]>
Commit | Line | Data |
---|---|---|
98069ff4 | 1 | DCCP protocol |
4886fcad | 2 | ============= |
98069ff4 | 3 | |
98069ff4 IM |
4 | |
5 | Contents | |
6 | ======== | |
98069ff4 IM |
7 | - Introduction |
8 | - Missing features | |
9 | - Socket options | |
4886fcad GR |
10 | - Sysctl variables |
11 | - IOCTLs | |
12 | - Other tunables | |
98069ff4 IM |
13 | - Notes |
14 | ||
4886fcad | 15 | |
98069ff4 IM |
16 | Introduction |
17 | ============ | |
98069ff4 | 18 | Datagram Congestion Control Protocol (DCCP) is an unreliable, connection |
e333b3ed GR |
19 | oriented protocol designed to solve issues present in UDP and TCP, particularly |
20 | for real-time and multimedia (streaming) traffic. | |
21 | It divides into a base protocol (RFC 4340) and plugable congestion control | |
22 | modules called CCIDs. Like plugable TCP congestion control, at least one CCID | |
23 | needs to be enabled in order for the protocol to function properly. In the Linux | |
24 | implementation, this is the TCP-like CCID2 (RFC 4341). Additional CCIDs, such as | |
25 | the TCP-friendly CCID3 (RFC 4342), are optional. | |
26 | For a brief introduction to CCIDs and suggestions for choosing a CCID to match | |
27 | given applications, see section 10 of RFC 4340. | |
98069ff4 IM |
28 | |
29 | It has a base protocol and pluggable congestion control IDs (CCIDs). | |
30 | ||
ebe6f7e7 GR |
31 | DCCP is a Proposed Standard (RFC 2026), and the homepage for DCCP as a protocol |
32 | is at http://www.ietf.org/html.charters/dccp-charter.html | |
98069ff4 | 33 | |
4886fcad | 34 | |
98069ff4 IM |
35 | Missing features |
36 | ================ | |
ebe6f7e7 GR |
37 | The Linux DCCP implementation does not currently support all the features that are |
38 | specified in RFCs 4340...42. | |
98069ff4 | 39 | |
ddfe10b8 IM |
40 | The known bugs are at: |
41 | http://linux-net.osdl.org/index.php/TODO#DCCP | |
98069ff4 | 42 | |
ebe6f7e7 GR |
43 | For more up-to-date versions of the DCCP implementation, please consider using |
44 | the experimental DCCP test tree; instructions for checking this out are on: | |
45 | http://linux-net.osdl.org/index.php/DCCP_Testing#Experimental_DCCP_source_tree | |
46 | ||
47 | ||
98069ff4 IM |
48 | Socket options |
49 | ============== | |
00e4d116 GR |
50 | DCCP_SOCKOPT_SERVICE sets the service. The specification mandates use of |
51 | service codes (RFC 4340, sec. 8.1.2); if this socket option is not set, | |
52 | the socket will fall back to 0 (which means that no meaningful service code | |
126acd5b GR |
53 | is present). On active sockets this is set before connect(); specifying more |
54 | than one code has no effect (all subsequent service codes are ignored). The | |
55 | case is different for passive sockets, where multiple service codes (up to 32) | |
56 | can be set before calling bind(). | |
98069ff4 | 57 | |
7c559a9e GR |
58 | DCCP_SOCKOPT_GET_CUR_MPS is read-only and retrieves the current maximum packet |
59 | size (application payload size) in bytes, see RFC 4340, section 14. | |
60 | ||
d90ebcbf | 61 | DCCP_SOCKOPT_AVAILABLE_CCIDS is also read-only and returns the list of CCIDs |
69a6a0b3 GR |
62 | supported by the endpoint. The option value is an array of type uint8_t whose |
63 | size is passed as option length. The minimum array size is 4 elements, the | |
64 | value returned in the optlen argument always reflects the true number of | |
65 | built-in CCIDs. | |
d90ebcbf | 66 | |
b20a9c24 GR |
67 | DCCP_SOCKOPT_CCID is write-only and sets both the TX and RX CCIDs at the same |
