Pass send queue sizes from userspace to kernel

Update to handle kernel mlx4 ABI version 2: pass log_2 of send queue
WQE basic block size and log_2 of number of send queue basic blocks to
the kernel to avoid bugs caused by the kernel calculating a different
send queue WQE size.  This will also allow us to use multiple BBs per
WQE if we want to someday.

Signed-off-by: Roland Dreier <rolandd@cisco.com>

5 files changed, 51 insertions, 11 deletions

diff --git a/src/mlx4-abi.h b/src/mlx4-abi.h
index 7810913..97f5dcd 100644
--- a/src/mlx4-abi.h
+++ b/src/mlx4-abi.h
@@ -35,7 +35,8 @@
 
 #include <infiniband/kern-abi.h>
 
-#define MLX4_UVERBS_ABI_VERSION	1
+#define MLX4_UVERBS_MIN_ABI_VERSION	2
+#define MLX4_UVERBS_MAX_ABI_VERSION	2
 
 struct mlx4_alloc_ucontext_resp {
 	struct ibv_get_context_resp	ibv_resp;
@@ -83,6 +84,9 @@ struct mlx4_create_qp {
 	struct ibv_create_qp		ibv_cmd;
 	__u64				buf_addr;
 	__u64				db_addr;
+	__u8				log_sq_bb_count;
+	__u8				log_sq_stride;
+	__u8				reserved[6];
 };
 
 #endif /* MLX4_ABI_H */
diff --git a/src/mlx4.c b/src/mlx4.c
index 23577c1..ac522a7 100644
--- a/src/mlx4.c
+++ b/src/mlx4.c
@@ -250,9 +250,13 @@ static struct ibv_device *mlx4_driver_init(const char *uverbs_sys_path,
 	return NULL;
 
 found:
-	if (abi_version > MLX4_UVERBS_ABI_VERSION) {
-		fprintf(stderr, PFX "Fatal: ABI version %d of %s is too new (expected %d)\n",
-			abi_version, uverbs_sys_path, MLX4_UVERBS_ABI_VERSION);
+	if (abi_version < MLX4_UVERBS_MIN_ABI_VERSION ||
+	    abi_version > MLX4_UVERBS_MAX_ABI_VERSION) {
+		fprintf(stderr, PFX "Fatal: ABI version %d of %s is not supported "
+			"(min supported %d, max supported %d)\n",
+			abi_version, uverbs_sys_path,
+			MLX4_UVERBS_MIN_ABI_VERSION,
+			MLX4_UVERBS_MAX_ABI_VERSION);
 		return NULL;
 	}
 
diff --git a/src/mlx4.h b/src/mlx4.h
index 1e92b88..7978f9f 100644
--- a/src/mlx4.h
+++ b/src/mlx4.h
@@ -343,6 +343,8 @@ int mlx4_post_recv(struct ibv_qp *ibqp, struct ibv_recv_wr *wr,
 			  struct ibv_recv_wr **bad_wr);
 int mlx4_alloc_qp_buf(struct ibv_pd *pd, struct ibv_qp_cap *cap,
 		       enum ibv_qp_type type, struct mlx4_qp *qp);
+void mlx4_set_sq_sizes(struct mlx4_qp *qp, struct ibv_qp_cap *cap,
+		       enum ibv_qp_type type);
 struct mlx4_qp *mlx4_find_qp(struct mlx4_context *ctx, uint32_t qpn);
 int mlx4_store_qp(struct mlx4_context *ctx, uint32_t qpn, struct mlx4_qp *qp);
 void mlx4_clear_qp(struct mlx4_context *ctx, uint32_t qpn);
diff --git a/src/qp.c b/src/qp.c
index a4384f9..bff3f33 100644
--- a/src/qp.c
+++ b/src/qp.c
@@ -469,6 +469,32 @@ int mlx4_alloc_qp_buf(struct ibv_pd *pd, struct ibv_qp_cap *cap,
 	return 0;
 }
 
+void mlx4_set_sq_sizes(struct mlx4_qp *qp, struct ibv_qp_cap *cap,
+		       enum ibv_qp_type type)
+{
+	int wqe_size;
+
+	wqe_size = 1 << qp->sq.wqe_shift;
+	switch (type) {
+	case IBV_QPT_UD:
+		wqe_size -= sizeof (struct mlx4_wqe_datagram_seg);
+		break;
+
+	case IBV_QPT_UC:
+	case IBV_QPT_RC:
+		wqe_size -= sizeof (struct mlx4_wqe_raddr_seg);
+		break;
+
+	default:
+		break;
+	}
+
+	qp->sq.max_gs        = wqe_size / sizeof (struct mlx4_wqe_data_seg);
+	cap->max_send_sge    = qp->sq.max_gs;
+	qp->max_inline_data  = wqe_size - sizeof (struct mlx4_wqe_inline_seg);
+	cap->max_inline_data = qp->max_inline_data;
+}
+
 struct mlx4_qp *mlx4_find_qp(struct mlx4_context *ctx, uint32_t qpn)
 {
 	int tind = (qpn & (ctx->num_qps - 1)) >> ctx->qp_table_shift;
diff --git a/src/verbs.c b/src/verbs.c
index fa9674b..41914ac 100644
--- a/src/verbs.c
+++ b/src/verbs.c
@@ -390,8 +390,14 @@ struct ibv_qp *mlx4_create_qp(struct ibv_pd *pd, struct ibv_qp_init_attr *attr)
 
 	*qp->db = 0;
 
-	cmd.buf_addr = (uintptr_t) qp->buf.buf;
-	cmd.db_addr  = (uintptr_t) qp->db;
+	cmd.buf_addr	  = (uintptr_t) qp->buf.buf;
+	cmd.db_addr	  = (uintptr_t) qp->db;
+	cmd.log_sq_stride = qp->sq.wqe_shift;
+	for (cmd.log_sq_bb_count = 0;
+	     qp->sq.max > 1 << cmd.log_sq_bb_count;
+	     ++cmd.log_sq_bb_count)
+		; /* nothing */
+	memset(cmd.reserved, 0, sizeof cmd.reserved);
 
 	ret = ibv_cmd_create_qp(pd, &qp->ibv_qp, attr, &cmd.ibv_cmd, sizeof cmd,
 				&resp, sizeof resp);
@@ -402,11 +408,9 @@ struct ibv_qp *mlx4_create_qp(struct ibv_pd *pd, struct ibv_qp_init_attr *attr)
 	if (ret)
 		goto err_destroy;
 
-	qp->sq.max	    = attr->cap.max_send_wr;
-	qp->rq.max	    = attr->cap.max_recv_wr;
-	qp->sq.max_gs	    = attr->cap.max_send_sge;
-	qp->rq.max_gs	    = attr->cap.max_recv_sge;
-	qp->max_inline_data = attr->cap.max_inline_data;
+	qp->rq.max    = attr->cap.max_recv_wr;
+	qp->rq.max_gs = attr->cap.max_recv_sge;
+	mlx4_set_sq_sizes(qp, &attr->cap, attr->qp_type);
 
 	qp->doorbell_qpn    = htonl(qp->ibv_qp.qp_num << 8);
 	if (attr->sq_sig_all)