net/mlx5e: Refactor XDP counters

Separate the XDP counters into two sets:
(1) One set reside in the RQ stats, and they monitor XDP stats
in the RQ side.
(2) Another set is per XDP-SQ, and they monitor XDP stats that
are related to XDP transmit flow.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Eugenia Emantayev <eugenia@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>

drivers/net/ethernet/mellanox/mlx5/core/en.h

@@ -424,6 +424,7 @@ struct mlx5e_xdpsq {
 
 	/* read only */
 	struct mlx5_wq_cyc         wq;
+	struct mlx5e_xdpsq_stats  *stats;
 	void __iomem              *uar_map;
 	u32                        sqn;
 	struct device             *pdev;
@@ -619,6 +620,7 @@ struct mlx5e_channel_stats {
 	struct mlx5e_ch_stats ch;
 	struct mlx5e_sq_stats sq[MLX5E_MAX_NUM_TC];
 	struct mlx5e_rq_stats rq;
+	struct mlx5e_xdpsq_stats rq_xdpsq;
 } ____cacheline_aligned_in_smp;
 
 enum mlx5e_traffic_types {

drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c

@@ -106,8 +106,6 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi)
 	u16                       pi   = mlx5_wq_cyc_ctr2ix(wq, sq->pc);
 	struct mlx5e_tx_wqe      *wqe  = mlx5_wq_cyc_get_wqe(wq, pi);
 
-	struct mlx5e_rq *rq = container_of(sq, struct mlx5e_rq, xdpsq);
-
 	struct mlx5_wqe_ctrl_seg *cseg = &wqe->ctrl;
 	struct mlx5_wqe_eth_seg  *eseg = &wqe->eth;
 	struct mlx5_wqe_data_seg *dseg = wqe->data;
@@ -116,12 +114,12 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi)
 	dma_addr_t dma_addr  = xdpi->dma_addr;
 	unsigned int dma_len = xdpf->len;
 
-	struct mlx5e_rq_stats *stats = rq->stats;
+	struct mlx5e_xdpsq_stats *stats = sq->stats;
 
 	prefetchw(wqe);
 
 	if (unlikely(dma_len < MLX5E_XDP_MIN_INLINE || sq->hw_mtu < dma_len)) {
-		stats->xdp_drop++;
+		stats->err++;
 		return false;
 	}
 
@@ -131,7 +129,7 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi)
 			mlx5e_xmit_xdp_doorbell(sq);
 			sq->db.doorbell = false;
 		}
-		stats->xdp_tx_full++;
+		stats->full++;
 		return false;
 	}
 
@@ -160,7 +158,7 @@ bool mlx5e_xmit_xdp_frame(struct mlx5e_xdpsq *sq, struct mlx5e_xdp_info *xdpi)
 
 	sq->db.doorbell = true;
 
-	stats->xdp_tx++;
+	stats->xmit++;
 	return true;
 }
 
@@ -212,7 +210,7 @@ bool mlx5e_poll_xdpsq_cq(struct mlx5e_cq *cq)
 		} while (!last_wqe);
 	} while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
 
-	rq->stats->xdp_tx_cqe += i;
+	sq->stats->cqes += i;
 
 	mlx5_cqwq_update_db_record(&cq->wq);
 

drivers/net/ethernet/mellanox/mlx5/core/en_main.c

@@ -1001,6 +1001,7 @@ static int mlx5e_alloc_xdpsq(struct mlx5e_channel *c,
 	sq->uar_map   = mdev->mlx5e_res.bfreg.map;
 	sq->min_inline_mode = params->tx_min_inline_mode;
 	sq->hw_mtu    = MLX5E_SW2HW_MTU(params, params->sw_mtu);
+	sq->stats     = &c->priv->channel_stats[c->ix].rq_xdpsq;
 
 	param->wq.db_numa_node = cpu_to_node(c->cpu);
 	err = mlx5_wq_cyc_create(mdev, &param->wq, sqc_wq, wq, &sq->wq_ctrl);

