gianfar: Fix reported number of sent bytes to BQL

Fix the amount of sent bytes reported to BQL by reporting the
number of bytes on wire in the xmit routine, and recording that
value for each skb in order to be correctly confirmed on Tx
confirmation cleanup.

Reporting skb->len to BQL just before exiting xmit is not correct
due to possible insertions of TOE block and alignment bytes in the
skb->data, which are being stripped off by the controller before
transmission on wire.  This led to mismatch of (incorrectly)
reported bytes to BQL b/w xmit and Tx confirmation, resulting in
Tx timeout firing, for the h/w tx timestamping acceleration case.

There's no easy way to obtain the number of bytes on wire in the Tx
confirmation routine, so skb->cb is used to convey that information
from xmit to Tx confirmation, for now (as proposed by Eric). Revived
the currently unused GFAR_CB() construct for that purpose.
Signed-off-by: Claudiu Manoil <claudiu.manoil@freescale.com>

Cc: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/freescale/gianfar.c

@@ -2092,7 +2092,7 @@ static int gfar_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	int do_tstamp, do_csum, do_vlan;
 	u32 bufaddr;
 	unsigned long flags;
-	unsigned int nr_frags, nr_txbds, length, fcb_len = 0;
+	unsigned int nr_frags, nr_txbds, bytes_sent, fcb_len = 0;
 
 	rq = skb->queue_mapping;
 	tx_queue = priv->tx_queue[rq];
@@ -2147,7 +2147,10 @@ static int gfar_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	}
 
 	/* Update transmit stats */
-	tx_queue->stats.tx_bytes += skb->len;
+	bytes_sent = skb->len;
+	tx_queue->stats.tx_bytes += bytes_sent;
+	/* keep Tx bytes on wire for BQL accounting */
+	GFAR_CB(skb)->bytes_sent = bytes_sent;
 	tx_queue->stats.tx_packets++;
 
 	txbdp = txbdp_start = tx_queue->cur_tx;
@@ -2167,12 +2170,13 @@ static int gfar_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	} else {
 		/* Place the fragment addresses and lengths into the TxBDs */
 		for (i = 0; i < nr_frags; i++) {
+			unsigned int frag_len;
 			/* Point at the next BD, wrapping as needed */
 			txbdp = next_txbd(txbdp, base, tx_queue->tx_ring_size);
 
-			length = skb_shinfo(skb)->frags[i].size;
+			frag_len = skb_shinfo(skb)->frags[i].size;
 
-			lstatus = txbdp->lstatus | length |
+			lstatus = txbdp->lstatus | frag_len |
 				  BD_LFLAG(TXBD_READY);
 
 			/* Handle the last BD specially */
@@ -2182,7 +2186,7 @@ static int gfar_start_xmit(struct sk_buff *skb, struct net_device *dev)
 			bufaddr = skb_frag_dma_map(priv->dev,
 						   &skb_shinfo(skb)->frags[i],
 						   0,
-						   length,
+						   frag_len,
 						   DMA_TO_DEVICE);
 
 			/* set the TxBD length and buffer pointer */
@@ -2250,7 +2254,7 @@ static int gfar_start_xmit(struct sk_buff *skb, struct net_device *dev)
 		lstatus |= BD_LFLAG(TXBD_CRC | TXBD_READY) | skb_headlen(skb);
 	}
 
-	netdev_tx_sent_queue(txq, skb->len);
+	netdev_tx_sent_queue(txq, bytes_sent);
 
 	/* We can work in parallel with gfar_clean_tx_ring(), except
 	 * when modifying num_txbdfree. Note that we didn't grab the lock
@@ -2570,7 +2574,7 @@ static void gfar_clean_tx_ring(struct gfar_priv_tx_q *tx_queue)
 			bdp = next_txbd(bdp, base, tx_ring_size);
 		}
 
-		bytes_sent += skb->len;
+		bytes_sent += GFAR_CB(skb)->bytes_sent;
 
 		dev_kfree_skb_any(skb);
 

drivers/net/ethernet/freescale/gianfar.h

@@ -575,7 +575,7 @@ struct rxfcb {
 };
 
 struct gianfar_skb_cb {
-	int alignamount;
+	unsigned int bytes_sent; /* bytes-on-wire (i.e. no FCB) */
 };
 
 #define GFAR_CB(skb) ((struct gianfar_skb_cb *)((skb)->cb))