man7/packet.7

   1 .\" This man page is Copyright (C) 1999 Andi Kleen <ak@muc.de>.
   2 .\"
   3 .\" %%%LICENSE_START(VERBATIM_ONE_PARA)
   4 .\" Permission is granted to distribute possibly modified copies
   5 .\" of this page provided the header is included verbatim,
   6 .\" and in case of nontrivial modification author and date
   7 .\" of the modification is added to the header.
   8 .\" %%%LICENSE_END
   9 .\"
  10 .\" $Id: packet.7,v 1.13 2000/08/14 08:03:45 ak Exp $
  11 .\"
  12 .TH PACKET  7 2017-09-15 "Linux" "Linux Programmer's Manual"
  13 .SH NAME
  14 packet \- packet interface on device level
  15 .SH SYNOPSIS
  16 .nf
  17 .B #include <sys/socket.h>
  18 .B #include <linux/if_packet.h>
  19 .B #include <net/ethernet.h>     /* the L2 protocols */
  20 .PP
  21 .BI "packet_socket = socket(AF_PACKET, int " socket_type ", int "protocol );
  22 .fi
  23 .SH DESCRIPTION
  24 Packet sockets are used to receive or send raw packets at the device driver
  25 (OSI Layer 2) level.
  26 They allow the user to implement protocol modules in user space
  27 on top of the physical layer.
  28 .PP
  29 The
  30 .I socket_type
  31 is either
  32 .B SOCK_RAW
  33 for raw packets including the link-level header or
  34 .B SOCK_DGRAM
  35 for cooked packets with the link-level header removed.
  36 The link-level header information is available in a common format in a
  37 .IR sockaddr_ll
  38 structure.
  39 .I protocol
  40 is the IEEE 802.3 protocol number in network byte order.
  41 See the
  42 .I <linux/if_ether.h>
  43 include file for a list of allowed protocols.
  44 When protocol
  45 is set to
  46 .BR htons(ETH_P_ALL) ,
  47 then all protocols are received.
  48 All incoming packets of that protocol type will be passed to the packet
  49 socket before they are passed to the protocols implemented in the kernel.
  50 .PP
  51 In order to create a packet socket, a process must have the
  52 .B CAP_NET_RAW
  53 capability in the user namespace that governs its network namespace.
  54 .PP
  55 .B SOCK_RAW
  56 packets are passed to and from the device driver without any changes in
  57 the packet data.
  58 When receiving a packet, the address is still parsed and
  59 passed in a standard
  60 .I sockaddr_ll
  61 address structure.
  62 When transmitting a packet, the user-supplied buffer
  63 should contain the physical-layer header.
  64 That packet is then
  65 queued unmodified to the network driver of the interface defined by the
  66 destination address.
  67 Some device drivers always add other headers.
  68 .B SOCK_RAW
  69 is similar to but not compatible with the obsolete
  70 .B AF_INET/SOCK_PACKET
  71 of Linux 2.0.
  72 .PP
  73 .B SOCK_DGRAM
  74 operates on a slightly higher level.
  75 The physical header is removed before the packet is passed to the user.
  76 Packets sent through a
  77 .B SOCK_DGRAM
  78 packet socket get a suitable physical-layer header based on the
  79 information in the
  80 .I sockaddr_ll
  81 destination address before they are queued.
  82 .PP
  83 By default, all packets of the specified protocol type
  84 are passed to a packet socket.
  85 To get packets only from a specific interface use
  86 .BR bind (2)
  87 specifying an address in a
  88 .I struct sockaddr_ll
  89 to bind the packet socket to an interface.
  90 Fields used for binding are
  91 .IR sll_family
  92 (should be
  93 .BR AF_PACKET ),
  94 .IR sll_protocol ,
  95 and
  96 .IR sll_ifindex .
  97 .PP
  98 The
  99 .BR connect (2)
 100 operation is not supported on packet sockets.
 101 .PP
 102 When the
 103 .B MSG_TRUNC
 104 flag is passed to
 105 .BR recvmsg (2),
 106 .BR recv (2),
 107 or
 108 .BR recvfrom (2),
 109 the real length of the packet on the wire is always returned,
 110 even when it is longer than the buffer.
