net/mlx4_en: Use prefetch in tx path

mlx4_en_free_tx_desc() uses a prefetchw(&skb->users) to speed up
consume_skb()
prefetchw(&ring->tx_queue->dql) to speed up BQL update

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Eric Dumazet 2014-10-05 12:35:14 +03:00 committed by David S. Miller
parent 3d03641cb4
commit 29d40c9032

View File

@ -37,6 +37,7 @@
#include <linux/mlx4/qp.h> #include <linux/mlx4/qp.h>
#include <linux/skbuff.h> #include <linux/skbuff.h>
#include <linux/if_vlan.h> #include <linux/if_vlan.h>
#include <linux/prefetch.h>
#include <linux/vmalloc.h> #include <linux/vmalloc.h>
#include <linux/tcp.h> #include <linux/tcp.h>
#include <linux/ip.h> #include <linux/ip.h>
@ -267,6 +268,11 @@ static u32 mlx4_en_free_tx_desc(struct mlx4_en_priv *priv,
int nr_maps = tx_info->nr_maps; int nr_maps = tx_info->nr_maps;
int i; int i;
/* We do not touch skb here, so prefetch skb->users location
* to speedup consume_skb()
*/
prefetchw(&skb->users);
if (unlikely(timestamp)) { if (unlikely(timestamp)) {
struct skb_shared_hwtstamps hwts; struct skb_shared_hwtstamps hwts;
@ -385,6 +391,7 @@ static bool mlx4_en_process_tx_cq(struct net_device *dev,
if (!priv->port_up) if (!priv->port_up)
return true; return true;
prefetchw(&ring->tx_queue->dql.limit);
index = cons_index & size_mask; index = cons_index & size_mask;
cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor; cqe = mlx4_en_get_cqe(buf, index, priv->cqe_size) + factor;
ring_index = ring->cons & size_mask; ring_index = ring->cons & size_mask;
@ -722,6 +729,8 @@ netdev_tx_t mlx4_en_xmit(struct sk_buff *skb, struct net_device *dev)
} }
} }
prefetchw(&ring->tx_queue->dql);
/* Track current inflight packets for performance analysis */ /* Track current inflight packets for performance analysis */
AVG_PERF_COUNTER(priv->pstats.inflight_avg, AVG_PERF_COUNTER(priv->pstats.inflight_avg,
(u32) (ring->prod - ring->cons - 1)); (u32) (ring->prod - ring->cons - 1));