[dpdk-dev] [PATCH 1/3] mlx4: improve RX performance with better prefetching

[dpdk-dev] [PATCH 1/3] mlx4: improve RX performance with better prefetching

[Adrien Mazarguil]

From: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>

Prefetching initial bytes of mbuf structures earlier and in two cache lines
instead of one improves performance of mlx4_rx_burst(), which accesses the
mbuf->next field not present in the first 128 bytes.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
---
 drivers/net/mlx4/mlx4.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index cb08ee8..4cf2f7d 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -2824,6 +2824,12 @@ mlx4_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		assert(wr->num_sge == 1);
 		assert(elts_head < rxq->elts_n);
 		assert(rxq->elts_head < rxq->elts_n);
+		/*
+		 * Fetch initial bytes of packet descriptor into a
+		 * cacheline while allocating rep.
+		 */
+		rte_prefetch0(seg);
+		rte_prefetch0(&seg->cacheline1);
 		ret = rxq->if_cq->poll_length_flags(rxq->cq, NULL, NULL,
 						    &flags);
 		if (unlikely(ret < 0)) {
@@ -2861,11 +2867,6 @@ mlx4_rx_burst(void *dpdk_rxq, struct rte_mbuf **pkts, uint16_t pkts_n)
 		if (ret == 0)
 			break;
 		len = ret;
-		/*
-		 * Fetch initial bytes of packet descriptor into a
-		 * cacheline while allocating rep.
-		 */
-		rte_prefetch0(seg);
 		rep = __rte_mbuf_raw_alloc(rxq->mp);
 		if (unlikely(rep == NULL)) {
 			/*
-- 
2.1.0

[dpdk-dev] [PATCH 2/3] mlx4: do not complain about port state

[Adrien Mazarguil]

Port is considering inactive when the related netdevice is down. There is no
reason to warn about it (and confuse users) since it's automatically brought
up later by the PMD.

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 4cf2f7d..3b25492 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -4985,9 +4985,9 @@ mlx4_pci_devinit(struct rte_pci_driver *pci_drv, struct rte_pci_device *pci_dev)
 			goto port_error;
 		}
 		if (port_attr.state != IBV_PORT_ACTIVE)
-			WARN("bad state for port %d: \"%s\" (%d)",
-			     port, ibv_port_state_str(port_attr.state),
-			     port_attr.state);
+			DEBUG("port %d is not active: \"%s\" (%d)",
+			      port, ibv_port_state_str(port_attr.state),
+			      port_attr.state);
 
 		/* Allocate protection domain. */
 		pd = ibv_alloc_pd(ctx);
-- 
2.1.0

[dpdk-dev] [PATCH 3/3] mlx4: fix missing unlock in case of error

[Adrien Mazarguil]

Fixes: 7fae69eeff13 ("mlx4: new poll mode driver")

Signed-off-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx4/mlx4.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/net/mlx4/mlx4.c b/drivers/net/mlx4/mlx4.c
index 3b25492..691e0e5 100644
--- a/drivers/net/mlx4/mlx4.c
+++ b/drivers/net/mlx4/mlx4.c
@@ -3652,6 +3652,7 @@ mlx4_dev_start(struct rte_eth_dev *dev)
 			}
 		}
 		priv->started = 0;
+		priv_unlock(priv);
 		return -ret;
 	} while ((--r) && ((rxq = (*priv->rxqs)[++i]), i));
 	priv_dev_interrupt_handler_install(priv, dev);
-- 
2.1.0

Re: [dpdk-dev] [PATCH 1/3] mlx4: improve RX performance with better prefetching

[Thomas Monjalon]

Series applied, thanks