net/mlx5e: Add control path for SHAMPO feature

This commit introduces the control path infrastructure for SHAMPO feature. SHAMPO feature enables packet stitching by splitting packets to header and payload, the header is placed on a dedicated buffer and the payload on the RX ring, this allows stitching the data part of a flow together continuously in the receive buffer. SHAMPO feature is implemented as linked list striding RQ feature. To support packets splitting and payload stitching: - Enlarge the ICOSQ and the correspond CQ to support the header buffer memory regions. - Add support to create linked list striding RQ with SHAMPO feature set in the open_rq function. - Add deallocation function and corresponded calls for SHAMPO header buffer. - Add mlx5e_create_umr_klm_mkey to support KLM mkey for the header buffer. - Rename mlx5e_create_umr_mkey to mlx5e_create_umr_mtt_mkey. Signed-off-by: Ben Ben-Ishay <benishay@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
mariux64 · Oct 27, 2021 · e5ca8fb · e5ca8fb
1 parent d7b896a
commit e5ca8fb
Show file tree

Hide file tree

Showing 6 changed files with 400 additions and 30 deletions.
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en.h b/drivers/net/ethernet/mellanox/mlx5/core/en.h
@@ -79,6 +79,11 @@ struct page_pool;
 				 SKB_DATA_ALIGN(sizeof(struct skb_shared_info)))
 
 #define MLX5E_RX_MAX_HEAD (256)
+#define MLX5E_SHAMPO_LOG_MAX_HEADER_ENTRY_SIZE (9)
+#define MLX5E_SHAMPO_WQ_HEADER_PER_PAGE (PAGE_SIZE >> MLX5E_SHAMPO_LOG_MAX_HEADER_ENTRY_SIZE)
+#define MLX5E_SHAMPO_WQ_BASE_HEAD_ENTRY_SIZE (64)
+#define MLX5E_SHAMPO_WQ_RESRV_SIZE (64 * 1024)
+#define MLX5E_SHAMPO_WQ_BASE_RESRV_SIZE (4096)
 
 #define MLX5_MPWRQ_MIN_LOG_STRIDE_SZ(mdev) \
 	(6 + MLX5_CAP_GEN(mdev, cache_line_128byte)) /* HW restriction */
@@ -273,6 +278,10 @@ enum packet_merge {
 struct mlx5e_packet_merge_param {
 	enum packet_merge type;
 	u32 timeout;
+	struct {
+		u8 match_criteria_type;
+		u8 alignment_granularity;
+	} shampo;
 };
 
 struct mlx5e_params {
@@ -319,7 +328,8 @@ enum {
 	MLX5E_RQ_STATE_NO_CSUM_COMPLETE,
 	MLX5E_RQ_STATE_CSUM_FULL, /* cqe_csum_full hw bit is set */
 	MLX5E_RQ_STATE_FPGA_TLS, /* FPGA TLS enabled */
-	MLX5E_RQ_STATE_MINI_CQE_HW_STRIDX /* set when mini_cqe_resp_stride_index cap is used */
+	MLX5E_RQ_STATE_MINI_CQE_HW_STRIDX, /* set when mini_cqe_resp_stride_index cap is used */
+	MLX5E_RQ_STATE_SHAMPO, /* set when SHAMPO cap is used */
 };
 
 struct mlx5e_cq {
@@ -610,6 +620,7 @@ typedef struct sk_buff *
 			 struct mlx5e_wqe_frag_info *wi, u32 cqe_bcnt);
 typedef bool (*mlx5e_fp_post_rx_wqes)(struct mlx5e_rq *rq);
 typedef void (*mlx5e_fp_dealloc_wqe)(struct mlx5e_rq*, u16);
+typedef void (*mlx5e_fp_shampo_dealloc_hd)(struct mlx5e_rq*, u16, u16, bool);
 
 int mlx5e_rq_set_handlers(struct mlx5e_rq *rq, struct mlx5e_params *params, bool xsk);
 void mlx5e_rq_set_trap_handlers(struct mlx5e_rq *rq, struct mlx5e_params *params);
@@ -631,6 +642,19 @@ struct mlx5e_rq_frags_info {
 	u8 wqe_bulk;
 };
 
