From 850d8fd765079d223d47de89f1f03ab95d1036cb Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Thu, 10 Nov 2016 11:30:56 +0200 Subject: IB/mlx4: Handle IPv4 header when demultiplexing MAD When MAD arrives to the hypervisor, we need to identify which slave it should be sent by destination GID. When L3 protocol is IPv4 the GRH is replaced by an IPv4 header. This patch detects when IPv4 header needs to be parsed instead of GRH. Fixes: b6ffaeffaea4 ('mlx4: In RoCE allow guests to have multiple GIDS') Signed-off-by: Moni Shoua Signed-off-by: Daniel Jurgens Reviewed-by: Mark Bloch Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 5ad43a487745..467a4b476e29 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -2580,6 +2580,24 @@ void ib_dealloc_pd(struct ib_pd *pd); */ struct ib_ah *ib_create_ah(struct ib_pd *pd, struct ib_ah_attr *ah_attr); +/** + * ib_get_gids_from_rdma_hdr - Get sgid and dgid from GRH or IPv4 header + * work completion. + * @hdr: the L3 header to parse + * @net_type: type of header to parse + * @sgid: place to store source gid + * @dgid: place to store destination gid + */ +int ib_get_gids_from_rdma_hdr(const union rdma_network_hdr *hdr, + enum rdma_network_type net_type, + union ib_gid *sgid, union ib_gid *dgid); + +/** + * ib_get_rdma_header_version - Get the header version + * @hdr: the L3 header to parse + */ +int ib_get_rdma_header_version(const union rdma_network_hdr *hdr); + /** * ib_init_ah_from_wc - Initializes address handle attributes from a * work completion. @@ -3357,4 +3375,5 @@ int ib_sg_to_pages(struct ib_mr *mr, struct scatterlist *sgl, int sg_nents, void ib_drain_rq(struct ib_qp *qp); void ib_drain_sq(struct ib_qp *qp); void ib_drain_qp(struct ib_qp *qp); + #endif /* IB_VERBS_H */ -- cgit v1.2.3 From c226dc22ec4904340e3e14a536983cda3dbe7914 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Mon, 31 Oct 2016 12:15:20 +0200 Subject: net/mlx5: Report multi packet WQE capabilities Multi packet WQE enables sending multiple fix sized packets using a single WQE. The exposed field reports such HW support. Signed-off-by: Bodong Wang Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/linux/mlx5/mlx5_ifc.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 2632cb2caf10..0779ad2e8f51 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -576,7 +576,7 @@ struct mlx5_ifc_per_protocol_networking_offload_caps_bits { u8 self_lb_en_modifiable[0x1]; u8 reserved_at_9[0x2]; u8 max_lso_cap[0x5]; - u8 reserved_at_10[0x2]; + u8 multi_pkt_send_wqe[0x2]; u8 wqe_inline_mode[0x2]; u8 rss_ind_tbl_cap[0x4]; u8 reg_umr_sq[0x1]; -- cgit v1.2.3 From 191ded4a4d991acf17207e0b4370fef070bce3e9 Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Mon, 31 Oct 2016 12:15:21 +0200 Subject: IB/mlx5: Report mlx5 multi packet WQE caps during query The capabilities whether hardware support multi packet WQE or not is exposed to user space through query_device by uhw. Signed-off-by: Bodong Wang Reviewed-by: Matan Barak Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/mlx5-abi.