[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>

11 files changed, 27 insertions, 18 deletions

diff --git a/drivers/net/cris/eth_v10.c b/drivers/net/cris/eth_v10.c
index 98643801a3b0..7feb9c561147 100644
--- a/drivers/net/cris/eth_v10.c
+++ b/drivers/net/cris/eth_v10.c
@@ -1348,7 +1348,7 @@ e100_rx(struct net_device *dev)
 
 #ifdef ETHDEBUG
 		printk("head = 0x%x, data = 0x%x, tail = 0x%x, end = 0x%x\n",
-		  skb->head, skb->data, skb->tail, skb->end);
+		  skb->head, skb->data, skb_tail_pointer(skb), skb->end);
 		printk("copying packet to 0x%x.\n", skb_data_ptr);
 #endif
 
diff --git a/drivers/net/cxgb3/sge.c b/drivers/net/cxgb3/sge.c
index 610e4769efa4..c5faf1380e15 100644
--- a/drivers/net/cxgb3/sge.c
+++ b/drivers/net/cxgb3/sge.c
@@ -1325,13 +1325,13 @@ static void write_ofld_wr(struct adapter *adap, struct sk_buff *skb,
 	flits = skb_transport_offset(skb) / 8;
 	sgp = ndesc == 1 ? (struct sg_ent *)&d->flit[flits] : sgl;
 	sgl_flits = make_sgl(skb, sgp, skb_transport_header(skb),
-			     skb->tail - skb_transport_header(skb),
+			     skb->tail - skb->transport_header,
 			     adap->pdev);
 	if (need_skb_unmap()) {
 		setup_deferred_unmapping(skb, adap->pdev, sgp, sgl_flits);
 		skb->destructor = deferred_unmap_destructor;
 		((struct unmap_info *)skb->cb)->len = (skb->tail -
-						       skb_transport_header(skb));
+						       skb->transport_header);
 	}
 
 	write_wr_hdr_sgl(ndesc, skb, d, pidx, q, sgl, flits, sgl_flits,
@@ -1353,7 +1353,7 @@ static inline unsigned int calc_tx_descs_ofld(const struct sk_buff *skb)
 		return 1;	/* packet fits as immediate data */
 
 	flits = skb_transport_offset(skb) / 8;	/* headers */
-	if (skb->tail != skb_transport_header(skb))
+	if (skb->tail != skb->transport_header)
 		cnt++;
 	return flits_to_desc(flits + sgl_len(cnt));
 }
diff --git a/drivers/net/e1000/e1000_main.c b/drivers/net/e1000/e1000_main.c
index e86deb2ef823..e7c93f44f810 100644
--- a/drivers/net/e1000/e1000_main.c
+++ b/drivers/net/e1000/e1000_main.c
@@ -3304,7 +3304,7 @@ e1000_xmit_frame(struct sk_buff *skb, struct net_device *netdev)
 				 * NOTE: this is a TSO only workaround
 				 * if end byte alignment not correct move us
 				 * into the next dword */
-				if ((unsigned long)(skb->tail - 1) & 4)
+				if ((unsigned long)(skb_tail_pointer(skb) - 1) & 4)
 					break;
 				/* fall through */
 			case e1000_82571:
@@ -4388,7 +4388,7 @@ e1000_clean_rx_irq_ps(struct e1000_adapter *adapter,
 				PCI_DMA_FROMDEVICE);
 			vaddr = kmap_atomic(ps_page->ps_page[0],
 			                    KM_SKB_DATA_SOFTIRQ);
-			memcpy(skb->tail, vaddr, l1);
+			memcpy(skb_tail_pointer(skb), vaddr, l1);
 			kunmap_atomic(vaddr, KM_SKB_DATA_SOFTIRQ);
 			pci_dma_sync_single_for_device(pdev,
 				ps_page_dma->ps_page_dma[0],
diff --git a/drivers/net/ibm_emac/ibm_emac_core.c b/drivers/net/ibm_emac/ibm_emac_core.c
index b1ad62d89eb9..3d82d46f4998 100644
--- a/drivers/net/ibm_emac/ibm_emac_core.c
+++ b/drivers/net/ibm_emac/ibm_emac_core.c
@@ -1338,7 +1338,7 @@ static inline int emac_rx_sg_append(struct ocp_enet_private *dev, int slot)
 			dev_kfree_skb(dev->rx_sg_skb);
 			dev->rx_sg_skb = NULL;
 		} else {
-			cacheable_memcpy(dev->rx_sg_skb->tail,
+			cacheable_memcpy(skb_tail_pointer(dev->rx_sg_skb),
 					 dev->rx_skb[slot]->data, len);
 			skb_put(dev->rx_sg_skb, len);
 			emac_recycle_rx_skb(dev, slot, len);
diff --git a/drivers/net/macb.c b/drivers/net/macb.c
index 0c3649be0d05..98bf51afcee7 100644
--- a/drivers/net/macb.c
+++ b/drivers/net/macb.c
@@ -575,7 +575,8 @@ static int macb_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	int i;
 	dev_dbg(&bp->pdev->dev,
 		"start_xmit: len %u head %p data %p tail %p end %p\n",
-		skb->len, skb->head, skb->data, skb->tail, skb->end);
+		skb->len, skb->head, skb->data,
+		skb_tail_pointer(skb), skb->end);
 	dev_dbg(&bp->pdev->dev,
 		"data:");
 	for (i = 0; i < 16; i++)
diff --git a/drivers/net/pcmcia/nmclan_cs.c b/drivers/net/pcmcia/nmclan_cs.c
index ec0af65cd5d7..73da611fd536 100644
--- a/drivers/net/pcmcia/nmclan_cs.c
+++ b/drivers/net/pcmcia/nmclan_cs.c
@@ -1185,7 +1185,7 @@ static int mace_rx(struct net_device *dev, unsigned char RxCnt)
 	skb_reserve(skb, 2);
 	insw(ioaddr + AM2150_RCV, skb_put(skb, pkt_len), pkt_len>>1);
 	if (pkt_len & 1)
-	    *(skb->tail-1) = inb(ioaddr + AM2150_RCV);
+	    *(skb_tail_pointer(skb) - 1) = inb(ioaddr + AM2150_RCV);
 	skb->protocol = eth_type_trans(skb, dev);
 	
