net/mlx5: WQ, fixes for fragmented WQ buffers API

mlx5e netdevice used to calculate fragment edges by a call to
mlx5_wq_cyc_get_frag_size(). This calculation did not give the correct
indication for queues smaller than a PAGE_SIZE, (broken by default on
PowerPC, where PAGE_SIZE == 64KB).  Here it is replaced by the correct new
calls/API.

Since (TX/RX) Work Queues buffers are fragmented, here we introduce
changes to the API in core driver, so that it gets a stride index and
returns the index of last stride on same fragment, and an additional
wrapping function that returns the number of physically contiguous
strides that can be written contiguously to the work queue.

This obsoletes the following API functions, and their buggy
usage in EN driver:
* mlx5_wq_cyc_get_frag_size()
* mlx5_wq_cyc_ctr2fragix()

The new API improves modularity and hides the details of such
calculation for mlx5e netdevice and mlx5_ib rdma drivers.

New calculation is also more efficient, and improves performance
as follows:

Packet rate test: pktgen, UDP / IPv4, 64byte, single ring, 8K ring size.

Before: 16,477,619 pps
After:  17,085,793 pps

3.7% improvement

Fixes: 3a2f70331226 ("net/mlx5: Use order-0 allocations for all WQ types")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

1 files changed, 11 insertions, 11 deletions

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
index ae73ea992845..6dacaeba2fbf 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/en_tx.c
@@ -290,10 +290,9 @@ dma_unmap_wqe_err:
 
 static inline void mlx5e_fill_sq_frag_edge(struct mlx5e_txqsq *sq,
 					   struct mlx5_wq_cyc *wq,
-					   u16 pi, u16 frag_pi)
+					   u16 pi, u16 nnops)
 {
 	struct mlx5e_tx_wqe_info *edge_wi, *wi = &sq->db.wqe_info[pi];
-	u8 nnops = mlx5_wq_cyc_get_frag_size(wq) - frag_pi;
 
 	edge_wi = wi + nnops;
 
@@ -348,8 +347,8 @@ netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 	struct mlx5e_tx_wqe_info *wi;
 
 	struct mlx5e_sq_stats *stats = sq->stats;
+	u16 headlen, ihs, contig_wqebbs_room;
 	u16 ds_cnt, ds_cnt_inl = 0;
-	u16 headlen, ihs, frag_pi;
 	u8 num_wqebbs, opcode;
 	u32 num_bytes;
 	int num_dma;
@@ -386,9 +385,9 @@ netdev_tx_t mlx5e_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 	}
 
 	num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
-	frag_pi = mlx5_wq_cyc_ctr2fragix(wq, sq->pc);
-	if (unlikely(frag_pi + num_wqebbs > mlx5_wq_cyc_get_frag_size(wq))) {
-		mlx5e_fill_sq_frag_edge(sq, wq, pi, frag_pi);
+	contig_wqebbs_room = mlx5_wq_cyc_get_contig_wqebbs(wq, pi);
+	if (unlikely(contig_wqebbs_room < num_wqebbs)) {
+		mlx5e_fill_sq_frag_edge(sq, wq, pi, contig_wqebbs_room);
 		mlx5e_sq_fetch_wqe(sq, &wqe, &pi);
 	}
 
@@ -636,7 +635,7 @@ netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 	struct mlx5e_tx_wqe_info *wi;
 
 	struct mlx5e_sq_stats *stats = sq->stats;
-	u16 headlen, ihs, pi, frag_pi;
+	u16 headlen, ihs, pi, contig_wqebbs_room;
 	u16 ds_cnt, ds_cnt_inl = 0;
 	u8 num_wqebbs, opcode;
 	u32 num_bytes;
@@ -672,13 +671,14 @@ netdev_tx_t mlx5i_sq_xmit(struct mlx5e_txqsq *sq, struct sk_buff *skb,
 	}
 
 	num_wqebbs = DIV_ROUND_UP(ds_cnt, MLX5_SEND_WQEBB_NUM_DS);
-	frag_pi = mlx5_wq_cyc_ctr2fragix(wq, sq->pc);
-	if (unlikely(frag_pi + num_wqebbs > mlx5_wq_cyc_get_frag_size(wq))) {
+	pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
+	contig_wqebbs_room = mlx5_wq_cyc_get_contig_wqebbs(wq, pi);
+	if (unlikely(contig_wqebbs_room < num_wqebbs)) {
+		mlx5e_fill_sq_frag_edge(sq, wq, pi, contig_wqebbs_room);
 		pi = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
-		mlx5e_fill_sq_frag_edge(sq, wq, pi, frag_pi);
 	}
 
-	mlx5i_sq_fetch_wqe(sq, &wqe, &pi);
+	mlx5i_sq_fetch_wqe(sq, &wqe, pi);
 
 	/* fill wqe */
 	wi       = &sq->db.wqe_info[pi];