16 Datagram Congestion Control Protocol (DCCP) is an unreliable, connection
17 oriented protocol designed to solve issues present in UDP and TCP, particularly
18 for real-time and multimedia (streaming) traffic.
19 It divides into a base protocol (RFC 4340) and plugable congestion control
20 modules called CCIDs. Like plugable TCP congestion control, at least one CCID
21 needs to be enabled in order for the protocol to function properly. In the Linux
22 implementation, this is the TCP-like CCID2 (RFC 4341). Additional CCIDs, such as
23 the TCP-friendly CCID3 (RFC 4342), are optional.
24 For a brief introduction to CCIDs and suggestions for choosing a CCID to match
25 given applications, see section 10 of RFC 4340.
27 It has a base protocol and pluggable congestion control IDs (CCIDs).
29 DCCP is a Proposed Standard (RFC 2026), and the homepage for DCCP as a protocol
30 is at http://www.ietf.org/html.charters/dccp-charter.html
35 The Linux DCCP implementation does not currently support all the features that are
36 specified in RFCs 4340...42.
38 The known bugs are at:
39 http://linux-net.osdl.org/index.php/TODO#DCCP
41 For more up-to-date versions of the DCCP implementation, please consider using
42 the experimental DCCP test tree; instructions for checking this out are on:
43 http://linux-net.osdl.org/index.php/DCCP_Testing#Experimental_DCCP_source_tree
48 DCCP_SOCKOPT_QPOLICY_ID sets the dequeuing policy for outgoing packets. It takes
49 a policy ID as argument and can only be set before the connection (i.e. changes
50 during an established connection are not supported). Currently, two policies are
51 defined: the "simple" policy (DCCPQ_POLICY_SIMPLE), which does nothing special,
52 and a priority-based variant (DCCPQ_POLICY_PRIO). The latter allows to pass an
53 u32 priority value as ancillary data to sendmsg(), where higher numbers indicate
54 a higher packet priority (similar to SO_PRIORITY). This ancillary data needs to
55 be formatted using a cmsg(3) message header filled in as follows:
56 cmsg->cmsg_level = SOL_DCCP;
57 cmsg->cmsg_type = DCCP_SCM_PRIORITY;
58 cmsg->cmsg_len = CMSG_LEN(sizeof(uint32_t)); /* or CMSG_LEN(4) */
60 DCCP_SOCKOPT_QPOLICY_TXQLEN sets the maximum length of the output queue. A zero
61 value is always interpreted as unbounded queue length. If different from zero,
62 the interpretation of this parameter depends on the current dequeuing policy
63 (see above): the "simple" policy will enforce a fixed queue size by returning
64 EAGAIN, whereas the "prio" policy enforces a fixed queue length by dropping the
65 lowest-priority packet first. The default value for this parameter is
66 initialised from /proc/sys/net/dccp/default/tx_qlen.
68 DCCP_SOCKOPT_SERVICE sets the service. The specification mandates use of
69 service codes (RFC 4340, sec. 8.1.2); if this socket option is not set,
70 the socket will fall back to 0 (which means that no meaningful service code
71 is present). On active sockets this is set before connect(); specifying more
72 than one code has no effect (all subsequent service codes are ignored). The
73 case is different for passive sockets, where multiple service codes (up to 32)
74 can be set before calling bind().
76 DCCP_SOCKOPT_GET_CUR_MPS is read-only and retrieves the current maximum packet
77 size (application payload size) in bytes, see RFC 4340, section 14.
79 DCCP_SOCKOPT_AVAILABLE_CCIDS is also read-only and returns the list of CCIDs
80 supported by the endpoint (see include/linux/dccp.h for symbolic constants).
81 The caller needs to provide a sufficiently large (> 2) array of type uint8_t.
83 DCCP_SOCKOPT_CCID is write-only and sets both the TX and RX CCIDs at the same
84 time, combining the operation of the next two socket options. This option is
85 preferrable over the latter two, since often applications will use the same
86 type of CCID for both directions; and mixed use of CCIDs is not currently well
87 understood. This socket option takes as argument at least one uint8_t value, or
88 an array of uint8_t values, which must match available CCIDS (see above). CCIDs
89 must be registered on the socket before calling connect() or listen().
91 DCCP_SOCKOPT_TX_CCID is read/write. It returns the current CCID (if set) or sets
92 the preference list for the TX CCID, using the same format as DCCP_SOCKOPT_CCID.
