DPDK patches and discussions
 help / color / mirror / Atom feed
From: Balazs Nemeth <bnemeth@redhat.com>
To: bnemeth@redhat.com, dev@dpdk.org
Subject: [dpdk-dev] [PATCH v2 2/4] vhost: perform all mbuf allocations in one loop
Date: Wed,  7 Apr 2021 12:17:19 +0200	[thread overview]
Message-ID: <589dabc3c3a22a8e297e09cb554810aa811bc9a7.1617790501.git.bnemeth@redhat.com> (raw)
In-Reply-To: <f7b08832f86d2fb9a8834b2a19b13f26d3c6857b.1617723339.git.bnemeth@redhat.com>

Move allocation out further and perform all allocation in one loop. The
same goes for freeing packets. This is to prepare for use of bulk
versions of these functions.

Signed-off-by: Balazs Nemeth <bnemeth@redhat.com>
---
 lib/librte_vhost/virtio_net.c | 18 ++++++------------
 1 file changed, 6 insertions(+), 12 deletions(-)

diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c
index 666e7fdb8..496f750e3 100644
--- a/lib/librte_vhost/virtio_net.c
+++ b/lib/librte_vhost/virtio_net.c
@@ -2471,14 +2471,13 @@ virtio_dev_tx_packed(struct virtio_net *dev,
 	uint32_t remained = count;
 	uint16_t i;
 
+	for (i = 0; i < count; ++i)
+		pkts[i] = rte_pktmbuf_alloc(mbuf_pool);
+
 	do {
 		rte_prefetch0(&vq->desc_packed[vq->last_avail_idx]);
 
 		if (remained >= PACKED_BATCH_SIZE) {
-			vhost_for_each_try_unroll(i, 0, PACKED_BATCH_SIZE) {
-				pkts[pkt_idx + i] =
-					rte_pktmbuf_alloc(mbuf_pool);
-			}
 
 			if (!virtio_dev_tx_batch_packed(dev, vq,
 							&pkts[pkt_idx])) {
@@ -2486,19 +2485,11 @@ virtio_dev_tx_packed(struct virtio_net *dev,
 				remained -= PACKED_BATCH_SIZE;
 
 				continue;
-			} else {
-				vhost_for_each_try_unroll(i, 0,
-					PACKED_BATCH_SIZE) {
-					rte_pktmbuf_free(pkts[pkt_idx + i]);
-				}
 			}
 		}
 
-		pkts[pkt_idx] = rte_pktmbuf_alloc(mbuf_pool);
-
 		if (virtio_dev_tx_single_packed(dev, vq, mbuf_pool,
 						&pkts[pkt_idx])) {
-			rte_pktmbuf_free(pkts[pkt_idx]);
 			break;
 		}
 		pkt_idx++;
@@ -2506,6 +2497,9 @@ virtio_dev_tx_packed(struct virtio_net *dev,
 
 	} while (remained);
 
+	for (i = pkt_idx; i < count; ++i)
+		rte_pktmbuf_free(pkts[i]);
+
 	if (vq->shadow_used_idx) {
 		do_data_copy_dequeue(vq);
 
-- 
2.30.2


  parent reply	other threads:[~2021-04-07 10:19 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-06 15:44 [dpdk-dev] [PATCH 0/4] Use bulk alloc/free in virtio packed Balazs Nemeth
2021-04-06 15:44 ` [dpdk-dev] [PATCH 1/4] vhost: move allocation of mbuf outside of packet enqueue Balazs Nemeth
2021-04-07 10:17   ` [dpdk-dev] [PATCH v2 " Balazs Nemeth
2021-04-15 12:37     ` Maxime Coquelin
2021-04-16  8:18     ` [dpdk-dev] [PATCH v3] vhost: allocate and free packets in bulk Balazs Nemeth
2021-04-16  8:36       ` Maxime Coquelin
2021-04-16  9:05       ` David Marchand
2021-04-16  9:12         ` Balazs Nemeth
2021-04-16  9:41           ` Maxime Coquelin
2021-04-16  9:43           ` David Marchand
2021-04-16  9:48       ` [dpdk-dev] [PATCH v4] " Balazs Nemeth
2021-04-16 10:25         ` [dpdk-dev] [PATCH v5] vhost: allocate and free packets in bulk in Tx packed Balazs Nemeth
2021-04-16 11:14           ` David Marchand
2021-04-21  7:48           ` Maxime Coquelin
2021-04-28  3:16             ` Xia, Chenbo
2021-04-07 10:17   ` Balazs Nemeth [this message]
2021-04-15 15:30     ` [dpdk-dev] [PATCH v2 2/4] vhost: perform all mbuf allocations in one loop Maxime Coquelin
2021-04-07 10:17   ` [dpdk-dev] [PATCH v2 3/4] vhost: allocate and free packets in bulk Balazs Nemeth
2021-04-15 15:32     ` Maxime Coquelin
2021-04-15 15:45       ` David Marchand
2021-04-15 15:50         ` Maxime Coquelin
2021-04-15 15:58           ` Maxime Coquelin
2021-04-07 10:17   ` [dpdk-dev] [PATCH v2 4/4] vhost: remove unnecessary level of indirection Balazs Nemeth
2021-04-15 15:38     ` Maxime Coquelin
2021-04-06 15:44 ` [dpdk-dev] [PATCH 2/4] vhost: perform all mbuf allocations in one loop Balazs Nemeth
2021-04-06 15:44 ` [dpdk-dev] [PATCH 3/4] vhost: allocate and free packets in bulk Balazs Nemeth
2021-04-06 15:44 ` [dpdk-dev] [PATCH 4/4] vhost: remove unnecessary level of indirection Balazs Nemeth

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=589dabc3c3a22a8e297e09cb554810aa811bc9a7.1617790501.git.bnemeth@redhat.com \
    --to=bnemeth@redhat.com \
    --cc=dev@dpdk.org \
    /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).