+struct mlx5e_shampo_hd {
+	struct mlx5_core_mkey mkey;
+	struct mlx5e_dma_info *info;
+	struct page *last_page;
+	u16 hd_per_wq;
+	u16 hd_per_wqe;
+	unsigned long *bitmap;
+	u16 pi;
+	u16 ci;
+	__be32 key;
+	u64 last_addr;
+};
+
 struct mlx5e_rq {
 	/* data path */
 	union {
@@ -652,6 +676,7 @@ struct mlx5e_rq {
 			u8                     umr_in_progress;
 			u8                     umr_last_bulk;
 			u8                     umr_completed;
+			struct mlx5e_shampo_hd *shampo;
 		} mpwqe;
 	};
 	struct {
@@ -947,6 +972,7 @@ void mlx5e_build_ptys2ethtool_map(void);
 
 bool mlx5e_check_fragmented_striding_rq_cap(struct mlx5_core_dev *mdev);
 
+void mlx5e_shampo_dealloc_hd(struct mlx5e_rq *rq, u16 len, u16 start, bool close);
 void mlx5e_get_stats(struct net_device *dev, struct rtnl_link_stats64 *stats);
 void mlx5e_fold_sw_stats64(struct mlx5e_priv *priv, struct rtnl_link_stats64 *s);
 

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.c b/drivers/net/ethernet/mellanox/mlx5/core/en/params.c
@@ -139,6 +139,27 @@ u8 mlx5e_mpwqe_get_log_rq_size(struct mlx5e_params *params,
 	return params->log_rq_mtu_frames - log_pkts_per_wqe;
 }
 
+u8 mlx5e_shampo_get_log_hd_entry_size(struct mlx5_core_dev *mdev,
+				      struct mlx5e_params *params)
+{
+	return order_base_2(DIV_ROUND_UP(MLX5E_RX_MAX_HEAD, MLX5E_SHAMPO_WQ_BASE_HEAD_ENTRY_SIZE));
+}
+
+u8 mlx5e_shampo_get_log_rsrv_size(struct mlx5_core_dev *mdev,
+				  struct mlx5e_params *params)
+{
+	return order_base_2(MLX5E_SHAMPO_WQ_RESRV_SIZE / MLX5E_SHAMPO_WQ_BASE_RESRV_SIZE);
+}
+
+u8 mlx5e_shampo_get_log_pkt_per_rsrv(struct mlx5_core_dev *mdev,
+				     struct mlx5e_params *params)
+{
+	u32 resrv_size = BIT(mlx5e_shampo_get_log_rsrv_size(mdev, params)) *
+			 PAGE_SIZE;
+
+	return order_base_2(DIV_ROUND_UP(resrv_size, params->sw_mtu));
+}
+
 u8 mlx5e_mpwqe_get_log_stride_size(struct mlx5_core_dev *mdev,
 				   struct mlx5e_params *params,
 				   struct mlx5e_xsk_param *xsk)
@@ -443,6 +464,23 @@ static void mlx5e_build_common_cq_param(struct mlx5_core_dev *mdev,
 		MLX5_SET(cqc, cqc, cqe_sz, CQE_STRIDE_128_PAD);
 }
 
+static u32 mlx5e_shampo_get_log_cq_size(struct mlx5_core_dev *mdev,
+					struct mlx5e_params *params,
+					struct mlx5e_xsk_param *xsk)
+{
+	int rsrv_size = BIT(mlx5e_shampo_get_log_rsrv_size(mdev, params)) * PAGE_SIZE;
+	u16 num_strides = BIT(mlx5e_mpwqe_get_log_num_strides(mdev, params, xsk));
+	int pkt_per_rsrv = BIT(mlx5e_shampo_get_log_pkt_per_rsrv(mdev, params));
+	u8 log_stride_sz = mlx5e_mpwqe_get_log_stride_size(mdev, params, xsk);
+	int wq_size = BIT(mlx5e_mpwqe_get_log_rq_size(params, xsk));
+	int wqe_size = BIT(log_stride_sz) * num_strides;
+
+	/* +1 is for the case that the pkt_per_rsrv dont consume the reservation
+	 * so we get a filler cqe for the rest of the reservation.
+	 */
+	return order_base_2((wqe_size / rsrv_size) * wq_size * (pkt_per_rsrv + 1));
+}
+
 static void mlx5e_build_rx_cq_param(struct mlx5_core_dev *mdev,
 				    struct mlx5e_params *params,
 				    struct mlx5e_xsk_param *xsk,
@@ -454,9 +492,12 @@ static void mlx5e_build_rx_cq_param(struct mlx5_core_dev *mdev,
 
