aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorGerrit Renker <gerrit@erg.abdn.ac.uk>2010-08-29 15:23:13 -0400
committerDavid S. Miller <davem@davemloft.net>2010-08-30 16:45:27 -0400
commit4886fcad6e12572afbd230dfab1b268eace20d6d (patch)
treec0ac7dc930a8de315083ec3882ff78a325747e70
parent22b71c8f4f3db8df92f5e7b081c265bc56c0bd2f (diff)
dccp ccid-2: Share TCP's minimum RTO code
Using a fixed RTO_MIN of 0.2 seconds was found to cause problems for CCID-2 over 802.11g: at least once per session there was a spurious timeout. It helped to then increase the the value of RTO_MIN over this link. Since the problem is the same as in TCP, this patch makes the solution from commit "05bb1fad1cde025a864a90cfeb98dcbefe78a44a" "[TCP]: Allow minimum RTO to be configurable via routing metrics." available to DCCP. This avoids reinventing the wheel, so that e.g. the following works in the expected way now also for CCID-2: > ip route change 10.0.0.2 rto_min 800 dev ath0 Luckily this useful rto_min function was recently moved to net/tcp.h, which simplifies sharing code originating from TCP. Documentation also updated (plus minor whitespace fixes). Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
-rw-r--r--Documentation/networking/dccp.txt26
-rw-r--r--net/dccp/ccids/ccid2.c5
2 files changed, 23 insertions, 8 deletions
diff --git a/Documentation/networking/dccp.txt b/Documentation/networking/dccp.txt
index a62fdf7a6bf..cdb64922ba1 100644
--- a/Documentation/networking/dccp.txt
+++ b/Documentation/networking/dccp.txt
@@ -1,18 +1,20 @@
1DCCP protocol 1DCCP protocol
2============ 2=============
3 3
4 4
5Contents 5Contents
6======== 6========
7
8- Introduction 7- Introduction
9- Missing features 8- Missing features
10- Socket options 9- Socket options
10- Sysctl variables
11- IOCTLs
12- Other tunables
11- Notes 13- Notes
12 14
15
13Introduction 16Introduction
14============ 17============
15
16Datagram Congestion Control Protocol (DCCP) is an unreliable, connection 18Datagram Congestion Control Protocol (DCCP) is an unreliable, connection
17oriented protocol designed to solve issues present in UDP and TCP, particularly 19oriented protocol designed to solve issues present in UDP and TCP, particularly
18for real-time and multimedia (streaming) traffic. 20for real-time and multimedia (streaming) traffic.
@@ -29,9 +31,9 @@ It has a base protocol and pluggable congestion control IDs (CCIDs).
29DCCP is a Proposed Standard (RFC 2026), and the homepage for DCCP as a protocol 31DCCP is a Proposed Standard (RFC 2026), and the homepage for DCCP as a protocol
30is at http://www.ietf.org/html.charters/dccp-charter.html 32is at http://www.ietf.org/html.charters/dccp-charter.html
31 33
34
32Missing features 35Missing features
33================ 36================
34
35The Linux DCCP implementation does not currently support all the features that are 37The Linux DCCP implementation does not currently support all the features that are
36specified in RFCs 4340...42. 38specified in RFCs 4340...42.
37 39
@@ -45,7 +47,6 @@ http://linux-net.osdl.org/index.php/DCCP_Testing#Experimental_DCCP_source_tree
45 47
46Socket options 48Socket options
47============== 49==============
48
49DCCP_SOCKOPT_SERVICE sets the service. The specification mandates use of 50DCCP_SOCKOPT_SERVICE sets the service. The specification mandates use of
50service codes (RFC 4340, sec. 8.1.2); if this socket option is not set, 51service codes (RFC 4340, sec. 8.1.2); if this socket option is not set,
51the socket will fall back to 0 (which means that no meaningful service code 52the socket will fall back to 0 (which means that no meaningful service code
@@ -112,6 +113,7 @@ DCCP_SOCKOPT_CCID_TX_INFO
112On unidirectional connections it is useful to close the unused half-connection 113On unidirectional connections it is useful to close the unused half-connection
113via shutdown (SHUT_WR or SHUT_RD): this will reduce per-packet processing costs. 114via shutdown (SHUT_WR or SHUT_RD): this will reduce per-packet processing costs.
