[TCP]: Prevent sending past receiver window with TSO (at last skb)

With TSO it was possible to send past the receiver window when the skb
to be sent was the last in the write queue while the receiver window
is the limiting factor. One can notice that there's a loophole in the
tcp_mss_split_point that lacked a receiver window check for the
tcp_write_queue_tail() if also cwnd was smaller than the full skb.

Noticed by Thomas Gleixner <[e-mail]> in form of "Treason
uncloaked! Peer ... shrinks window .... Repaired." messages (the peer
didn't actually shrink its window as the message suggests, we had just
sent something past it without a permission to do so).

Signed-off-by: Ilpo Järvinen <[e-mail]>
Tested-by: Thomas Gleixner <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

net/ipv4/tcp_output.c

···
  * introducing MSS oddities to segment boundaries. In rare cases where
  * mss_now != mss_cache, we will request caller to create a small skb
  * per input skb which could be mostly avoided here (if desired).
+ *
+ * We explicitly want to create a request for splitting write queue tail
+ * to a small skb for Nagle purposes while avoiding unnecessary modulos,
+ * thus all the complexity (cwnd_len is always MSS multiple which we
+ * return whenever allowed by the other factors). Basically we need the
+ * modulo only when the receiver window alone is the limiting factor or
+ * when we would be allowed to send the split-due-to-Nagle skb fully.
  */
 static unsigned int tcp_mss_split_point(struct sock *sk, struct sk_buff *skb,
 					unsigned int mss_now, unsigned int cwnd)
···
 	if (likely(cwnd_len <= window && skb != tcp_write_queue_tail(sk)))
 		return cwnd_len;
 
-	if (skb == tcp_write_queue_tail(sk) && cwnd_len <= skb->len)
+	needed = min(skb->len, window);
+
+	if (skb == tcp_write_queue_tail(sk) && cwnd_len <= needed)
 		return cwnd_len;
 
-	needed = min(skb->len, window);
 	return needed - needed % mss_now;
 }