 	switch (params->rq_wq_type) {
 	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
-		log_cq_size = mlx5e_mpwqe_get_log_rq_size(params, xsk) +
-			mlx5e_mpwqe_get_log_num_strides(mdev, params, xsk);
 		hw_stridx = MLX5_CAP_GEN(mdev, mini_cqe_resp_stride_index);
+		if (params->packet_merge.type == MLX5E_PACKET_MERGE_SHAMPO)
+			log_cq_size = mlx5e_shampo_get_log_cq_size(mdev, params, xsk);
+		else
+			log_cq_size = mlx5e_mpwqe_get_log_rq_size(params, xsk) +
+				mlx5e_mpwqe_get_log_num_strides(mdev, params, xsk);
 		break;
 	default: /* MLX5_WQ_TYPE_CYCLIC */
 		log_cq_size = params->log_rq_mtu_frames;
@@ -511,6 +552,22 @@ int mlx5e_build_rq_param(struct mlx5_core_dev *mdev,
 		MLX5_SET(wq, wq, log_wqe_stride_size,
 			 log_wqe_stride_size - MLX5_MPWQE_LOG_STRIDE_SZ_BASE);
 		MLX5_SET(wq, wq, log_wq_sz, mlx5e_mpwqe_get_log_rq_size(params, xsk));
+		if (params->packet_merge.type == MLX5E_PACKET_MERGE_SHAMPO) {
+			MLX5_SET(wq, wq, shampo_enable, true);
+			MLX5_SET(wq, wq, log_reservation_size,
+				 mlx5e_shampo_get_log_rsrv_size(mdev, params));
+			MLX5_SET(wq, wq,
+				 log_max_num_of_packets_per_reservation,
+				 mlx5e_shampo_get_log_pkt_per_rsrv(mdev, params));
+			MLX5_SET(wq, wq, log_headers_entry_size,
+				 mlx5e_shampo_get_log_hd_entry_size(mdev, params));
+			MLX5_SET(rqc, rqc, reservation_timeout,
+				 params->packet_merge.timeout);
+			MLX5_SET(rqc, rqc, shampo_match_criteria_type,
+				 params->packet_merge.shampo.match_criteria_type);
+			MLX5_SET(rqc, rqc, shampo_no_match_alignment_granularity,
+				 params->packet_merge.shampo.alignment_granularity);
+		}
 		break;
 	}
 	default: /* MLX5_WQ_TYPE_CYCLIC */
@@ -611,17 +668,80 @@ static u8 mlx5e_get_rq_log_wq_sz(void *rqc)
 	return MLX5_GET(wq, wq, log_wq_sz);
 }
 