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h index f5d0f4e83b59..93d6b9fe8e78 100644 --- a/include/uapi/rdma/mlx5-abi.h +++ b/include/uapi/rdma/mlx5-abi.h @@ -129,6 +129,8 @@ struct mlx5_ib_query_device_resp { __u32 response_length; struct mlx5_ib_tso_caps tso_caps; struct mlx5_ib_rss_caps rss_caps; + __u32 mlx5_ib_support_multi_pkt_send_wqes; + __u32 reserved; }; struct mlx5_ib_create_cq { -- cgit v1.2.3 From 7e43a2a5bae39fedaa7cce21d637e0c8d96d8e54 Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Mon, 31 Oct 2016 12:16:44 +0200 Subject: IB/mlx5: Report mlx5 CQE compression caps during query 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 Reviewed-by: Matan Barak Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/mlx5-abi.h | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'include') diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h index 93d6b9fe8e78..6649d13a2dbb 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; __u32 mlx5_ib_support_multi_pkt_send_wqes; __u32 reserved; }; -- cgit v1.2.3 From 1cbe6fc86ccfe05a910be4883da7c7bd28c190fe Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Mon, 31 Oct 2016 12:16:45 +0200 Subject: IB/mlx5: Add support for CQE compressing 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 Reviewed-by: Matan Barak Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/mlx5-abi.h | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h index 6649d13a2dbb..b0a41c8dc1fb 100644 --- a/include/uapi/rdma/mlx5-abi.h +++ b/include/uapi/rdma/mlx5-abi.h @@ -149,7 +149,9 @@ 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 { -- cgit v1.2.3 From 0dbf3332b7b683db33a385a3ce9baab157e3ff9a Mon Sep 17 00:00:00 2001 From: Moses Reuben Date: Mon, 14 Nov 2016 19:04:47 +0200 Subject: IB/core: Add flow spec tunneling support In order to support tunneling, that can be used by the QP, both struct ib_flow_spec_tunnel and struct ib_flow_tunnel_filter can be used to more IP or UDP based tunneling protocols (e.g NVGRE, GRE, etc). IB_FLOW_SPEC_VXLAN_TUNNEL type flow specification is added to use this functionality and match specific Vxlan packets. In similar to IPv6, we check overflow of the vni value by comparing with the maximum size. Signed-off-by: Moses Reuben Reviewed-by: Maor Gottlieb Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 19 ++++++++++++++++++- 1 file changed, 18 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 467a4b476e29..b8213818de89 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -1599,7 +1599,8 @@ enum ib_flow_spec_type { IB_FLOW_SPEC_IPV6 = 0x31, /* L4 headers*/ IB_FLOW_SPEC_TCP = 0x40, - IB_FLOW_SPEC_UDP = 0x41 + IB_FLOW_SPEC_UDP = 0x41, + IB_FLOW_SPEC_VXLAN_TUNNEL = 0x50, }; #define IB_FLOW_SPEC_LAYER_MASK 0xF0 #define IB_FLOW_SPEC_SUPPORT_LAYERS 4 @@ -1707,6 +1708,21 @@ struct ib_flow_spec_tcp_udp { struct ib_flow_tcp_udp_filter mask; }; +struct ib_flow_tunnel_filter { + __be32 tunnel_id; + u8 real_sz[0]; +}; + +/* ib_flow_spec_tunnel describes the Vxlan tunnel + * the tunnel_id from val has the vni value + */ +struct ib_flow_spec_tunnel { + enum ib_flow_spec_type type; + u16 size; + struct ib_flow_tunnel_filter val; + struct ib_flow_tunnel_filter mask; +}; + union ib_flow_spec { struct { enum ib_flow_spec_type type; @@ -1717,6 +1733,7 @@ union ib_flow_spec { struct ib_flow_spec_ipv4 ipv4; struct ib_flow_spec_tcp_udp tcp_udp; struct ib_flow_spec_ipv6 ipv6; + struct ib_flow_spec_tunnel tunnel; }; struct ib_flow_attr { -- cgit v1.2.3 From 76bd23b34204cad78f48aec4cef38869a66aa594 Mon Sep 17 00:00:00 2001 From: Moses Reuben Date: Mon, 14 Nov 2016 19:04:48 +0200 Subject: IB/core: Align structure ib_flow_spec_type Aligned the structure ib_flow_spec_type indentation, after adding a new definition. Signed-off-by: Moses Reuben Reviewed-by: Maor Gottlieb Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index b8213818de89..4bc748fddcac 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -1592,14 +1592,14 @@ enum ib_flow_attr_type { /* Supported steering header types */ enum ib_flow_spec_type { /* L2 headers*/ - IB_FLOW_SPEC_ETH = 0x20, - IB_FLOW_SPEC_IB = 0x22, + IB_FLOW_SPEC_ETH = 0x20, + IB_FLOW_SPEC_IB = 0x22, /* L3 header*/ - IB_FLOW_SPEC_IPV4 = 0x30, - IB_FLOW_SPEC_IPV6 = 0x31, + IB_FLOW_SPEC_IPV4 = 0x30, + IB_FLOW_SPEC_IPV6 = 0x31, /* L4 headers*/ - IB_FLOW_SPEC_TCP = 0x40, - IB_FLOW_SPEC_UDP = 0x41, + IB_FLOW_SPEC_TCP = 0x40, + IB_FLOW_SPEC_UDP = 0x41, IB_FLOW_SPEC_VXLAN_TUNNEL = 0x50, }; #define IB_FLOW_SPEC_LAYER_MASK 0xF0 -- cgit v1.2.3 From a0cb4c759af12943806564294fa53ab08cb7cf93 Mon Sep 17 00:00:00 2001 From: Moses Reuben Date: Mon, 14 Nov 2016 19:04:49 +0200 Subject: IB/uverbs: Add support for Vxlan protocol Add ib_uverbs_flow_spec_tunnel to define the rule to match Vxlan, the type, size, reserved fields are identical to rest of the protocols, and are used to identify the spec. The tunnel id is the vni value of the Vxlan protocol, and it is used as part of the steering rule, it is limited by the mask. The steering rule configured on the hardware does a match according to vni and other protocols. In the same way as rest of the protocols that we match, the uniq field's of each protocol are represented on the val and the mask. Signed-off-by: Moses Reuben Reviewed-by: Maor Gottlieb Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/ib_user_verbs.h | 17 +++++++++++++++++ 1 file changed, 17 insertions(+) (limited to 'include') diff --git a/include/uapi/rdma/ib_user_verbs.h b/include/uapi/rdma/ib_user_verbs.h index 25225ebbc7d5..90ba5e88ec00 100644 --- a/include/uapi/rdma/ib_user_verbs.h +++ b/include/uapi/rdma/ib_user_verbs.h @@ -908,6 +908,23 @@ struct ib_uverbs_flow_spec_ipv6 { struct ib_uverbs_flow_ipv6_filter mask; }; +struct ib_uverbs_flow_tunnel_filter { + __be32 tunnel_id; +}; + +struct ib_uverbs_flow_spec_tunnel { + union { + struct ib_uverbs_flow_spec_hdr hdr; + struct { + __u32 type; + __u16 size; + __u16 reserved; + }; + }; + struct ib_uverbs_flow_tunnel_filter val; + struct ib_uverbs_flow_tunnel_filter mask; +}; + struct ib_uverbs_flow_attr { __u32 type; __u16 size; -- cgit v1.2.3 From fbf46860b19ddb485f00bef1ad1a43aabc9f71ad Mon Sep 17 00:00:00 2001 From: Moses Reuben Date: Mon, 14 Nov 2016 19:04:51 +0200 Subject: IB/core: Introduce inner flow steering For a tunneled packet which contains external and internal headers, we refer to the external headers as "outer fields" and the internal headers as "inner fields". Example of a tunneled packet: { L2 | L3 | L4 | tunnel header | L2 | L3 | l4 | data } | | | | | | | { outer fields }{ inner fields } This patch introduces a new flag for flow steering rules - IB_FLOW_SPEC_INNER - which specifies that the rule applies to the inner fields, rather than to the outer fields of the packet. Signed-off-by: Moses Reuben Reviewed-by: Maor Gottlieb Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 4bc748fddcac..6d0dd6525e14 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -1601,9 +1601,10 @@ enum ib_flow_spec_type { IB_FLOW_SPEC_TCP = 0x40, IB_FLOW_SPEC_UDP = 0x41, IB_FLOW_SPEC_VXLAN_TUNNEL = 0x50, + IB_FLOW_SPEC_INNER = 0x100, }; #define IB_FLOW_SPEC_LAYER_MASK 0xF0 -#define IB_FLOW_SPEC_SUPPORT_LAYERS 4 +#define IB_FLOW_SPEC_SUPPORT_LAYERS 8 /* Flow steering rule priority is set according to it's domain. * Lower domain value means higher priority. @@ -1631,7 +1632,7 @@ struct ib_flow_eth_filter { }; struct ib_flow_spec_eth { - enum ib_flow_spec_type type; + u32 