DPDK patches and discussions
 help / color / mirror / Atom feed
From: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
To: dev@dpdk.org
Cc: Thomas Monjalon <thomas.monjalon@6wind.com>,
	Adrien Mazarguil <adrien.mazarguil@6wind.com>
Subject: [dpdk-dev] [PATCH 7/7] net/mlx5: remove inefficient prefetching
Date: Thu, 24 Nov 2016 17:03:36 +0100	[thread overview]
Message-ID: <9c689005d303b22bd4704d226aff4782a5f5a00b.1479995764.git.nelio.laranjeiro@6wind.com> (raw)
In-Reply-To: <cover.1479995764.git.nelio.laranjeiro@6wind.com>
In-Reply-To: <cover.1479995764.git.nelio.laranjeiro@6wind.com>

Prefetching completion queue entries is inefficient because too few CPU
cycles are spent before their use, which results into cache misses anyway.

Signed-off-by: Nelio Laranjeiro <nelio.laranjeiro@6wind.com>
Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>
---
 drivers/net/mlx5/mlx5_rxtx.c | 21 ---------------------
 1 file changed, 21 deletions(-)

diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 4b8c197..9f74fd4 100644
--- a/drivers/net/mlx5/mlx5_rxtx.c
+++ b/drivers/net/mlx5/mlx5_rxtx.c
@@ -348,23 +348,6 @@ mlx5_tx_dbrec(struct txq *txq)
 }
 
 /**
- * Prefetch a CQE.
- *
- * @param txq
- *   Pointer to TX queue structure.
- * @param cqe_ci
- *   CQE consumer index.
- */
-static inline void
-tx_prefetch_cqe(struct txq *txq, uint16_t ci)
-{
-	volatile struct mlx5_cqe *cqe;
-
-	cqe = &(*txq->cqes)[ci & ((1 << txq->cqe_n) - 1)];
-	rte_prefetch0(cqe);
-}
-
-/**
  * DPDK callback for TX.
  *
  * @param dpdk_txq
@@ -395,8 +378,6 @@ mlx5_tx_burst(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	if (unlikely(!pkts_n))
 		return 0;
 	/* Prefetch first packet cacheline. */
-	tx_prefetch_cqe(txq, txq->cq_ci);
-	tx_prefetch_cqe(txq, txq->cq_ci + 1);
 	rte_prefetch0(*pkts);
 	/* Start processing. */
 	txq_complete(txq);
@@ -733,7 +714,6 @@ mlx5_tx_burst_mpw(void *dpdk_txq, struct rte_mbuf **pkts, uint16_t pkts_n)
 	if (unlikely(!pkts_n))
 		return 0;
 	/* Prefetch first packet cacheline. */
-	tx_prefetch_cqe(txq, txq->cq_ci);
 	rte_prefetch0(tx_mlx5_wqe(txq, txq->wqe_ci));
 	rte_prefetch0(tx_mlx5_wqe(txq, txq->wqe_ci + 1));
 	/* Start processing. */
@@ -938,7 +918,6 @@ mlx5_tx_burst_mpw_inline(void *dpdk_txq, struct rte_mbuf **pkts,
 	if (unlikely(!pkts_n))
 		return 0;
 	/* Prefetch first packet cacheline. */
-	tx_prefetch_cqe(txq, txq->cq_ci);
 	rte_prefetch0(tx_mlx5_wqe(txq, txq->wqe_ci));
 	rte_prefetch0(tx_mlx5_wqe(txq, txq->wqe_ci + 1));
 	/* Start processing. */
-- 
2.1.4

  parent reply	other threads:[~2016-11-24 16:04 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-24 16:03 [dpdk-dev] [PATCH 0/7] net/mlx5: improve single core performance Nelio Laranjeiro
2016-11-24 16:03 ` [dpdk-dev] [PATCH 1/7] net/mlx5: prepare Tx vectorization Nelio Laranjeiro
2016-11-24 16:03 ` [dpdk-dev] [PATCH 2/7] net/mlx5: use work queue buffer as a raw buffer Nelio Laranjeiro
2016-11-24 16:03 ` [dpdk-dev] [PATCH 3/7] net/mlx5: use vector types to speed up processing Nelio Laranjeiro
2016-11-24 16:03 ` [dpdk-dev] [PATCH 4/7] net/mlx5: fix missing inline attributes Nelio Laranjeiro
2016-11-24 16:03 ` [dpdk-dev] [PATCH 5/7] net/mlx5: move static prototype Nelio Laranjeiro
2016-11-24 16:03 ` [dpdk-dev] [PATCH 6/7] net/mlx5: optimize copy of Ethernet header Nelio Laranjeiro
2016-11-24 16:03 ` Nelio Laranjeiro [this message]
2017-01-05 14:13 ` [dpdk-dev] [PATCH 0/7] net/mlx5: improve single core performance Ferruh Yigit
2017-01-18  7:20   ` Thomas Monjalon
2017-01-18  9:23     ` Nélio Laranjeiro
2017-01-18 10:01     ` [dpdk-dev] [PATCH] net/mlx5: fix 32bits compilation issue Nelio Laranjeiro
2017-01-18 10:13       ` Thomas Monjalon

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=9c689005d303b22bd4704d226aff4782a5f5a00b.1479995764.git.nelio.laranjeiro@6wind.com \
    --to=nelio.laranjeiro@6wind.com \
    --cc=adrien.mazarguil@6wind.com \
    --cc=dev@dpdk.org \
    --cc=thomas.monjalon@6wind.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).