-static u8 mlx5e_build_icosq_log_wq_sz(struct mlx5e_params *params,
+/* This function calculates the maximum number of headers entries that are needed
+ * per WQE, the formula is based on the size of the reservations and the
+ * restriction we have about max packets for reservation that is equal to max
+ * headers per reservation.
+ */
+u32 mlx5e_shampo_hd_per_wqe(struct mlx5_core_dev *mdev,
+			    struct mlx5e_params *params,
+			    struct mlx5e_rq_param *rq_param)
+{
+	int resv_size = BIT(mlx5e_shampo_get_log_rsrv_size(mdev, params)) * PAGE_SIZE;
+	u16 num_strides = BIT(mlx5e_mpwqe_get_log_num_strides(mdev, params, NULL));
+	int pkt_per_resv = BIT(mlx5e_shampo_get_log_pkt_per_rsrv(mdev, params));
+	u8 log_stride_sz = mlx5e_mpwqe_get_log_stride_size(mdev, params, NULL);
+	int wqe_size = BIT(log_stride_sz) * num_strides;
+	u32 hd_per_wqe;
+
+	/* Assumption: hd_per_wqe % 8 == 0. */
+	hd_per_wqe = (wqe_size / resv_size) * pkt_per_resv;
+	mlx5_core_dbg(mdev, "%s hd_per_wqe = %d rsrv_size = %d wqe_size = %d pkt_per_resv = %d\n",
+		      __func__, hd_per_wqe, resv_size, wqe_size, pkt_per_resv);
+	return hd_per_wqe;
+}
+
+/* This function calculates the maximum number of headers entries that are needed
+ * for the WQ, this value is uesed to allocate the header buffer in HW, thus
+ * must be a pow of 2.
+ */
+u32 mlx5e_shampo_hd_per_wq(struct mlx5_core_dev *mdev,
+			   struct mlx5e_params *params,
+			   struct mlx5e_rq_param *rq_param)
+{
+	void *wqc = MLX5_ADDR_OF(rqc, rq_param->rqc, wq);
+	int wq_size = BIT(MLX5_GET(wq, wqc, log_wq_sz));
+	u32 hd_per_wqe, hd_per_wq;
+
+	hd_per_wqe = mlx5e_shampo_hd_per_wqe(mdev, params, rq_param);
+	hd_per_wq = roundup_pow_of_two(hd_per_wqe * wq_size);
+	return hd_per_wq;
+}
+
+static u32 mlx5e_shampo_icosq_sz(struct mlx5_core_dev *mdev,
+				 struct mlx5e_params *params,
+				 struct mlx5e_rq_param *rq_param)
+{
+	int max_num_of_umr_per_wqe, max_hd_per_wqe, max_klm_per_umr, rest;
+	void *wqc = MLX5_ADDR_OF(rqc, rq_param->rqc, wq);
+	int wq_size = BIT(MLX5_GET(wq, wqc, log_wq_sz));
+	u32 wqebbs;
+
+	max_klm_per_umr = MLX5E_MAX_KLM_PER_WQE(mdev);
+	max_hd_per_wqe = mlx5e_shampo_hd_per_wqe(mdev, params, rq_param);
+	max_num_of_umr_per_wqe = max_hd_per_wqe / max_klm_per_umr;
+	rest = max_hd_per_wqe % max_klm_per_umr;
+	wqebbs = MLX5E_KLM_UMR_WQEBBS(max_klm_per_umr) * max_num_of_umr_per_wqe;
+	if (rest)
+		wqebbs += MLX5E_KLM_UMR_WQEBBS(rest);
+	wqebbs *= wq_size;
+	return wqebbs;
+}
+
+static u8 mlx5e_build_icosq_log_wq_sz(struct mlx5_core_dev *mdev,
+				      struct mlx5e_params *params,
 				      struct mlx5e_rq_param *rqp)
 {
-	switch (params->rq_wq_type) {
-	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
-		return max_t(u8, MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE,
-			     order_base_2(MLX5E_UMR_WQEBBS) +
-			     mlx5e_get_rq_log_wq_sz(rqp->rqc));
-	default: /* MLX5_WQ_TYPE_CYCLIC */
+	u32 wqebbs;
+
+	/* MLX5_WQ_TYPE_CYCLIC */
+	if (params->rq_wq_type != MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ)
 		return MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE;
-	}
+
+	wqebbs = MLX5E_UMR_WQEBBS * BIT(mlx5e_get_rq_log_wq_sz(rqp->rqc));
+	if (params->packet_merge.type == MLX5E_PACKET_MERGE_SHAMPO)
+		wqebbs += mlx5e_shampo_icosq_sz(mdev, params, rqp);
+	return max_t(u8, MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE, order_base_2(wqebbs));
 }
 
 static u8 mlx5e_build_async_icosq_log_wq_sz(struct mlx5_core_dev *mdev)
@@ -688,7 +808,7 @@ int mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
 	if (err)
 		return err;
 
