bnxt_en: Allow some TX packets to be unprocessed in NAPI

The driver's current logic will always free all the TX SKBs up to
txr->tx_hw_cons within NAPI.  In the next patches, we'll be adding
logic to handle TX timestamp completion and we may need to hold
some remaining TX SKBs if we don't have the timestamp completions
yet.

Modify __bnxt_poll_work_done() to clear each event bit separately to
allow bnapi->tx_int() to decide whether to clear BNXT_TX_CMP_EVENT or
not.  bnapi->tx_int() will not clear BNXT_TX_CMP_EVENT if some TX
SKBs are held waiting for TX timestamps.  Note that legacy chips will
never hold any SKBs this way.  The SKB is always deferred to the PTP
worker slow path to retrieve the timestamp from firmware.  On the new
P7 chips, the timestamp is returned by the hardware directly and we
can retrieve it directly from NAPI.

Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

drivers/net/ethernet/broadcom/bnxt/bnxt.c

@@ -770,7 +770,8 @@ static netdev_tx_t bnxt_start_xmit(struct sk_buff *skb, struct net_device *dev)
 	return NETDEV_TX_OK;
 }
 
-static void __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
+/* Returns true if some remaining TX packets not processed. */
+static bool __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
 			  int budget)
 {
 	struct netdev_queue *txq = netdev_get_tx_queue(bp->dev, txr->txq_index);
@@ -793,7 +794,7 @@ static void __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
 
 		if (unlikely(!skb)) {
 			bnxt_sched_reset_txr(bp, txr, cons);
-			return;
+			return false;
 		}
 
 		tx_pkts++;
@@ -842,18 +843,22 @@ static void __bnxt_tx_int(struct bnxt *bp, struct bnxt_tx_ring_info *txr,
 	__netif_txq_completed_wake(txq, tx_pkts, tx_bytes,
 				   bnxt_tx_avail(bp, txr), bp->tx_wake_thresh,
 				   READ_ONCE(txr->dev_state) == BNXT_DEV_STATE_CLOSING);
+
+	return false;
 }
 
 static void bnxt_tx_int(struct bnxt *bp, struct bnxt_napi *bnapi, int budget)
 {
 	struct bnxt_tx_ring_info *txr;
+	bool more = false;
 	int i;
 
 	bnxt_for_each_napi_tx(i, bnapi, txr) {
 		if (txr->tx_hw_cons != RING_TX(bp, txr->tx_cons))
-			__bnxt_tx_int(bp, txr, budget);
+			more |= __bnxt_tx_int(bp, txr, budget);
 	}
-	bnapi->events &= ~BNXT_TX_CMP_EVENT;
+	if (!more)
+		bnapi->events &= ~BNXT_TX_CMP_EVENT;
 }
 
 static struct page *__bnxt_alloc_rx_page(struct bnxt *bp, dma_addr_t *mapping,
@@ -2950,8 +2955,10 @@ static int __bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 		}
 	}
 
-	if (event & BNXT_REDIRECT_EVENT)
+	if (event & BNXT_REDIRECT_EVENT) {
 		xdp_do_flush();
+		event &= ~BNXT_REDIRECT_EVENT;
+	}
 
 	if (event & BNXT_TX_EVENT) {
 		struct bnxt_tx_ring_info *txr = bnapi->tx_ring[0];
@@ -2961,6 +2968,7 @@ static int __bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,
 		wmb();
 
 		bnxt_db_write_relaxed(bp, &txr->tx_db, prod);
+		event &= ~BNXT_TX_EVENT;
 	}
 
 	cpr->cp_raw_cons = raw_cons;
@@ -2978,13 +2986,14 @@ static void __bnxt_poll_work_done(struct bnxt *bp, struct bnxt_napi *bnapi,
 		struct bnxt_rx_ring_info *rxr = bnapi->rx_ring;
 
 		bnxt_db_write(bp, &rxr->rx_db, rxr->rx_prod);
+		bnapi->events &= ~BNXT_RX_EVENT;
 	}
 	if (bnapi->events & BNXT_AGG_EVENT) {
 		struct bnxt_rx_ring_info *rxr = bnapi->rx_ring;
 
 		bnxt_db_write(bp, &rxr->rx_agg_db, rxr->rx_agg_prod);
+		bnapi->events &= ~BNXT_AGG_EVENT;
 	}
-	bnapi->events &= BNXT_TX_CMP_EVENT;
 }
 
 static int bnxt_poll_work(struct bnxt *bp, struct bnxt_cp_ring_info *cpr,