tg3: switch to build_skb() infrastructure

This is very similar to bnx2x conversion, but simpler since no special
alignement is required, so goal was not to reduce skb truesize.

Using build_skb() reduces cache line misses in the driver, since we
use cache hot skb instead of cold ones. Number of in-flight sk_buff
structures is lower, they are more likely recycled in SLUB caches
while still hot.
Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
CC: Matt Carlson <mcarlson@broadcom.com>
CC: Michael Chan <mchan@broadcom.com>
CC: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/tg3.h

@@ -2662,9 +2662,13 @@ struct tg3_hw_stats {
 /* 'mapping' is superfluous as the chip does not write into
  * the tx/rx post rings so we could just fetch it from there.
  * But the cache behavior is better how we are doing it now.
+ *
+ * This driver uses new build_skb() API :
+ * RX ring buffer contains pointer to kmalloc() data only,
+ * skb are built only after Hardware filled the frame.
  */
 struct ring_info {
-	struct sk_buff			*skb;
+	u8				*data;
 	DEFINE_DMA_UNMAP_ADDR(mapping);
 };