68 | time, combining the operation of the next two socket options. This option is | |
69 | preferrable over the latter two, since often applications will use the same | |
70 | type of CCID for both directions; and mixed use of CCIDs is not currently well | |
71 | understood. This socket option takes as argument at least one uint8_t value, or | |
72 | an array of uint8_t values, which must match available CCIDS (see above). CCIDs | |
73 | must be registered on the socket before calling connect() or listen(). | |
74 | ||
75 | DCCP_SOCKOPT_TX_CCID is read/write. It returns the current CCID (if set) or sets | |
76 | the preference list for the TX CCID, using the same format as DCCP_SOCKOPT_CCID. | |
77 | Please note that the getsockopt argument type here is `int', not uint8_t. | |
78 | ||
79 | DCCP_SOCKOPT_RX_CCID is analogous to DCCP_SOCKOPT_TX_CCID, but for the RX CCID. | |
80 | ||
b8599d20 GR |
81 | DCCP_SOCKOPT_SERVER_TIMEWAIT enables the server (listening socket) to hold |
82 | timewait state when closing the connection (RFC 4340, 8.3). The usual case is | |
83 | that the closing server sends a CloseReq, whereupon the client holds timewait | |
84 | state. When this boolean socket option is on, the server sends a Close instead | |
85 | and will enter TIMEWAIT. This option must be set after accept() returns. | |
86 | ||
6f4e5fff GR |
87 | DCCP_SOCKOPT_SEND_CSCOV and DCCP_SOCKOPT_RECV_CSCOV are used for setting the |
88 | partial checksum coverage (RFC 4340, sec. 9.2). The default is that checksums | |
89 | always cover the entire packet and that only fully covered application data is | |
90 | accepted by the receiver. Hence, when using this feature on the sender, it must | |
91 | be enabled at the receiver, too with suitable choice of CsCov. | |
92 | ||
93 | DCCP_SOCKOPT_SEND_CSCOV sets the sender checksum coverage. Values in the | |
94 | range 0..15 are acceptable. The default setting is 0 (full coverage), | |
95 | values between 1..15 indicate partial coverage. | |
2bfd754d | 96 | DCCP_SOCKOPT_RECV_CSCOV is for the receiver and has a different meaning: it |
6f4e5fff GR |
97 | sets a threshold, where again values 0..15 are acceptable. The default |
98 | of 0 means that all packets with a partial coverage will be discarded. | |
99 | Values in the range 1..15 indicate that packets with minimally such a | |
100 | coverage value are also acceptable. The higher the number, the more | |
2bfd754d GR |
101 | restrictive this setting (see [RFC 4340, sec. 9.2.1]). Partial coverage |
102 | settings are inherited to the child socket after accept(). | |
6f4e5fff | 103 | |
f2645101 GR |
104 | The following two options apply to CCID 3 exclusively and are getsockopt()-only. |
105 | In either case, a TFRC info struct (defined in <linux/tfrc.h>) is returned. | |
106 | DCCP_SOCKOPT_CCID_RX_INFO | |
107 | Returns a `struct tfrc_rx_info' in optval; the buffer for optval and | |
108 | optlen must be set to at least sizeof(struct tfrc_rx_info). | |
109 | DCCP_SOCKOPT_CCID_TX_INFO | |
110 | Returns a `struct tfrc_tx_info' in optval; the buffer for optval and | |
111 | optlen must be set to at least sizeof(struct tfrc_tx_info). | |
112 | ||
8e8c71f1 GR |
113 | On unidirectional connections it is useful to close the unused half-connection |
114 | via shutdown (SHUT_WR or SHUT_RD): this will reduce per-packet processing costs. | |
f2645101 | 115 | |
4886fcad | 116 | |
2e2e9e92 GR |
117 | Sysctl variables |
118 | ================ | |
