[ofa-general] [PATCH] IB/mlx4: Add checksum offload support

[Eli Cohen]

IB/mlx4: Add checksum offload support

ConnectX devices support checksum generation and verification
of TCP,UDP/IP packets. This patch checks if the HCA supports
this and sets a capability flag.

Signed-off-by: Eli Cohen <eli at mellnaox.co.il>
Signed-off-by: Ali Ayub <ali at mellnaox.co.il>
---
 drivers/infiniband/hw/mlx4/cq.c   |    9 +++++++++
 drivers/infiniband/hw/mlx4/main.c |    2 ++
 drivers/infiniband/hw/mlx4/qp.c   |    3 +++
 drivers/net/mlx4/fw.c             |    3 +++
 include/linux/mlx4/cq.h           |    4 ++--
 include/linux/mlx4/qp.h           |    2 ++
 6 files changed, 21 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/mlx4/cq.c b/drivers/infiniband/hw/mlx4/cq.c
index 7360bba..561c526 100644
--- a/drivers/infiniband/hw/mlx4/cq.c
+++ b/drivers/infiniband/hw/mlx4/cq.c
@@ -309,6 +309,11 @@ static int mlx4_ib_poll_one(struct mlx4_ib_cq *cq,
 	int is_error;
 	u32 g_mlpath_rqpn;
 	u16 wqe_ctr;
+	__be32 status;
+
+#define CSUM_MASK_BITS cpu_to_be32(0x13c00000)
+#define CSUM_VAL_BITS  cpu_to_be32(0x10400000)
+#define CSUM_MASK2_BITS cpu_to_be32(0x0c000000)
 
 	cqe = next_cqe_sw(cq);
 	if (!cqe)
@@ -434,6 +439,10 @@ static int mlx4_ib_poll_one(struct mlx4_ib_cq *cq,
 		wc->dlid_path_bits = (g_mlpath_rqpn >> 24) & 0x7f;
 		wc->wc_flags	  |= g_mlpath_rqpn & 0x80000000 ? IB_WC_GRH : 0;
 		wc->pkey_index     = be32_to_cpu(cqe->immed_rss_invalid) & 0x7f;
+		status = cqe->ipoib_status;
+		wc->csum_ok = (status & CSUM_MASK_BITS) == CSUM_VAL_BITS &&
+			      (status & CSUM_MASK2_BITS) &&
+			      cqe->checksum == 0xffff;
 	}
 
 	return 0;
diff --git a/drivers/infiniband/hw/mlx4/main.c b/drivers/infiniband/hw/mlx4/main.c
index 96a39b5..ef5e9db 100644
--- a/drivers/infiniband/hw/mlx4/main.c
+++ b/drivers/infiniband/hw/mlx4/main.c
@@ -99,6 +99,8 @@ static int mlx4_ib_query_device(struct ib_device *ibdev,
 		props->device_cap_flags |= IB_DEVICE_AUTO_PATH_MIG;
 	if (dev->dev->caps.flags & MLX4_DEV_CAP_FLAG_UD_AV_PORT)
 		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_UD_IP_CSUM;
 
 	props->vendor_id	   = be32_to_cpup((__be32 *) (out_mad->data + 36)) &
 		0xffffff;
diff --git a/drivers/infiniband/hw/mlx4/qp.c b/drivers/infiniband/hw/mlx4/qp.c
index 958e205..8c1b29e 100644
--- a/drivers/infiniband/hw/mlx4/qp.c
+++ b/drivers/infiniband/hw/mlx4/qp.c
@@ -1436,6 +1436,9 @@ int mlx4_ib_post_send(struct ib_qp *ibqp, struct ib_send_wr *wr,
 			 cpu_to_be32(MLX4_WQE_CTRL_CQ_UPDATE) : 0) |
 			(wr->send_flags & IB_SEND_SOLICITED ?
 			 cpu_to_be32(MLX4_WQE_CTRL_SOLICITED) : 0) |
+			((wr->send_flags & IB_SEND_IP_CSUM) ?
+			 cpu_to_be32(MLX4_WQE_CTRL_IP_CSUM |
+				     MLX4_WQE_CTRL_TCP_UDP_CSUM) : 0) |
 			qp->sq_signal_bits;
 
 		if (wr->opcode == IB_WR_SEND_WITH_IMM ||
diff --git a/drivers/net/mlx4/fw.c b/drivers/net/mlx4/fw.c
index 61dc495..63edf48 100644
--- a/drivers/net/mlx4/fw.c
+++ b/drivers/net/mlx4/fw.c
@@ -730,6 +730,9 @@ int mlx4_INIT_HCA(struct mlx4_dev *dev, struct mlx4_init_hca_param *param)
 	MLX4_PUT(inbox, (u8) (PAGE_SHIFT - 12), INIT_HCA_UAR_PAGE_SZ_OFFSET);
 	MLX4_PUT(inbox, param->log_uar_sz,      INIT_HCA_LOG_UAR_SZ_OFFSET);
 
+	if (dev->caps.flags & MLX4_DEV_CAP_FLAG_IPOIB_CSUM)
+		*(inbox + INIT_HCA_FLAGS_OFFSET / 4) |= cpu_to_be32(1 << 3);
+
 	err = mlx4_cmd(dev, mailbox->dma, 0, 0, MLX4_CMD_INIT_HCA, 10000);
 
 	if (err)
diff --git a/include/linux/mlx4/cq.h b/include/linux/mlx4/cq.h
index 0181e0a..5fdc859 100644
--- a/include/linux/mlx4/cq.h
+++ b/include/linux/mlx4/cq.h
@@ -45,11 +45,11 @@ struct mlx4_cqe {
 	u8			sl;
 	u8			reserved1;
 	__be16			rlid;
-	u32			reserved2;
+	__be32			ipoib_status;
 	__be32			byte_cnt;
 	__be16			wqe_index;
 	__be16			checksum;
-	u8			reserved3[3];
+	u8			reserved2[3];
 	u8			owner_sr_opcode;
 };
 
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index 09a2230..f30e965 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -162,6 +162,8 @@ enum {
 	MLX4_WQE_CTRL_FENCE	= 1 << 6,
 	MLX4_WQE_CTRL_CQ_UPDATE	= 3 << 2,
 	MLX4_WQE_CTRL_SOLICITED	= 1 << 1,
+	MLX4_WQE_CTRL_IP_CSUM           = 1 << 4,
+	MLX4_WQE_CTRL_TCP_UDP_CSUM      = 1 << 5,
 };
 
 struct mlx4_wqe_ctrl_seg {
-- 
1.5.4.2