Enable FACK congestion avoidance

$cat /proc/sys/net/ipv4/tcp_fack
1
$
tcp_fack
Enable FACK congestion avoidance and fast retransmission.
The value is not used, if tcp_sack is not enabled.

http://www.psc.edu/networking/ftp/papers/Fack.9608.ps

grepping through the linux kernel source 2.6.32

proc.c:210:     SNMP_MIB_ITEM("TCPFACKReorder", LINUX_MIB_TCPFACKREORDER),
sysctl_net_ipv4.c:475:          .procname       = "tcp_fack",
tcp_input.c:847: * Packet counting of FACK is based on in-order assumptions, therefore TCP
tcp_input.c:1029: *        A''. Its FACK modfication, head until snd.fack is lost.
tcp_minisocks.c:412:            newtp->fackets_out = 0;
tcp_vegas.c:16: *     using fine-grained timers, NewReno, and FACK.
tcp_output.c:762:/* When a modification to fackets out becomes necessary, we need to check
tcp_output.c:763: * skb is counted to fackets_out or not.