93 Please note that the getsockopt argument type here is `int', not uint8_t.
95 DCCP_SOCKOPT_RX_CCID is analogous to DCCP_SOCKOPT_TX_CCID, but for the RX CCID.
97 DCCP_SOCKOPT_SERVER_TIMEWAIT enables the server (listening socket) to hold
98 timewait state when closing the connection (RFC 4340, 8.3). The usual case is
99 that the closing server sends a CloseReq, whereupon the client holds timewait
100 state. When this boolean socket option is on, the server sends a Close instead
101 and will enter TIMEWAIT. This option must be set after accept() returns.
103 DCCP_SOCKOPT_SEND_CSCOV and DCCP_SOCKOPT_RECV_CSCOV are used for setting the
104 partial checksum coverage (RFC 4340, sec. 9.2). The default is that checksums
105 always cover the entire packet and that only fully covered application data is
106 accepted by the receiver. Hence, when using this feature on the sender, it must
107 be enabled at the receiver, too with suitable choice of CsCov.
109 DCCP_SOCKOPT_SEND_CSCOV sets the sender checksum coverage. Values in the
110 range 0..15 are acceptable. The default setting is 0 (full coverage),
111 values between 1..15 indicate partial coverage.
112 DCCP_SOCKOPT_RECV_CSCOV is for the receiver and has a different meaning: it
113 sets a threshold, where again values 0..15 are acceptable. The default
114 of 0 means that all packets with a partial coverage will be discarded.
115 Values in the range 1..15 indicate that packets with minimally such a
116 coverage value are also acceptable. The higher the number, the more
117 restrictive this setting (see [RFC 4340, sec. 9.2.1]). Partial coverage
118 settings are inherited to the child socket after accept().
120 The following two options apply to CCID 3 exclusively and are getsockopt()-only.
121 In either case, a TFRC info struct (defined in <linux/tfrc.h>) is returned.
122 DCCP_SOCKOPT_CCID_RX_INFO
123 Returns a `struct tfrc_rx_info' in optval; the buffer for optval and
124 optlen must be set to at least sizeof(struct tfrc_rx_info).
125 DCCP_SOCKOPT_CCID_TX_INFO
126 Returns a `struct tfrc_tx_info' in optval; the buffer for optval and
127 optlen must be set to at least sizeof(struct tfrc_tx_info).
129 On unidirectional connections it is useful to close the unused half-connection
130 via shutdown (SHUT_WR or SHUT_RD): this will reduce per-packet processing costs.
134 Several DCCP default parameters can be managed by the following sysctls
135 (sysctl net.dccp.default or /proc/sys/net/dccp/default):
138 The number of active connection initiation retries (the number of
139 Requests minus one) before timing out. In addition, it also governs
140 the behaviour of the other, passive side: this variable also sets
141 the number of times DCCP repeats sending a Response when the initial
142 handshake does not progress from RESPOND to OPEN (i.e. when no Ack
143 is received after the initial Request). This value should be greater
144 than 0, suggested is less than 10. Analogue of tcp_syn_retries.
147 How often a DCCP Response is retransmitted until the listening DCCP
148 side considers its connecting peer dead. Analogue of tcp_retries1.
151 The number of times a general DCCP packet is retransmitted. This has
152 importance for retransmitted acknowledgments and feature negotiation,
153 data packets are never retransmitted. Analogue of tcp_retries2.
156 Default CCID for the sender-receiver half-connection. Depending on the
157 choice of CCID, the Send Ack Vector feature is enabled automatically.
160 Default CCID for the receiver-sender half-connection; see tx_ccid.
163 The initial sequence window (sec. 7.5.2) of the sender. This influences
164 the local ackno validity and the remote seqno validity windows (7.5.1).
167 The size of the transmit buffer in packets. A value of 0 corresponds
168 to an unbounded transmit buffer.
170 sync_ratelimit = 125 ms
171 The timeout between subsequent DCCP-Sync packets sent in response to
172 sequence-invalid packets on the same socket (RFC 4340, 7.5.4). The unit
173 of this parameter is milliseconds; a value of 0 disables rate-limiting.
178 Works as in udp(7): returns in the `int' argument pointer the size of
179 the next pending datagram in bytes, or 0 when no datagram is pending.
184 DCCP does not travel through NAT successfully at present on many boxes. This is
185 because the checksum covers the pseudo-header as per TCP and UDP. Linux NAT
186 support for DCCP has been added.