 111 .SS Address types
 112 The
 113 .I sockaddr_ll
 114 structure is a device-independent physical-layer address.
 115 .PP
 116 .in +4n
 117 .EX
 118 struct sockaddr_ll {
 119     unsigned short sll_family;   /* Always AF_PACKET */
 120     unsigned short sll_protocol; /* Physical-layer protocol */
 121     int            sll_ifindex;  /* Interface number */
 122     unsigned short sll_hatype;   /* ARP hardware type */
 123     unsigned char  sll_pkttype;  /* Packet type */
 124     unsigned char  sll_halen;    /* Length of address */
 125     unsigned char  sll_addr[8];  /* Physical-layer address */
 126 };
 127 .EE
 128 .in
 129 .PP
 130 The fields of this structure are as follows:
 131 .IP * 3
 132 .I sll_protocol
 133 is the standard ethernet protocol type in network byte order as defined
 134 in the
 135 .I <linux/if_ether.h>
 136 include file.
 137 It defaults to the socket's protocol.
 138 .IP *
 139 .I sll_ifindex
 140 is the interface index of the interface
 141 (see
 142 .BR netdevice (7));
 143 0 matches any interface (only permitted for binding).
 144 .I sll_hatype
 145 is an ARP type as defined in the
 146 .I <linux/if_arp.h>
 147 include file.
 148 .IP *
 149 .I sll_pkttype
 150 contains the packet type.
 151 Valid types are
 152 .B PACKET_HOST
 153 for a packet addressed to the local host,
 154 .B PACKET_BROADCAST
 155 for a physical-layer broadcast packet,
 156 .B PACKET_MULTICAST
 157 for a packet sent to a physical-layer multicast address,
 158 .B PACKET_OTHERHOST
 159 for a packet to some other host that has been caught by a device driver
 160 in promiscuous mode, and
 161 .B PACKET_OUTGOING
 162 for a packet originating from the local host that is looped back to a packet
 163 socket.
 164 These types make sense only for receiving.
 165 .IP *
 166 .I sll_addr
 167 and
 168 .I sll_halen
 169 contain the physical-layer (e.g., IEEE 802.3) address and its length.
 170 The exact interpretation depends on the device.
 171 .PP
 172 When you send packets, it is enough to specify
 173 .IR sll_family ,
 174 .IR sll_addr ,
 175 .IR sll_halen ,
 176 .IR sll_ifindex ,
 177 and
 178 .IR sll_protocol .
 179 The other fields should be 0.
 180 .I sll_hatype
 181 and
 182 .I sll_pkttype
 183 are set on received packets for your information.
 184 .SS Socket options
 185 Packet socket options are configured by calling
 186 .BR setsockopt (2)
 187 with level
 188 .BR SOL_PACKET .
 189 .TP
 190 .BR PACKET_ADD_MEMBERSHIP
 191 .PD 0
 192 .TP
 193 .BR PACKET_DROP_MEMBERSHIP
 194 .PD
 195 Packet sockets can be used to configure physical-layer multicasting
 196 and promiscuous mode.
 197 .B PACKET_ADD_MEMBERSHIP
 198 adds a binding and
 199 .B PACKET_DROP_MEMBERSHIP
 200 drops it.
 201 They both expect a
 202 .I packet_mreq
 203 structure as argument:
 204 .IP
 205 .in +4n
 206 .EX
 207 struct packet_mreq {
 208     int            mr_ifindex;    /* interface index */
 209     unsigned short mr_type;       /* action */
 210     unsigned short mr_alen;       /* address length */
 211     unsigned char  mr_address[8]; /* physical-layer address */
 212 };
 213 .EE
 214 .in
 215 .IP
 216 .I mr_ifindex
 217 contains the interface index for the interface whose status
 218 should be changed.
 219 The
 220 .I mr_type
 221 field specifies which action to perform.
 222 .B PACKET_MR_PROMISC
 223 enables receiving all packets on a shared medium (often known as
 224 "promiscuous mode"),
 225 .B PACKET_MR_MULTICAST
 226 binds the socket to the physical-layer multicast group specified in
 227 .I mr_address
 228 and
 229 .IR mr_alen ,
 230 and
 231 .B PACKET_MR_ALLMULTI
 232 sets the socket up to receive all multicast packets arriving at
 233 the interface.