drivers/net/ethernet/mellanox/mlx5/core/en_stats.c

@@ -60,9 +60,10 @@ static const struct counter_desc sw_stats_desc[] = {
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_csum_unnecessary_inner) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_drop) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_redirect) },
-	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx) },
-	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_cqe) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_xmit) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_full) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_err) },
+	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, rx_xdp_tx_cqe) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_csum_none) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_csum_partial) },
 	{ MLX5E_DECLARE_STAT(struct mlx5e_sw_stats, tx_csum_partial_inner) },
@@ -129,6 +130,7 @@ void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv)
 	for (i = 0; i < priv->profile->max_nch(priv->mdev); i++) {
 		struct mlx5e_channel_stats *channel_stats =
 			&priv->channel_stats[i];
+		struct mlx5e_xdpsq_stats *xdpsq_stats = &channel_stats->rq_xdpsq;
 		struct mlx5e_rq_stats *rq_stats = &channel_stats->rq;
 		struct mlx5e_ch_stats *ch_stats = &channel_stats->ch;
 		int j;
@@ -144,9 +146,10 @@ void mlx5e_grp_sw_update_stats(struct mlx5e_priv *priv)
 		s->rx_csum_unnecessary_inner += rq_stats->csum_unnecessary_inner;
 		s->rx_xdp_drop     += rq_stats->xdp_drop;
 		s->rx_xdp_redirect += rq_stats->xdp_redirect;
-		s->rx_xdp_tx += rq_stats->xdp_tx;
-		s->rx_xdp_tx_cqe  += rq_stats->xdp_tx_cqe;
-		s->rx_xdp_tx_full += rq_stats->xdp_tx_full;
+		s->rx_xdp_tx_xmit  += xdpsq_stats->xmit;
+		s->rx_xdp_tx_full  += xdpsq_stats->full;
+		s->rx_xdp_tx_err   += xdpsq_stats->err;
+		s->rx_xdp_tx_cqe   += xdpsq_stats->cqes;
 		s->rx_wqe_err   += rq_stats->wqe_err;
 		s->rx_mpwqe_filler_cqes    += rq_stats->mpwqe_filler_cqes;
 		s->rx_mpwqe_filler_strides += rq_stats->mpwqe_filler_strides;
@@ -1129,9 +1132,6 @@ static const struct counter_desc rq_stats_desc[] = {
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, csum_none) },
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_drop) },
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_redirect) },
-	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_tx) },
-	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_tx_cqe) },
-	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, xdp_tx_full) },
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, lro_packets) },
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, lro_bytes) },
 	{ MLX5E_DECLARE_RX_STAT(struct mlx5e_rq_stats, removed_vlan_packets) },
@@ -1171,6 +1171,13 @@ static const struct counter_desc sq_stats_desc[] = {
 	{ MLX5E_DECLARE_TX_STAT(struct mlx5e_sq_stats, cqe_err) },
 };
 
+static const struct counter_desc rq_xdpsq_stats_desc[] = {
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, xmit) },
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, full) },
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, err) },
+	{ MLX5E_DECLARE_RQ_XDPSQ_STAT(struct mlx5e_xdpsq_stats, cqes) },
+};
+
 static const struct counter_desc ch_stats_desc[] = {
 	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, events) },
 	{ MLX5E_DECLARE_CH_STAT(struct mlx5e_ch_stats, poll) },
@@ -1181,6 +1188,7 @@ static const struct counter_desc ch_stats_desc[] = {
 
 #define NUM_RQ_STATS			ARRAY_SIZE(rq_stats_desc)
 #define NUM_SQ_STATS			ARRAY_SIZE(sq_stats_desc)
+#define NUM_RQ_XDPSQ_STATS		ARRAY_SIZE(rq_xdpsq_stats_desc)
 #define NUM_CH_STATS			ARRAY_SIZE(ch_stats_desc)
 
 static int mlx5e_grp_channels_get_num_stats(struct mlx5e_priv *priv)
@@ -1189,7 +1197,8 @@ static int mlx5e_grp_channels_get_num_stats(struct mlx5e_priv *priv)
 
 	return (NUM_RQ_STATS * max_nch) +
 	       (NUM_CH_STATS * max_nch) +
-	       (NUM_SQ_STATS * max_nch * priv->max_opened_tc);
+	       (NUM_SQ_STATS * max_nch * priv->max_opened_tc) +
+	       (NUM_RQ_XDPSQ_STATS * max_nch);
 }
 
 static int mlx5e_grp_channels_fill_strings(struct mlx5e_priv *priv, u8 *data,
@@ -1203,9 +1212,14 @@ static int mlx5e_grp_channels_fill_strings(struct mlx5e_priv *priv, u8 *data,
 			sprintf(data + (idx++) * ETH_GSTRING_LEN,
 				ch_stats_desc[j].format, i);
 
