[ofa-general] [PATCH 11/16 v4] IB/mlx4: Add LSO support to mlx4

Eli Cohen eli at mellanox.co.il
Wed Jan 30 08:31:35 PST 2008


IB/mlx4: Add LSO support to mlx4

Signed-off-by: Eli Cohen <eli at mellnaox.co.il>
---
 drivers/infiniband/hw/mlx4/cq.c   |    3 ++
 drivers/infiniband/hw/mlx4/main.c |    4 +++
 drivers/infiniband/hw/mlx4/qp.c   |   52 +++++++++++++++++++++++++++++++++---
 drivers/net/mlx4/fw.c             |    9 ++++++
 drivers/net/mlx4/fw.h             |    1 +
 drivers/net/mlx4/main.c           |    1 +
 include/linux/mlx4/device.h       |    1 +
 include/linux/mlx4/qp.h           |    5 +++
 8 files changed, 71 insertions(+), 5 deletions(-)

diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index 539c69c..75fc2b3 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -409,6 +409,9 @@ static int mlx4_ib_poll_one(struct mlx4_ib_cq *cq,
 		case MLX4_OPCODE_BIND_MW:
 			wc->opcode    = IB_WC_BIND_MW;
 			break;
+		case MLX4_OPCODE_LSO:
+			wc->opcode    = IB_WC_LSO;
+			break;
 		}
 	} else {
 		wc->byte_len = be32_to_cpu(cqe->byte_cnt);
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 8ce94a1..2dd0de3 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -101,6 +101,8 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 		props->device_cap_flags |= IB_DEVICE_UD_AV_PORT_ENFORCE;
 	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
 		props->device_cap_flags |= IB_DEVICE_IP_CSUM;
+	if (dev->dev->caps.max_gso_sz)
+		props->device_cap_flags |= IB_DEVICE_TCP_TSO;
 
 	props->vendor_id	   = be32_to_cpup((__be32 *) (out_mad->data + 36)) &
 		0xffffff;
@@ -616,6 +618,8 @@ static void *mlx4_ib_add(struct mlx4_dev *dev)
 
 	if (ibdev->dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
 		ibdev->ib_dev.flags |= IB_DEVICE_IP_CSUM;
+	if (ibdev->dev->caps.max_gso_sz)
+		ibdev->ib_dev.flags |= IB_DEVICE_TCP_TSO;
 
 	if (init_node_data(ibdev))
 		goto err_map;
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index a04e931..fc4811c 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -69,6 +69,7 @@ enum {
 
 static const __be32 mlx4_ib_opcode[] = {
 	[IB_WR_SEND]			= __constant_cpu_to_be32(MLX4_OPCODE_SEND),
+	[IB_WR_LSO]			= __constant_cpu_to_be32(MLX4_OPCODE_LSO),
 	[IB_WR_SEND_WITH_IMM]		= __constant_cpu_to_be32(MLX4_OPCODE_SEND_IMM),
 	[IB_WR_RDMA_WRITE]		= __constant_cpu_to_be32(MLX4_OPCODE_RDMA_WRITE),
 	[IB_WR_RDMA_WRITE_WITH_IMM]	= __constant_cpu_to_be32(MLX4_OPCODE_RDMA_WRITE_IMM),
@@ -243,6 +244,7 @@ static int set_kernel_sq_size(struct mlx4_ib_dev *dev, struct ib_qp_init_attr *i
 {
 	struct ib_qp_cap *cap = &init_attr->cap;
 	enum ib_qp_type type = init_attr->qp_type;
+	int reserve = 0;
 
 	/* Sanity check SQ size before proceeding */
 	if (cap->max_send_wr	 > dev->dev->caps.max_wqes  ||
@@ -259,15 +261,18 @@ static int set_kernel_sq_size(struct mlx4_ib_dev *dev, struct ib_qp_init_attr *i
 	    cap->max_send_sge + 2 > dev->dev->caps.max_sq_sg)
 		return -EINVAL;
 
