[PATCH rdma-next 1/2] IB/mlx5: Report mlx5 CQE compression capabilities when querying device

[PATCH rdma-next 1/2] IB/mlx5: Report mlx5 CQE compression capabilities when querying device

[Leon Romanovsky]

From: Bodong Wang <bodong-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

The capabilities include:
- Max number of compressed and aggregated CQEs in a single session,
  while zero means unsupported.
- For Responder, there are two formats of mini CQE: mini CQE with Rx
  hash and mini CQE with checksum. They're mutual exclusive.

Signed-off-by: Bodong Wang <bodong-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Matan Barak <matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/main.c | 10 ++++++++++
 include/uapi/rdma/mlx5-abi.h      | 12 ++++++++++++
 2 files changed, 22 insertions(+)

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index 2217477..3a8d4b7 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -669,6 +669,16 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 			1 << MLX5_CAP_GEN(dev->mdev, log_max_rq);
 	}
 
+	if (field_avail(typeof(resp), cqe_comp_caps, uhw->outlen)) {
+		resp.cqe_comp_caps.max_num =
+			MLX5_CAP_GEN(dev->mdev, cqe_compression) ?
+			MLX5_CAP_GEN(dev->mdev, cqe_compression_max_num) : 0;
+		resp.cqe_comp_caps.supported_format =
+			MLX5_IB_CQE_RES_FORMAT_HASH |
+			MLX5_IB_CQE_RES_FORMAT_CSUM;
+		resp.response_length += sizeof(resp.cqe_comp_caps);
+	}
+
 	if (uhw->outlen) {
 		err = ib_copy_to_udata(uhw, &resp, resp.response_length);
 
diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h
index f5d0f4e..24a7435 100644
--- a/include/uapi/rdma/mlx5-abi.h
+++ b/include/uapi/rdma/mlx5-abi.h
@@ -124,11 +124,23 @@ struct mlx5_ib_rss_caps {
 	__u8 reserved[7];
 };
 
+enum mlx5_ib_cqe_comp_res_format {
+	MLX5_IB_CQE_RES_FORMAT_HASH	= 1 << 0,
+	MLX5_IB_CQE_RES_FORMAT_CSUM	= 1 << 1,
+	MLX5_IB_CQE_RES_RESERVED	= 1 << 2,
+};
+
+struct mlx5_ib_cqe_comp_caps {
+	__u32 max_num;
+	__u32 supported_format; /* enum mlx5_ib_cqe_comp_res_format */
+};
+
 struct mlx5_ib_query_device_resp {
 	__u32	comp_mask;
 	__u32	response_length;
 	struct	mlx5_ib_tso_caps tso_caps;
 	struct	mlx5_ib_rss_caps rss_caps;
+	struct	mlx5_ib_cqe_comp_caps cqe_comp_caps;
 };
 
 struct mlx5_ib_create_cq {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH rdma-next 2/2] IB/mlx5: Add support for CQE compressing

[Leon Romanovsky]

From: Bodong Wang <bodong-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

CQE compressing reduces PCI overhead by coalescing and compressing
multiple CQEs into a single merged CQE. Successful compressing
improves message rate especially for small packet traffic.

CQE compressing is supported for all 64B CQE formats (with certain
limitations) generated by RQ/Responder or by SQ/Requestor.

Signed-off-by: Bodong Wang <bodong-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Reviewed-by: Matan Barak <matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/cq.c | 30 +++++++++++++++++++++++++++++-
 include/uapi/rdma/mlx5-abi.h    |  4 +++-
 2 files changed, 32 insertions(+), 2 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/cq.c b/drivers/infiniband/hw/mlx5/cq.c
index 79d017b..9ac36ef 100644
--- a/drivers/infiniband/hw/mlx5/cq.c
+++ b/drivers/infiniband/hw/mlx5/cq.c
@@ -731,7 +731,7 @@ static int create_cq_user(struct mlx5_ib_dev *dev, struct ib_udata *udata,
 			  int entries, u32 **cqb,
 			  int *cqe_size, int *index, int *inlen)
 {
-	struct mlx5_ib_create_cq ucmd;
+	struct mlx5_ib_create_cq ucmd = {};
 	size_t ucmdlen;
 	int page_shift;
 	__be64 *pas;
@@ -792,8 +792,36 @@ static int create_cq_user(struct mlx5_ib_dev *dev, struct ib_udata *udata,
 
 	*index = to_mucontext(context)->uuari.uars[0].index;
 
+	if (ucmd.cqe_comp_en == 1) {
+		if (unlikely((*cqe_size != 64) ||
+			     !MLX5_CAP_GEN(dev->mdev, cqe_compression))) {
+			err = -EOPNOTSUPP;
+			mlx5_ib_warn(dev, "CQE compression is not supported for size %d!\n",
+				     *cqe_size);
+			goto err_cqb;
+		}
+
+		if (unlikely(!ucmd.cqe_comp_res_format ||
+			     !(ucmd.cqe_comp_res_format <
+			       MLX5_IB_CQE_RES_RESERVED) ||
+			     (ucmd.cqe_comp_res_format &
+			      (ucmd.cqe_comp_res_format - 1)))) {
+			err = -EOPNOTSUPP;
+			mlx5_ib_warn(dev, "CQE compression res format %d is not supported!\n",
+				     ucmd.cqe_comp_res_format);
+			goto err_cqb;
+		}
+
+		MLX5_SET(cqc, cqc, cqe_comp_en, 1);
+		MLX5_SET(cqc, cqc, mini_cqe_res_format,
+			 ilog2(ucmd.cqe_comp_res_format));
+	}
+
 	return 0;
 
+err_cqb:
+	kfree(cqb);
+
 err_db:
 	mlx5_ib_db_unmap_user(to_mucontext(context), &cq->db);
 
diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h
index 24a7435..208710b 100644
--- a/include/uapi/rdma/mlx5-abi.h
+++ b/include/uapi/rdma/mlx5-abi.h
@@ -147,7 +147,9 @@ struct mlx5_ib_create_cq {
 	__u64	buf_addr;
 	__u64	db_addr;
 	__u32	cqe_size;
-	__u32	reserved; /* explicit padding (optional on i386) */
+	__u8    cqe_comp_en;
+	__u8    cqe_comp_res_format;
+	__u16	reserved; /* explicit padding (optional on i386) */
 };
 
 struct mlx5_ib_create_cq_resp {
-- 
2.7.4

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html