.\" $OpenBSD: frame.4,v 1.3 2025/05/16 02:50:12 dlg Exp $ .\" .\" Copyright (c) 2024 David Gwynne .\" .\" Permission to use, copy, modify, and distribute this software for any .\" purpose with or without fee is hereby granted, provided that the above .\" copyright notice and this permission notice appear in all copies. .\" .\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES .\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF .\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR .\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES .\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN .\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF .\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. .\" .Dd $Mdocdate: May 16 2025 $ .Dt FRAME 4 .Os .Sh NAME .Nm frame .Nd frame protocol family .Sh SYNOPSIS .Cd "pseudo-device af_frame" .Pp .In sys/types.h .In net/frame.h .Sh DESCRIPTION The .Nm protocol family provides an interface for sending and receiving low level network interface frames through the normal .Xr socket 2 mechanisms. .Pp The .Nm protocol family supports the .Dv SOCK_DGRAM socket type. Only root may create .Nm protocol family sockets. .Pp .Nm protocol family sockets are designed as an alternative to .Xr bpf 4 for handling low data and packet rate communication protocols. Rather than filtering every frame entering the system before the network stack, like .Xr bpf 4 , processing of the .Nm protocol family runs after the built in protocol handlers in the kernel, thus avoiding the overhead. For this reason, it is not possible to handle IPv4 or IPv6 packets with .Nm protocol sockets because the kernel network stack consumes them before the receive handling for .Nm sockets is run. .Pp Sockets can be created in the .Nm protocol family by using .Dv AF_FRAME as the .Fa domain argument to .Xr socket 2 . The type of interface, as per .In net/if_types.h , is specified as the socket .Fa protocol . Currently only Ethernet interfaces are supported. .Pp Sockets bound to the .Nm family use the following address structure: .Bd -literal -offset indent #define FRAME_ADDRLEN 8 #define FRAME_DATALEN 32 struct sockaddr_frame { uint8_t sfrm_len; uint8_t sfrm_family; uint16_t sfrm_proto; unsigned int sfrm_ifindex; uint8_t sfrm_addr[FRAME_ADDRLEN]; char sfrm_ifname[IFNAMSIZ]; uint8_t sfrm_data[FRAME_DATALEN]; }; .Ed .Pp The interface used for transmitting or receiving frames with a .Nm domain socket may be specified by using an interface index with .Fa sfrm_ifindex , or by name with .Fa sfrm_ifname . The use of other .Vt struct sockaddr_frame fields depends on the type of interface. .Ss Ethernet frame sockets A .Nm socket for use with Ethernet interfaces can be created using .Dv IFT_ETHER as the .Fa protocol argument to .Xr socket 2 : .Bd -literal -offset indent int sock = socket(AF_FRAME, SOCK_DGRAM, IFT_ETHER); .Ed .Pp The Ethernet protocol is specified with .Fa sfrm_proto in network byte order. Ethernet addresses are specified using the first 6 bytes of .Fa sfrm_addr . .Pp Ethernet .Nm sockets may receive frames on all interfaces by specifying 0 for .Va sfrm_ifindex when using .Xr bind 2 . Similarly, a .Dq wildcard local address of all zeros can be specified in .Fa sfrm_addr . .Pp An interface and address must be specified when sending Ethernet frames. .Pp Ethernet sockets support the following .Nm socket options using .Dv IFT_ETHER as the .Fa level argument with .Xr setsockopt 2 and .Xr getsockopt 2 : .Bl -tag -width Ds .It Dv FRAME_RECVDSTADDR Ft int Enable or disable the reception of the Ethernet destination address as a .Vt struct ether_addr control message for frames received with .Xr recvmsg 2 . .It Dv FRAME_RECVPRIO Ft int Enable or disable the reception of the mbuf packet priority field as an .Vt int sized control message for frames received with .Xr recvmsg 2 . .It Dv FRAME_ADD_MEMBERSHIP Ft struct frame_mreq Configure an Ethernet interface to enable reception of frames destined to the specified multicast Ethernet address. .Bd -literal -offset indent struct frame_mreq { unsigned int fmr_ifindex; uint8_t fmr_addr[FRAME_ADDRLEN]; char fmr_ifname[IFNAMSIZ]; }; .Ed .Pp An interface must be specified using either .Fa fmr_ifindex or .Fa fmr_ifname . The multicast Ethernet address is specified in the first 6 bytes of .Fa fmr_addr . .It Dv FRAME_DEL_MEMBERSHIP Ft struct frame_mreq Configure an Ethernet interface to disable reception of frames destined to the specified multicast Ethernet address. .It Dv FRAME_SENDPRIO Ft int Specify an mbuf priority value between .Dv IF_HDRPRIO_MIN .Pq 0 and .Dv IF_HDRPRIO_MAX .Pq 7 for frames sent with the Ethernet .Nm socket, or .Dv IF_HDRPRIO_PACKET to use the existing mbuf priority value. The default is .Dv IF_HDRPRIO_PACKET . .El .Sh EXAMPLES To receive LLDP frames on the em0 Ethernet interface: .Bd -literal -offset indent struct sockaddr_frame sfrm = { .sfrm_family = AF_FRAME, .sfrm_ifname = "em0", .sfrm_proto = htons(ETHERTYPE_LLDP), }; struct frame_mreq fmr = { .fmr_ifname = "em0", .fmr_addr = { 0x01, 0x80, 0xc2, 0x00, 0x00, 0x0e }, }; int sock; sock = socket(AF_FRAME, SOCK_DGRAM, IFT_ETHER); if (sock == -1) err(1, "ethernet frame socket"); if (bind(sock, (struct sockaddr *)&sfrm, sizeof(sfrm)) == -1) err(1, "bind"); if (setsockopt(sock, IFT_ETHER, FRAME_ADD_MEMBERSHIP, &fmr, sizeof(fmr)) == -1) err(1, "join lldp multicast group"); for (;;) { socklen_t sfrmlen = sizeof(sfrm); uint8_t frame[2048]; ssize_t rv; rv = recvfrom(sock, frame, sizeof(frame), 0, (struct sockaddr *)&sfrm, &sfrmlen); if (rv == -1) err(1, "lldp recv"); printf("received %zd bytes from %s", rv, ether_ntoa((struct ether_addr *)sfrm.sfrm_addr)); } .Ed .Sh SEE ALSO .Xr socket 2 , .Xr netintro 4 .Sh HISTORY .Nm domain sockets appeared in .Ox 7.7 . .\" The .\" .Ox .\" implementation was influenced by the Linux .\" .Dv AF_PACKET .\" .Dq packet interface on device level .\" socket interface, but is not compatible with it. .Sh AUTHORS .An David Gwynne Aq Mt dlg@openbsd.org .