not help to split 2812load between several servers. 2813\begin{NB} 2814It is a pretty commonly held belief that it is useful to split load between 2815several servers with NAT. This is a mistake. All you get from this 2816is the requirement that the router keep the state of all the TCP connections 2817going via it. Well, if the router is so powerful, run apache on it. 8) 2818\end{NB} 2819 2820The second feature: it does not touch packet payload, 2821does not try to ``improve'' broken protocols by looking 2822through its data and mangling it. It mangles IP addresses, 2823only IP addresses and nothing but IP addresses. 2824This also, is not missing any functionality. 2825 2826To resume: if you need to compress address space or keep 2827active FTP clients happy, your choice is not route NAT but masquerading, 2828port forwarding, NAPT etc. 2829\begin{NB} 2830By the way, you may also want to look at 2831http://www.suse.com/\~mha/HyperNews/get/linux-ip-nat.html 2832\end{NB} 2833 2834 2835\paragraph{How it works.} 2836Some part of the address space is reserved for dummy addresses 2837which will look for all the world like some host addresses 2838inside your network. No other hosts may use these addresses, 2839however other routers may also be configured to translate them. 2840\begin{NB} 2841A great advantage of route NAT is that it may be used not 2842only in stub networks but in environments with arbitrarily complicated 2843structure. It does not firewall, it {\em forwards.} 2844\end{NB} 2845These addresses are selected by the \verb|ip route| command 2846(sec.\ref{IP-ROUTE-ADD}, p.\pageref{IP-ROUTE-ADD}). F.e.\ 2847\begin{verbatim} 2848 ip route add nat 192.203.80.144 via 193.233.7.83 2849\end{verbatim} 2850states that the single address 192.203.80.144 is a dummy NAT address. 2851For all the world it looks like a host address inside our network. 2852For neighbouring hosts and routers it looks like the local address 2853of the translating router. The router answers ARP for it, advertises 2854this address as routed via it, {\em et al\/}. When the router 2855receives a packet destined for 192.203.80.144, it replaces 2856this address with 193.233.7.83 which is the address of some real 2857host and forwards the packet. If you need to remap 2858blocks of addresses, you may use a command like: 2859\begin{verbatim} 2860 ip route add nat 192.203.80.192/26 via 193.233.7.64 2861\end{verbatim} 2862This command will map a block of 63 addresses 192.203.80.192-255 to 2863193.233.7.64-127. 2864 2865When an internal host (193.233.7.83 in the example above) 2866sends something to the outer world and these packets are forwarded 2867by our router, it should translate the source address 193.233.7.83 2868into 192.203.80.144. This task is solved by setting a special 2869policy rule (sec.\ref{IP-RULE-ADD}, p.\pageref{IP-RULE-ADD}): 2870\begin{verbatim} 2871 ip rule add prio 320 from 193.233.7.83 nat 192.203.80.144 2872\end{verbatim} 2873This rule says that the source address 193.233.7.83 2874should be translated into 192.203.80.144 before forwarding. 2875It is important that the address after the \verb|nat| keyword 2876is some NAT address, declared by {\tt ip route add nat}. 2877If it is just a random address the router will not map to it. 2878\begin{NB} 2879The exception is when the address is a local address of this 2880router (or 0.0.0.0) and masquerading is configured in the linux-2.2 2881kernel. In this case the router will masquerade the packets as this address. 2882If 0.0.0.0 is selected, the result is equivalent to one 2883obtained with firewalling rules. Otherwise, you have the way 2884to order Linux to masquerade to this fixed address. 