 234 .IP
 235 In addition, the traditional ioctls
 236 .BR SIOCSIFFLAGS ,
 237 .BR SIOCADDMULTI ,
 238 .B SIOCDELMULTI
 239 can be used for the same purpose.
 240 .TP
 241 .BR PACKET_AUXDATA " (since Linux 2.6.21)"
 242 .\" commit 8dc4194474159660d7f37c495e3fc3f10d0db8cc
 243 If this binary option is enabled, the packet socket passes a metadata
 244 structure along with each packet in the
 245 .BR recvmsg (2)
 246 control field.
 247 The structure can be read with
 248 .BR cmsg (3).
 249 It is defined as
 250 .IP
 251 .in +4n
 252 .EX
 253 struct tpacket_auxdata {
 254     __u32 tp_status;
 255     __u32 tp_len;      /* packet length */
 256     __u32 tp_snaplen;  /* captured length */
 257     __u16 tp_mac;
 258     __u16 tp_net;
 259     __u16 tp_vlan_tci;
 260     __u16 tp_vlan_tpid; /* Since Linux 3.14; earlier, these
 261                            were unused padding bytes */
 262 .\" commit a0cdfcf39362410d5ea983f4daf67b38de129408 added tp_vlan_tpid
 263 };
 264 .EE
 265 .in
 266 .TP
 267 .BR PACKET_FANOUT " (since Linux 3.1)"
 268 .\" commit dc99f600698dcac69b8f56dda9a8a00d645c5ffc
 269 To scale processing across threads, packet sockets can form a fanout
 270 group.
 271 In this mode, each matching packet is enqueued onto only one
 272 socket in the group.
 273 A socket joins a fanout group by calling
 274 .BR setsockopt (2)
 275 with level
 276 .B SOL_PACKET
 277 and option
 278 .BR PACKET_FANOUT .
 279 Each network namespace can have up to 65536 independent groups.
 280 A socket selects a group by encoding the ID in the first 16 bits of
 281 the integer option value.
 282 The first packet socket to join a group implicitly creates it.
 283 To successfully join an existing group, subsequent packet sockets
 284 must have the same protocol, device settings, fanout mode and
 285 flags (see below).
 286 Packet sockets can leave a fanout group only by closing the socket.
 287 The group is deleted when the last socket is closed.
 288 .IP
 289 Fanout supports multiple algorithms to spread traffic between sockets,
 290 as follows:
 291 .RS
 292 .IP * 3
 293 The default mode,
 294 .BR PACKET_FANOUT_HASH ,
 295 sends packets from the same flow to the same socket to maintain
 296 per-flow ordering.
 297 For each packet, it chooses a socket by taking the packet flow hash
 298 modulo the number of sockets in the group, where a flow hash is a hash
 299 over network-layer address and optional transport-layer port fields.
 300 .IP *
 301 The load-balance mode
 302 .BR PACKET_FANOUT_LB
 303 implements a round-robin algorithm.
 304 .IP *
 305 .BR PACKET_FANOUT_CPU
 306 selects the socket based on the CPU that the packet arrived on.
 307 .IP *
 308 .BR PACKET_FANOUT_ROLLOVER
 309 processes all data on a single socket, moving to the next when one
 310 becomes backlogged.
 311 .IP *
 312 .BR PACKET_FANOUT_RND
 313 selects the socket using a pseudo-random number generator.
 314 .IP *
 315 .BR PACKET_FANOUT_QM
 316 .\" commit 2d36097d26b5991d71a2cf4a20c1a158f0f1bfcd
 317 (available since Linux 3.14)
 318 selects the socket using the recorded queue_mapping of the received skb.
 319 .RE
 320 .IP
 321 Fanout modes can take additional options.
 322 IP fragmentation causes packets from the same flow to have different
 323 flow hashes.
 324 The flag
 325 .BR PACKET_FANOUT_FLAG_DEFRAG ,
 326 if set, causes packets to be defragmented before fanout is applied, to
 327 preserve order even in this case.
 328 Fanout mode and options are communicated in the second 16 bits of the
 329 integer option value.
