tcp: assign skb hash after tcp_event_data_sent
authorYuchung Cheng <ycheng@google.com>
Mon, 11 Jan 2021 23:05:52 +0000 (15:05 -0800)
committerJakub Kicinski <kuba@kernel.org>
Thu, 14 Jan 2021 03:36:51 +0000 (19:36 -0800)
Move skb_set_hash_from_sk s.t. it's called after instead of before
tcp_event_data_sent is called. This enables congestion control
modules to change the socket hash right before restarting from
idle (via the TX_START congestion event).

Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Link: https://lore.kernel.org/r/20210111230552.2704579-1-ycheng@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
net/ipv4/tcp_output.c

index f322e79..899d053 100644 (file)
@@ -1319,7 +1319,6 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
        skb_orphan(skb);
        skb->sk = sk;
        skb->destructor = skb_is_tcp_pure_ack(skb) ? __sock_wfree : tcp_wfree;
-       skb_set_hash_from_sk(skb, sk);
        refcount_add(skb->truesize, &sk->sk_wmem_alloc);
 
        skb_set_dst_pending_confirm(skb, sk->sk_dst_pending_confirm);
@@ -1390,6 +1389,7 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb,
                              tcp_skb_pcount(skb));
 
        tp->segs_out += tcp_skb_pcount(skb);
+       skb_set_hash_from_sk(skb, sk);
        /* OK, its time to fill skb_shinfo(skb)->gso_{segs|size} */
        skb_shinfo(skb)->gso_segs = tcp_skb_pcount(skb);
        skb_shinfo(skb)->gso_size = tcp_skb_mss(skb);