2885NAT mechanism used in linux-2.4 is more flexible than 2886masquerading, so that this feature has lost meaning and disabled. 2887\end{NB} 2888 2889If the network has non-trivial internal structure, it is 2890useful and even necessary to add rules disabling translation 2891when a packet does not leave this network. Let us return to the 2892example from sec.\ref{IP-RULE-SHOW} (p.\pageref{IP-RULE-SHOW}). 2893\begin{verbatim} 2894300: from 193.233.7.83 to 193.233.7.0/24 lookup main 2895310: from 193.233.7.83 to 192.203.80.0/24 lookup main 2896320: from 193.233.7.83 lookup inr.ruhep map-to 192.203.80.144 2897\end{verbatim} 2898This block of rules causes normal forwarding when 2899packets from 193.233.7.83 do not leave networks 193.233.7/24 2900and 192.203.80/24. Also, if the \verb|inr.ruhep| table does not 2901contain a route to the destination (which means that the routing 2902domain owning addresses from 192.203.80/24 is dead), no translation 2903will occur. Otherwise, the packets are translated. 2904 2905\paragraph{How to only translate selected ports:} 2906If you only want to translate selected ports (f.e.\ http) 2907and leave the rest intact, you may use \verb|ipchains| 2908to \verb|fwmark| a class of packets. 2909Suppose you did and all the packets from 193.233.7.83 2910destined for port 80 are marked with marker 0x1234 in input fwchain. 2911In this case you may replace rule \#320 with: 2912\begin{verbatim} 2913320: from 193.233.7.83 fwmark 1234 lookup main map-to 192.203.80.144 2914\end{verbatim} 2915and translation will only be enabled for outgoing http requests. 2916 2917\section{Example: minimal host setup} 2918\label{EXAMPLE-SETUP} 2919 2920The following script gives an example of a fault safe 2921setup of IP (and IPv6, if it is compiled into the kernel) 2922in the common case of a node attached to a single broadcast 2923network. A more advanced script, which may be used both on multihomed 2924hosts and on routers, is described in the following 2925section. 2926 2927The utilities used in the script may be found in the 2928directory ftp://ftp.inr.ac.ru/ip-routing/: 2929\begin{enumerate} 2930\item \verb|ip| --- package \verb|iproute2|. 2931\item \verb|arping| --- package \verb|iputils|. 2932\item \verb|rdisc| --- package \verb|iputils|. 2933\end{enumerate} 2934\begin{NB} 2935It also refers to a DHCP client, \verb|dhcpcd|. I should refrain from 2936recommending a good DHCP client to use. All that I can 2937say is that ISC \verb|dhcp-2.0b1pl6| patched with the patch that 2938can be found in the \verb|dhcp.bootp.rarp| subdirectory of 2939the same ftp site {\em does\/} work, 2940at least on Ethernet and Token Ring. 2941\end{NB} 2942 2943\begin{verbatim} 2944#! /bin/bash 2945\end{verbatim} 2946\begin{flushleft} 2947\# {\bf Usage: \verb|ifone ADDRESS[/PREFIX-LENGTH] [DEVICE]|}\\ 2948\# {\bf Parameters:}\\ 2949\# \$1 --- Static IP address, optionally followed by prefix length.\\ 2950\# \$2 --- Device name. If it is missing, \verb|eth0| is asssumed.\\ 2951\# F.e. \verb|ifone 193.233.7.90| 2952\end{flushleft} 2953\begin{verbatim} 2954dev=$2 2955: ${dev:=eth0} 2956ipaddr= 2957\end{verbatim} 2958\# Parse IP address, splitting prefix length. 2959\begin{verbatim} 2960if [ "$1" != "" ]; then 2961 ipaddr=${1%/*} 2962 if [ "$1" != "$ipaddr" ]; then 2963 pfxlen=${1#*/} 2964 fi 2965 : ${pfxlen:=24} 2966fi 2967pfx="${ipaddr}/${pfxlen}" 2968\end{verbatim} 2969 2970\begin{flushleft} 2971\# {\bf Step 0} --- enable loopback.\\ 2972\#\\ 2973\# This step is necessary on any networked box before attempt\\ 2974\# to configure any other device.