[PATCH] net/mlx5/hws: fix send queue drain on FW WQE destroy

Maayan Kashani mkashani at nvidia.com
Sun Apr 27 13:28:21 CEST 2025


Queue sync operation was skipped on rule destroy.
Unlike on fw wqe rule create in which both fence and notify_hw
are set to true, on destroy fence was set to false causing
previous queue operation to be stuck in the queue forever.
Example:
   rule_a - HW rule, rule_b - FW WQE rule.
Sequence:
   rule_a destroy, burst=1 (HW rule put to queue but no DB)
   rule_b destroy, burst=0 (FW WQE rule cmd but no queue sync)
Outcome:
   rule_a is stuck forever in the queue - no completion.

Fixes: 338aaf911665 ("net/mlx5/hws: add send FW match STE using gen WQE")
Cc: stable at dpdk.org

Signed-off-by: Alex Vesker <valex at nvidia.com>
Signed-off-by: Maayan Kashani <mkashani at nvidia.com>
---
 drivers/net/mlx5/hws/mlx5dr_send.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/mlx5/hws/mlx5dr_send.c b/drivers/net/mlx5/hws/mlx5dr_send.c
index e121c7f7ed5..d01fc7ef2ca 100644
--- a/drivers/net/mlx5/hws/mlx5dr_send.c
+++ b/drivers/net/mlx5/hws/mlx5dr_send.c
@@ -339,7 +339,7 @@ void mlx5dr_send_stes_fw(struct mlx5dr_send_engine *queue,
 	pdn = ctx->pd_num;
 
 	/* Writing through FW can't HW fence, therefore we drain the queue */
-	if (send_attr->fence)
+	if (send_attr->fence || send_attr->notify_hw)
 		mlx5dr_send_queue_action(ctx,
 					 queue_id,
 					 MLX5DR_SEND_QUEUE_ACTION_DRAIN_SYNC);
-- 
2.21.0



More information about the dev mailing list