-	if (init_attr->create_flags & QP_CREATE_LSO)
+	if (init_attr->create_flags & QP_CREATE_LSO) {
 		qp->flags |= MLX4_QP_LSO;
+		reserve = 64;
+	}
 
 	qp->sq.wqe_shift = ilog2(roundup_pow_of_two(max(cap->max_send_sge *
-							sizeof (struct mlx4_wqe_data_seg),
+							sizeof (struct mlx4_wqe_data_seg) +
+								reserve,
 							cap->max_inline_data +
 							sizeof (struct mlx4_wqe_inline_seg)) +
 						    send_wqe_overhead(type)));
-	qp->sq.max_gs    = ((1 << qp->sq.wqe_shift) - send_wqe_overhead(type)) /
+	qp->sq.max_gs    = ((1 << qp->sq.wqe_shift) - reserve - send_wqe_overhead(type)) /
 		sizeof (struct mlx4_wqe_data_seg);
 
 	/*
@@ -755,9 +760,11 @@ static int __mlx4_ib_modify_qp(struct ib_qp *ibqp,
 		}
 	}
 
-	if (ibqp->qp_type == IB_QPT_GSI || ibqp->qp_type == IB_QPT_SMI ||
-	    ibqp->qp_type == IB_QPT_UD)
+	if (ibqp->qp_type == IB_QPT_GSI || ibqp->qp_type == IB_QPT_SMI)
 		context->mtu_msgmax = (IB_MTU_4096 << 5) | 11;
+	else if (ibqp->qp_type == IB_QPT_UD)
+		context->mtu_msgmax = (IB_MTU_4096 << 5) |
+			ilog2(dev->dev->caps.max_gso_sz);
 	else if (attr_mask & IB_QP_PATH_MTU) {
 		if (attr->path_mtu < IB_MTU_256 || attr->path_mtu > IB_MTU_4096) {
 			printk(KERN_ERR "path MTU (%u) is invalid\n",
@@ -1274,6 +1281,28 @@ static void __set_data_seg(struct mlx4_wqe_data_seg *dseg, struct ib_sge *sg)
 	dseg->addr       = cpu_to_be64(sg->addr);
 }
 
+static int build_lso_seg(struct mlx4_lso_seg *wqe, struct ib_send_wr *wr,
+			 struct mlx4_ib_qp *qp, int *lso_seg_len)
+{
+	int halign;
+
+	halign = ALIGN(wr->wr.ud.hlen, 16);
+	if (unlikely(!(qp->flags & MLX4_QP_LSO) && wr->num_sge > qp->sq.max_gs - (halign >> 4)))
+                return -EINVAL;
+
+	memcpy(wqe->header, wr->wr.ud.header, wr->wr.ud.hlen);
+
+	/* make sure LSO header is written before
+	   overwriting stamping */
+	wmb();
+
+	wqe->mss_hdr_size = cpu_to_be32(((wr->wr.ud.mss - wr->wr.ud.hlen)
+					 << 16) | wr->wr.ud.hlen);
+
+	*lso_seg_len = halign;
+	return 0;
+}
+
 int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 		      struct ib_send_wr **bad_wr)
 {
@@ -1364,6 +1393,19 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			set_datagram_seg(wqe, wr);
 			wqe  += sizeof (struct mlx4_wqe_datagram_seg);
 			size += sizeof (struct mlx4_wqe_datagram_seg) / 16;
+
+			if (wr->opcode == IB_WR_LSO) {
+				int hlen;
+
+				err = build_lso_seg(wqe, wr, qp, &hlen);
+				if (err) {
+					*bad_wr = wr;
+					goto out;
+				}
+				wqe += hlen;
+				size += hlen >> 4;
+			}
+
 			break;
 