\\ 2975\end{flushleft} 2976\begin{verbatim} 2977ip link set up dev lo 2978ip addr add 127.0.0.1/8 dev lo brd + scope host 2979\end{verbatim} 2980\begin{flushleft} 2981\# IPv6 autoconfigure themself on loopback.\\ 2982\#\\ 2983\# If user gave loopback as device, we add the address as alias and exit. 2984\end{flushleft} 2985\begin{verbatim} 2986if [ "$dev" = "lo" ]; then 2987 if [ "$ipaddr" != "" -a "$ipaddr" != "127.0.0.1" ]; then 2988 ip address add $ipaddr dev $dev 2989 exit $? 2990 fi 2991 exit 0 2992fi 2993\end{verbatim} 2994 2995\noindent\# {\bf Step 1} --- enable device \verb|$dev| 2996 2997\begin{verbatim} 2998if ! ip link set up dev $dev ; then 2999 echo "Cannot enable interface $dev. Aborting." 1>&2 3000 exit 1 3001fi 3002\end{verbatim} 3003\begin{flushleft} 3004\# The interface is \verb|UP|. IPv6 started stateless autoconfiguration itself,\\ 3005\# and its configuration finishes here. However,\\ 3006\# IP still needs some static preconfigured address. 3007\end{flushleft} 3008\begin{verbatim} 3009if [ "$ipaddr" = "" ]; then 3010 echo "No address for $dev is configured, trying DHCP..." 1>&2 3011 dhcpcd 3012 exit $? 3013fi 3014\end{verbatim} 3015 3016\begin{flushleft} 3017\# {\bf Step 2} --- IP Duplicate Address Detection~\cite{RFC-DHCP}.\\ 3018\# Send two probes and wait for result for 3 seconds.\\ 3019\# If the interface opens slower f.e.\ due to long media detection,\\ 3020\# you want to increase the timeout.\\ 3021\end{flushleft} 3022\begin{verbatim} 3023if ! arping -q -c 2 -w 3 -D -I $dev $ipaddr ; then 3024 echo "Address $ipaddr is busy, trying DHCP..." 1>&2 3025 dhcpcd 3026 exit $? 3027fi 3028\end{verbatim} 3029\begin{flushleft} 3030\# OK, the address is unique, we may add it on the interface.\\ 3031\#\\ 3032\# {\bf Step 3} --- Configure the address on the interface. 3033\end{flushleft} 3034 3035\begin{verbatim} 3036if ! ip address add $pfx brd + dev $dev; then 3037 echo "Failed to add $pfx on $dev, trying DHCP..." 1>&2 3038 dhcpcd 3039 exit $? 3040fi 3041\end{verbatim} 3042 3043\noindent\# {\bf Step 4} --- Announce our presence on the link. 3044\begin{verbatim} 3045arping -A -c 1 -I $dev $ipaddr 3046noarp=$? 3047( sleep 2; 3048 arping -U -c 1 -I $dev $ipaddr ) >& /dev/null </dev/null & 3049\end{verbatim} 3050 3051\begin{flushleft} 3052\# {\bf Step 5} (optional) --- Add some control routes.\\ 3053\#\\ 3054\# 1. Prohibit link local multicast addresses.\\ 3055\# 2. Prohibit link local (alias, limited) broadcast.\\ 3056\# 3. Add default multicast route. 3057\end{flushleft} 3058\begin{verbatim} 3059ip route add unreachable 224.0.0.0/24 3060ip route add unreachable 255.255.255.255 3061if [ `ip link ls $dev | grep -c MULTICAST` -ge 1 ]; then 3062 ip route add 224.0.0.0/4 dev $dev scope global 3063fi 3064\end{verbatim} 3065 3066\begin{flushleft} 3067\# {\bf Step 6} --- Add fallback default route with huge metric.\\ 3068\# If a proxy ARP server is present on the interface, we will be\\ 3069\# able to talk to all the Internet without further configuration.\\ 3070\# It is not so cheap though and we still hope that this route\\ 3071\# will be overridden by more correct one by rdisc.