From: Jiawen Wu <jiawenwu@trustnetic.com>
To: dev@dpdk.org
Cc: zaiyuwang@trustnetic.com, Jiawen Wu <jiawenwu@trustnetic.com>,
stable@dpdk.org
Subject: [PATCH 04/19] net/ngbe: reduce memory size of ring descriptors
Date: Mon, 27 Oct 2025 11:15:27 +0800 [thread overview]
Message-ID: <20251027031542.10512-5-jiawenwu@trustnetic.com> (raw)
In-Reply-To: <20251027031542.10512-1-jiawenwu@trustnetic.com>
The memory of ring descriptors was allocated in size of the maximum ring
size. It seems not friendly to our hardware on some domestic platforms.
Change it to allocate in size of the real ring size.
Fixes: 43b7e5ea60ac ("net/ngbe: support Rx queue setup/release")
Cc: stable@dpdk.org
Signed-off-by: Jiawen Wu <jiawenwu@trustnetic.com>
---
drivers/net/ngbe/ngbe_rxtx.c | 20 +++++++-------------
1 file changed, 7 insertions(+), 13 deletions(-)
diff --git a/drivers/net/ngbe/ngbe_rxtx.c b/drivers/net/ngbe/ngbe_rxtx.c
index a60421293b..03ada844bf 100644
--- a/drivers/net/ngbe/ngbe_rxtx.c
+++ b/drivers/net/ngbe/ngbe_rxtx.c
@@ -2058,13 +2058,9 @@ ngbe_dev_tx_queue_setup(struct rte_eth_dev *dev,
if (txq == NULL)
return -ENOMEM;
- /*
- * Allocate Tx ring hardware descriptors. A memzone large enough to
- * handle the maximum ring size is allocated in order to allow for
- * resizing in later calls to the queue setup function.
- */
+ /* Allocate Tx ring hardware descriptors. */
tz = rte_eth_dma_zone_reserve(dev, "tx_ring", queue_idx,
- sizeof(struct ngbe_tx_desc) * NGBE_RING_DESC_MAX,
+ sizeof(struct ngbe_tx_desc) * nb_desc,
NGBE_ALIGN, socket_id);
if (tz == NULL) {
ngbe_tx_queue_release(txq);
@@ -2324,6 +2320,7 @@ ngbe_dev_rx_queue_setup(struct rte_eth_dev *dev,
uint16_t len;
struct ngbe_adapter *adapter = ngbe_dev_adapter(dev);
uint64_t offloads;
+ uint32_t size;
PMD_INIT_FUNC_TRACE();
hw = ngbe_dev_hw(dev);
@@ -2357,13 +2354,10 @@ ngbe_dev_rx_queue_setup(struct rte_eth_dev *dev,
rxq->rx_deferred_start = rx_conf->rx_deferred_start;
rxq->offloads = offloads;
- /*
- * Allocate Rx ring hardware descriptors. A memzone large enough to
- * handle the maximum ring size is allocated in order to allow for
- * resizing in later calls to the queue setup function.
- */
+ /* Allocate Rx ring hardware descriptors. */
+ size = (nb_desc + RTE_PMD_NGBE_RX_MAX_BURST) * sizeof(struct ngbe_rx_desc);
rz = rte_eth_dma_zone_reserve(dev, "rx_ring", queue_idx,
- RX_RING_SZ, NGBE_ALIGN, socket_id);
+ size, NGBE_ALIGN, socket_id);
if (rz == NULL) {
ngbe_rx_queue_release(rxq);
return -ENOMEM;
@@ -2373,7 +2367,7 @@ ngbe_dev_rx_queue_setup(struct rte_eth_dev *dev,
/*
* Zero init all the descriptors in the ring.
*/
- memset(rz->addr, 0, RX_RING_SZ);
+ memset(rz->addr, 0, size);
rxq->rdt_reg_addr = NGBE_REG_ADDR(hw, NGBE_RXWP(rxq->reg_idx));
rxq->rdh_reg_addr = NGBE_REG_ADDR(hw, NGBE_RXRP(rxq->reg_idx));
--
2.48.1
next prev parent reply other threads:[~2025-10-27 3:16 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-27 3:15 [PATCH 00/19] Wangxun Fixes Jiawen Wu
2025-10-27 3:15 ` [PATCH 01/19] net/txgbe: fix hardware statistic rx_l3_l4_xsum_error Jiawen Wu
2025-10-27 3:15 ` [PATCH 02/19] net/ngbe: " Jiawen Wu
2025-10-27 3:15 ` [PATCH 03/19] net/txgbe: reduce memory size of ring descriptors Jiawen Wu
2025-10-27 3:15 ` Jiawen Wu [this message]
2025-10-27 3:15 ` [PATCH 05/19] net/txgbe: fix VF Rx buffer size in config register Jiawen Wu
2025-10-27 3:15 ` [PATCH 06/19] net/ngbe: " Jiawen Wu
2025-10-27 3:15 ` [PATCH 07/19] net/txgbe: remove duplicate txq assignment Jiawen Wu
2025-10-27 3:15 ` [PATCH 08/19] net/txgbe: add device arguments for FDIR Jiawen Wu
2025-10-27 3:15 ` [PATCH 09/19] net/txgbe: fix the maxinum number of FDIR filter Jiawen Wu
2025-10-27 3:15 ` [PATCH 10/19] net/txgbe: fix FDIR mode is not be cleared Jiawen Wu
2025-10-27 3:15 ` [PATCH 11/19] net/txgbe: fix FDIR drop action for L4 match packets Jiawen Wu
2025-10-27 3:15 ` [PATCH 12/19] net/txgbe: fix to create FDIR filter for tunnel SCTP packet Jiawen Wu
2025-10-27 3:15 ` [PATCH 13/19] net/txgbe: filter FDIR match flex bytes for tunnel packets Jiawen Wu
2025-10-27 3:15 ` [PATCH 14/19] net/txgbe: fix FDIR rule raw relative for L3 packets Jiawen Wu
2025-10-27 3:15 ` [PATCH 15/19] net/txgbe: fix FDIR input mask Jiawen Wu
2025-10-27 3:15 ` [PATCH 16/19] net/txgbe: switch to use FDIR when ntuple filter is full Jiawen Wu
2025-10-27 3:15 ` [PATCH 17/19] net/txgbe: fix VF-PF message for ntuple flow filter Jiawen Wu
2025-10-27 3:15 ` [PATCH 18/19] net/txgbe: switch to use FDIR on VF Jiawen Wu
2025-10-27 3:15 ` [PATCH 19/19] net/txgbe: remove unsupported flow action mark Jiawen Wu
2025-10-27 16:51 ` [PATCH 00/19] Wangxun Fixes Stephen Hemminger
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=20251027031542.10512-5-jiawenwu@trustnetic.com \
--to=jiawenwu@trustnetic.com \
--cc=dev@dpdk.org \
--cc=stable@dpdk.org \
--cc=zaiyuwang@trustnetic.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).