 330 The flag
 331 .BR PACKET_FANOUT_FLAG_ROLLOVER
 332 enables the roll over mechanism as a backup strategy: if the
 333 original fanout algorithm selects a backlogged socket, the packet
 334 rolls over to the next available one.
 335 .TP
 336 .BR PACKET_LOSS " (with " PACKET_TX_RING )
 337 When a malformed packet is encountered on a transmit ring,
 338 the default is to reset its
 339 .I tp_status
 340 to
 341 .BR TP_STATUS_WRONG_FORMAT
 342 and abort the transmission immediately.
 343 The malformed packet blocks itself and subsequently enqueued packets from
 344 being sent.
 345 The format error must be fixed, the associated
 346 .I tp_status
 347 reset to
 348 .BR TP_STATUS_SEND_REQUEST ,
 349 and the transmission process restarted via
 350 .BR send (2).
 351 However, if
 352 .BR PACKET_LOSS
 353 is set, any malformed packet will be skipped, its
 354 .I tp_status
 355 reset to
 356 .BR TP_STATUS_AVAILABLE ,
 357 and the transmission process continued.
 358 .TP
 359 .BR PACKET_RESERVE " (with " PACKET_RX_RING )
 360 By default, a packet receive ring writes packets immediately following the
 361 metadata structure and alignment padding.
 362 This integer option reserves additional headroom.
 363 .TP
 364 .BR PACKET_RX_RING
 365 Create a memory-mapped ring buffer for asynchronous packet reception.
 366 The packet socket reserves a contiguous region of application address
 367 space, lays it out into an array of packet slots and copies packets
 368 (up to
 369 .IR tp_snaplen )
 370 into subsequent slots.
 371 Each packet is preceded by a metadata structure similar to
 372 .IR tpacket_auxdata .
 373 The protocol fields encode the offset to the data
 374 from the start of the metadata header.
 375 .I tp_net
 376 stores the offset to the network layer.
 377 If the packet socket is of type
 378 .BR SOCK_DGRAM ,
 379 then
 380 .I tp_mac
 381 is the same.
 382 If it is of type
 383 .BR SOCK_RAW ,
 384 then that field stores the offset to the link-layer frame.
 385 Packet socket and application communicate the head and tail of the ring
 386 through the
 387 .I tp_status
 388 field.
 389 The packet socket owns all slots with
 390 .I tp_status
 391 equal to
 392 .BR TP_STATUS_KERNEL .
 393 After filling a slot, it changes the status of the slot to transfer
 394 ownership to the application.
 395 During normal operation, the new
 396 .I tp_status
 397 value has at least the
 398 .BR TP_STATUS_USER
 399 bit set to signal that a received packet has been stored.
 400 When the application has finished processing a packet, it transfers
 401 ownership of the slot back to the socket by setting
 402 .I tp_status
 403 equal to
 404 .BR TP_STATUS_KERNEL .
 405 .IP
 406 Packet sockets implement multiple variants of the packet ring.
 407 The implementation details are described in
 408 .IR Documentation/networking/packet_mmap.txt
 409 in the Linux kernel source tree.
 410 .TP
 411 .BR PACKET_STATISTICS
 412 Retrieve packet socket statistics in the form of a structure
 413 .IP
 414 .in +4n
 415 .EX
 416 struct tpacket_stats {
 417     unsigned int tp_packets;  /* Total packet count */
 418     unsigned int tp_drops;    /* Dropped packet count */
 419 };
 420 .EE
 421 .in
 422 .IP
 423 Receiving statistics resets the internal counters.
 424 The statistics structure differs when using a ring of variant
 425 .BR TPACKET_V3 .
 426 .TP
 427 .BR PACKET_TIMESTAMP " (with " PACKET_RX_RING "; since Linux 2.6.36)"
 428 .\" commit 614f60fa9d73a9e8fdff3df83381907fea7c5649
 429 The packet receive ring always stores a timestamp in the metadata header.
 430 By default, this is a software generated timestamp generated when the
 431 packet is copied into the ring.
 432 This integer option selects the type of timestamp.
 433 Besides the default, it support the two hardware formats described in
 434 .IR Documentation/networking/timestamping.txt
 435 in the Linux kernel source tree.
