tcp: refresh skb timestamp at retransmit time
authorEric Dumazet <edumazet@google.com>
Tue, 10 May 2016 03:55:16 +0000 (20:55 -0700)
committerDavid S. Miller <davem@davemloft.net>
Tue, 10 May 2016 19:58:41 +0000 (15:58 -0400)
In the very unlikely case __tcp_retransmit_skb() can not use the cloning
done in tcp_transmit_skb(), we need to refresh skb_mstamp before doing
the copy and transmit, otherwise TCP TS val will be an exact copy of
original transmit.

Fixes: 7faee5c0d514 ("tcp: remove TCP_SKB_CB(skb)->when")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Acked-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp_output.c

index 441ae9da3a233fd96385853197fd3938959449e3..79a03b87a7714b1ade63aa78ce80101a182e2597 100644 (file)
@@ -2640,8 +2640,10 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_buff *skb)
         */
        if (unlikely((NET_IP_ALIGN && ((unsigned long)skb->data & 3)) ||
                     skb_headroom(skb) >= 0xFFFF)) {
-               struct sk_buff *nskb = __pskb_copy(skb, MAX_TCP_HEADER,
-                                                  GFP_ATOMIC);
+               struct sk_buff *nskb;
+
+               skb_mstamp_get(&skb->skb_mstamp);
+               nskb = __pskb_copy(skb, MAX_TCP_HEADER, GFP_ATOMIC);
                err = nskb ? tcp_transmit_skb(sk, nskb, 0, GFP_ATOMIC) :
                             -ENOBUFS;
        } else {