From: Devendra Singh Rawat <dsinghrawat@marvell.com>
To: <stable@dpdk.org>
Cc: <rmody@marvell.com>, <palok@marvell.com>,
Devendra Singh Rawat <dsinghrawat@marvell.com>
Subject: [PATCH 20.11] net/qede: fix Rx bulk mbuf allocation
Date: Wed, 16 Mar 2022 17:33:06 +0530 [thread overview]
Message-ID: <20220316120306.561108-1-dsinghrawat@marvell.com> (raw)
[ upstream commit f65c7fbceca91b54200ca3dc5d27f2292e5d829f ]
qede_alloc_rx_bulk_mbufs() was trimming the number of requested
mbufs count to QEDE_MAX_BULK_ALLOC_COUNT.
The Rx callback was ignorant of this trimming and it was always
resetting the number of empty RX BD ring slots to 0.
This resulted in Rx BD ring getting into an inconsistent
state and ultimately the application fails to receive any traffic.
The fix trims the number of requested mbufs count before
making call to qede_alloc_rx_bulk_mbufs().
After qede_alloc_rx_bulk_mbufs() returns successfully, the
number of empty Rx BD ring slots are decremented by the
correct count.
Fixes: 8f2312474529 ("net/qede: fix performance bottleneck in Rx path")
Signed-off-by: Devendra Singh Rawat <dsinghrawat@marvell.com>
Signed-off-by: Rasesh Mody <rmody@marvell.com>
---
drivers/net/qede/qede_rxtx.c | 49 ++++++++++++++++++------------------
1 file changed, 24 insertions(+), 25 deletions(-)
diff --git a/drivers/net/qede/qede_rxtx.c b/drivers/net/qede/qede_rxtx.c
index 122e9290ed..f357a8f258 100644
--- a/drivers/net/qede/qede_rxtx.c
+++ b/drivers/net/qede/qede_rxtx.c
@@ -46,17 +46,14 @@ static inline int qede_alloc_rx_bulk_mbufs(struct qede_rx_queue *rxq, int count)
int i, ret = 0;
uint16_t idx;
- if (count > QEDE_MAX_BULK_ALLOC_COUNT)
- count = QEDE_MAX_BULK_ALLOC_COUNT;
+ idx = rxq->sw_rx_prod & NUM_RX_BDS(rxq);
ret = rte_mempool_get_bulk(rxq->mb_pool, obj_p, count);
if (unlikely(ret)) {
PMD_RX_LOG(ERR, rxq,
"Failed to allocate %d rx buffers "
"sw_rx_prod %u sw_rx_cons %u mp entries %u free %u",
- count,
- rxq->sw_rx_prod & NUM_RX_BDS(rxq),
- rxq->sw_rx_cons & NUM_RX_BDS(rxq),
+ count, idx, rxq->sw_rx_cons & NUM_RX_BDS(rxq),
rte_mempool_avail_count(rxq->mb_pool),
rte_mempool_in_use_count(rxq->mb_pool));
return -ENOMEM;
@@ -1542,25 +1539,26 @@ qede_recv_pkts_regular(void *p_rxq, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
uint8_t bitfield_val;
#endif
uint8_t offset, flags, bd_num;
-
+ uint16_t count = 0;
/* Allocate buffers that we used in previous loop */
if (rxq->rx_alloc_count) {
- if (unlikely(qede_alloc_rx_bulk_mbufs(rxq,
- rxq->rx_alloc_count))) {
+ count = rxq->rx_alloc_count > QEDE_MAX_BULK_ALLOC_COUNT ?
+ QEDE_MAX_BULK_ALLOC_COUNT : rxq->rx_alloc_count;
+
+ if (unlikely(qede_alloc_rx_bulk_mbufs(rxq, count))) {
struct rte_eth_dev *dev;
PMD_RX_LOG(ERR, rxq,
- "New buffer allocation failed,"
- "dropping incoming packetn");
+ "New buffers allocation failed,"
+ "dropping incoming packets\n");
dev = &rte_eth_devices[rxq->port_id];
- dev->data->rx_mbuf_alloc_failed +=
- rxq->rx_alloc_count;
- rxq->rx_alloc_errors += rxq->rx_alloc_count;
+ dev->data->rx_mbuf_alloc_failed += count;
+ rxq->rx_alloc_errors += count;
return 0;
}
qede_update_rx_prod(qdev, rxq);
- rxq->rx_alloc_count = 0;
+ rxq->rx_alloc_count -= count;
}
hw_comp_cons = rte_le_to_cpu_16(*rxq->hw_cons_ptr);
@@ -1729,7 +1727,7 @@ qede_recv_pkts_regular(void *p_rxq, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
}
/* Request number of buffers to be allocated in next loop */
- rxq->rx_alloc_count = rx_alloc_count;
+ rxq->rx_alloc_count += rx_alloc_count;
rxq->rcv_pkts += rx_pkt;
rxq->rx_segs += rx_pkt;
@@ -1769,25 +1767,26 @@ qede_recv_pkts(void *p_rxq, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
struct qede_agg_info *tpa_info = NULL;
uint32_t rss_hash;
int rx_alloc_count = 0;
-
+ uint16_t count = 0;
/* Allocate buffers that we used in previous loop */
if (rxq->rx_alloc_count) {
- if (unlikely(qede_alloc_rx_bulk_mbufs(rxq,
- rxq->rx_alloc_count))) {
+ count = rxq->rx_alloc_count > QEDE_MAX_BULK_ALLOC_COUNT ?
+ QEDE_MAX_BULK_ALLOC_COUNT : rxq->rx_alloc_count;
+
+ if (unlikely(qede_alloc_rx_bulk_mbufs(rxq, count))) {
struct rte_eth_dev *dev;
PMD_RX_LOG(ERR, rxq,
- "New buffer allocation failed,"
- "dropping incoming packetn");
+ "New buffers allocation failed,"
+ "dropping incoming packets\n");
dev = &rte_eth_devices[rxq->port_id];
- dev->data->rx_mbuf_alloc_failed +=
- rxq->rx_alloc_count;
- rxq->rx_alloc_errors += rxq->rx_alloc_count;
+ dev->data->rx_mbuf_alloc_failed += count;
+ rxq->rx_alloc_errors += count;
return 0;
}
qede_update_rx_prod(qdev, rxq);
- rxq->rx_alloc_count = 0;
+ rxq->rx_alloc_count -= count;
}
hw_comp_cons = rte_le_to_cpu_16(*rxq->hw_cons_ptr);
@@ -2026,7 +2025,7 @@ qede_recv_pkts(void *p_rxq, struct rte_mbuf **rx_pkts, uint16_t nb_pkts)
}
/* Request number of buffers to be allocated in next loop */
- rxq->rx_alloc_count = rx_alloc_count;
+ rxq->rx_alloc_count += rx_alloc_count;
rxq->rcv_pkts += rx_pkt;
--
2.18.2
reply other threads:[~2022-03-16 12:04 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20220316120306.561108-1-dsinghrawat@marvell.com \
--to=dsinghrawat@marvell.com \
--cc=palok@marvell.com \
--cc=rmody@marvell.com \
--cc=stable@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).