NETINTRO(4)

NETINTRO(4) 💖

okay bestie, welcome to networking — like, the entire introduction to networking facilities. this is the manual page, section 4, and it's giving "you're about to learn how computers gossip with each other" energy. grab a snack 🍓

---

✨ SYNOPSIS ✨

#include <sys/types.h>
#include <sys/time.h>
#include <sys/socket.h>
#include <net/if.h>
#include <net/route.h>

look at those includes, they're iconic, they're literally the starter pack 🎀

---

💅 DESCRIPTION 💅

so this whole section is the general intro to all the networking stuff your system has going on. the docs in this part of section 4 split into three little cliques, no cap:

• protocol families (aka domains)
• protocols
• network interfaces

every network protocol is ride-or-die with a specific protocol family. a protocol family is like the group chat that gives basic services to the protocol implementation so it can actually function in a specific network environment. those services can include packet fragmentation and reassembly, routing, addressing, and basic transport. a protocol family may support multiple methods of addressing, though ngl the current protocol implementations do not. a protocol family is normally made up of a bunch of protocols, one per socket(2) type. and no, it's not required that a protocol family support all socket types. a protocol family can totally contain multiple protocols supporting the same socket abstraction. slay.

a protocol supports one of the socket abstractions detailed in socket(2). you can reach a specific protocol either by creating a socket of the appropriate type and protocol family, or by requesting the protocol explicitly when you create a socket. protocols normally accept only one type of address format, usually decided by the addressing structure that's inherent in the design of the protocol family / network architecture. certain semantics of the basic socket abstractions are protocol specific. all protocols are expected to support the basic model for their particular socket type, but they may also provide non-standard facilities or extensions to a mechanism. for example — and this is such a good example — a protocol supporting the SOCK_STREAM abstraction may allow more than one byte of out-of-band data to be transmitted per out-of-band message. the way that works?? obsessed. 😌

a network interface is similar to a device interface. network interfaces make up the lowest layer of the networking subsystem, and they're the ones actually interacting with the real transport hardware. an interface may support one or more protocol families and/or address formats. the SYNOPSIS section of each network interface entry gives you a sample specification of the related drivers for use in providing a system description to the config(8) program. and the DIAGNOSTICS section lists the messages that may show up on the console and/or in the system error log, /var/log/messages (peep syslogd(8)), because of errors in device operation. so if your console starts screaming, that's where you look 🔍

---

🌈 PROTOCOLS 🌈

right now the system supports the Internet protocols, the Xerox Network Systems(tm) protocols, and some of the ISO OSI protocols. raw socket interfaces are provided to the IP protocol layer of the Internet, and to the IDP protocol of Xerox NS. for more info on the support for each protocol family, go consult the appropriate manual pages in this section. iykyk 📚

---

🦋 ADDRESSING 🦋

every protocol family has an address format attached to it. all network addresses stick to a general structure called a sockaddr, described below. but each protocol imposes finer and more specific structure on top, generally renaming the variant — and that's discussed in the protocol family manual page we alluded to above. periodt.

struct sockaddr {
    u_char      sa_len;
    u_char      sa_family;
    char        sa_data[14];
};

the field sa_len contains the total length of the structure, which may exceed 16 bytes. the following address values for sa_family are known to the system (and, bestie, additional formats are defined for possible future implementation):

#define    AF_UNIX      1    /* local to host (pipes, portals) */
#define    AF_INET      2    /* internetwork: UDP, TCP, etc. */
#define    AF_NS        6    /* Xerox NS protocols */
#define    AF_CCITT     10   /* CCITT protocols, X.25 etc */
#define    AF_HYLINK    15   /* NSC Hyperchannel */
#define    AF_ISO       18   /* ISO protocols */

those defines are literally the main characters of this page 💖

---

👑 ROUTING 👑

FreeBSD provides some packet routing facilities. the kernel keeps a routing information database, and that's what gets used when selecting the appropriate network interface to transmit packets. so yeah, the kernel is out here making decisions, main character energy fr.

a user process (or possibly multiple co-operating processes) maintains this database by sending messages over a special kind of socket. this supplants the fixed size ioctl(2) used in earlier releases. old news, we don't do that anymore 💅

this facility is described in route(4).

