IPoIB/cm: Reduce connected mode TX object size

Since IPoIB connected mode does not NETIF_F_SG, we only have one DMA
mapping per send, so we don't need a mapping[] array.  Define a new
struct with a single u64 mapping member and use it for the CM tx_ring.
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Roland Dreier <rolandd@cisco.com>

drivers/infiniband/ulp/ipoib/ipoib.h

@@ -157,6 +157,11 @@ struct ipoib_tx_buf {
 	u64		mapping[MAX_SKB_FRAGS + 1];
 };
 
+struct ipoib_cm_tx_buf {
+	struct sk_buff *skb;
+	u64		mapping;
+};
+
 struct ib_cm_id;
 
 struct ipoib_cm_data {
@@ -215,7 +220,7 @@ struct ipoib_cm_tx {
 	struct net_device   *dev;
 	struct ipoib_neigh  *neigh;
 	struct ipoib_path   *path;
-	struct ipoib_tx_buf *tx_ring;
+	struct ipoib_cm_tx_buf *tx_ring;
 	unsigned	     tx_head;
 	unsigned	     tx_tail;
 	unsigned long	     flags;

drivers/infiniband/ulp/ipoib/ipoib_cm.c

@@ -703,7 +703,7 @@ static inline int post_send(struct ipoib_dev_priv *priv,
 void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_tx *tx)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
-	struct ipoib_tx_buf *tx_req;
+	struct ipoib_cm_tx_buf *tx_req;
 	u64 addr;
 
 	if (unlikely(skb->len > tx->mtu)) {
@@ -734,7 +734,7 @@ void ipoib_cm_send(struct net_device *dev, struct sk_buff *skb, struct ipoib_cm_
 		return;
 	}
 
-	tx_req->mapping[0] = addr;
+	tx_req->mapping = addr;
 
 	if (unlikely(post_send(priv, tx, tx->tx_head & (ipoib_sendq_size - 1),
 			       addr, skb->len))) {
@@ -759,7 +759,7 @@ void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
 	struct ipoib_dev_priv *priv = netdev_priv(dev);
 	struct ipoib_cm_tx *tx = wc->qp->qp_context;
 	unsigned int wr_id = wc->wr_id & ~IPOIB_OP_CM;
-	struct ipoib_tx_buf *tx_req;
+	struct ipoib_cm_tx_buf *tx_req;
 	unsigned long flags;
 
 	ipoib_dbg_data(priv, "cm send completion: id %d, status: %d\n",
@@ -773,7 +773,7 @@ void ipoib_cm_handle_tx_wc(struct net_device *dev, struct ib_wc *wc)
 
 	tx_req = &tx->tx_ring[wr_id];
 
-	ib_dma_unmap_single(priv->ca, tx_req->mapping[0], tx_req->skb->len, DMA_TO_DEVICE);
+	ib_dma_unmap_single(priv->ca, tx_req->mapping, tx_req->skb->len, DMA_TO_DEVICE);
 
 	/* FIXME: is this right? Shouldn't we only increment on success? */
 	++dev->stats.tx_packets;
@@ -1143,7 +1143,7 @@ static int ipoib_cm_tx_init(struct ipoib_cm_tx *p, u32 qpn,
 static void ipoib_cm_tx_destroy(struct ipoib_cm_tx *p)
 {
 	struct ipoib_dev_priv *priv = netdev_priv(p->dev);
-	struct ipoib_tx_buf *tx_req;
+	struct ipoib_cm_tx_buf *tx_req;
 	unsigned long flags;
 	unsigned long begin;
 
@@ -1171,7 +1171,7 @@ static void ipoib_cm_tx_destroy(struct ipoib_cm_tx *p)
 
 	while ((int) p->tx_tail - (int) p->tx_head < 0) {
 		tx_req = &p->tx_ring[p->tx_tail & (ipoib_sendq_size - 1)];
-		ib_dma_unmap_single(priv->ca, tx_req->mapping[0], tx_req->skb->len,
+		ib_dma_unmap_single(priv->ca, tx_req->mapping, tx_req->skb->len,
 				    DMA_TO_DEVICE);
 		dev_kfree_skb_any(tx_req->skb);
 		++p->tx_tail;