type; u16 size; struct ib_flow_eth_filter val; struct ib_flow_eth_filter mask; @@ -1645,7 +1646,7 @@ struct ib_flow_ib_filter { }; struct ib_flow_spec_ib { - enum ib_flow_spec_type type; + u32 type; u16 size; struct ib_flow_ib_filter val; struct ib_flow_ib_filter mask; @@ -1670,7 +1671,7 @@ struct ib_flow_ipv4_filter { }; struct ib_flow_spec_ipv4 { - enum ib_flow_spec_type type; + u32 type; u16 size; struct ib_flow_ipv4_filter val; struct ib_flow_ipv4_filter mask; @@ -1688,7 +1689,7 @@ struct ib_flow_ipv6_filter { }; struct ib_flow_spec_ipv6 { - enum ib_flow_spec_type type; + u32 type; u16 size; struct ib_flow_ipv6_filter val; struct ib_flow_ipv6_filter mask; @@ -1702,7 +1703,7 @@ struct ib_flow_tcp_udp_filter { }; struct ib_flow_spec_tcp_udp { - enum ib_flow_spec_type type; + u32 type; u16 size; struct ib_flow_tcp_udp_filter val; struct ib_flow_tcp_udp_filter mask; @@ -1717,7 +1718,7 @@ struct ib_flow_tunnel_filter { * the tunnel_id from val has the vni value */ struct ib_flow_spec_tunnel { - enum ib_flow_spec_type type; + u32 type; u16 size; struct ib_flow_tunnel_filter val; struct ib_flow_tunnel_filter mask; @@ -1725,7 +1726,7 @@ struct ib_flow_spec_tunnel { union ib_flow_spec { struct { - enum ib_flow_spec_type type; + u32 type; u16 size; }; struct ib_flow_spec_eth eth; -- cgit v1.2.3 From c90ea9d8e51196d9c528e57d9ab09ee7d41f0ba0 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Wed, 23 Nov 2016 08:23:22 +0200 Subject: IB/core: Change ib_resolve_eth_dmac to use it in create AH The function ib_resolve_eth_dmac() requires struct qp_attr * and qp_attr_mask as parameters while the function might be useful to resolve dmac for address handles. This patch changes the signature of the function so it can be used in the flow of creating an address handle. Signed-off-by: Moni Shoua Reviewed-by: Yishai Hadas Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 6d0dd6525e14..0c6f973e407c 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -3394,4 +3394,6 @@ void ib_drain_rq(struct ib_qp *qp); void ib_drain_sq(struct ib_qp *qp); void ib_drain_qp(struct ib_qp *qp); +int ib_resolve_eth_dmac(struct ib_device *device, + struct ib_ah_attr *ah_attr); #endif /* IB_VERBS_H */ -- cgit v1.2.3 From 6ad279c5a2e55bf2bd100b4222090d4717de88d5 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Wed, 23 Nov 2016 08:23:23 +0200 Subject: IB/mlx5: Report that device has udata response in create_ah To make mlx5 user driver aware of whether kernel driver returns dmac in user data response add a new flag that will be returned back to user-space through alloc_ucontext. Signed-off-by: Moni Shoua Reviewed-by: Yishai Hadas Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/mlx5-abi.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include') diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h index b0a41c8dc1fb..ac28729a1245 100644 --- a/include/uapi/rdma/mlx5-abi.h +++ b/include/uapi/rdma/mlx5-abi.h @@ -82,6 +82,7 @@ enum mlx5_ib_alloc_ucontext_resp_mask { enum mlx5_user_cmds_supp_uhw { MLX5_USER_CMDS_SUPP_UHW_QUERY_DEVICE = 1 << 0, + MLX5_USER_CMDS_SUPP_UHW_CREATE_AH = 1 << 1, }; struct mlx5_ib_alloc_ucontext_resp { -- cgit v1.2.3 From 477864c8fcd953e5a988073ca5be18bb7fd93410 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Wed, 23 Nov 2016 08:23:24 +0200 Subject: IB/core: Let create_ah return extended response to user Add struct ib_udata to the signature of create_ah callback that is implemented by IB device drivers. This allows HW