okay bestie this page is long so it comes in 2 parts and this was part 1 — more parts follow, don't go anywhere 🫶✨

✨ INTERFACES ✨

Okay bestie, this is the part where it gets real. 💅

Every network interface in a system is basically a lil path where messages get sent and received, right? And usually there's an actual hardware device attached to it — except some iconic exceptions like the loopback interface, lo(4), which is out here living its best life with zero hardware, no cap. 🦋

---

🎀 The ioctl() Calls That Do The Thing

So here's the tea: these ioctl(2) calls are how you poke at network interfaces. You make the ioctl() on a socket (typically of type SOCK_DGRAM) in whatever domain you want. Most of the requests that were supported in earlier releases take an ifreq structure as their parameter. And this structure looks like this:

struct  ifreq {
#define    IFNAMSIZ    16
    char    ifr_name[IFNAMSIZ];        /* if name, e.g. "en0" */
    union {
        struct    sockaddr ifru_addr;
        struct    sockaddr ifru_dstaddr;
        struct    sockaddr ifru_broadaddr;
        struct    ifreq_buffer ifru_buffer;
        short     ifru_flags[2];
        short     ifru_index;
        int       ifru_metric;
        int       ifru_mtu;
        int       ifru_phys;
        int       ifru_media;
        caddr_t   ifru_data;
        int       ifru_cap[2];
    } ifr_ifru;
#define ifr_addr      ifr_ifru.ifru_addr      /* address */
#define ifr_dstaddr   ifr_ifru.ifru_dstaddr   /* other end of p-to-p link */
#define ifr_broadaddr ifr_ifru.ifru_broadaddr /* broadcast address */
#define ifr_buffer    ifr_ifru.ifru_buffer    /* user supplied buffer with its length */
#define ifr_flags     ifr_ifru.ifru_flags[0]  /* flags (low 16 bits) */
#define ifr_flagshigh ifr_ifru.ifru_flags[1]  /* flags (high 16 bits) */
#define ifr_metric    ifr_ifru.ifru_metric    /* metric */
#define ifr_mtu       ifr_ifru.ifru_mtu       /* mtu */
#define ifr_phys      ifr_ifru.ifru_phys      /* physical wire */
#define ifr_media     ifr_ifru.ifru_media     /* physical media */
#define ifr_data      ifr_ifru.ifru_data      /* for use by interface */
#define ifr_reqcap    ifr_ifru.ifru_cap[0]    /* requested capabilities */
#define ifr_curcap    ifr_ifru.ifru_cap[1]    /* current capabilities */
#define ifr_index     ifr_ifru.ifru_index     /* interface index */
};

Obsessed with a good union, honestly. 🫶

---

🌸 Requests That Still Use The Classic ifreq Struct

The ioctl() requests to obtain addresses, plus the ones that both set and retrieve other data, are still fully supported and they use the ifreq structure. Slay:

• SIOCGIFADDR — Get the interface address for a protocol family.
• SIOCGIFDSTADDR — Get the point-to-point address for a protocol family and interface.
• SIOCGIFBRDADDR — Get the broadcast address for a protocol family and interface.
• SIOCSIFCAP — Attempt to set the enabled capabilities field for the interface to the value of the ifr_reqcap field of the ifreq structure. ⚠️ Heads up bestie: depending on the particular interface features, some capabilities may appear hard-coded to enabled, or toggling a capability may affect the status of other ones. The supported capabilities field is read-only, and the ifr_curcap field is unused by this call.
• SIOCGIFCAP — Get the interface capabilities fields. The values for supported and enabled capabilities get returned in the ifr_reqcap and ifr_curcap fields of the ifreq structure, respectively.
• SIOCGIFDESCR — Get the interface description, returned in the buffer field of the ifru_buffer struct. The user supplied buffer length should be defined in the length field of the ifru_buffer struct passed in as parameter, and the length would include the terminating nul character. If there's not enough space to hold the interface length, no copy would be done and the buffer field of ifru_buffer would be set to NULL. The kernel will store the buffer length in the length field upon return, regardless whether the buffer itself is sufficient to hold the data. 👑
• SIOCSIFDESCR — Set the interface description to the value of the buffer field of the ifru_buffer struct, with the length field specifying its length (counting the terminating nul).
• SIOCSIFFLAGS — Set the interface flags field. If the interface is marked down, any processes currently routing packets through the interface are notified; some interfaces may be reset so that incoming packets are no longer received. When marked up again, the interface is reinitialized. 💀
• SIOCGIFFLAGS — Get the interface flags.
• SIOCSIFMETRIC — Set the interface routing metric. The metric is used only by user-level routers, fyi.
• SIOCGIFMETRIC — Get the interface metric.
• SIOCIFCREATE — Attempt to create the specified interface. If the interface name is given without a unit number, the system will attempt to create a new interface with an arbitrary unit number. On successful return the ifr_name field will contain the new interface name. ✨
• SIOCIFDESTROY — Attempt to destroy the specified interface. 💥

