tcp: evaluate packet losses upon RTT change
authorYuchung Cheng <ycheng@google.com>
Thu, 7 Dec 2017 19:33:33 +0000 (11:33 -0800)
committerDavid S. Miller <davem@davemloft.net>
Fri, 8 Dec 2017 19:14:11 +0000 (14:14 -0500)
commit6065fd0d179b96ddc488c76542349bcb148a95fd
tree39472b7c1cc30a0ce316278fc8854101245b97b7
parent428aec5e69fa17d223e1495f395833c50770f7ae
tcp: evaluate packet losses upon RTT change

RACK skips an ACK unless it advances the most recently delivered
TX timestamp (rack.mstamp). Since RACK also uses the most recent
RTT to decide if a packet is lost, RACK should still run the
loss detection whenever the most recent RTT changes. For example,
an ACK that does not advance the timestamp but triggers the cwnd
undo due to reordering, would then use the most recent (higher)
RTT measurement to detect further losses.

Signed-off-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Neal Cardwell <ncardwell@google.com>
Reviewed-by: Priyaranjan Jha <priyarjha@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp_recovery.c