drivers to return extra data to the userspace library. This patch prepares the ground for mlx5 driver to resolve destination mac address for a given GID and return it to userspace. This patch was previously submitted by Knut Omang as a part of the patch set to support Oracle's Infiniband HCA (SIF). Signed-off-by: Knut Omang Signed-off-by: Moni Shoua Reviewed-by: Yishai Hadas Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 0c6f973e407c..73417a22ee4d 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -1951,7 +1951,8 @@ struct ib_device { struct ib_udata *udata); int (*dealloc_pd)(struct ib_pd *pd); struct ib_ah * (*create_ah)(struct ib_pd *pd, - struct ib_ah_attr *ah_attr); + struct ib_ah_attr *ah_attr, + struct ib_udata *udata); int (*modify_ah)(struct ib_ah *ah, struct ib_ah_attr *ah_attr); int (*query_ah)(struct ib_ah *ah, -- cgit v1.2.3 From 5097e71f3edafad3e7d8d4f9c4a137d9aad0fae2 Mon Sep 17 00:00:00 2001 From: Moni Shoua Date: Wed, 23 Nov 2016 08:23:25 +0200 Subject: IB/mlx5: Use kernel driver to help userspace create ah Resolving a MAC address for a given IP address in userspace is inefficient. This patch lets mlx5 user driver using the kernel driver to resolve the mac and get the answer in the private section of the response. Signed-off-by: Moni Shoua Reviewed-by: Yishai Hadas Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/mlx5-abi.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include') diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h index ac28729a1245..3ebf3db24c34 100644 --- a/include/uapi/rdma/mlx5-abi.h +++ b/include/uapi/rdma/mlx5-abi.h @@ -249,6 +249,12 @@ struct mlx5_ib_create_wq { __u32 reserved; }; +struct mlx5_ib_create_ah_resp { + __u32 response_length; + __u8 dmac[ETH_ALEN]; + __u8 reserved[6]; +}; + struct mlx5_ib_create_wq_resp { __u32 response_length; __u32 reserved; -- cgit v1.2.3 From d949167d68b304c0a00331cf33ef49a29b65d85f Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Thu, 1 Dec 2016 13:43:13 +0200 Subject: IB/mlx5: Report mlx5 packet pacing capabilities when querying device Enable mlx5 based hardware to report packet pacing capabilities from kernel to user space. Packet pacing allows to limit the rate to any number between the maximum and minimum, based on user settings. The capabilities are exposed to user space through query_device by uhw. The following capabilities are reported: 1. The maximum and minimum rate limit in kbps supported by packet pacing. 2. Bitmap showing which QP types are supported by packet pacing operation. Signed-off-by: Bodong Wang Reviewed-by: Matan Barak Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/mlx5-abi.h | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'include') diff --git a/include/uapi/rdma/mlx5-abi.h b/include/uapi/rdma/mlx5-abi.h index 3ebf3db24c34..fae6cdaeb56d 100644 --- a/include/uapi/rdma/mlx5-abi.h +++ b/include/uapi/rdma/mlx5-abi.h @@ -136,12 +136,25 @@ struct mlx5_ib_cqe_comp_caps { __u32 supported_format; /* enum mlx5_ib_cqe_comp_res_format */ }; +struct mlx5_packet_pacing_caps { + __u32 qp_rate_limit_min; + __u32 qp_rate_limit_max; /* In kpbs */ + + /* Corresponding bit will be set if qp type from + * 'enum ib_qp_type' is supported, e.g. + * supported_qpts |= 1 << IB_QPT_RAW_PACKET + */ + __u32 supported_qpts; + __u32 reserved; +}; + 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_packet_pacing_caps packet_pacing_caps; __u32 mlx5_ib_support_multi_pkt_send_wqes; __u32 reserved; }; -- cgit v1.2.3 From 528e5a1bd3f0e9b760cb3a1062fce7513712a15d Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Thu, 1 Dec 2016 13:43:14 +0200 Subject: IB/core: Support rate limit for packet pacing Add new member rate_limit to ib_qp_attr which holds the packet pacing rate in kbps, 0 means unlimited. IB_QP_RATE_LIMIT is added to ib_attr_mask and could be used by RAW QPs when changing QP state from RTR to RTS, RTS to RTS. Signed-off-by: Bodong Wang Reviewed-by: Matan Barak Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/rdma/ib_verbs.