[v4,11/14] crypto/mlx5: add WQE set initialization

Message ID 20210509160507.224644-12-matan@nvidia.com (mailing list archive)
State Changes Requested, archived
Delegated to: akhil goyal
Headers
Series drivers: introduce mlx5 crypto PMD |

Checks

Context Check Description
ci/checkpatch success coding style OK

Commit Message

Matan Azrad May 9, 2021, 4:05 p.m. UTC
From: Suanming Mou <suanmingm@nvidia.com>

Currently, HW handles the WQEs much faster than the software,
Using the constant WQE set layout can initialize most of the WQE
segments in advanced, and software only needs to configure very
limited segments in datapath. This accelerates the software WQE
organize in datapath.

This commit initializes the fixed WQE set segments.

Signed-off-by: Suanming Mou <suanmingm@nvidia.com>
Signed-off-by: Matan Azrad <matan@nvidia.com>
---
 drivers/crypto/mlx5/mlx5_crypto.c | 83 +++++++++++++++++++++++++++++--
 drivers/crypto/mlx5/mlx5_crypto.h | 10 +++-
 2 files changed, 87 insertions(+), 6 deletions(-)
  

Patch

diff --git a/drivers/crypto/mlx5/mlx5_crypto.c b/drivers/crypto/mlx5/mlx5_crypto.c
index 1a7317aef5..913276a4f3 100644
--- a/drivers/crypto/mlx5/mlx5_crypto.c
+++ b/drivers/crypto/mlx5/mlx5_crypto.c
@@ -269,6 +269,69 @@  mlx5_crypto_qp2rts(struct mlx5_crypto_qp *qp)
 	return 0;
 }
 
+static void
+mlx5_crypto_qp_init(struct mlx5_crypto_priv *priv, struct mlx5_crypto_qp *qp)
+{
+	uint32_t i;
+
+	for (i = 0 ; i < qp->entries_n; i++) {
+		struct mlx5_wqe_cseg *cseg = RTE_PTR_ADD(qp->umem_buf, i *
+							 priv->wqe_set_size);
+		struct mlx5_wqe_umr_cseg *ucseg = (struct mlx5_wqe_umr_cseg *)
+								     (cseg + 1);
+		struct mlx5_wqe_umr_bsf_seg *bsf =
+			(struct mlx5_wqe_umr_bsf_seg *)(RTE_PTR_ADD(cseg,
+						       priv->umr_wqe_size)) - 1;
+		struct mlx5_wqe_rseg *rseg;
+
+		/* Init UMR WQE. */
+		cseg->sq_ds = rte_cpu_to_be_32((qp->qp_obj->id << 8) |
+					 (priv->umr_wqe_size / MLX5_WSEG_SIZE));
+		cseg->flags = RTE_BE32(MLX5_COMP_ONLY_FIRST_ERR <<
+				       MLX5_COMP_MODE_OFFSET);
+		cseg->misc = rte_cpu_to_be_32(qp->mkey[i]->id);
+		ucseg->if_cf_toe_cq_res = RTE_BE32(1u << MLX5_UMRC_IF_OFFSET);
+		ucseg->mkey_mask = RTE_BE64(1u << 0); /* Mkey length bit. */
+		ucseg->ko_to_bs = rte_cpu_to_be_32
+			((RTE_ALIGN(priv->max_segs_num, 4u) <<
+			 MLX5_UMRC_KO_OFFSET) | (4 << MLX5_UMRC_TO_BS_OFFSET));
+		bsf->keytag = priv->keytag;
+		/* Init RDMA WRITE WQE. */
+		cseg = RTE_PTR_ADD(cseg, priv->umr_wqe_size);
+		cseg->flags = RTE_BE32((MLX5_COMP_ALWAYS <<
+				      MLX5_COMP_MODE_OFFSET) |
+				      MLX5_WQE_CTRL_INITIATOR_SMALL_FENCE);
+		rseg = (struct mlx5_wqe_rseg *)(cseg + 1);
+		rseg->rkey = rte_cpu_to_be_32(qp->mkey[i]->id);
+	}
+}
+
+static int
+mlx5_crypto_indirect_mkeys_prepare(struct mlx5_crypto_priv *priv,
+				  struct mlx5_crypto_qp *qp)
+{
+	struct mlx5_umr_wqe *umr;
+	uint32_t i;
+	struct mlx5_devx_mkey_attr attr = {
+		.pd = priv->pdn,
+		.umr_en = 1,
+		.crypto_en = 1,
+		.set_remote_rw = 1,
+		.klm_num = RTE_ALIGN(priv->max_segs_num, 4),
+	};
+
+	for (umr = (struct mlx5_umr_wqe *)qp->umem_buf, i = 0;
+	   i < qp->entries_n; i++, umr = RTE_PTR_ADD(umr, priv->wqe_set_size)) {
+		attr.klm_array = (struct mlx5_klm *)&umr->kseg[0];
+		qp->mkey[i] = mlx5_devx_cmd_mkey_create(priv->ctx, &attr);
+		if (!qp->mkey[i]) {
+			DRV_LOG(ERR, "Failed to allocate indirect mkey.");
+			return -1;
+		}
+	}
+	return 0;
+}
+
 static int
 mlx5_crypto_queue_pair_setup(struct rte_cryptodev *dev, uint16_t qp_id,
 			     const struct rte_cryptodev_qp_conf *qp_conf,
@@ -279,7 +342,7 @@  mlx5_crypto_queue_pair_setup(struct rte_cryptodev *dev, uint16_t qp_id,
 	struct mlx5_crypto_qp *qp;
 	uint16_t log_nb_desc = rte_log2_u32(qp_conf->nb_descriptors);
 	uint32_t umem_size = RTE_BIT32(log_nb_desc) *
-			      MLX5_CRYPTO_WQE_SET_SIZE +
+			      priv->wqe_set_size +
 			      sizeof(*qp->db_rec) * 2;
 	uint32_t alloc_size = sizeof(*qp);
 	struct mlx5_devx_cq_attr cq_attr = {
@@ -287,7 +350,9 @@  mlx5_crypto_queue_pair_setup(struct rte_cryptodev *dev, uint16_t qp_id,
 	};
 