 436 .TP
 437 .BR PACKET_TX_RING " (since Linux 2.6.31)"
 438 .\" commit 69e3c75f4d541a6eb151b3ef91f34033cb3ad6e1
 439 Create a memory-mapped ring buffer for packet transmission.
 440 This option is similar to
 441 .BR PACKET_RX_RING
 442 and takes the same arguments.
 443 The application writes packets into slots with
 444 .I tp_status
 445 equal to
 446 .BR TP_STATUS_AVAILABLE
 447 and schedules them for transmission by changing
 448 .I tp_status
 449 to
 450 .BR TP_STATUS_SEND_REQUEST .
 451 When packets are ready to be transmitted, the application calls
 452 .BR send (2)
 453 or a variant thereof.
 454 The
 455 .I buf
 456 and
 457 .I len
 458 fields of this call are ignored.
 459 If an address is passed using
 460 .BR sendto (2)
 461 or
 462 .BR sendmsg (2),
 463 then that overrides the socket default.
 464 On successful transmission, the socket resets
 465 .I tp_status
 466 to
 467 .BR TP_STATUS_AVAILABLE .
 468 It immediately aborts the transmission on error unless
 469 .BR PACKET_LOSS
 470 is set.
 471 .TP
 472 .BR PACKET_VERSION " (with " PACKET_RX_RING "; since Linux 2.6.27)"
 473 .\" commit bbd6ef87c544d88c30e4b762b1b61ef267a7d279
 474 By default,
 475 .BR PACKET_RX_RING
 476 creates a packet receive ring of variant
 477 .BR TPACKET_V1 .
 478 To create another variant, configure the desired variant by setting this
 479 integer option before creating the ring.
 480 .TP
 481 .BR PACKET_QDISC_BYPASS " (since Linux 3.14)"
 482 .\" commit d346a3fae3ff1d99f5d0c819bf86edf9094a26a1
 483 By default, packets sent through packet sockets pass through the kernel's
 484 qdisc (traffic control) layer, which is fine for the vast majority of use
 485 cases.
 486 For traffic generator appliances using packet sockets
 487 that intend to brute-force flood the network\(emfor example,
 488 to test devices under load in a similar
 489 fashion to pktgen\(emthis layer can be bypassed by setting
 490 this integer option to 1.
 491 A side effect is that packet buffering in the qdisc layer is avoided,
 492 which will lead to increased drops when network
 493 device transmit queues are busy;
 494 therefore, use at your own risk.
 495 .SS Ioctls
 496 .B SIOCGSTAMP
 497 can be used to receive the timestamp of the last received packet.
 498 Argument is a
 499 .I struct timeval
 500 variable.
 501 .\" FIXME Document SIOCGSTAMPNS
 502 .PP
 503 In addition, all standard ioctls defined in
 504 .BR netdevice (7)
 505 and
 506 .BR socket (7)
 507 are valid on packet sockets.
 508 .SS Error handling
 509 Packet sockets do no error handling other than errors occurred
 510 while passing the packet to the device driver.
 511 They don't have the concept of a pending error.
 512 .SH ERRORS
 513 .TP
 514 .B EADDRNOTAVAIL
 515 Unknown multicast group address passed.
 516 .TP
 517 .B EFAULT
 518 User passed invalid memory address.
 519 .TP
 520 .B EINVAL
 521 Invalid argument.
 522 .TP
 523 .B EMSGSIZE
 524 Packet is bigger than interface MTU.
 525 .TP
 526 .B ENETDOWN
 527 Interface is not up.
 528 .TP
 529 .B ENOBUFS
 530 Not enough memory to allocate the packet.
 531 .TP
 532 .B ENODEV
 533 Unknown device name or interface index specified in interface address.
 534 .TP
 535 .B ENOENT
 536 No packet received.
 537 .TP
 538 .B ENOTCONN
 539 No interface address passed.
 540 .TP
 541 .B ENXIO
 542 Interface address contained an invalid interface index.
 543 .TP
 544 .B EPERM
 545 User has insufficient privileges to carry out this operation.
 546 .PP
 547 In addition, other errors may be generated by the low-level driver.
 548 .SH VERSIONS
 549 .B AF_PACKET
 550 is a new feature in Linux 2.2.
