Skip to contentSkip to conversation

net/mana: implement RX CQE coalescing

PATCH Applied [email protected] 2 participants 3 messages May 6, 2023 – May 19, 2023 Network drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net/mana: implement RX CQE coalescing

From: Long Li <[email protected]>

With RX coalescing, one CQE entry can be used to indicate up to 4 packets on the receive queue. This saves processing time and PCI bandwidth over the CQ.

---
 drivers/net/mana/mana.h |  4 +++
 drivers/net/mana/rx.c   | 71 ++++++++++++++++++++++++++++++++++-------
 2 files changed, 64 insertions(+), 11 deletions(-)
drivers/net/mana/mana.h +4 -0
index be88537f1a..dfeda531eb 100644
@@ -429,6 +429,10 @@ struct mana_rxq {
 	struct mana_gdma_queue gdma_cq;
 	struct gdma_comp *gdma_comp_buf;
 
+	uint32_t comp_buf_len;
+	uint32_t comp_buf_idx;
+	uint32_t backlog_idx;
+
 	struct mana_stats stats;
 	struct mana_mr_btree mr_btree;
 
drivers/net/mana/rx.c +60 -11

128 lines not on this page · raw message

Signature · 1 line
2.34.1
Applied
#1 lore raw
Quoted 8 lines · On 5/6/2023 2:33 AM, [email protected] wrote:

From: Long Li <[email protected]>

With RX coalescing, one CQE entry can be used to indicate up to 4 packets on the receive queue. This saves processing time and PCI bandwidth over the CQ.

Applied to dpdk-next-net/main, thanks.

Dropped stable tag in next-net, can add back again based on discussion with LTS maintainers.

#2 lore raw
In reply to [email protected]
Quoted 9 lines · On 5/6/2023 2:33 AM, [email protected] wrote:

From: Long Li <[email protected]>

With RX coalescing, one CQE entry can be used to indicate up to 4 packets on the receive queue. This saves processing time and PCI bandwidth over the CQ.

Backport to stable tree is requested but this is not a fix patch, cc'ed stable tree maintainers for comment.

stable tag removed while merging to next-net for now.