net: bcmgenet: simplify __bcmgenet_tx_reclaim()

1. Use c_index and ring->c_index to determine how many TxCBs/TxBDs are
   ready for cleanup
   - c_index = the current value of TDMA_CONS_INDEX
   - TDMA_CONS_INDEX is HW-incremented and auto-wraparound (0x0-0xFFFF)
   - ring->c_index = __bcmgenet_tx_reclaim() cleaned up to this point on
     the previous invocation

2. Add bcmgenet_tx_ring->clean_ptr
   - index of the next TxCB to be cleaned
   - incremented as TxCBs/TxBDs are processed
   - value always in range [ring->cb_ptr, ring->end_ptr]

3. Fix incrementing of dev->stats.tx_packets
   - should be incremented only when tx_cb_ptr->skb != NULL

These changes simplify __bcmgenet_tx_reclaim(). Furthermore, Tx ring size
can now be any value.

With the old code, Tx ring size had to be a power-of-2:
   num_tx_bds = ring->size;
   c_index &= (num_tx_bds - 1);
   last_c_index &= (num_tx_bds - 1);

Signed-off-by: Petri Gynther <pgynther@google.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/genet/bcmgenet.c

@@ -978,39 +978,32 @@ static unsigned int __bcmgenet_tx_reclaim(struct net_device *dev,
 					  struct bcmgenet_tx_ring *ring)
 {
 	struct bcmgenet_priv *priv = netdev_priv(dev);
-	int last_tx_cn, last_c_index, num_tx_bds;
 	struct enet_cb *tx_cb_ptr;
 	struct netdev_queue *txq;
 	unsigned int pkts_compl = 0;
-	unsigned int bds_compl;
 	unsigned int c_index;
+	unsigned int txbds_ready;
+	unsigned int txbds_processed = 0;
 
 	/* Compute how many buffers are transmitted since last xmit call */
 	c_index = bcmgenet_tdma_ring_readl(priv, ring->index, TDMA_CONS_INDEX);
-	txq = netdev_get_tx_queue(dev, ring->queue);
+	c_index &= DMA_C_INDEX_MASK;
 
-	last_c_index = ring->c_index;
+	if (likely(c_index >= ring->c_index))
-	num_tx_bds = ring->size;
+		txbds_ready = c_index - ring->c_index;
-
-	c_index &= (num_tx_bds - 1);
-
-	if (c_index >= last_c_index)
-		last_tx_cn = c_index - last_c_index;
 	else
-		last_tx_cn = num_tx_bds - last_c_index + c_index;
+		txbds_ready = (DMA_C_INDEX_MASK + 1) - ring->c_index + c_index;
 
 	netif_dbg(priv, tx_done, dev,
-		  "%s ring=%d index=%d last_tx_cn=%d last_index=%d\n",
+		  "%s ring=%d old_c_index=%u c_index=%u txbds_ready=%u\n",
-		  __func__, ring->index,
+		  __func__, ring->index, ring->c_index, c_index, txbds_ready);
-		  c_index, last_tx_cn, last_c_index);
 
 	/* Reclaim transmitted buffers */
-	while (last_tx_cn-- > 0) {
+	while (txbds_processed < txbds_ready) {
-		tx_cb_ptr = ring->cbs + last_c_index;
+		tx_cb_ptr = &priv->tx_cbs[ring->clean_ptr];
-		bds_compl = 0;
 		if (tx_cb_ptr->skb) {
 			pkts_compl++;
-			bds_compl = skb_shinfo(tx_cb_ptr->skb)->nr_frags + 1;
+			dev->stats.tx_packets++;
 			dev->stats.tx_bytes += tx_cb_ptr->skb->len;
 			dma_unmap_single(&dev->dev,
 					 dma_unmap_addr(tx_cb_ptr, dma_addr),
@@ -1026,20 +1019,23 @@ static unsigned int __bcmgenet_tx_reclaim(struct net_device *dev,
 				       DMA_TO_DEVICE);
 			dma_unmap_addr_set(tx_cb_ptr, dma_addr, 0);
 		}
-		dev->stats.tx_packets++;
-		ring->free_bds += bds_compl;
 
-		last_c_index++;
+		txbds_processed++;
-		last_c_index &= (num_tx_bds - 1);
+		if (likely(ring->clean_ptr < ring->end_ptr))
+			ring->clean_ptr++;
+		else
+			ring->clean_ptr = ring->cb_ptr;
 	}
 
+	ring->free_bds += txbds_processed;
+	ring->c_index = (ring->c_index + txbds_processed) & DMA_C_INDEX_MASK;
+
 	if (ring->free_bds > (MAX_SKB_FRAGS + 1)) {
+		txq = netdev_get_tx_queue(dev, ring->queue);
 		if (netif_tx_queue_stopped(txq))
 			netif_tx_wake_queue(txq);
 	}
 
-	ring->c_index = c_index;
-
 	return pkts_compl;
 }
 
@@ -1734,6 +1730,7 @@ static void bcmgenet_init_tx_ring(struct bcmgenet_priv *priv,
 	}
 	ring->cbs = priv->tx_cbs + start_ptr;
 	ring->size = size;
+	ring->clean_ptr = start_ptr;
 	ring->c_index = 0;
 	ring->free_bds = size;
 	ring->write_ptr = start_ptr;

drivers/net/ethernet/broadcom/genet/bcmgenet.h

@@ -525,6 +525,7 @@ struct bcmgenet_tx_ring {
 	unsigned int	queue;		/* queue index */
 	struct enet_cb	*cbs;		/* tx ring buffer control block*/
 	unsigned int	size;		/* size of each tx ring */
+	unsigned int    clean_ptr;      /* Tx ring clean pointer */
 	unsigned int	c_index;	/* last consumer index of each ring*/
 	unsigned int	free_bds;	/* # of free bds for each ring */
 	unsigned int	write_ptr;	/* Tx ring write pointer SW copy */