enic: support skb->xmit_more

Check and update posted_index only when skb->xmit_more is 0 or tx queue is full.

v2:
use txq_map instead of skb_get_queue_mapping(skb)
Signed-off-by: Govindarajulu Varadarajan <_govind@gmx.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/cisco/enic/enic_main.c

@@ -533,6 +533,7 @@ static netdev_tx_t enic_hard_start_xmit(struct sk_buff *skb,
 	struct vnic_wq *wq;
 	unsigned long flags;
 	unsigned int txq_map;
+	struct netdev_queue *txq;
 
 	if (skb->len <= 0) {
 		dev_kfree_skb_any(skb);
@@ -541,6 +542,7 @@ static netdev_tx_t enic_hard_start_xmit(struct sk_buff *skb,
 
 	txq_map = skb_get_queue_mapping(skb) % enic->wq_count;
 	wq = &enic->wq[txq_map];
+	txq = netdev_get_tx_queue(netdev, txq_map);
 
 	/* Non-TSO sends must fit within ENIC_NON_TSO_MAX_DESC descs,
 	 * which is very likely.  In the off chance it's going to take
@@ -558,7 +560,7 @@ static netdev_tx_t enic_hard_start_xmit(struct sk_buff *skb,
 
 	if (vnic_wq_desc_avail(wq) <
 	    skb_shinfo(skb)->nr_frags + ENIC_DESC_MAX_SPLITS) {
-		netif_tx_stop_queue(netdev_get_tx_queue(netdev, txq_map));
+		netif_tx_stop_queue(txq);
 		/* This is a hard error, log it */
 		netdev_err(netdev, "BUG! Tx ring full when queue awake!\n");
 		spin_unlock_irqrestore(&enic->wq_lock[txq_map], flags);
@@ -568,7 +570,9 @@ static netdev_tx_t enic_hard_start_xmit(struct sk_buff *skb,
 	enic_queue_wq_skb(enic, wq, skb);
 
 	if (vnic_wq_desc_avail(wq) < MAX_SKB_FRAGS + ENIC_DESC_MAX_SPLITS)
-		netif_tx_stop_queue(netdev_get_tx_queue(netdev, txq_map));
+		netif_tx_stop_queue(txq);
+	if (!skb->xmit_more || netif_xmit_stopped(txq))
+		vnic_wq_doorbell(wq);
 
 	spin_unlock_irqrestore(&enic->wq_lock[txq_map], flags);
 

drivers/net/ethernet/cisco/enic/vnic_wq.h

@@ -104,6 +104,17 @@ static inline void *vnic_wq_next_desc(struct vnic_wq *wq)
 	return wq->to_use->desc;
 }
 
+static inline void vnic_wq_doorbell(struct vnic_wq *wq)
+{
+	/* Adding write memory barrier prevents compiler and/or CPU
+	 * reordering, thus avoiding descriptor posting before
+	 * descriptor is initialized. Otherwise, hardware can read
+	 * stale descriptor fields.
+	 */
+	wmb();
+	iowrite32(wq->to_use->index, &wq->ctrl->posted_index);
+}
+
 static inline void vnic_wq_post(struct vnic_wq *wq,
 	void *os_buf, dma_addr_t dma_addr,
 	unsigned int len, int sop, int eop,
@@ -122,15 +133,6 @@ static inline void vnic_wq_post(struct vnic_wq *wq,
 	buf->wr_id = wrid;
 
 	buf = buf->next;
-	if (eop) {
-		/* Adding write memory barrier prevents compiler and/or CPU
-		 * reordering, thus avoiding descriptor posting before
-		 * descriptor is initialized. Otherwise, hardware can read
-		 * stale descriptor fields.
-		 */
-		wmb();
-		iowrite32(buf->index, &wq->ctrl->posted_index);
-	}
 	wq->to_use = buf;
 
 	wq->ring.desc_avail -= desc_skip_cnt;