\\ 3072\# Do not make this step if the device is not ARPable,\\ 3073\# because dead nexthop detection does not work on them. 3074\end{flushleft} 3075\begin{verbatim} 3076if [ "$noarp" = "0" ]; then 3077 ip ro add default dev $dev metric 30000 scope global 3078fi 3079\end{verbatim} 3080 3081\begin{flushleft} 3082\# {\bf Step 7} --- Restart router discovery and exit. 3083\end{flushleft} 3084\begin{verbatim} 3085killall -HUP rdisc || rdisc -fs 3086exit 0 3087\end{verbatim} 3088 3089 3090\section{Example: {\protect\tt ifcfg} --- interface address management} 3091\label{EXAMPLE-IFCFG} 3092 3093This is a simplistic script replacing one option of \verb|ifconfig|, 3094namely, IP address management. It not only adds 3095addresses, but also carries out Duplicate Address Detection~\cite{RFC-DHCP}, 3096sends unsolicited ARP to update the caches of other hosts sharing 3097the interface, adds some control routes and restarts Router Discovery 3098when it is necessary. 3099 3100I strongly recommend using it {\em instead\/} of \verb|ifconfig| both 3101on hosts and on routers. 3102 3103\begin{verbatim} 3104#! /bin/bash 3105\end{verbatim} 3106\begin{flushleft} 3107\# {\bf Usage: \verb?ifcfg DEVICE[:ALIAS] [add|del] ADDRESS[/LENGTH] [PEER]?}\\ 3108\# {\bf Parameters:}\\ 3109\# ---Device name. It may have alias suffix, separated by colon.\\ 3110\# ---Command: add, delete or stop.\\ 3111\# ---IP address, optionally followed by prefix length.\\ 3112\# ---Optional peer address for pointopoint interfaces.\\ 3113\# F.e. \verb|ifcfg eth0 193.233.7.90/24| 3114 3115\noindent\# This function determines, whether it is router or host.\\ 3116\# It returns 0, if the host is apparently not router. 3117\end{flushleft} 3118\begin{verbatim} 3119CheckForwarding () { 3120 local sbase fwd 3121 sbase=/proc/sys/net/ipv4/conf 3122 fwd=0 3123 if [ -d $sbase ]; then 3124 for dir in $sbase/*/forwarding; do 3125 fwd=$[$fwd + `cat $dir`] 3126 done 3127 else 3128 fwd=2 3129 fi 3130 return $fwd 3131} 3132\end{verbatim} 3133\begin{flushleft} 3134\# This function restarts Router Discovery.\\ 3135\end{flushleft} 3136\begin{verbatim} 3137RestartRDISC () { 3138 killall -HUP rdisc || rdisc -fs 3139} 3140\end{verbatim} 3141\begin{flushleft} 3142\# Calculate ABC "natural" mask length\\ 3143\# Arg: \$1 = dotquad address 3144\end{flushleft} 3145\begin{verbatim} 3146ABCMaskLen () { 3147 local class; 3148 class=${1%%.*} 3149 if [ $class -eq 0 -o $class -ge 224 ]; then return 0 3150 elif [ $class -ge 192 ]; then return 24 3151 elif [ $class -ge 128 ]; then return 16 3152 else return 8 ; fi 3153} 3154\end{verbatim} 3155 3156 3157\begin{flushleft} 3158\# {\bf MAIN()}\\ 3159\#\\ 3160\# Strip alias suffix separated by colon. 3161\end{flushleft} 3162\begin{verbatim} 3163label="label $1" 3164ldev=$1 3165dev=${1%:*} 3166if [ "$dev" = "" -o "$1" = "help" ]; then 3167 echo "Usage: ifcfg DEV [[add|del [ADDR[/LEN]] [PEER] | stop]" 1>&2 3168 echo " add - add new address" 1>&2 3169 echo " del - delete address" 1>&2 3170 echo " stop - completely disable IP" 1>&2 3171 exit 1 3172fi 3173shift 3174 3175CheckForwarding 3176fwd=$? 3177\end{verbatim} 3178\begin{flushleft} 3179\# Parse command. If it is ``stop'', flush and exit. 3180\end{flushleft} 3181\begin{verbatim} 3182deleting=0 3183case "$1" in 3184add) shift ;; 3185stop) 3186 if [ "$ldev" != "$dev" ]; then 3187 echo "Cannot stop alias $ldev" 1>&2 3188 exit 1; 3189 fi 3190 ip -4 addr flush dev $dev $label || exit 1 3191 if [ $fwd -eq 0 ]; then RestartRDISC; fi 3192 exit 0 ;; 3193del*) 3194 deleting=1; shift ;; 3195*) 3196esac 3197\end{verbatim} 3198\begin{flushleft} 3199\# Parse prefix, split prefix length, separated by slash. 3200\end{flushleft} 3201\begin{verbatim} 3202ipaddr= 3203pfxlen= 3204if [ "$1" != "" ]; then 3205 ipaddr=${1%/*} 3206 if [ "$1" != "$ipaddr" ]; then 3207 pfxlen=${1#*/} 3208 fi 3209 if [ "$ipaddr" = "" ]; then 3210 echo "$1 is bad IP address." 1>&2 3211 exit 1 3212 fi 3213fi 3214shift 3215\end{verbatim} 3216\begin{flushleft} 3217\# If peer address is present, prefix length is 32.