RECV(2)

πŸŽ€ recv(2) πŸŽ€

okay bestie, sit down, hydrate, and let's talk about the family of syscalls that gets your data: recv(2). This is the read side of socket life, the part where the bytes come to you. πŸ’…


✨ NAME ✨

recv, recvfrom, recvmsg, recvmmsg β€” receive message(s) from a socket 🫢


πŸ“š LIBRARY πŸ“š

Standard C Library (libc, -lc) πŸ’–


🌸 SYNOPSIS 🌸

#include <sys/socket.h>

ssize_t
recv(int s, void *buf, size_t len, int flags);

ssize_t
recvfrom(int s, void *buf, size_t len, int flags,
    struct sockaddr * restrict from, socklen_t * restrict fromlen);

ssize_t
recvmsg(int s, struct msghdr *msg, int flags);

ssize_t
recvmmsg(int s, struct mmsghdr * restrict msgvec, size_t vlen,
    int flags, const struct timespec * restrict timeout);

like, four whole functions, all doing the "give me the message" thing. slay. πŸ’«


πŸ¦‹ DESCRIPTION πŸ¦‹

The recvfrom(), recvmsg(), and recvmmsg() system calls are used to receive messages from a socket, and β€” plot twist β€” they work whether or not the socket is connection-oriented. no gatekeeping here. ✨

If from is not a null pointer and the socket is not connection-oriented, the source address of the message gets filled in. The fromlen argument is a value-result argument: you initialize it to the size of the buffer associated with from, and on return it's been modified to show the actual size of the address that got stored there. it's giving "I changed while you were gone." 😭

The recv() function is normally used only on a connected socket (see connect(2)) and is identical to recvfrom() with a null pointer passed as its from argument. twins, basically. πŸ‘―

The recvmmsg() function is used to receive multiple messages in one call. Their number is supplied by vlen. The messages are placed in the buffers described by the msgvec vector, after reception. The size of each received message is placed in the msg_len field of each element of the vector. If timeout is NULL the call blocks until the data is available for each supplied message buffer. Otherwise it waits for data for the specified amount of time. If the timeout expired and there is no data received, a value 0 is returned. The ppoll(2) system call is used to implement the timeout mechanism, before first receive is performed. multi-tasking icon behavior. πŸ‘‘

The recv(), recvfrom() and recvmsg() return the length of the message on successful completion, whereas recvmmsg() returns the number of received messages. If a message is too long to fit in the supplied buffer, excess bytes may be discarded depending on the type of socket the message is received from (see socket(2)). yeah, they just… fall off. πŸ’€

If no messages are available at the socket, the receive call waits for a message to arrive β€” unless the socket is non-blocking (see fcntl(2)), in which case the value -1 is returned and the global variable errno is set to EAGAIN. The receive calls except recvmmsg() normally return any data available, up to the requested amount, rather than waiting for receipt of the full amount requested; this behavior is affected by the socket-level options SO_RCVLOWAT and SO_RCVTIMEO described in getsockopt(2). The recvmmsg() function implements this behaviour for each message in the vector. so it does the little "take what you can get" thing, but per-message. πŸ’–

The select(2) system call may be used to determine when more data arrives. ✨

The flags argument to a recv() function is formed by or'ing one or more of the values:

  • MSG_OOB β€” process out-of-band data
  • MSG_PEEK β€” peek at incoming message
  • MSG_WAITALL β€” wait for full request or error
  • MSG_DONTWAIT β€” do not block
  • MSG_CMSG_CLOEXEC β€” set received fds close-on-exec
  • MSG_WAITFORONE β€” do not block after receiving the first message (only for recvmmsg())

Now let's unpack those, because they're lowkey the whole personality of this call:

The MSG_OOB flag requests receipt of out-of-band data that would not be received in the normal data stream. Some protocols place expedited data at the head of the normal data queue, and thus this flag cannot be used with such protocols. ⚠️

The MSG_PEEK flag causes the receive operation to return data from the beginning of the receive queue without removing that data from the queue. Thus, a subsequent receive call will return the same data. it's a sneak peek, the message stays seated. πŸ‘€

The MSG_WAITALL flag requests that the operation block until the full request is satisfied. However, the call may still return less data than requested if a signal is caught, an error or disconnect occurs, or the next data to be received is of a different type than that returned. so "wait for all" comes with fine print, bestie. πŸ™ˆ

The MSG_DONTWAIT flag requests the call to return when it would block otherwise. If no data is available, errno is set to EAGAIN. This flag is not available in ANSI X3.159-1989 ("ANSI C89") or ISO/IEC 9899:1999 ("ISO C99") compilation mode. 🚫

