[TCP]: Fix tcp_tso_should_defer() when limit>=65536



That's >= a full sized TSO frame, so we should always
return 0 in that case.

Based upon a report and initial patch from Lachlan
Andrew, final patch suggested by Herbert Xu.

Signed-off-by: David S. Miller <davem@davemloft.net>

parent c1274376

net/ipv4/tcp_output.c

+4 −0

Original line number	Diff line number	Diff line
		@@ -1036,6 +1036,10 @@ static int tcp_tso_should_defer(struct sock sk, struct tcp_sock tp, struct sk_

		limit = min(send_win, cong_win);

		/* If a full-sized TSO skb can be sent, do it. */
		if (limit >= 65536)
		return 0;

		if (sysctl_tcp_tso_win_divisor) {
		u32 chunk = min(tp->snd_wnd, tp->snd_cwnd * tp->mss_cache);