{ "mode": "man", "parameter": "udp", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/udp/7/json", "generated": "2026-08-14T04:13:11Z", "synopsis": "", "sections": { "NAME": { "content": "udp - User Datagram Protocol for IPv4\n", "subsections": [] }, "SYNOPSIS": { "content": "", "subsections": [ { "name": "#include ", "content": "" }, { "name": "#include ", "content": "" }, { "name": "#include ", "content": "udpsocket = socket(AFINET, SOCKDGRAM, 0);\n" } ] }, "DESCRIPTION": { "content": "This is an implementation of the User Datagram Protocol described in RFC 768. It implements\na connectionless, unreliable datagram packet service. Packets may be reordered or duplicated\nbefore they arrive. UDP generates and checks checksums to catch transmission errors.\n\nWhen a UDP socket is created, its local and remote addresses are unspecified. Datagrams can\nbe sent immediately using sendto(2) or sendmsg(2) with a valid destination address as an ar‐\ngument. When connect(2) is called on the socket, the default destination address is set and\ndatagrams can now be sent using send(2) or write(2) without specifying a destination address.\nIt is still possible to send to other destinations by passing an address to sendto(2) or\nsendmsg(2). In order to receive packets, the socket can be bound to a local address first by\nusing bind(2). Otherwise, the socket layer will automatically assign a free local port out\nof the range defined by /proc/sys/net/ipv4/iplocalportrange and bind the socket to IN‐\nADDRANY.\n\nAll receive operations return only one packet. When the packet is smaller than the passed\nbuffer, only that much data is returned; when it is bigger, the packet is truncated and the\nMSGTRUNC flag is set. MSGWAITALL is not supported.\n\nIP options may be sent or received using the socket options described in ip(7). They are\nprocessed by the kernel only when the appropriate /proc parameter is enabled (but still\npassed to the user even when it is turned off). See ip(7).\n\nWhen the MSGDONTROUTE flag is set on sending, the destination address must refer to a local\ninterface address and the packet is sent only to that interface.\n\nBy default, Linux UDP does path MTU (Maximum Transmission Unit) discovery. This means the\nkernel will keep track of the MTU to a specific target IP address and return EMSGSIZE when a\nUDP packet write exceeds it. When this happens, the application should decrease the packet\nsize. Path MTU discovery can be also turned off using the IPMTUDISCOVER socket option or\nthe /proc/sys/net/ipv4/ipnopmtudisc file; see ip(7) for details. When turned off, UDP\nwill fragment outgoing UDP packets that exceed the interface MTU. However, disabling it is\nnot recommended for performance and reliability reasons.\n", "subsections": [ { "name": "Address format", "content": "UDP uses the IPv4 sockaddrin address format described in ip(7).\n" }, { "name": "Error handling", "content": "All fatal errors will be passed to the user as an error return even when the socket is not\nconnected. This includes asynchronous errors received from the network. You may get an er‐\nror for an earlier packet that was sent on the same socket. This behavior differs from many\nother BSD socket implementations which don't pass any errors unless the socket is connected.\nLinux's behavior is mandated by RFC 1122.\n\nFor compatibility with legacy code, in Linux 2.0 and 2.2 it was possible to set the SOBSD‐\nCOMPAT SOLSOCKET option to receive remote errors only when the socket has been connected\n(except for EPROTO and EMSGSIZE). Locally generated errors are always passed. Support for\nthis socket option was removed in later kernels; see socket(7) for further information.\n\nWhen the IPRECVERR option is enabled, all errors are stored in the socket error queue, and\ncan be received by recvmsg(2) with the MSGERRQUEUE flag set.\n" }, { "name": "/proc interfaces", "content": "System-wide UDP parameter settings can be accessed by files in the directory\n/proc/sys/net/ipv4/.\n\nudpmem (since Linux 2.6.25)\nThis is a vector of three integers governing the number of pages allowed for queueing\nby all UDP sockets.\n\nmin Below this number of pages, UDP is not bothered about its memory appetite.\nWhen the amount of memory allocated by UDP exceeds this number, UDP starts to\nmoderate memory usage.\n\npressure\nThis value was introduced to follow the format of tcpmem (see tcp(7)).\n\nmax Number of pages allowed for queueing by all UDP sockets.\n\nDefaults values for these three items are calculated at boot time from the amount of\navailable memory.