man7/packet.7

   1 .\" This man page is Copyright (C) 1999 Andi Kleen <ak@muc.de>.
   2 .\" Permission is granted to distribute possibly modified copies
   3 .\" of this page provided the header is included verbatim,
   4 .\" and in case of nontrivial modification author and date
   5 .\" of the modification is added to the header.
   6 .\" $Id: packet.7,v 1.13 2000/08/14 08:03:45 ak Exp $
   7 .TH PACKET  7 1999-04-29 "Linux Man Page" "Linux Programmer's Manual"
   8 .SH NAME
   9 packet, PF_PACKET \- packet interface on device level.
  10 .SH SYNOPSIS
  11 .nf
  12 .B #include <sys/socket.h>
  13 .br
  14 .B #include <netpacket/packet.h>
  15 .br
  16 .B #include <net/ethernet.h>    /* the L2 protocols */
  17 .sp
  18 .BI "packet_socket = socket(PF_PACKET, int " socket_type ", int "protocol );
  19 .fi
  20 .SH DESCRIPTION
  21 Packet sockets are used to receive or send raw packets at the device driver
  22 (OSI Layer 2)
  23 level. They allow the user to implement protocol modules in user space
  24 on top of the physical layer.
  25
  26 The
  27 .I socket_type
  28 is either
  29 .B SOCK_RAW
  30 for raw packets including the link level header or
  31 .B SOCK_DGRAM
  32 for cooked packets with the link level header removed. The link level
  33 header information is available in a common format in a
  34 .BR sockaddr_ll .
  35 .I protocol
  36 is the IEEE 802.3 protocol number in network order. See the
  37 .B <linux/if_ether.h>
  38 include file for a list of allowed protocols. When protocol
  39 is set to
  40 .B htons(ETH_P_ALL)
  41 then all protocols are received.
  42 All incoming packets of that protocol type will be passed to the packet
  43 socket before they are passed to the protocols implemented in the kernel.
  44
  45 Only processes with effective uid 0 or the
  46 .B CAP_NET_RAW
  47 capability may open packet sockets.
  48
  49 .B SOCK_RAW
  50 packets are passed to and from the device driver without any changes in
  51 the packet data.  When receiving a packet, the address is still parsed and
  52 passed in a standard
  53 .B sockaddr_ll
  54 address structure.  When transmitting a packet, the user supplied buffer
  55 should contain the physical layer header.  That packet is then
  56 queued unmodified to the network driver of the interface defined by the
  57 destination address. Some device drivers always add other headers.
  58 .B SOCK_RAW
  59 is similar to but not compatible with the obsolete
  60 .B SOCK_PACKET
  61 of Linux 2.0.
  62
  63 .B SOCK_DGRAM
  64 operates on a slightly higher level. The physical header is removed before
  65 the packet is passed to the user.  Packets sent through a
  66 .B SOCK_DGRAM
  67 packet socket get a suitable physical layer header based on the information
  68 in the
  69 .B sockaddr_ll
  70 destination address before they are queued.
  71
  72 By default all packets of the specified protocol type
  73 are passed to a packet socket. To only get packets from a specific interface
  74 use
  75 .BR bind (2)
  76 specifying an address in a
  77 .B struct sockaddr_ll
  78 to bind the packet socket to an interface. Only the
  79 .B sll_protocol
  80 and the
  81 .B sll_ifindex
  82 address fields are used for purposes of binding.
  83
  84 The
  85 .BR connect (2)
  86 operation is not supported on packet sockets.
  87
  88 When the
  89 .B MSG_TRUNC
  90 flag is passed to
  91 .BR recvmsg (2),
  92 .BR recv (2),
  93 .BR recvfrom (2)
  94 the real length of the packet on the wire is always returned, even when it
  95 is longer than the buffer.
  96
  97 .SH "ADDRESS TYPES"
  98 The sockaddr_ll is a device independent physical layer address.
  99
 100 .RS
 101 .nf
 102 .ta 4n 20n 35n
 103 struct sockaddr_ll {
 104         unsigned short  sll_family;     /* Always AF_PACKET */
 105         unsigned short  sll_protocol;   /* Physical layer protocol */
 106         int     sll_ifindex;    /* Interface number */
 107         unsigned short  sll_hatype;     /* Header type */
 108         unsigned char   sll_pkttype;    /* Packet type */
 109         unsigned char   sll_halen;      /* Length of address */
 110         unsigned char   sll_addr[8];    /* Physical layer address */
 111 };
 112 .ta
 113 .fi
 114 .RE
 115
 116 .B sll_protocol
 117 is the standard ethernet protocol type in network order as defined
 118 in the
 119 .B linux/if_ether.h
 120 include file.  It defaults to the socket's protocol.