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include') diff --git a/include/rdma/ib_verbs.h b/include/rdma/ib_verbs.h index 73417a22ee4d..8029d2a51f14 100644 --- a/include/rdma/ib_verbs.h +++ b/include/rdma/ib_verbs.h @@ -1102,6 +1102,7 @@ enum ib_qp_attr_mask { IB_QP_RESERVED2 = (1<<22), IB_QP_RESERVED3 = (1<<23), IB_QP_RESERVED4 = (1<<24), + IB_QP_RATE_LIMIT = (1<<25), }; enum ib_qp_state { @@ -1151,6 +1152,7 @@ struct ib_qp_attr { u8 rnr_retry; u8 alt_port_num; u8 alt_timeout; + u32 rate_limit; }; enum ib_wr_opcode { -- cgit v1.2.3 From 189aba99e70030cfb56bd8f199bc5b077a1bc6ff Mon Sep 17 00:00:00 2001 From: Bodong Wang Date: Thu, 1 Dec 2016 13:43:15 +0200 Subject: IB/uverbs: Extend modify_qp and support packet pacing An new uverbs command ib_uverbs_ex_modify_qp is added to support more QP attributes. User driver should choose to call the legacy/extended API based on input mask. IB_USER_LAST_QP_ATTR_MASK is added to indicated the maximum bit position which supports legacy ib_uverbs_modify_qp. IB_USER_LEGACY_LAST_QP_ATTR_MASK indicates the maximum bit position which supports ib_uverbs_ex_modify_qp, the value of this mask should be updated if new mask is added later. Along with this change, rate_limit is supported by the extended command, user driver could use it to control packet packing. Signed-off-by: Bodong Wang Reviewed-by: Matan Barak Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- include/uapi/rdma/ib_user_verbs.h | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) (limited to 'include') diff --git a/include/uapi/rdma/ib_user_verbs.h b/include/uapi/rdma/ib_user_verbs.h index 90ba5e88ec00..dfdfe4e92d31 100644 --- a/include/uapi/rdma/ib_user_verbs.h +++ b/include/uapi/rdma/ib_user_verbs.h @@ -37,6 +37,7 @@ #define IB_USER_VERBS_H #include +#include /* * Increment this value if any changes that break userspace ABI @@ -93,6 +94,7 @@ enum { IB_USER_VERBS_EX_CMD_QUERY_DEVICE = IB_USER_VERBS_CMD_QUERY_DEVICE, IB_USER_VERBS_EX_CMD_CREATE_CQ = IB_USER_VERBS_CMD_CREATE_CQ, IB_USER_VERBS_EX_CMD_CREATE_QP = IB_USER_VERBS_CMD_CREATE_QP, + IB_USER_VERBS_EX_CMD_MODIFY_QP = IB_USER_VERBS_CMD_MODIFY_QP, IB_USER_VERBS_EX_CMD_CREATE_FLOW = IB_USER_VERBS_CMD_THRESHOLD, IB_USER_VERBS_EX_CMD_DESTROY_FLOW, IB_USER_VERBS_EX_CMD_CREATE_WQ, @@ -545,6 +547,14 @@ enum { IB_UVERBS_CREATE_QP_SUP_COMP_MASK = IB_UVERBS_CREATE_QP_MASK_IND_TABLE, }; +enum { + IB_USER_LEGACY_LAST_QP_ATTR_MASK = IB_QP_DEST_QPN +}; + +enum { + IB_USER_LAST_QP_ATTR_MASK = IB_QP_RATE_LIMIT +}; + struct ib_uverbs_ex_create_qp { __u64 user_handle; __u32 pd_handle; @@ -684,9 +694,20 @@ struct ib_uverbs_modify_qp { __u64 driver_data[0]; }; +struct ib_uverbs_ex_modify_qp { + struct ib_uverbs_modify_qp base; + __u32 rate_limit; + __u32 reserved; +}; + struct ib_uverbs_modify_qp_resp { }; +struct ib_uverbs_ex_modify_qp_resp { + __u32 comp_mask; + __u32 response_length; +}; + struct ib_uverbs_destroy_qp { __u64 response; __u32 qp_handle; -- cgit v1.2.3