 	alloc_size = RTE_ALIGN(alloc_size, RTE_CACHE_LINE_SIZE);
-	alloc_size += sizeof(struct rte_crypto_op *) * RTE_BIT32(log_nb_desc);
+	alloc_size += (sizeof(struct rte_crypto_op *) +
+		       sizeof(struct mlx5_devx_obj *)) *
+		       RTE_BIT32(log_nb_desc);
 	qp = rte_zmalloc_socket(__func__, alloc_size, RTE_CACHE_LINE_SIZE,
 				socket_id);
 	if (qp == NULL) {
@@ -331,8 +396,7 @@  mlx5_crypto_queue_pair_setup(struct rte_cryptodev *dev, uint16_t qp_id,
 	attr.wq_umem_id = qp->umem_obj->umem_id;
 	attr.wq_umem_offset = 0;
 	attr.dbr_umem_id = qp->umem_obj->umem_id;
-	attr.dbr_address = RTE_BIT64(log_nb_desc) *
-			   MLX5_CRYPTO_WQE_SET_SIZE;
+	attr.dbr_address = RTE_BIT64(log_nb_desc) * priv->wqe_set_size;
 	qp->qp_obj = mlx5_devx_cmd_create_qp(priv->ctx, &attr);
 	if (qp->qp_obj == NULL) {
 		DRV_LOG(ERR, "Failed to create QP(%u).", rte_errno);
@@ -341,8 +405,17 @@  mlx5_crypto_queue_pair_setup(struct rte_cryptodev *dev, uint16_t qp_id,
 	qp->db_rec = RTE_PTR_ADD(qp->umem_buf, (uintptr_t)attr.dbr_address);
 	if (mlx5_crypto_qp2rts(qp))
 		goto error;
-	qp->ops = (struct rte_crypto_op **)RTE_ALIGN((uintptr_t)(qp + 1),
+	qp->mkey = (struct mlx5_devx_obj **)RTE_ALIGN((uintptr_t)(qp + 1),
 							   RTE_CACHE_LINE_SIZE);
+	qp->ops = (struct rte_crypto_op **)(qp->mkey + RTE_BIT32(log_nb_desc));
+	qp->entries_n = 1 << log_nb_desc;
+	if (mlx5_crypto_indirect_mkeys_prepare(priv, qp)) {
+		DRV_LOG(ERR, "Cannot allocate indirect memory regions.");
+		rte_errno = ENOMEM;
+		goto error;
+	}
+	mlx5_crypto_qp_init(priv, qp);
+	qp->priv = priv;
 	dev->data->queue_pairs[qp_id] = qp;
 	return 0;
 error:
diff --git a/drivers/crypto/mlx5/mlx5_crypto.h b/drivers/crypto/mlx5/mlx5_crypto.h
index 81452bd700..52fcf5217f 100644
--- a/drivers/crypto/mlx5/mlx5_crypto.h
+++ b/drivers/crypto/mlx5/mlx5_crypto.h
@@ -16,7 +16,6 @@ 
 
 #define MLX5_CRYPTO_DEK_HTABLE_SZ (1 << 11)
 #define MLX5_CRYPTO_KEY_LENGTH 80
-#define MLX5_CRYPTO_WQE_SET_SIZE 1024
 
 struct mlx5_crypto_priv {
 	TAILQ_ENTRY(mlx5_crypto_priv) next;
@@ -24,6 +23,7 @@  struct mlx5_crypto_priv {
 	struct rte_pci_device *pci_dev;
 	struct rte_cryptodev *crypto_dev;
 	void *uar; /* User Access Region. */
+	volatile uint64_t *uar_addr;
 	uint32_t pdn; /* Protection Domain number. */
 	uint32_t max_segs_num; /* Maximum supported data segs. */
 	struct ibv_pd *pd;
@@ -40,13 +40,21 @@  struct mlx5_crypto_priv {
 };
 
 struct mlx5_crypto_qp {
+	struct mlx5_crypto_priv *priv;
 	struct mlx5_devx_cq cq_obj;
 	struct mlx5_devx_obj *qp_obj;
+	struct rte_cryptodev_stats stats;
 	struct mlx5dv_devx_umem *umem_obj;
 	void *umem_buf;
 	volatile uint32_t *db_rec;
 	struct rte_crypto_op **ops;
+	struct mlx5_devx_obj **mkey; /* WQE's indirect mekys. */
 	struct mlx5_mr_ctrl mr_ctrl;
+	uint8_t *wqe;
+	uint16_t entries_n;
+	uint16_t pi;
+	uint16_t ci;
+	uint16_t db_pi;
 };
 
 struct mlx5_crypto_dek {