net/mvpp2: remove debug log on fast-path
authorLiron Himi <lironh@marvell.com>
Wed, 27 Jan 2021 16:09:16 +0000 (18:09 +0200)
committerFerruh Yigit <ferruh.yigit@intel.com>
Fri, 29 Jan 2021 17:16:10 +0000 (18:16 +0100)
In case of non-ip frame the current code reached the 'default'
case which result with function call to log a msg.
Those kind of calls should not be performed on fast-path.

The performance for this kind of frames increased by 50%

Fixes: acab7d58c81b ("net/mvpp2: convert to dynamic logging")
Cc: stable@dpdk.org
Signed-off-by: Liron Himi <lironh@marvell.com>
drivers/net/mvpp2/mrvl_ethdev.c

index 2f5b2f2..a564752 100644 (file)
@@ -2171,7 +2171,6 @@ mrvl_desc_to_packet_type_and_offset(struct pp2_ppio_desc *desc,
                *l4_offset = *l3_offset + MRVL_ARP_LENGTH;
                break;
        default:
-               MRVL_LOG(DEBUG, "Failed to recognise l3 packet type");
                break;
        }
 
@@ -2183,7 +2182,6 @@ mrvl_desc_to_packet_type_and_offset(struct pp2_ppio_desc *desc,
                packet_type |= RTE_PTYPE_L4_UDP;
                break;
        default:
-               MRVL_LOG(DEBUG, "Failed to recognise l4 packet type");
                break;
        }
 
@@ -2253,10 +2251,9 @@ mrvl_rx_pkt_burst(void *rxq, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
 
        ret = pp2_ppio_recv(q->priv->ppio, q->priv->rxq_map[q->queue_id].tc,
                            q->priv->rxq_map[q->queue_id].inq, descs, &nb_pkts);
-       if (unlikely(ret < 0)) {
-               MRVL_LOG(ERR, "Failed to receive packets");
+       if (unlikely(ret < 0))
                return 0;
-       }
+
        mrvl_port_bpool_size[bpool->pp2_id][bpool->id][core_id] -= nb_pkts;
 
        for (i = 0; i < nb_pkts; i++) {
@@ -2319,21 +2316,13 @@ mrvl_rx_pkt_burst(void *rxq, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
 
                if (unlikely(num <= q->priv->bpool_min_size ||
                             (!rx_done && num < q->priv->bpool_init_size))) {
-                       ret = mrvl_fill_bpool(q, MRVL_BURST_SIZE);
-                       if (ret)
-                               MRVL_LOG(ERR, "Failed to fill bpool");
+                       mrvl_fill_bpool(q, MRVL_BURST_SIZE);
                } else if (unlikely(num > q->priv->bpool_max_size)) {
                        int i;
                        int pkt_to_remove = num - q->priv->bpool_init_size;
                        struct rte_mbuf *mbuf;
                        struct pp2_buff_inf buff;
 
-                       MRVL_LOG(DEBUG,
-                               "port-%d:%d: bpool %d oversize - remove %d buffers (pool size: %d -> %d)",
-                               bpool->pp2_id, q->priv->ppio->port_id,
-                               bpool->id, pkt_to_remove, num,
-                               q->priv->bpool_init_size);
-
                        for (i = 0; i < pkt_to_remove; i++) {
                                ret = pp2_bpool_get_buff(hif, bpool, &buff);
                                if (ret)
@@ -2526,12 +2515,8 @@ mrvl_tx_pkt_burst(void *txq, struct rte_mbuf **tx_pkts, uint16_t nb_pkts)
                                       sq, q->queue_id, 0);
 
        sq_free_size = MRVL_PP2_TX_SHADOWQ_SIZE - sq->size - 1;
-       if (unlikely(nb_pkts > sq_free_size)) {
-               MRVL_LOG(DEBUG,
-                       "No room in shadow queue for %d packets! %d packets will be sent.",
-                       nb_pkts, sq_free_size);
+       if (unlikely(nb_pkts > sq_free_size))
                nb_pkts = sq_free_size;
-       }
 
        for (i = 0; i < nb_pkts; i++) {
                struct rte_mbuf *mbuf = tx_pkts[i];
@@ -2648,10 +2633,6 @@ mrvl_tx_sg_pkt_burst(void *txq, struct rte_mbuf **tx_pkts,
                 */
                if (unlikely(total_descs > sq_free_size)) {
                        total_descs -= nb_segs;
-                       RTE_LOG(DEBUG, PMD,
-                               "No room in shadow queue for %d packets! "
-                               "%d packets will be sent.\n",
-                               nb_pkts, i);
                        break;
                }