tcp: limit GSO packets to half cwnd

In DC world, GSO packets initially cooked by tcp_sendmsg() are usually
big, as sk_pacing_rate is high.

When network is congested, cwnd can be smaller than the GSO packets
found in socket write queue. tcp_write_xmit() splits GSO packets
using the available cwnd, and we end up sending a single GSO packet,
consuming all available cwnd.

With GRO aggregation on the receiver, we might handle a single GRO
packet, sending back a single ACK.

1) This single ACK might be lost
   TLP or RTO are forced to attempt a retransmit.
2) This ACK releases a full cwnd, sender sends another big GSO packet,
   in a ping pong mode.

This behavior does not fill the pipes in the best way, because of
scheduling artifacts.

Make sure we always have at least two GSO packets in flight.

This allows us to safely increase GRO efficiency without risking
spurious retransmits.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 0b88158..eb73a1d 100644 (file)
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1562,7 +1562,7 @@ static unsigned int tcp_mss_split_point(const struct sock *sk,
 static inline unsigned int tcp_cwnd_test(const struct tcp_sock *tp,
                                         const struct sk_buff *skb)
 {
-       u32 in_flight, cwnd;
+       u32 in_flight, cwnd, halfcwnd;
 
        /* Don't be strict about the congestion window for the final FIN.  */
        if ((TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN) &&
@@ -1571,10 +1571,14 @@ static inline unsigned int tcp_cwnd_test(const struct tcp_sock *tp,
 
        in_flight = tcp_packets_in_flight(tp);
        cwnd = tp->snd_cwnd;
-       if (in_flight < cwnd)
-               return (cwnd - in_flight);
+       if (in_flight >= cwnd)
+               return 0;
 
-       return 0;
+       /* For better scheduling, ensure we have at least
+        * 2 GSO packets in flight.
+        */
+       halfcwnd = max(cwnd >> 1, 1U);
+       return min(halfcwnd, cwnd - in_flight);
 }
 
 /* Initialize TSO state of a skb.