[tcp] Record ts_recent on first received packet

Commit 6861304 ("[tcp] Handle out-of-order received packets")
introduced a regression in which ts_recent would not be updated until
the first packet is received in the ESTABLISHED state, i.e. the
timestamp from the SYN+ACK packet would be ignored.  This causes the
connection to be dropped by strictly-conforming TCP peers, such as
FreeBSD.

Fix by delaying the timestamp window check until after processing the
received SYN flag.

Reported-by: winders@sonnet.com
Signed-off-by: Michael Brown <mcb30@ipxe.org>

src/net/tcp.c

 	struct tcp_options options;
 	size_t hlen;
 	uint16_t csum;
+	uint32_t start_seq;
 	uint32_t seq;
 	uint32_t ack;
 	uint32_t win;
 	
 	/* Parse parameters from header and strip header */
 	tcp = tcp_demux ( ntohs ( tcphdr->dest ) );
-	seq = ntohl ( tcphdr->seq );
+	seq = start_seq = ntohl ( tcphdr->seq );
 	ack = ntohl ( tcphdr->ack );
 	win = ntohs ( tcphdr->win );
 	flags = tcphdr->flags;
 		goto discard;
 	}
 
-	/* Update timestamp, if applicable */
-	if ( options.tsopt && tcp_in_window ( tcp->rcv_ack, seq, seq_len ) )
-		tcp->ts_recent = ntohl ( options.tsopt->tsval );
-
 	/* Handle ACK, if present */
 	if ( flags & TCP_ACK ) {
 		if ( ( rc = tcp_rx_ack ( tcp, ack, win ) ) != 0 ) {
 			goto discard;
 	}
 
+	/* Update timestamp, if applicable */
+	if ( options.tsopt &&
+	     tcp_in_window ( tcp->rcv_ack, start_seq, seq_len ) ) {
+		tcp->ts_recent = ntohl ( options.tsopt->tsval );
+	}
+
 	/* Enqueue received data */
 	tcp_rx_enqueue ( tcp, seq, flags, iob_disown ( iobuf ) );