 121 .B sll_ifindex
 122 is the interface index of the interface
 123 (see
 124 .BR netdevice (7));
 125 0 matches any interface (only legal for binding).
 126 .B sll_hatype
 127 is a ARP type as defined in the
 128 .B linux/if_arp.h
 129 include file.
 130 .B sll_pkttype
 131 contains the packet type. Valid types are
 132 .B PACKET_HOST
 133 for a packet addressed to the local host,
 134 .B PACKET_BROADCAST
 135 for a physical layer broadcast packet,
 136 .B PACKET_MULTICAST
 137 for a packet sent to a physical layer multicast address,
 138 .B PACKET_OTHERHOST
 139 for a packet to some other host that has been caught by a device driver
 140 in promiscuous mode, and
 141 .B PACKET_OUTGOING
 142 for a packet originated from the local host that is looped back to a packet
 143 socket. These types make only sense for receiving.
 144 .B sll_addr
 145 and
 146 .B sll_halen
 147 contain the physical layer (e.g. IEEE 802.3) address and its length. The
 148 exact interpretation depends on the device.
 149
 150 When you send packets it is enough to specify
 151 .BR sll_family ,
 152 .BR sll_addr ,
 153 .BR sll_halen ,
 154 .BR sll_ifindex .
 155 The other fields should be 0.
 156 .B sll_hatype
 157 and
 158 .B sll_pkttype
 159 are set on received packets for your information.
 160 For bind only
 161 .B sll_protocol
 162 and
 163 .B sll_ifindex
 164 are used.
 165
 166 .SH "SOCKET OPTIONS"
 167 Packet sockets can be used to configure physical layer multicasting
 168 and promiscuous mode. It works by calling
 169 .BR setsockopt (2)
 170 on a packet socket for SOL_PACKET and one of the options
 171 .B PACKET_ADD_MEMBERSHIP
 172 to add a binding or
 173 .B PACKET_DROP_MEMBERSHIP
 174 to drop it.
 175 They both expect a
 176 .B packet_mreq
 177 structure as argument:
 178
 179 .RS
 180 .nf
 181 .ta 4n 20n 35n
 182 struct packet_mreq
 183 {
 184         int     mr_ifindex;     /* interface index */
 185         unsigned short  mr_type;        /* action */
 186         unsigned short  mr_alen;        /* address length */
 187         unsigned char   mr_address[8];  /* physical layer address */
 188 };
 189 .ta
 190 .fi
 191 .RE
 192
 193 .B mr_ifindex
 194 contains the interface index for the interface whose status
 195 should be changed.
 196 The
 197 .B mr_type
 198 parameter specifies which action to perform.
 199 .B PACKET_MR_PROMISC
 200 enables receiving all packets on a shared medium - often known as
 201 ``promiscuous mode'',
 202 .B PACKET_MR_MULTICAST
 203 binds the socket to the physical layer multicast group specified in
 204 .B mr_address
 205 and
 206 .BR mr_alen ,
 207 and
 208 .B PACKET_MR_ALLMULTI
 209 sets the socket up to receive all multicast packets arriving at the interface.
 210
 211 In addition the traditional ioctls
 212 .B SIOCSIFFLAGS,
 213 .B SIOCADDMULTI,
 214 .B SIOCDELMULTI
 215 can be used for the same purpose.
 216
 217
 218 .SH IOCTLS
 219 .B SIOCGSTAMP
 220 can be used to receive the time stamp of the last received packet. Argument
 221 is a
 222 .B struct timeval.
 223
 224 In addition all standard ioctls defined in
 225 .BR netdevice (7)
 226 and
 227 .BR socket (7)
 228 are valid on packet sockets.
 229
 230 .SH "ERROR HANDLING"
 231 Packet sockets do no error handling other than errors occurred while passing
 232 the packet to the device driver. They don't have the concept of a pending
 233 error.
 234
 235 .SH COMPATIBILITY
 236 In Linux 2.0, the only way to get a packet socket was by calling
 237 .BI "socket(PF_INET, SOCK_PACKET, " protocol )\fR.
 238 This is still supported but strongly deprecated.
 239 The main difference between the two methods is that
 240 .B SOCK_PACKET
 241 uses the old
 242 .B struct sockaddr_pkt
 243 to specify an interface, which doesn't provide physical layer independence.
 244
 245 .RS
 246 .nf
 247 .ta 4n 20n 35n
 248 struct sockaddr_pkt
 249 {
 250         unsigned short  spkt_family;
 251         unsigned char   spkt_device[14];
 252         unsigned short  spkt_protocol;
 253 };
 254 .ta
 255 .fi
 256 .RE
 257
 258 .B spkt_family
 259 contains
 260 the device type,
 261 .B spkt_protocol
 262 is the IEEE 802.3 protocol type as defined in
 263 .B <sys/if_ether.h>
 264 and
 265 .B spkt_device
 266 is the device name as a null terminated string, e.g. eth0.