-	icosq_log_wq_sz = mlx5e_build_icosq_log_wq_sz(params, &cparam->rq);
+	icosq_log_wq_sz = mlx5e_build_icosq_log_wq_sz(mdev, params, &cparam->rq);
 	async_icosq_log_wq_sz = mlx5e_build_async_icosq_log_wq_sz(mdev);
 
 	mlx5e_build_sq_param(mdev, params, &cparam->txq_sq);

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/params.h b/drivers/net/ethernet/mellanox/mlx5/core/en/params.h
@@ -111,6 +111,18 @@ bool mlx5e_rx_mpwqe_is_linear_skb(struct mlx5_core_dev *mdev,
 				  struct mlx5e_xsk_param *xsk);
 u8 mlx5e_mpwqe_get_log_rq_size(struct mlx5e_params *params,
 			       struct mlx5e_xsk_param *xsk);
+u8 mlx5e_shampo_get_log_hd_entry_size(struct mlx5_core_dev *mdev,
+				      struct mlx5e_params *params);
+u8 mlx5e_shampo_get_log_rsrv_size(struct mlx5_core_dev *mdev,
+				  struct mlx5e_params *params);
+u8 mlx5e_shampo_get_log_pkt_per_rsrv(struct mlx5_core_dev *mdev,
+				     struct mlx5e_params *params);
+u32 mlx5e_shampo_hd_per_wqe(struct mlx5_core_dev *mdev,
+			    struct mlx5e_params *params,
+			    struct mlx5e_rq_param *rq_param);
+u32 mlx5e_shampo_hd_per_wq(struct mlx5_core_dev *mdev,
+			   struct mlx5e_params *params,
+			   struct mlx5e_rq_param *rq_param);
 u8 mlx5e_mpwqe_get_log_stride_size(struct mlx5_core_dev *mdev,
 				   struct mlx5e_params *params,
 				   struct mlx5e_xsk_param *xsk);

diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en/tir.c b/drivers/net/ethernet/mellanox/mlx5/core/en/tir.c
@@ -79,15 +79,21 @@ void mlx5e_tir_builder_build_packet_merge(struct mlx5e_tir_builder *builder,
 	if (builder->modify)
 		MLX5_SET(modify_tir_in, builder->in, bitmask.packet_merge, 1);
 
-	if (pkt_merge_param->type == MLX5E_PACKET_MERGE_NONE)
-		return;
-
-	MLX5_SET(tirc, tirc, packet_merge_mask,
-		 MLX5_TIRC_PACKET_MERGE_MASK_IPV4_LRO |
-		 MLX5_TIRC_PACKET_MERGE_MASK_IPV6_LRO);
-	MLX5_SET(tirc, tirc, lro_max_ip_payload_size,
-		 (MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ - rough_max_l2_l3_hdr_sz) >> 8);
-	MLX5_SET(tirc, tirc, lro_timeout_period_usecs, pkt_merge_param->timeout);
+	switch (pkt_merge_param->type) {
+	case MLX5E_PACKET_MERGE_LRO:
+		MLX5_SET(tirc, tirc, packet_merge_mask,
+			 MLX5_TIRC_PACKET_MERGE_MASK_IPV4_LRO |
+			 MLX5_TIRC_PACKET_MERGE_MASK_IPV6_LRO);
+		MLX5_SET(tirc, tirc, lro_max_ip_payload_size,
+			 (MLX5E_PARAMS_DEFAULT_LRO_WQE_SZ - rough_max_l2_l3_hdr_sz) >> 8);
+		MLX5_SET(tirc, tirc, lro_timeout_period_usecs, pkt_merge_param->timeout);
+		break;
+	case MLX5E_PACKET_MERGE_SHAMPO:
+		MLX5_SET(tirc, tirc, packet_merge_mask, MLX5_TIRC_PACKET_MERGE_MASK_SHAMPO);
+		break;
+	default:
+		break;
+	}
 }
 
 static int mlx5e_hfunc_to_hw(u8 hfunc)