From: Wathsala Vithanage <wathsala.vithanage@arm.com>
To: Thomas Monjalon <thomas@monjalon.net>,
Jakub Grajciar <jgrajcia@cisco.com>,
Ferruh Yigit <ferruh.yigit@intel.com>
Cc: dev@dpdk.org, nd@arm.com,
Wathsala Vithanage <wathsala.vithanage@arm.com>,
Liangxing Wang <liangxing.wang@arm.com>,
Ruifeng Wang <ruifeng.wang@arm.com>
Subject: [PATCH V2] net/memif: fix extra mbuf refcnt update in zero copy Tx
Date: Wed, 14 Feb 2024 00:36:16 +0000 [thread overview]
Message-ID: <20240214003616.1382823-1-wathsala.vithanage@arm.com> (raw)
In-Reply-To: <20231208023801.3156065-1-liangxing.wang@arm.com>
The refcnt update of stored mbufs in memif driver is redundant since
those mbufs are only freed in eth_memif_tx_zc(). No other place can
free those stored mbufs quietly. By removing this redundant update
single core dpdk memif performance can be improved by 7.5%.
testpmd stats on Arm Neoverse N1 (Ampere Altra)
+-----------------------------+-----------------------+
| | With refcnt update | Without refcnt update |
+--------+--------------------+-----------------------+
| Rx-pps | 2748851 | 2955487 |
+--------+--------------------+-----------------------+
| Tx-pps | 2748812 | 2955436 |
+--------+--------------------+-----------------------+
Fixes: 43b815d88188 ("net/memif: support zero-copy slave")
Cc: jgrajcia@cisco.com
Signed-off-by: Liangxing Wang <liangxing.wang@arm.com>
Signed-off-by: Wathsala Vithanage <wathsala.vithanage@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
---
.mailmap | 2 ++
drivers/net/memif/rte_eth_memif.c | 6 ------
2 files changed, 2 insertions(+), 6 deletions(-)
diff --git a/.mailmap b/.mailmap
index de339562f4..00a642fcc3 100644
--- a/.mailmap
+++ b/.mailmap
@@ -791,6 +791,7 @@ Liang Ma <liangma@bytedance.com> <liangma@liangbit.com> <liang.j.ma@intel.com>
Liang-Min Larry Wang <liang-min.wang@intel.com>
Liang Xu <liang.xu@cinfotech.cn>
Liang Zhang <zhangliang@bigo.sg>
+Liangxing Wang <liangxing.wang@arm.com>
Li Feng <fengli@smartx.com>
Li Han <han.li1@zte.com.cn>
Lihong Ma <lihongx.ma@intel.com>
@@ -1506,6 +1507,7 @@ Vlad Zolotarov <vladz@cloudius-systems.com>
Vlastimil Kosar <kosar@rehivetech.com>
Volodymyr Fialko <vfialko@marvell.com>
Vu Pham <vuhuong@mellanox.com>
+Wathsala Vithanage <wathsala.vithanage@arm.com>
Wajeeh Atrash <atrwajee@amazon.com>
Walter Heymans <walter.heymans@corigine.com>
Wang Sheng-Hui <shhuiw@gmail.com>
diff --git a/drivers/net/memif/rte_eth_memif.c b/drivers/net/memif/rte_eth_memif.c
index 7cc8c0da91..962d390b90 100644
--- a/drivers/net/memif/rte_eth_memif.c
+++ b/drivers/net/memif/rte_eth_memif.c
@@ -265,8 +265,6 @@ memif_free_stored_mbufs(struct pmd_process_private *proc_private, struct memif_q
cur_tail = __atomic_load_n(&ring->tail, __ATOMIC_ACQUIRE);
while (mq->last_tail != cur_tail) {
RTE_MBUF_PREFETCH_TO_FREE(mq->buffers[(mq->last_tail + 1) & mask]);
- /* Decrement refcnt and free mbuf. (current segment) */
- rte_mbuf_refcnt_update(mq->buffers[mq->last_tail & mask], -1);
rte_pktmbuf_free_seg(mq->buffers[mq->last_tail & mask]);
mq->last_tail++;
}
@@ -825,10 +823,6 @@ memif_tx_one_zc(struct pmd_process_private *proc_private, struct memif_queue *mq
next_in_chain:
/* store pointer to mbuf to free it later */
mq->buffers[slot & mask] = mbuf;
- /* Increment refcnt to make sure the buffer is not freed before server
- * receives it. (current segment)
- */
- rte_mbuf_refcnt_update(mbuf, 1);
/* populate descriptor */
d0 = &ring->desc[slot & mask];
d0->length = rte_pktmbuf_data_len(mbuf);
--
2.25.1
next prev parent reply other threads:[~2024-02-14 0:37 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-08 2:38 [PATCH v1] net/memif: remove " Liangxing Wang
2023-12-08 13:44 ` Ferruh Yigit
2024-02-08 1:06 ` Ferruh Yigit
2024-02-14 0:36 ` Wathsala Vithanage [this message]
2024-02-14 11:09 ` [PATCH V2] net/memif: fix " Ferruh Yigit
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=20240214003616.1382823-1-wathsala.vithanage@arm.com \
--to=wathsala.vithanage@arm.com \
--cc=dev@dpdk.org \
--cc=ferruh.yigit@intel.com \
--cc=jgrajcia@cisco.com \
--cc=liangxing.wang@arm.com \
--cc=nd@arm.com \
--cc=ruifeng.wang@arm.com \
--cc=thomas@monjalon.net \
/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).