-	for (i = 0; i < max_nch; i++)
+	for (i = 0; i < max_nch; i++) {
 		for (j = 0; j < NUM_RQ_STATS; j++)
-			sprintf(data + (idx++) * ETH_GSTRING_LEN, rq_stats_desc[j].format, i);
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				rq_stats_desc[j].format, i);
+		for (j = 0; j < NUM_RQ_XDPSQ_STATS; j++)
+			sprintf(data + (idx++) * ETH_GSTRING_LEN,
+				rq_xdpsq_stats_desc[j].format, i);
+	}
 
 	for (tc = 0; tc < priv->max_opened_tc; tc++)
 		for (i = 0; i < max_nch; i++)
@@ -1229,11 +1243,16 @@ static int mlx5e_grp_channels_fill_stats(struct mlx5e_priv *priv, u64 *data,
 				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].ch,
 						     ch_stats_desc, j);
 
-	for (i = 0; i < max_nch; i++)
+	for (i = 0; i < max_nch; i++) {
 		for (j = 0; j < NUM_RQ_STATS; j++)
 			data[idx++] =
 				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].rq,
 						     rq_stats_desc, j);
+		for (j = 0; j < NUM_RQ_XDPSQ_STATS; j++)
+			data[idx++] =
+				MLX5E_READ_CTR64_CPU(&priv->channel_stats[i].rq_xdpsq,
+						     rq_xdpsq_stats_desc, j);
+	}
 
 	for (tc = 0; tc < priv->max_opened_tc; tc++)
 		for (i = 0; i < max_nch; i++)

drivers/net/ethernet/mellanox/mlx5/core/en_stats.h

@@ -44,6 +44,7 @@
 #define MLX5E_DECLARE_STAT(type, fld) #fld, offsetof(type, fld)
 #define MLX5E_DECLARE_RX_STAT(type, fld) "rx%d_"#fld, offsetof(type, fld)
 #define MLX5E_DECLARE_TX_STAT(type, fld) "tx%d_"#fld, offsetof(type, fld)
+#define MLX5E_DECLARE_RQ_XDPSQ_STAT(type, fld) "rx%d_xdp_tx_"#fld, offsetof(type, fld)
 #define MLX5E_DECLARE_CH_STAT(type, fld) "ch%d_"#fld, offsetof(type, fld)
 
 struct counter_desc {
@@ -71,9 +72,10 @@ struct mlx5e_sw_stats {
 	u64 rx_csum_unnecessary_inner;
 	u64 rx_xdp_drop;
 	u64 rx_xdp_redirect;
-	u64 rx_xdp_tx;
-	u64 rx_xdp_tx_cqe;
+	u64 rx_xdp_tx_xmit;
 	u64 rx_xdp_tx_full;
+	u64 rx_xdp_tx_err;
+	u64 rx_xdp_tx_cqe;
 	u64 tx_csum_none;
 	u64 tx_csum_partial;
 	u64 tx_csum_partial_inner;
@@ -180,9 +182,6 @@ struct mlx5e_rq_stats {
 	u64 removed_vlan_packets;
 	u64 xdp_drop;
 	u64 xdp_redirect;
-	u64 xdp_tx;
-	u64 xdp_tx_cqe;
-	u64 xdp_tx_full;
 	u64 wqe_err;
 	u64 mpwqe_filler_cqes;
 	u64 mpwqe_filler_strides;
@@ -227,6 +226,13 @@ struct mlx5e_sq_stats {
 	u64 cqe_err;
 };
 
+struct mlx5e_xdpsq_stats {
+	u64 xmit;
+	u64 full;
+	u64 err;
+	u64 cqes;
+};
+
 struct mlx5e_ch_stats {
 	u64 events;
 	u64 poll;