[PATCH rdma-next 0/3] RSS related capabilities

[PATCH rdma-next 0/3] RSS related capabilities

[Leon Romanovsky]

Hi Doug,

This series exposes the RSS related capabilities to user space domain.
It comes as supplementary part to previous series of RSS that was
already accepted.

The series includes three patches:
Patch #1: Exposes the related properties from the IB core layer.
Patch #2: The uverbs layer queries the related properties from the
              core layer and exposes to user space area.
Patch #3: Exposes the related capabilities from the mlx5 driver.

Available in the "topic/rss_cap" topic branch of this git repo:
git://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git

Or for browsing:
https://git.kernel.org/cgit/linux/kernel/git/leon/linux-rdma.git/log/?h=topic/rss_cap

Thanks,
  Yishai & Leon

Yishai Hadas (3):
  IB/core: Expose RSS related capabilities
  IB/uverbs: Expose RSS related capabilities
  IB/mlx5: Expose RSS related capabilities

 drivers/infiniband/core/uverbs_cmd.c | 17 +++++++++++++++++
 drivers/infiniband/hw/mlx5/main.c    | 31 +++++++++++++++++++++++++++++++
 drivers/infiniband/hw/mlx5/user.h    |  7 +++++++
 include/rdma/ib_verbs.h              | 12 ++++++++++++
 include/uapi/rdma/ib_user_verbs.h    | 14 ++++++++++++++
 5 files changed, 81 insertions(+)

--
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 1/3] IB/core: Expose RSS related capabilities

[Leon Romanovsky]

From: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Expose RSS related capabilities, it includes both direct ones (i.e.
struct ib_rss_caps) and max_wq_type_rq which may be used in both
RSS and non RSS flows.

Specifically,
supported_qpts:
- QP types that support RSS on the device.

max_rwq_indirection_tables:
- Max number of receive work queue indirection tables that
  could be opened on the device.

max_rwq_indirection_table_size:
- Max size of a receive work queue indirection table.

max_wq_type_rq:
- Max number of work queues of receive type that
  could be opened on the device.

Signed-off-by: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 include/rdma/ib_verbs.h | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h
index 8e90dd2..eacbde7 100644
--- a/include/rdma/ib_verbs.h
+++ b/include/rdma/ib_verbs.h
@@ -261,6 +261,16 @@ struct ib_odp_caps {
 	} per_transport_caps;
 };

+struct ib_rss_caps {
+	/* Corresponding bit will be set if qp type from
+	 * 'enum ib_qp_type' is supported, e.g.
+	 * supported_qpts |= 1 << IB_QPT_UD
+	 */
+	u32 supported_qpts;
+	u32 max_rwq_indirection_tables;
+	u32 max_rwq_indirection_table_size;
+};
+
 enum ib_cq_creation_flags {
 	IB_CQ_FLAGS_TIMESTAMP_COMPLETION   = 1 << 0,
 	IB_CQ_FLAGS_IGNORE_OVERRUN	   = 1 << 1,
@@ -318,6 +328,8 @@ struct ib_device_attr {
 	struct ib_odp_caps	odp_caps;
 	uint64_t		timestamp_mask;
 	uint64_t		hca_core_clock; /* in KHZ */
+	struct ib_rss_caps	rss_caps;
+	u32			max_wq_type_rq;
 };

 enum ib_mtu {
--
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 2/3] IB/uverbs: Expose RSS related capabilities

[Leon Romanovsky]

From: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Query RSS related attributes and return them to user-space via the
extended query device uverbs command.

It includes both direct ones (i.e. struct ib_uverbs_rss_caps) and
max_wq_type_rq which may be used in both RSS and non RSS flows.