114 115
116
115Sysctl variables 117Sysctl variables
116================ 118================
117Several DCCP default parameters can be managed by the following sysctls 119Several DCCP default parameters can be managed by the following sysctls
@@ -155,15 +157,27 @@ sync_ratelimit = 125 ms
155 sequence-invalid packets on the same socket (RFC 4340, 7.5.4). The unit 157 sequence-invalid packets on the same socket (RFC 4340, 7.5.4). The unit
156 of this parameter is milliseconds; a value of 0 disables rate-limiting. 158 of this parameter is milliseconds; a value of 0 disables rate-limiting.
157 159
160
158IOCTLS 161IOCTLS
159====== 162======
160FIONREAD 163FIONREAD
161 Works as in udp(7): returns in the `int' argument pointer the size of 164 Works as in udp(7): returns in the `int' argument pointer the size of
162 the next pending datagram in bytes, or 0 when no datagram is pending. 165 the next pending datagram in bytes, or 0 when no datagram is pending.
163 166
167
168Other tunables
169==============
170Per-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
177
178
164Notes 179Notes
165===== 180=====
166
167DCCP does not travel through NAT successfully at present on many boxes. This is 181DCCP does not travel through NAT successfully at present on many boxes. This is
168because the checksum covers the pseudo-header as per TCP and UDP. Linux NAT 182because the checksum covers the pseudo-header as per TCP and UDP. Linux NAT
169support for DCCP has been added. 183support for DCCP has been added.
diff --git a/net/dccp/ccids/ccid2.c b/net/dccp/ccids/ccid2.c
index b9c942a09c9..dc18172b1e5 100644
--- a/net/dccp/ccids/ccid2.c
+++ b/net/dccp/ccids/ccid2.c
@@ -325,8 +325,9 @@ static void ccid2_rtt_estimator(struct sock *sk, const long mrtt)
325 hc->tx_srtt = m << 3; 325 hc->tx_srtt = m << 3;
326 hc->tx_mdev = m << 1; 326 hc->tx_mdev = m << 1;
327 327
328 hc->tx_mdev_max = max(TCP_RTO_MIN, hc->tx_mdev); 328 hc->tx_mdev_max = max(hc->tx_mdev, tcp_rto_min(sk));
329 hc->tx_rttvar = hc->tx_mdev_max; 329 hc->tx_rttvar = hc->tx_mdev_max;
330
330 hc->tx_rtt_seq = dccp_sk(sk)->dccps_gss; 331 hc->tx_rtt_seq = dccp_sk(sk)->dccps_gss;
331 } else { 332 } else {
332 /* Update scaled SRTT as SRTT += 1/8 * (m - SRTT) */ 333 /* Update scaled SRTT as SRTT += 1/8 * (m - SRTT) */
@@ -367,7 +368,7 @@ static void ccid2_rtt_estimator(struct sock *sk, const long mrtt)
367 hc->tx_rttvar -= (hc->tx_rttvar - 368 hc->tx_rttvar -= (hc->tx_rttvar -
368 hc->tx_mdev_max) >> 2; 369 hc->tx_mdev_max) >> 2;
369 hc->tx_rtt_seq = dccp_sk(sk)->dccps_gss; 370 hc->tx_rtt_seq = dccp_sk(sk)->dccps_gss;
370 hc->tx_mdev_max = TCP_RTO_MIN; 371 hc->tx_mdev_max = tcp_rto_min(sk);
371 } 372 }
372 } 373 }
373 374