---

🦋 The Two Requests That Use A Whole New Structure

There are two requests that make use of a new structure, and they are that girl:

• SIOCAIFADDR — An interface may have more than one address associated with it in some protocols. This request provides a means to add additional addresses (or modify characteristics of the primary address if the default address for the address family is specified). Rather than making separate calls to set destination or broadcast addresses, or network masks (now an integral feature of multiple protocols), a separate structure is used to specify all three facets simultaneously (see below). One would use a slightly tailored version of this struct specific to each family (replacing each sockaddr by one of the family-specific type). Where the sockaddr itself is larger than the default size, one needs to modify the ioctl() identifier itself to include the total size, as described in ioctl().
• SIOCDIFADDR — This requests deletes the specified address from the list associated with an interface. It also uses the ifaliasreq structure to allow for the possibility of protocols allowing multiple masks or destination addresses, and also adopts the convention that specification of the default address means to delete the first address for the interface belonging to the address family in which the original socket was opened.
• SIOCGIFCONF — Get interface configuration list. This request takes an ifconf structure (see below) as a value-result parameter. The ifc_len field should be initially set to the size of the buffer pointed to by ifc_buf. On return it will contain the length, in bytes, of the configuration list.
• SIOCIFGCLONERS — Get list of clonable interfaces. This request takes an if_clonereq structure (see below) as a value-result parameter. The ifcr_count field should be set to the number of IFNAMSIZ sized strings that can be fit in the buffer pointed to by ifcr_buffer. On return, ifcr_total will be set to the number of clonable interfaces and the buffer pointed to by ifcr_buffer will be filled with the names of clonable interfaces aligned on IFNAMSIZ boundaries.

---

💖 The Structures, In All Their Glory

/*
* Structure used in SIOCAIFADDR request.
*/
struct ifaliasreq {
        char    ifra_name[IFNAMSIZ];   /* if name, e.g. "en0" */
        struct  sockaddr        ifra_addr;
        struct  sockaddr        ifra_broadaddr;
        struct  sockaddr        ifra_mask;
};

/*
* Structure used in SIOCGIFCONF request.
* Used to retrieve interface configuration
* for machine (useful for programs which
* must know all networks accessible).
*/
struct ifconf {
    int   ifc_len;              /* size of associated buffer */
    union {
        caddr_t    ifcu_buf;
        struct     ifreq *ifcu_req;
    } ifc_ifcu;
#define ifc_buf ifc_ifcu.ifcu_buf /* buffer address */
#define ifc_req ifc_ifcu.ifcu_req /* array of structures returned */
};

/* Structure used in SIOCIFGCLONERS request. */
struct if_clonereq {
        int     ifcr_total;     /* total cloners (out) */
        int     ifcr_count;     /* room for this many in user buffer */
        char    *ifcr_buffer;   /* buffer for cloner names */
};

/* Structure used in SIOCGIFDESCR and SIOCSIFDESCR requests */
struct ifreq_buffer {
        size_t  length;         /* length of the buffer */
        void   *buffer;         /* pointer to userland space buffer */
};

Look at all these lil structs living their truth. 🌈

---

👑 SEE ALSO

ioctl(2), socket(2), intro(4), config(8), routed(8), ifnet(9)

---

🌸 HISTORY

The netintro manual appeared in 4.3BSD-Tahoe. An absolute classic, fr fr. ⭐

---

And that's a wrap on this one, bestie! 🎀 This page came to you from Debian, dated January 26, 2012, filed under NETINTRO(4) — go forth and configure those interfaces like the main character you are. 💅✨