 551 Earlier Linux versions supported only
 552 .BR SOCK_PACKET .
 553 .PP
 554 .SH NOTES
 555 For portable programs it is suggested to use
 556 .B AF_PACKET
 557 via
 558 .BR pcap (3);
 559 although this covers only a subset of the
 560 .B AF_PACKET
 561 features.
 562 .PP
 563 The
 564 .B SOCK_DGRAM
 565 packet sockets make no attempt to create or parse the IEEE 802.2 LLC
 566 header for a IEEE 802.3 frame.
 567 When
 568 .B ETH_P_802_3
 569 is specified as protocol for sending the kernel creates the
 570 802.3 frame and fills out the length field; the user has to supply the LLC
 571 header to get a fully conforming packet.
 572 Incoming 802.3 packets are not multiplexed on the DSAP/SSAP protocol
 573 fields; instead they are supplied to the user as protocol
 574 .B ETH_P_802_2
 575 with the LLC header prefixed.
 576 It is thus not possible to bind to
 577 .BR ETH_P_802_3 ;
 578 bind to
 579 .B ETH_P_802_2
 580 instead and do the protocol multiplex yourself.
 581 The default for sending is the standard Ethernet DIX
 582 encapsulation with the protocol filled in.
 583 .PP
 584 Packet sockets are not subject to the input or output firewall chains.
 585 .SS Compatibility
 586 In Linux 2.0, the only way to get a packet socket was with the call:
 587 .PP
 588     socket(AF_INET, SOCK_PACKET, protocol)
 589 .PP
 590 This is still supported, but deprecated and strongly discouraged.
 591 The main difference between the two methods is that
 592 .B SOCK_PACKET
 593 uses the old
 594 .I struct sockaddr_pkt
 595 to specify an interface, which doesn't provide physical-layer
 596 independence.
 597 .PP
 598 .in +4n
 599 .EX
 600 struct sockaddr_pkt {
 601     unsigned short spkt_family;
 602     unsigned char  spkt_device[14];
 603     unsigned short spkt_protocol;
 604 };
 605 .EE
 606 .in
 607 .PP
 608 .I spkt_family
 609 contains
 610 the device type,
 611 .I spkt_protocol
 612 is the IEEE 802.3 protocol type as defined in
 613 .I <sys/if_ether.h>
 614 and
 615 .I spkt_device
 616 is the device name as a null-terminated string, for example, eth0.
 617 .PP
 618 This structure is obsolete and should not be used in new code.
 619 .SH BUGS
 620 The IEEE 802.2/803.3 LLC handling could be considered as a bug.
 621 .PP
 622 Socket filters are not documented.
 623 .PP
 624 The
 625 .B MSG_TRUNC
 626 .BR recvmsg (2)
 627 extension is an ugly hack and should be replaced by a control message.
 628 There is currently no way to get the original destination address of
 629 packets via
 630 .BR SOCK_DGRAM .
 631 .\" .SH CREDITS
 632 .\" This man page was written by Andi Kleen with help from Matthew Wilcox.
 633 .\" AF_PACKET in Linux 2.2 was implemented
 634 .\" by Alexey Kuznetsov, based on code by Alan Cox and others.
 635 .SH SEE ALSO
 636 .BR socket (2),
 637 .BR pcap (3),
 638 .BR capabilities (7),
 639 .BR ip (7),
 640 .BR raw (7),
 641 .BR socket (7)
 642 .PP
 643 RFC\ 894 for the standard IP Ethernet encapsulation.
 644 RFC\ 1700 for the IEEE 802.3 IP encapsulation.
 645 .PP
 646 The
 647 .I <linux/if_ether.h>
 648 include file for physical-layer protocols.
 649 .PP
 650 The Linux kernel source tree.
 651 .IR /Documentation/networking/filter.txt
 652 describes how to apply Berkeley Packet Filters to packet sockets.
 653 .IR /tools/testing/selftests/net/psock_tpacket.c
 654 contains example source code for all available versions of
 655 .BR PACKET_RX_RING
 656 and
 657 .BR PACKET_TX_RING .