 	netif_rx(skb); /* Send the packet to the upper (protocol) layers. */
diff --git a/drivers/net/s2io.c b/drivers/net/s2io.c
index 46ebf141ee5a..600d3ff347fc 100644
--- a/drivers/net/s2io.c
+++ b/drivers/net/s2io.c
@@ -2195,7 +2195,7 @@ static int fill_rxd_3buf(struct s2io_nic *nic, struct RxD_t *rxdp, struct \
 	frag_list->next = NULL;
 	tmp = (void *)ALIGN((long)frag_list->data, ALIGN_SIZE + 1);
 	frag_list->data = tmp;
-	frag_list->tail = tmp;
+	skb_reset_tail_pointer(frag_list);
 
 	/* Buffer-2 receives L4 data payload */
 	((struct RxD3*)rxdp)->Buffer2_ptr = pci_map_single(nic->pdev,
@@ -2349,7 +2349,7 @@ static int fill_rx_buffers(struct s2io_nic *nic, int ring_no)
 			tmp += ALIGN_SIZE;
 			tmp &= ~ALIGN_SIZE;
 			skb->data = (void *) (unsigned long)tmp;
-			skb->tail = (void *) (unsigned long)tmp;
+			skb_reset_tail_pointer(skb);
 
 			if (!(((struct RxD3*)rxdp)->Buffer0_ptr))
 				((struct RxD3*)rxdp)->Buffer0_ptr =
diff --git a/drivers/net/tulip/uli526x.c b/drivers/net/tulip/uli526x.c
index 9a5850fa644a..e46f4cb02c15 100644
--- a/drivers/net/tulip/uli526x.c
+++ b/drivers/net/tulip/uli526x.c
@@ -829,7 +829,9 @@ static void uli526x_rx_packet(struct net_device *dev, struct uli526x_board_info
 					!= NULL) ) {
 					/* size less than COPY_SIZE, allocate a rxlen SKB */
 					skb_reserve(skb, 2); /* 16byte align */
-					memcpy(skb_put(skb, rxlen), rxptr->rx_skb_ptr->tail, rxlen);
+					memcpy(skb_put(skb, rxlen),
+					       skb_tail_pointer(rxptr->rx_skb_ptr),
+					       rxlen);
 					uli526x_reuse_skb(db, rxptr->rx_skb_ptr);
 				} else
 					skb_put(skb, rxlen);
@@ -1175,7 +1177,10 @@ static void uli526x_reuse_skb(struct uli526x_board_info *db, struct sk_buff * sk
 
