dccp ccid-3: use per-route RTO or TCP RTO as fallback

This makes RTAX_RTO_MIN also available to CCID-3, replacing the compile-time
RTO lower bound with a per-route tunable value.

The original Kconfig option solved the problem that a very low RTT (in the
order of HZ) can trigger too frequent and unnecessary reductions of the
sending rate.

This tunable does not affect the initial RTO value of 2 seconds specified in
RFC 5348, section 4.2 and Appendix B. But like the hardcoded Kconfig value,
it allows to adapt to network conditions.

The same effect as the original Kconfig option of 100ms is now achieved by

> ip route replace to unicast 192.168.0.0/24 rto_min 100j dev eth0

(assuming HZ=1000).

Signed-off-by: Gerrit Renker <gerrit@erg.abdn.ac.uk>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Gerrit Renker 2010-08-29 19:23:14 +00:00 committed by David S. Miller
parent 4886fcad6e
commit 89858ad143
4 changed files with 9 additions and 38 deletions

View file

@ -174,6 +174,9 @@ Per-route rto_min support
> ip route change 10.0.0.0/24 rto_min 250j dev wlan0
> ip route add 10.0.0.254/32 rto_min 800j dev wlan0
> ip route show dev wlan0
CCID-3 also supports the rto_min setting: it is used to define the lower
bound for the expiry of the nofeedback timer. This can be useful on LANs
with very low RTTs (e.g., loopback, Gbit ethernet).
Notes

View file

@ -47,37 +47,6 @@ config IP_DCCP_CCID3_DEBUG
If in doubt, say N.
config IP_DCCP_CCID3_RTO
int "Use higher bound for nofeedback timer"
default 100
depends on IP_DCCP_CCID3 && EXPERIMENTAL
---help---
Use higher lower bound for nofeedback timer expiration.
The TFRC nofeedback timer normally expires after the maximum of 4
RTTs and twice the current send interval (RFC 3448, 4.3). On LANs
with a small RTT this can mean a high processing load and reduced
performance, since then the nofeedback timer is triggered very
frequently.
This option enables to set a higher lower bound for the nofeedback
value. Values in units of milliseconds can be set here.
A value of 0 disables this feature by enforcing the value specified
in RFC 3448. The following values have been suggested as bounds for
experimental use:
* 16-20ms to match the typical multimedia inter-frame interval
* 100ms as a reasonable compromise [default]
* 1000ms corresponds to the lower TCP RTO bound (RFC 2988, 2.4)
The default of 100ms is a compromise between a large value for
efficient DCCP implementations, and a small value to avoid disrupting
the network in times of congestion.
The purpose of the nofeedback timer is to slow DCCP down when there
is serious network congestion: experimenting with larger values should
therefore not be performed on WANs.
config IP_DCCP_TFRC_LIB
def_bool y if IP_DCCP_CCID3

View file

@ -460,13 +460,12 @@ static void ccid3_hc_tx_packet_recv(struct sock *sk, struct sk_buff *skb)
sk->sk_write_space(sk);
/*
* Update timeout interval for the nofeedback timer.
* We use a configuration option to increase the lower bound.
* This can help avoid triggering the nofeedback timer too
* often ('spinning') on LANs with small RTTs.
* Update timeout interval for the nofeedback timer. In order to control
* rate halving on networks with very low RTTs (<= 1 ms), use per-route
* tunable RTAX_RTO_MIN value as the lower bound.
*/
hc->tx_t_rto = max_t(u32, 4 * hc->tx_rtt, (CONFIG_IP_DCCP_CCID3_RTO *
(USEC_PER_SEC / 1000)));
hc->tx_t_rto = max_t(u32, 4 * hc->tx_rtt,
USEC_PER_SEC/HZ * tcp_rto_min(sk));
/*
* Schedule no feedback timer to expire in
* max(t_RTO, 2 * s/X) = max(t_RTO, 2 * t_ipi)

View file

@ -42,7 +42,7 @@
#include "lib/tfrc.h"
#include "../ccid.h"
/* Two seconds as per RFC 3448 4.2 */
/* Two seconds as per RFC 5348, 4.2 */
#define TFRC_INITIAL_TIMEOUT (2 * USEC_PER_SEC)
/* In usecs - half the scheduling granularity as per RFC3448 4.6 */