 		case IB_QPT_SMI:
diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c
index 736942f..7b426ff 100644
--- a/drivers/net/mlx4/fw.c
+++ b/drivers/net/mlx4/fw.c
@@ -133,6 +133,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 #define QUERY_DEV_CAP_MAX_AV_OFFSET		0x27
 #define QUERY_DEV_CAP_MAX_REQ_QP_OFFSET		0x29
 #define QUERY_DEV_CAP_MAX_RES_QP_OFFSET		0x2b
+#define QUERY_DEV_CAP_MAX_GSO_OFFSET		0x2d
 #define QUERY_DEV_CAP_MAX_RDMA_OFFSET		0x2f
 #define QUERY_DEV_CAP_RSZ_SRQ_OFFSET		0x33
 #define QUERY_DEV_CAP_ACK_DELAY_OFFSET		0x35
@@ -215,6 +216,13 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	dev_cap->max_requester_per_qp = 1 << (field & 0x3f);
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_RES_QP_OFFSET);
 	dev_cap->max_responder_per_qp = 1 << (field & 0x3f);
+	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_GSO_OFFSET);
+	field &= 0x1f;
+	if (!field)
+		dev_cap->max_gso_sz = 0;
+	else
+		dev_cap->max_gso_sz = 1 << field;
+
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_MAX_RDMA_OFFSET);
 	dev_cap->max_rdma_global = 1 << (field & 0x3f);
 	MLX4_GET(field, outbox, QUERY_DEV_CAP_ACK_DELAY_OFFSET);
@@ -377,6 +385,7 @@ int mlx4_QUERY_DEV_CAP(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 		 dev_cap->max_sq_desc_sz, dev_cap->max_sq_sg);
 	mlx4_dbg(dev, "Max RQ desc size: %d, max RQ S/G: %d\n",
 		 dev_cap->max_rq_desc_sz, dev_cap->max_rq_sg);
+	mlx4_dbg(dev, "Max GSO size: %d\n", dev_cap->max_gso_sz);
 
 	dump_dev_cap_flags(dev, dev_cap->flags);
 
diff --git a/drivers/net/mlx4/fw.h b/drivers/net/mlx4/fw.h
index 7e1dd9e..ad5abf3 100644
--- a/drivers/net/mlx4/fw.h
+++ b/drivers/net/mlx4/fw.h
@@ -96,6 +96,7 @@ struct mlx4_dev_cap {
 	u8  bmme_flags;
 	u32 reserved_lkey;
 	u64 max_icm_sz;
+	int max_gso_sz;
 };
 
 struct mlx4_adapter {
diff --git a/drivers/net/mlx4/main.c b/drivers/net/mlx4/main.c
index 89b3f0b..ed2c648 100644
--- a/drivers/net/mlx4/main.c
+++ b/drivers/net/mlx4/main.c
@@ -159,6 +159,7 @@ static int mlx4_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap)
 	dev->caps.page_size_cap	     = ~(u32) (dev_cap->min_page_sz - 1);
 	dev->caps.flags		     = dev_cap->flags;
 	dev->caps.stat_rate_support  = dev_cap->stat_rate_support;
+	dev->caps.max_gso_sz	     = dev_cap->max_gso_sz;
 
 	return 0;
 }
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 222815d..856570f 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -181,6 +181,7 @@ struct mlx4_caps {
 	u32			flags;
 	u16			stat_rate_support;
 	u8			port_width_cap[MLX4_MAX_PORTS + 1];
+	int			max_gso_sz;
 };
 
 struct mlx4_buf_list {
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index b4eb921..0bac8e8 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -215,6 +215,11 @@ struct mlx4_wqe_datagram_seg {
 	__be32			reservd[2];
 };
 
+struct mlx4_lso_seg {
+	__be32                  mss_hdr_size;
+	__be32                  header[0];
+};
+
 struct mlx4_wqe_bind_seg {
 	__be32			flags1;
 	__be32			flags2;
-- 
1.5.3.8






More information about the general mailing list