\n\nudprmemmin (integer; default value: PAGESIZE; since Linux 2.6.25)\nMinimal size, in bytes, of receive buffers used by UDP sockets in moderation. Each\nUDP socket is able to use the size for receiving data, even if total pages of UDP\nsockets exceed udpmem pressure.\n\nudpwmemmin (integer; default value: PAGESIZE; since Linux 2.6.25)\nMinimal size, in bytes, of send buffer used by UDP sockets in moderation. Each UDP\nsocket is able to use the size for sending data, even if total pages of UDP sockets\nexceed udpmem pressure.\n" }, { "name": "Socket options", "content": "To set or get a UDP socket option, call getsockopt(2) to read or setsockopt(2) to write the\noption with the option level argument set to IPPROTOUDP. Unless otherwise noted, optval is\na pointer to an int.\n\nFollowing is a list of UDP-specific socket options. For details of some other socket options\nthat are also applicable for UDP sockets, see socket(7).\n\nUDPCORK (since Linux 2.5.44)\nIf this option is enabled, then all data output on this socket is accumulated into a\nsingle datagram that is transmitted when the option is disabled. This option should\nnot be used in code intended to be portable.\n\nUDPSEGMENT (since Linux 4.18)\nEnables UDP segmentation offload. Segmentation offload reduces send(2) cost by trans‐\nferring multiple datagrams worth of data as a single large packet through the kernel\ntransmit path, even when that exceeds MTU. As late as possible, the large packet is\nsplit by segment size into a series of datagrams. This segmentation offload step is\ndeferred to hardware if supported, else performed in software. This option takes a\nvalue in the range [0, USHRTMAX] that sets the segment size: the size of datagram\npayload, excluding the UDP header. The segment size must be chosen such that at most\n64 datagrams are sent in a single call and that the datagrams after segmentation meet\nthe same MTU rules that apply to datagrams sent without this option. Segmentation of‐\nfload depends on checksum offload, as datagram checksums are computed after segmenta‐\ntion. The option may also be set for individual sendmsg(2) calls by passing it as a\ncmsg(3). A value of zero disables the feature. This option should not be used in\ncode intended to be portable.\n\nUDPGRO (since Linux 5.0)\nEnables UDP receive offload. If enabled, the socket may receive multiple datagrams\nworth of data as a single large buffer, together with a cmsg(3) that holds the segment\nsize. This option is the inverse of segmentation offload. It reduces receive cost by\nhandling multiple datagrams worth of data as a single large packet in the kernel re‐\nceive path, even when that exceeds MTU. This option should not be used in code in‐\ntended to be portable.\n" }, { "name": "Ioctls", "content": "These ioctls can be accessed using ioctl(2). The correct syntax is:\n\nint value;\nerror = ioctl(udpsocket, ioctltype, &value);\n\nFIONREAD (SIOCINQ)\nGets a pointer to an integer as argument. Returns the size of the next pending data‐\ngram in the integer in bytes, or 0 when no datagram is pending. Warning: Using FION‐\nREAD, it is impossible to distinguish the case where no datagram is pending from the\ncase where the next pending datagram contains zero bytes of data. It is safer to use\nselect(2), poll(2), or epoll(7) to distinguish these cases.\n\nTIOCOUTQ (SIOCOUTQ)\nReturns the number of data bytes in the local send queue. Supported only with Linux\n2.4 and above.\n\nIn addition, all ioctls documented in ip(7) and socket(7) are supported.\n" } ] }, "ERRORS": { "content": "All errors documented for socket(7) or ip(7) may be returned by a send or receive on a UDP\nsocket.\n\nECONNREFUSED\nNo receiver was associated with the destination address. This might be caused by a\nprevious packet sent over the socket.\n", "subsections": [] }, "VERSIONS": { "content": "IPRECVERR is a new feature in Linux 2.2.\n", "subsections": [] }, "SEE ALSO": { "content": "ip(7), raw(7), socket(7), udplite(7)\n\nThe kernel source file Documentation/networking/ip-sysctl.txt.\n\nRFC 768 for the User Datagram Protocol.\nRFC 1122 for the host requirements.\nRFC 1191 for a description of path MTU discovery.\n\nLinux man-pages 6.7 2023-10-31 udp(7)", "subsections": [] } }, "summary": "udp - User Datagram Protocol for IPv4", "flags": [], "examples": [], "see_also": [ { "name": "ip", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/ip/7/json" }, { "name": "raw", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/raw/7/json" }, { "name": "socket", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/socket/7/json" }, { "name": "udplite", "section": "7", "url": "https://www.chedong.com/phpMan.php/man/udplite/7/json" } ] }