 267
 268 This structure is obsolete and should not be used in new code.
 269
 270 .SH NOTES
 271 For portable programs it is suggested to use
 272 .B PF_PACKET
 273 via
 274 .BR pcap (3);
 275 although this only covers a subset of the
 276 .B PF_PACKET
 277 features.
 278
 279 The
 280 .B SOCK_DGRAM
 281 packet sockets make no attempt to create or parse the IEEE 802.2 LLC header
 282 for a IEEE 802.3 frame.
 283 When
 284 .B ETH_P_802_3
 285 is specified as protocol for sending the kernel creates the
 286 802.3 frame and fills out the length field; the user has to supply the LLC
 287 header to get a fully conforming packet. Incoming 802.3 packets are not
 288 multiplexed on the DSAP/SSAP protocol fields; instead they are supplied to the
 289 user as protocol
 290 .B ETH_P_802_2
 291 with the LLC header prepended. It is thus not possible to bind to
 292 .B ETH_P_802_3;
 293 bind to
 294 .B ETH_P_802_2
 295 instead and do the protocol multiplex yourself.
 296 The default for sending is the standard Ethernet DIX
 297 encapsulation with the protocol filled in.
 298
 299 Packet sockets are not subject to the input or output firewall chains.
 300
 301 .SH ERRORS
 302 .TP
 303 .B ENETDOWN
 304 Interface is not up.
 305
 306 .TP
 307 .B ENOTCONN
 308 No interface address passed.
 309
 310 .TP
 311 .B ENODEV
 312 Unknown device name or interface index specified in interface address.
 313
 314 .TP
 315 .B EMSGSIZE
 316 Packet is bigger than interface MTU.
 317
 318 .TP
 319 .B ENOBUFS
 320 Not enough memory to allocate the packet.
 321
 322 .TP
 323 .B EFAULT
 324 User passed invalid memory address.
 325
 326 .TP
 327 .B EINVAL
 328 Invalid argument.
 329
 330 .TP
 331 .B ENXIO
 332 Interface address contained illegal interface index.
 333
 334 .TP
 335 .B EPERM
 336 User has insufficient privileges to carry out this operation.
 337
 338 .TP
 339 .B EADDRNOTAVAIL
 340 Unknown multicast group address passed.
 341
 342 .TP
 343 .B ENOENT
 344 No packet received.
 345
 346 In addition other errors may be generated by the low-level driver.
 347 .SH VERSIONS
 348 .B PF_PACKET
 349 is a new feature in Linux 2.2. Earlier Linux versions supported only
 350 .B SOCK_PACKET.
 351
 352 .SH BUGS
 353 glibc 2.1 does not have a define for
 354 .B SOL_PACKET.
 355 The suggested workaround is to use
 356 .RS
 357 .nf
 358 #ifndef SOL_PACKET
 359 #define SOL_PACKET 263
 360 #endif
 361 .fi
 362 .RE
 363 This is fixed in later glibc versions and also does not occur on libc5 systems.
 364
 365 The IEEE 802.2/803.3 LLC handling could be considered as a bug.
 366
 367 Socket filters are not documented.
 368
 369 The
 370 .I MSG_TRUNC
 371 recvmsg extension is an ugly hack and should be replaced by a control message.
 372 There is currently no way to get the original destination address of
 373 packets via SOCK_DGRAM.
 374
 375 .SH HISTORICAL NOTE
 376 The include file
 377 .I <netpacket/packet.h>
 378 is present since glibc2.1. Older systems need
 379 .sp
 380 .nf
 381 .B #include <asm/types.h>
 382 .br
 383 .B #include <linux/if_packet.h>
 384 .br
 385 .B #include <linux/if_ether.h>  /* The L2 protocols */
 386 .br
 387 .fi
 388 .\" .SH CREDITS
 389 .\" This man page was written by Andi Kleen with help from Matthew Wilcox.
 390 .\" PF_PACKET in Linux 2.2 was implemented
 391 .\" by Alexey Kuznetsov, based on code by Alan Cox and others.
 392 .SH "SEE ALSO"
 393 .BR socket (2),
 394 .BR pcap (3),
 395 .BR capabilities (7),
 396 .BR ip (7),
 397 .BR raw (7),
 398 .BR socket (7)
 399
 400 RFC 894 for the standard IP Ethernet encapsulation.
 401
 402 RFC 1700 for the IEEE 802.3 IP encapsulation.
 403
 404 The
 405 .I <linux/if_ether.h>
 406 include file for physical layer protocols.