Signed-off-by: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 drivers/infiniband/core/uverbs_cmd.c | 17 +++++++++++++++++
 include/uapi/rdma/ib_user_verbs.h    | 14 ++++++++++++++
 2 files changed, 31 insertions(+)

diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c
index f664731..f44e675 100644
--- a/drivers/infiniband/core/uverbs_cmd.c
+++ b/drivers/infiniband/core/uverbs_cmd.c
@@ -4173,6 +4173,23 @@ int ib_uverbs_ex_query_device(struct ib_uverbs_file *file,
 
 	resp.device_cap_flags_ex = attr.device_cap_flags;
 	resp.response_length += sizeof(resp.device_cap_flags_ex);
+
+	if (ucore->outlen < resp.response_length + sizeof(resp.rss_caps))
+		goto end;
+
+	resp.rss_caps.supported_qpts = attr.rss_caps.supported_qpts;
+	resp.rss_caps.max_rwq_indirection_tables =
+		attr.rss_caps.max_rwq_indirection_tables;
+	resp.rss_caps.max_rwq_indirection_table_size =
+		attr.rss_caps.max_rwq_indirection_table_size;
+
+	resp.response_length += sizeof(resp.rss_caps);
+
+	if (ucore->outlen < resp.response_length + sizeof(resp.max_wq_type_rq))
+		goto end;
+
+	resp.max_wq_type_rq = attr.max_wq_type_rq;
+	resp.response_length += sizeof(resp.max_wq_type_rq);
 end:
 	err = ib_copy_to_udata(ucore, &resp, resp.response_length);
 	return err;
diff --git a/include/uapi/rdma/ib_user_verbs.h b/include/uapi/rdma/ib_user_verbs.h
index 7f035f4b..a9ebb47 100644
--- a/include/uapi/rdma/ib_user_verbs.h
+++ b/include/uapi/rdma/ib_user_verbs.h
@@ -224,6 +224,17 @@ struct ib_uverbs_odp_caps {
 	__u32 reserved;
 };
 
+struct ib_uverbs_rss_caps {
+	/* Corresponding bit will be set if qp type from
+	 * 'enum ib_qp_type' is supported, e.g.
+	 * supported_qpts |= 1 << IB_QPT_UD
+	 */
+	__u32 supported_qpts;
+	__u32 max_rwq_indirection_tables;
+	__u32 max_rwq_indirection_table_size;
+	__u32 reserved;
+};
+
 struct ib_uverbs_ex_query_device_resp {
 	struct ib_uverbs_query_device_resp base;
 	__u32 comp_mask;
@@ -232,6 +243,9 @@ struct ib_uverbs_ex_query_device_resp {
 	__u64 timestamp_mask;
 	__u64 hca_core_clock; /* in KHZ */
 	__u64 device_cap_flags_ex;
+	struct ib_uverbs_rss_caps rss_caps;
+	__u32  max_wq_type_rq;
+	__u32 reserved;
 };
 
 struct ib_uverbs_query_port {
-- 
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 3/3] IB/mlx5: Expose RSS related capabilities

[Leon Romanovsky]

From: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>

Expose RSS related capabilities on both IB and vendor channels.

In addition to the IB capabilities the driver reports some extra
capabilities on its vendor channel:
- Bit mask of the supported types of hash functions.
- Bit mask of the supported RX fields that can participate
  in the RX hashing.

Those capabilities are applicable only when the link layer
is Ethernet.

Signed-off-by: Yishai Hadas <yishaih-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Signed-off-by: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
---
 drivers/infiniband/hw/mlx5/main.c | 31 +++++++++++++++++++++++++++++++
 drivers/infiniband/hw/mlx5/user.h |  7 +++++++
 2 files changed, 38 insertions(+)

diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c
index f02a975..36706f7 100644
--- a/drivers/infiniband/hw/mlx5/main.c
+++ b/drivers/infiniband/hw/mlx5/main.c
@@ -531,6 +531,26 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 				resp.response_length += sizeof(resp.tso_caps);
 			}
 		}
+
+		if (field_avail(typeof(resp), rss_caps, uhw->outlen)) {
+			resp.rss_caps.rx_hash_function =
+						MLX5_RX_HASH_FUNC_TOEPLITZ;
+			resp.rss_caps.rx_hash_fields_mask =
+						MLX5_RX_HASH_SRC_IPV4 |
+						MLX5_RX_HASH_DST_IPV4 |
+						MLX5_RX_HASH_SRC_IPV6 |
+						MLX5_RX_HASH_DST_IPV6 |
+						MLX5_RX_HASH_SRC_PORT_TCP |
+						MLX5_RX_HASH_DST_PORT_TCP |
+						MLX5_RX_HASH_SRC_PORT_UDP |
+						MLX5_RX_HASH_DST_PORT_UDP;
+			resp.response_length += sizeof(resp.rss_caps);
+		}
+	} else {
+		if (field_avail(typeof(resp), tso_caps, uhw->outlen))
+			resp.response_length += sizeof(resp.tso_caps);
+		if (field_avail(typeof(resp), rss_caps, uhw->outlen))
+			resp.response_length += sizeof(resp.rss_caps);
 	}
 
 	if (MLX5_CAP_GEN(mdev, ipoib_basic_offloads)) {
@@ -594,6 +614,17 @@ static int mlx5_ib_query_device(struct ib_device *ibdev,
 	if (!mlx5_core_is_pf(mdev))
 		props->device_cap_flags |= IB_DEVICE_VIRTUAL_FUNCTION;
 
+	if (mlx5_ib_port_link_layer(ibdev, 1) ==
+	    IB_LINK_LAYER_ETHERNET) {
+		props->rss_caps.max_rwq_indirection_tables =
+			1 << MLX5_CAP_GEN(dev->mdev, log_max_rqt);
+		props->rss_caps.max_rwq_indirection_table_size =
+			1 << MLX5_CAP_GEN(dev->mdev, log_max_rqt_size);
+		props->rss_caps.supported_qpts = 1 << IB_QPT_RAW_PACKET;
+		props->max_wq_type_rq =
+			1 << MLX5_CAP_GEN(dev->mdev, log_max_rq);
+	}
+
 	if (uhw->outlen) {
 		err = ib_copy_to_udata(uhw, &resp, resp.response_length);
 
diff --git a/drivers/infiniband/hw/mlx5/user.h b/drivers/infiniband/hw/mlx5/user.h
index 188dac4..0e49d5b 100644
--- a/drivers/infiniband/hw/mlx5/user.h
+++ b/drivers/infiniband/hw/mlx5/user.h
@@ -121,10 +121,17 @@ struct mlx5_ib_tso_caps {
 	__u32 supported_qpts;
 };
 
+struct mlx5_ib_rss_caps {
+	__u64 rx_hash_fields_mask; /* enum mlx5_rx_hash_fields */
+	__u8 rx_hash_function; /* enum mlx5_rx_hash_function_flags */
+	__u8 reserved[7];
+};
+
 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_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

Re: [PATCH rdma-next 0/3] RSS related capabilities

[Sagi Grimberg]

Series looks good to me,

Reviewed-by: Sagi Grimberg <sagi-NQWnxTmZq1alnMjI0IkVqw@public.gmane.org>

Thanks Yishai!
--
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

Re: [PATCH rdma-next 0/3] RSS related capabilities

[Doug Ledford]

[-- Attachment #1.1: Type: text/plain, Size: 916 bytes --]

On 8/28/2016 4:28 AM, Leon Romanovsky wrote:
> Hi Doug,
> 
> This series exposes the RSS related capabilities to user space domain.
> It comes as supplementary part to previous series of RSS that was
> already accepted.
> 
> The series includes three patches:
> Patch #1: Exposes the related properties from the IB core layer.
> Patch #2: The uverbs layer queries the related properties from the
>               core layer and exposes to user space area.
> Patch #3: Exposes the related capabilities from the mlx5 driver.
> 
> Available in the "topic/rss_cap" topic branch of this git repo:
> git://git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git
> 
> Or for browsing:
> https://git.kernel.org/cgit/linux/kernel/git/leon/linux-rdma.git/log/?h=topic/rss_cap

Series applied, thanks.


-- 
Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
    GPG Key ID: 0E572FDD


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 884 bytes --]