\\ 3218\# Otherwise, if prefix length was not given, guess it. 3219\end{flushleft} 3220\begin{verbatim} 3221peer=$1 3222if [ "$peer" != "" ]; then 3223 if [ "$pfxlen" != "" -a "$pfxlen" != "32" ]; then 3224 echo "Peer address with non-trivial netmask." 1>&2 3225 exit 1 3226 fi 3227 pfx="$ipaddr peer $peer" 3228else 3229 if [ "$pfxlen" = "" ]; then 3230 ABCMaskLen $ipaddr 3231 pfxlen=$? 3232 fi 3233 pfx="$ipaddr/$pfxlen" 3234fi 3235if [ "$ldev" = "$dev" -a "$ipaddr" != "" ]; then 3236 label= 3237fi 3238\end{verbatim} 3239\begin{flushleft} 3240\# If deletion was requested, delete the address and restart RDISC 3241\end{flushleft} 3242\begin{verbatim} 3243if [ $deleting -ne 0 ]; then 3244 ip addr del $pfx dev $dev $label || exit 1 3245 if [ $fwd -eq 0 ]; then RestartRDISC; fi 3246 exit 0 3247fi 3248\end{verbatim} 3249\begin{flushleft} 3250\# Start interface initialization.\\ 3251\#\\ 3252\# {\bf Step 0} --- enable device \verb|$dev| 3253\end{flushleft} 3254\begin{verbatim} 3255if ! ip link set up dev $dev ; then 3256 echo "Error: cannot enable interface $dev." 1>&2 3257 exit 1 3258fi 3259if [ "$ipaddr" = "" ]; then exit 0; fi 3260\end{verbatim} 3261\begin{flushleft} 3262\# {\bf Step 1} --- IP Duplicate Address Detection~\cite{RFC-DHCP}.\\ 3263\# Send two probes and wait for result for 3 seconds.\\ 3264\# If the interface opens slower f.e.\ due to long media detection,\\ 3265\# you want to increase the timeout.\\ 3266\end{flushleft} 3267\begin{verbatim} 3268if ! arping -q -c 2 -w 3 -D -I $dev $ipaddr ; then 3269 echo "Error: some host already uses address $ipaddr on $dev." 1>&2 3270 exit 1 3271fi 3272\end{verbatim} 3273\begin{flushleft} 3274\# OK, the address is unique. We may add it to the interface.\\ 3275\#\\ 3276\# {\bf Step 2} --- Configure the address on the interface. 3277\end{flushleft} 3278\begin{verbatim} 3279if ! ip address add $pfx brd + dev $dev $label; then 3280 echo "Error: failed to add $pfx on $dev." 1>&2 3281 exit 1 3282fi 3283\end{verbatim} 3284\noindent\# {\bf Step 3} --- Announce our presence on the link 3285\begin{verbatim} 3286arping -q -A -c 1 -I $dev $ipaddr 3287noarp=$? 3288( sleep 2 ; 3289 arping -q -U -c 1 -I $dev $ipaddr ) >& /dev/null </dev/null & 3290\end{verbatim} 3291\begin{flushleft} 3292\# {\bf Step 4} (optional) --- Add some control routes.\\ 3293\#\\ 3294\# 1. Prohibit link local multicast addresses.\\ 3295\# 2. Prohibit link local (alias, limited) broadcast.\\ 3296\# 3. Add default multicast route. 3297\end{flushleft} 3298\begin{verbatim} 3299ip route add unreachable 224.0.0.0/24 >& /dev/null 3300ip route add unreachable 255.255.255.255 >& /dev/null 3301if [ `ip link ls $dev | grep -c MULTICAST` -ge 1 ]; then 3302 ip route add 224.0.0.0/4 dev $dev scope global >& /dev/null 3303fi 3304\end{verbatim} 3305\begin{flushleft} 3306\# {\bf Step 5} --- Add fallback default route with huge metric.\\ 3307\# If a proxy ARP server is present on the interface, we will be\\ 3308\# able to talk to all the Internet without further configuration.\\ 3309\# Do not make this step on router or if the device is not ARPable.\\ 3310\# because dead nexthop detection does not work on them. 3311\end{flushleft} 3312\begin{verbatim} 3313if [ $fwd -eq 0 ]; then 3314 if [ $noarp -eq 0 ]; then 3315 ip ro append default dev $dev metric 30000 scope global 3316 elif [ "$peer" != "" ]; then 3317 if ping -q -c 2 -w 4 $peer ; then 3318 ip ro append default via $peer dev $dev metric 30001 3319 fi 3320 fi 3321 RestartRDISC 3322fi 3323 3324exit 0 3325\end{verbatim} 3326\begin{flushleft} 3327\# End of {\bf MAIN()} 3328\end{flushleft} 3329 3330 3331\end{document} 3332