be2net: support TX batching using skb->xmit_more flag

This patch uses skb->xmit_more flag to batch TX requests.
TX is flushed either when xmit_more is false or there is
no more space in the TXQ.

Skyhawk-R and BEx chips require an even number of wrbs to be posted.
So, when a batch of TX requests is accumulated, the last header wrb
may need to be fixed with an extra dummy wrb.

This patch refactors be_xmit() routine as a sequence of be_xmit_enqueue()
and be_xmit_flush() calls. The Tx completion code is also
updated to be able to unmap/free a batch of skbs rather than a single
skb.
Signed-off-by: Sathya Perla <sathya.perla@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/emulex/benet/be.h

@@ -243,7 +243,6 @@ struct be_tx_stats {
 	u64 tx_bytes;
 	u64 tx_pkts;
 	u64 tx_reqs;
-	u64 tx_wrbs;
 	u64 tx_compl;
 	ulong tx_jiffies;
 	u32 tx_stops;
@@ -266,6 +265,9 @@ struct be_tx_obj {
 	/* Remember the skbs that were transmitted */
 	struct sk_buff *sent_skb_list[TX_Q_LEN];
 	struct be_tx_stats stats;
+	u16 pend_wrb_cnt;	/* Number of WRBs yet to be given to HW */
+	u16 last_req_wrb_cnt;	/* wrb cnt of the last req in the Q */
+	u16 last_req_hdr;	/* index of the last req's hdr-wrb */
 } ____cacheline_aligned_in_smp;
 
 /* Struct to remember the pages posted for rx frags */

drivers/net/ethernet/emulex/benet/be_ethtool.c

@@ -193,8 +193,6 @@ static const struct be_ethtool_stat et_tx_stats[] = {
 	{DRVSTAT_TX_INFO(tx_pkts)},
 	/* Number of skbs queued for trasmission by the driver */
 	{DRVSTAT_TX_INFO(tx_reqs)},
-	/* Number of TX work request blocks DMAed to HW */
-	{DRVSTAT_TX_INFO(tx_wrbs)},
 	/* Number of times the TX queue was stopped due to lack
 	 * of spaces in the TXQ.
 	 */

drivers/net/ethernet/emulex/benet/be_hw.h

@@ -311,6 +311,11 @@ struct amap_eth_hdr_wrb {
 	u8 vlan_tag[16];
 } __packed;
 
+#define TX_HDR_WRB_COMPL		1		/* word 2 */
+#define TX_HDR_WRB_EVT			(1 << 1)	/* word 2 */
+#define TX_HDR_WRB_NUM_SHIFT		13		/* word 2: bits 13:17 */
+#define TX_HDR_WRB_NUM_MASK		0x1F		/* word 2: bits 13:17 */
+
 struct be_eth_hdr_wrb {
 	u32 dw[4];
 };