tcp: fix FIONREAD/SIOCINQ
authorEric Dumazet <edumazet@google.com>
Thu, 18 Oct 2012 09:14:12 +0000 (09:14 +0000)
committerDavid S. Miller <davem@davemloft.net>
Thu, 18 Oct 2012 19:34:31 +0000 (15:34 -0400)
tcp_ioctl() tries to take into account if tcp socket received a FIN
to report correct number bytes in receive queue.

But its flaky because if the application ate the last skb,
we return 1 instead of 0.

Correct way to detect that FIN was received is to test SOCK_DONE.

Reported-by: Elliot Hughes <enh@google.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv4/tcp.c

index f32c02e2a54346cf4e120d39e17d4d1b5e966189..b7c2f439b54f92a9f9b036dccbc56206c4a6d53c 100644 (file)
@@ -549,14 +549,12 @@ int tcp_ioctl(struct sock *sk, int cmd, unsigned long arg)
                         !tp->urg_data ||
                         before(tp->urg_seq, tp->copied_seq) ||
                         !before(tp->urg_seq, tp->rcv_nxt)) {
-                       struct sk_buff *skb;
 
                        answ = tp->rcv_nxt - tp->copied_seq;
 
-                       /* Subtract 1, if FIN is in queue. */
-                       skb = skb_peek_tail(&sk->sk_receive_queue);
-                       if (answ && skb)
-                               answ -= tcp_hdr(skb)->fin;
+                       /* Subtract 1, if FIN was received */
+                       if (answ && sock_flag(sk, SOCK_DONE))
+                               answ--;
                } else
                        answ = tp->urg_seq - tp->copied_seq;
                release_sock(sk);