 	if (!(rxptr->rdes0 & cpu_to_le32(0x80000000))) {
 		rxptr->rx_skb_ptr = skb;
-		rxptr->rdes2 = cpu_to_le32( pci_map_single(db->pdev, skb->tail, RX_ALLOC_SIZE, PCI_DMA_FROMDEVICE) );
+		rxptr->rdes2 = cpu_to_le32(pci_map_single(db->pdev,
+							  skb_tail_pointer(skb),
+							  RX_ALLOC_SIZE,
+							  PCI_DMA_FROMDEVICE));
 		wmb();
 		rxptr->rdes0 = cpu_to_le32(0x80000000);
 		db->rx_avail_cnt++;
@@ -1339,7 +1344,10 @@ static void allocate_rx_buffer(struct uli526x_board_info *db)
 		if ( ( skb = dev_alloc_skb(RX_ALLOC_SIZE) ) == NULL )
 			break;
 		rxptr->rx_skb_ptr = skb; /* FIXME (?) */
-		rxptr->rdes2 = cpu_to_le32( pci_map_single(db->pdev, skb->tail, RX_ALLOC_SIZE, PCI_DMA_FROMDEVICE) );
+		rxptr->rdes2 = cpu_to_le32(pci_map_single(db->pdev,
+							  skb_tail_pointer(skb),
+							  RX_ALLOC_SIZE,
+							  PCI_DMA_FROMDEVICE));
 		wmb();
 		rxptr->rdes0 = cpu_to_le32(0x80000000);
 		rxptr = rxptr->next_rx_desc;
diff --git a/drivers/net/wan/hdlc_fr.c b/drivers/net/wan/hdlc_fr.c
index b747228c7198..aeb2789adf26 100644
--- a/drivers/net/wan/hdlc_fr.c
+++ b/drivers/net/wan/hdlc_fr.c
@@ -533,7 +533,7 @@ static void fr_lmi_send(struct net_device *dev, int fullrep)
 		skb->protocol = __constant_htons(NLPID_CCITT_ANSI_LMI);
 		fr_hard_header(&skb, LMI_CCITT_ANSI_DLCI);
 	}
-	data = skb->tail;
+	data = skb_tail_pointer(skb);
 	data[i++] = LMI_CALLREF;
 	data[i++] = dce ? LMI_STATUS : LMI_STATUS_ENQUIRY;
 	if (lmi == LMI_ANSI)
diff --git a/drivers/net/wan/lmc/lmc_main.c b/drivers/net/wan/lmc/lmc_main.c
index d4851465c83b..b731f3aae0df 100644
--- a/drivers/net/wan/lmc/lmc_main.c
+++ b/drivers/net/wan/lmc/lmc_main.c
@@ -1636,7 +1636,7 @@ static int lmc_rx (struct net_device *dev) /*fold00*/
             if (nsb) {
                 sc->lmc_rxq[i] = nsb;
                 nsb->dev = dev;
-                sc->lmc_rxring[i].buffer1 = virt_to_bus (nsb->tail);
+                sc->lmc_rxring[i].buffer1 = virt_to_bus(skb_tail_pointer(nsb));
             }
             sc->failed_recv_alloc = 1;
             goto skip_packet;
@@ -1679,7 +1679,7 @@ static int lmc_rx (struct net_device *dev) /*fold00*/
             if (nsb) {
                 sc->lmc_rxq[i] = nsb;
                 nsb->dev = dev;
-                sc->lmc_rxring[i].buffer1 = virt_to_bus (nsb->tail);
+                sc->lmc_rxring[i].buffer1 = virt_to_bus(skb_tail_pointer(nsb));
                 /* Transferred to 21140 below */
             }
             else {
diff --git a/drivers/net/wireless/hostap/hostap_80211_rx.c b/drivers/net/wireless/hostap/hostap_80211_rx.c
index 35a3a50724fe..5e3e9e262706 100644
--- a/drivers/net/wireless/hostap/hostap_80211_rx.c
+++ b/drivers/net/wireless/hostap/hostap_80211_rx.c
@@ -922,7 +922,7 @@ void hostap_80211_rx(struct net_device *dev, struct sk_buff *skb,
 		if (frag != 0)
 			flen -= hdrlen;
 
-		if (frag_skb->tail + flen > frag_skb->end) {
+		if (skb_tail_pointer(frag_skb) + flen > frag_skb->end) {
 			printk(KERN_WARNING "%s: host decrypted and "
 			       "reassembled frame did not fit skb\n",
 			       dev->name);