[SK_BUFF]: Convert skb->tail to sk_buff_data_t

So that it is also an offset from skb->head, reduces its size from 8 to 4 bytes
on 64bit architectures, allowing us to combine the 4 bytes hole left by the
layer headers conversion, reducing struct sk_buff size to 256 bytes, i.e. 4
64byte cachelines, and since the sk_buff slab cache is SLAB_HWCACHE_ALIGN...
:-)

Many calculations that previously required that skb->{transport,network,
mac}_header be first converted to a pointer now can be done directly, being
meaningful as offsets or pointers.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv4/ipvs/ip_vs_ftp.c

@@ -162,7 +162,7 @@ static int ip_vs_ftp_out(struct ip_vs_app *app, struct ip_vs_conn *cp,
 		iph = ip_hdr(*pskb);
 		th = (struct tcphdr *)&(((char *)iph)[iph->ihl*4]);
 		data = (char *)th + (th->doff << 2);
-		data_limit = (*pskb)->tail;
+		data_limit = skb_tail_pointer(*pskb);
 
 		if (ip_vs_ftp_get_addrport(data, data_limit,
 					   SERVER_STRING,
@@ -269,7 +269,7 @@ static int ip_vs_ftp_in(struct ip_vs_app *app, struct ip_vs_conn *cp,
 	   the length of the header in 32-bit multiples, it is accurate
 	   to calculate data address by th+HLEN*4 */
 	data = data_start = (char *)th + (th->doff << 2);
-	data_limit = (*pskb)->tail;
+	data_limit = skb_tail_pointer(*pskb);
 
 	while (data <= data_limit - 6) {
 		if (strnicmp(data, "PASV\r\n", 6) == 0) {