DPDK patches and discussions
 help / color / mirror / Atom feed
From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: tiwei.bie@intel.com, zhihong.wang@intel.com, dev@dpdk.org
Cc: Maxime Coquelin <maxime.coquelin@redhat.com>
Subject: [dpdk-dev] [PATCH v2 5/7] vhost: improve prefetching in dequeue path
Date: Sat, 23 Jun 2018 09:11:25 +0200	[thread overview]
Message-ID: <20180623071127.22999-6-maxime.coquelin@redhat.com> (raw)
In-Reply-To: <20180623071127.22999-1-maxime.coquelin@redhat.com>

This is an optimization to prefetch next buffer while the
current one is being processed.

Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
---
 lib/librte_vhost/virtio_net.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c
index 1ab1edd67..0980efe53 100644
--- a/lib/librte_vhost/virtio_net.c
+++ b/lib/librte_vhost/virtio_net.c
@@ -895,7 +895,13 @@ copy_desc_to_mbuf(struct virtio_net *dev, struct vhost_virtqueue *vq,
 			buf_iova = buf_vec[vec_idx].buf_iova;
 			buf_len = buf_vec[vec_idx].buf_len;
 
-			rte_prefetch0((void *)(uintptr_t)buf_addr);
+			/*
+			 * Prefecth desc n + 1 buffer while
+			 * desc n buffer is processed.
+			 */
+			if (vec_idx + 1 < nr_vec)
+				rte_prefetch0((void *)(uintptr_t)
+						buf_vec[vec_idx + 1].buf_addr);
 
 			buf_offset = 0;
 			buf_avail  = buf_len;
@@ -1127,6 +1133,8 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
 		if (likely(dev->dequeue_zero_copy == 0))
 			update_shadow_used_ring(vq, head_idx, 0);
 
+		rte_prefetch0((void *)(uintptr_t)buf_vec[0].buf_addr);
+
 		pkts[i] = rte_pktmbuf_alloc(mbuf_pool);
 		if (unlikely(pkts[i] == NULL)) {
 			RTE_LOG(ERR, VHOST_DATA,
-- 
2.14.4

  parent reply	other threads:[~2018-06-23  7:11 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-23  7:11 [dpdk-dev] [PATCH v2 0/7] vhost: generalize buffer vectors Maxime Coquelin
2018-06-23  7:11 ` [dpdk-dev] [PATCH v2 1/7] vhost: use shadow used ring in dequeue path Maxime Coquelin
2018-06-23  7:11 ` [dpdk-dev] [PATCH v2 2/7] vhost: make gpa to hpa failure an error Maxime Coquelin
2018-06-23  7:11 ` [dpdk-dev] [PATCH v2 3/7] vhost: use buffer vectors in dequeue path Maxime Coquelin
2018-06-23  7:11 ` [dpdk-dev] [PATCH v2 4/7] vhost: translate iovas at vectors fill time Maxime Coquelin
2018-06-25  2:21   ` Tiwei Bie
2018-06-25  7:19     ` Maxime Coquelin
2018-06-25 10:31       ` Tiwei Bie
2018-06-27  8:38         ` Maxime Coquelin
2018-06-23  7:11 ` Maxime Coquelin [this message]
2018-06-23  7:11 ` [dpdk-dev] [PATCH v2 6/7] vhost: prefetch first descriptor in dequeue path Maxime Coquelin
2018-06-23  7:11 ` [dpdk-dev] [PATCH v2 7/7] vhost: improve prefetching in enqueue path Maxime Coquelin

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=20180623071127.22999-6-maxime.coquelin@redhat.com \
    --to=maxime.coquelin@redhat.com \
    --cc=dev@dpdk.org \
    --cc=tiwei.bie@intel.com \
    --cc=zhihong.wang@intel.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).