geneve: Use GRO cells infrastructure.

Geneve can benefit from GRO at the device level in a manner similar
to other tunnels, especially as hardware offloads are still emerging.

After this patch, aggregated frames are seen on the tunnel interface.
Single stream throughput nearly doubles in ideal circumstances (on
old hardware).

Signed-off-by: Jesse Gross <jesse@nicira.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c
index 3908a22..68b0f03 100644
--- a/drivers/net/geneve.c
+++ b/drivers/net/geneve.c
@@ -16,6 +16,7 @@
 #include <linux/etherdevice.h>
 #include <linux/hash.h>
 #include <net/dst_metadata.h>
+#include <net/gro_cells.h>
 #include <net/rtnetlink.h>
 #include <net/geneve.h>
 #include <net/protocol.h>
@@ -58,6 +59,7 @@
 	struct list_head   next;	/* geneve's per namespace list */
 	__be16		   dst_port;
 	bool		   collect_md;
+	struct gro_cells   gro_cells;
 };
 
 struct geneve_sock {
@@ -199,7 +201,7 @@
 	stats->rx_bytes += skb->len;
 	u64_stats_update_end(&stats->syncp);
 
-	netif_rx(skb);
+	gro_cells_receive(&geneve->gro_cells, skb);
 	return;
 drop:
 	/* Consume bad packet */
@@ -209,14 +211,27 @@
 /* Setup stats when device is created */
 static int geneve_init(struct net_device *dev)
 {
+	struct geneve_dev *geneve = netdev_priv(dev);
+	int err;
+
 	dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats);
 	if (!dev->tstats)
 		return -ENOMEM;
+
+	err = gro_cells_init(&geneve->gro_cells, dev);
+	if (err) {
+		free_percpu(dev->tstats);
+		return err;
+	}
+
 	return 0;
 }
 
 static void geneve_uninit(struct net_device *dev)
 {
+	struct geneve_dev *geneve = netdev_priv(dev);
+
+	gro_cells_destroy(&geneve->gro_cells);
 	free_percpu(dev->tstats);
 }