The MSG_WAITFORONE flag sets MSG_DONTWAIT after the first message has been received. This flag is only relevant for recvmmsg(). 🧚

πŸͺž the msghdr situation πŸͺž

The recvmsg() system call uses a msghdr structure to minimize the number of directly supplied arguments. This structure has the following form, as defined in <sys/socket.h>:

struct msghdr {
        void            *msg_name;      /* optional address */
        socklen_t        msg_namelen;   /* size of address */
        struct iovec    *msg_iov;       /* scatter/gather array */
        int              msg_iovlen;    /* # elements in msg_iov */
        void            *msg_control;   /* ancillary data, see below */
        socklen_t        msg_controllen;/* ancillary data buffer len */
        int              msg_flags;     /* flags on received message */
};

Here msg_name and msg_namelen specify the source address if the socket is unconnected; msg_name may be given as a null pointer if no names are desired or required. The msg_iov and msg_iovlen arguments describe scatter gather locations, as discussed in read(2). The msg_control argument, which has length msg_controllen, points to a buffer for other protocol control related messages or other miscellaneous ancillary data. The messages are of the form:

struct cmsghdr {
        socklen_t  cmsg_len;    /* data byte count, including hdr */
        int        cmsg_level;  /* originating protocol */
        int        cmsg_type;   /* protocol-specific type */
/* followed by
        u_char     cmsg_data[]; */
};

As an example, one could use this to learn of changes in the data-stream in XNS/SPP, or in ISO, to obtain user-connection-request data by requesting a recvmsg() with no data buffer provided immediately after an accept() system call. ✨

With AF_UNIX domain sockets, ancillary data can be used to pass file descriptors and process credentials. See unix(4) for details. 🫢

The msg_flags field is set on return according to the message received:

  • MSG_EOR β€” indicates end-of-record; the data returned completed a record (generally used with sockets of type SOCK_SEQPACKET).
  • MSG_TRUNC β€” indicates that the trailing portion of a datagram was discarded because the datagram was larger than the buffer supplied.
  • MSG_CTRUNC β€” indicates that some control data were discarded due to lack of space in the buffer for ancillary data.
  • MSG_OOB β€” is returned to indicate that expedited or out-of-band data were received.

🧺 the mmsghdr situation 🧺

The recvmmsg() system call uses the mmsghdr structure, defined as follows in the <sys/socket.h> header:

struct mmsghdr {
        struct msghdr    msg_hdr;       /* message header */
        ssize_t          msg_len;       /* message length */
};

On data reception the msg_len field is updated to the length of the received message. ✨


πŸ’― RETURN VALUES πŸ’―

These calls except recvmmsg() return the number of bytes received. recvmmsg() returns the number of messages received. A value of -1 is returned if an error occurred. πŸ’…


🚨 ERRORS 🚨

The calls fail if:

  • EBADF β€” The argument s is an invalid descriptor.
  • ECONNRESET β€” The remote socket end is forcibly closed.
  • ENOTCONN β€” The socket is associated with a connection-oriented protocol and has not been connected (see connect(2) and accept(2)).
  • ENOTSOCK β€” The argument s does not refer to a socket.
  • EMSGSIZE β€” The recvmsg() system call was used to receive rights (file descriptors) that were in flight on the connection. However, the receiving program did not have enough free file descriptor slots to accept them. In this case the descriptors are closed, any pending data can be returned by another call to recvmsg().
  • EAGAIN β€” The socket is marked non-blocking and the receive operation would block, or a receive timeout had been set and the timeout expired before data were received.
  • EINTR β€” The receive was interrupted by delivery of a signal before any data were available.
  • EFAULT β€” The receive buffer pointer(s) point outside the process's address space.
fr fr, if you get EMSGSIZE those file descriptors are gone. not "paused." closed. respect the stakes. 😭

πŸ‘€ SEE ALSO πŸ‘€

fcntl(2), getsockopt(2), read(2), select(2), socket(2), CMSG_DATA(3), unix(4)


πŸ“œ HISTORY πŸ“œ

The recv() function appeared in 4.2BSD. The recvmmsg() function appeared in FreeBSD 11.0. the lore is deep. 🌈


okay and that's the whole page bestie, no cap. this version was brought to you by Debian, dated August 19, 2018 β€” and yes, that footer is the vibe check passing. πŸ’–πŸ¦‹πŸŒΈ go receive your messages. πŸ’…βœ¨

πŸ“¦ where this page comes from