119 | Several DCCP default parameters can be managed by the following sysctls | |
120 | (sysctl net.dccp.default or /proc/sys/net/dccp/default): | |
121 | ||
122 | request_retries | |
123 | The number of active connection initiation retries (the number of | |
124 | Requests minus one) before timing out. In addition, it also governs | |
125 | the behaviour of the other, passive side: this variable also sets | |
126 | the number of times DCCP repeats sending a Response when the initial | |
127 | handshake does not progress from RESPOND to OPEN (i.e. when no Ack | |
128 | is received after the initial Request). This value should be greater | |
129 | than 0, suggested is less than 10. Analogue of tcp_syn_retries. | |
130 | ||
131 | retries1 | |
132 | How often a DCCP Response is retransmitted until the listening DCCP | |
133 | side considers its connecting peer dead. Analogue of tcp_retries1. | |
134 | ||
135 | retries2 | |
136 | The number of times a general DCCP packet is retransmitted. This has | |
137 | importance for retransmitted acknowledgments and feature negotiation, | |
138 | data packets are never retransmitted. Analogue of tcp_retries2. | |
139 | ||
2e2e9e92 | 140 | tx_ccid = 2 |
0049bab5 GR |
141 | Default CCID for the sender-receiver half-connection. Depending on the |
142 | choice of CCID, the Send Ack Vector feature is enabled automatically. | |
2e2e9e92 GR |
143 | |
144 | rx_ccid = 2 | |
0049bab5 | 145 | Default CCID for the receiver-sender half-connection; see tx_ccid. |
2e2e9e92 GR |
146 | |
147 | seq_window = 100 | |
792b4878 GR |
148 | The initial sequence window (sec. 7.5.2) of the sender. This influences |
149 | the local ackno validity and the remote seqno validity windows (7.5.1). | |
2e2e9e92 | 150 | |
82e3ab9d IM |
151 | tx_qlen = 5 |
152 | The size of the transmit buffer in packets. A value of 0 corresponds | |
153 | to an unbounded transmit buffer. | |
154 | ||
a94f0f97 GR |
155 | sync_ratelimit = 125 ms |
156 | The timeout between subsequent DCCP-Sync packets sent in response to | |
157 | sequence-invalid packets on the same socket (RFC 4340, 7.5.4). The unit | |
158 | of this parameter is milliseconds; a value of 0 disables rate-limiting. | |
159 | ||
4886fcad | 160 | |
c2814901 GR |
161 | IOCTLS |
162 | ====== | |
163 | FIONREAD | |
164 | Works as in udp(7): returns in the `int' argument pointer the size of | |
165 | the next pending datagram in bytes, or 0 when no datagram is pending. | |
166 | ||
4886fcad GR |
167 | |
168 | Other tunables | |
169 | ============== | |
170 | Per-route rto_min support | |
171 | CCID-2 supports the RTAX_RTO_MIN per-route setting for the minimum value | |
172 | of the RTO timer. This setting can be modified via the 'rto_min' option | |
173 | of iproute2; for example: | |
174 | > ip route change 10.0.0.0/24 rto_min 250j dev wlan0 | |
175 | > ip route add 10.0.0.254/32 rto_min 800j dev wlan0 | |
176 | > ip route show dev wlan0 | |
89858ad1 GR |
177 | CCID-3 also supports the rto_min setting: it is used to define the lower |
178 | bound for the expiry of the nofeedback timer. This can be useful on LANs | |
179 | with very low RTTs (e.g., loopback, Gbit ethernet). | |
4886fcad GR |
180 | |
181 | ||
98069ff4 IM |
182 | Notes |
183 | ===== | |
ddfe10b8 | 184 | DCCP does not travel through NAT successfully at present on many boxes. This is |
126acd5b | 185 | because the checksum covers the pseudo-header as per TCP and UDP. Linux NAT |
ddfe10b8 | 186 | support for DCCP has been added. |