net/mlx5e: Enable adaptive-TX moderation

Add support for adaptive TX moderation. This greatly reduces TX interrupt
rate and increases bandwidth, mostly for TCP bandwidth over ARM
architecture (below). There is a slight single stream TCP with very large
message sizes degradation (x86). In this case if there's any moderation on
transmitted packets the bandwidth would reduce due to hitting TCP output limit.
Since this is a synthetic case, this is still worth doing.

Performance improvement (ConnectX-4Lx 40GbE, ARM)
TCP 64B bandwidth with 1-50 streams increased 6-35%.
TCP 64B bandwidth with 100-500 streams increased 20-70%.

Performance improvement (ConnectX-5 100GbE, x86)
Bandwidth: increased up to 40% (1024B with 10s of streams).
Interrupt rate: reduced up to 50% (1024B with 1000s of streams).

Performance degradation (ConnectX-5 100GbE, x86)
Bandwidth: up to 10% decrease single stream TCP (1MB message size from
51Gb/s to 47Gb/s).

Signed-off-by: Tal Gilboa <talgi@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 19 insertions, 5 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
index 1b286e1985d2..d67adf70a97b 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_dim.c
@@ -33,16 +33,30 @@
 #include <linux/net_dim.h>
 #include "en.h"
 
+static void
+mlx5e_complete_dim_work(struct net_dim *dim, struct net_dim_cq_moder moder,
+			struct mlx5_core_dev *mdev, struct mlx5_core_cq *mcq)
+{
+	mlx5_core_modify_cq_moderation(mdev, mcq, moder.usec, moder.pkts);
+	dim->state = NET_DIM_START_MEASURE;
+}
+
 void mlx5e_rx_dim_work(struct work_struct *work)
 {
-	struct net_dim *dim = container_of(work, struct net_dim,
-					   work);
+	struct net_dim *dim = container_of(work, struct net_dim, work);
 	struct mlx5e_rq *rq = container_of(dim, struct mlx5e_rq, dim);
 	struct net_dim_cq_moder cur_moder =
 		net_dim_get_rx_moderation(dim->mode, dim->profile_ix);
 
-	mlx5_core_modify_cq_moderation(rq->mdev, &rq->cq.mcq,
-				       cur_moder.usec, cur_moder.pkts);
+	mlx5e_complete_dim_work(dim, cur_moder, rq->mdev, &rq->cq.mcq);
+}
 
-	dim->state = NET_DIM_START_MEASURE;
+void mlx5e_tx_dim_work(struct work_struct *work)
+{
+	struct net_dim *dim = container_of(work, struct net_dim, work);
+	struct mlx5e_txqsq *sq = container_of(dim, struct mlx5e_txqsq, dim);
+	struct net_dim_cq_moder cur_moder =
+		net_dim_get_tx_moderation(dim->mode, dim->profile_ix);
+
+	mlx5e_complete_dim_work(dim, cur_moder, sq->cq.mdev, &sq->cq.mcq);
 }