X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5_rxtx.h;h=5708c2a7a7ea41247a2c730f8678c116a3cbb666;hb=4ec6960aec264172ac22f89c31ec4b43234bb264;hp=41838203473fc55b0fd4d31457c8313d038d26cc;hpb=87011737b715ff7c3d4054777f702315d31b4982;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h index 4183820347..5708c2a7a7 100644 --- a/drivers/net/mlx5/mlx5_rxtx.h +++ b/drivers/net/mlx5/mlx5_rxtx.h @@ -34,31 +34,36 @@ #ifndef RTE_PMD_MLX5_RXTX_H_ #define RTE_PMD_MLX5_RXTX_H_ +#include #include /* Verbs header. */ /* ISO C doesn't support unnamed structs/unions, disabling -pedantic. */ #ifdef PEDANTIC -#pragma GCC diagnostic ignored "-pedantic" +#pragma GCC diagnostic ignored "-Wpedantic" #endif #include +#include #ifdef PEDANTIC -#pragma GCC diagnostic error "-pedantic" +#pragma GCC diagnostic error "-Wpedantic" #endif /* DPDK headers don't like -pedantic. */ #ifdef PEDANTIC -#pragma GCC diagnostic ignored "-pedantic" +#pragma GCC diagnostic ignored "-Wpedantic" #endif #include #include +#include #ifdef PEDANTIC -#pragma GCC diagnostic error "-pedantic" +#pragma GCC diagnostic error "-Wpedantic" #endif #include "mlx5_utils.h" #include "mlx5.h" +#include "mlx5_autoconf.h" #include "mlx5_defs.h" +#include "mlx5_prm.h" struct mlx5_rxq_stats { unsigned int idx; /**< Mapping index. */ @@ -79,112 +84,250 @@ struct mlx5_txq_stats { uint64_t odropped; /**< Total of packets not sent when TX ring full. */ }; -/* RX element (scattered packets). */ -struct rxq_elt_sp { - struct ibv_recv_wr wr; /* Work Request. */ - struct ibv_sge sges[MLX5_PMD_SGE_WR_N]; /* Scatter/Gather Elements. */ - struct rte_mbuf *bufs[MLX5_PMD_SGE_WR_N]; /* SGEs buffers. */ -}; - -/* RX element. */ -struct rxq_elt { - struct ibv_recv_wr wr; /* Work Request. */ - struct ibv_sge sge; /* Scatter/Gather Element. */ - /* mbuf pointer is derived from WR_ID(wr.wr_id).offset. */ +/* Flow director queue structure. */ +struct fdir_queue { + struct ibv_qp *qp; /* Associated RX QP. */ + struct ibv_exp_rwq_ind_table *ind_table; /* Indirection table. */ + struct ibv_exp_wq *wq; /* Work queue. */ + struct ibv_cq *cq; /* Completion queue. */ }; struct priv; +/* Compressed CQE context. */ +struct rxq_zip { + uint16_t ai; /* Array index. */ + uint16_t ca; /* Current array index. */ + uint16_t na; /* Next array index. */ + uint16_t cq_ci; /* The next CQE. */ + uint32_t cqe_cnt; /* Number of CQEs. */ +}; + /* RX queue descriptor. */ struct rxq { + unsigned int csum:1; /* Enable checksum offloading. */ + unsigned int csum_l2tun:1; /* Same for L2 tunnels. */ + unsigned int vlan_strip:1; /* Enable VLAN stripping. */ + unsigned int crc_present:1; /* CRC must be subtracted. */ + unsigned int sges_n:2; /* Log 2 of SGEs (max buffers per packet). */ + unsigned int cqe_n:4; /* Log 2 of CQ elements. */ + unsigned int elts_n:4; /* Log 2 of Mbufs. */ + unsigned int port_id:8; + unsigned int rss_hash:1; /* RSS hash result is enabled. */ + unsigned int :9; /* Remaining bits. */ + volatile uint32_t *rq_db; + volatile uint32_t *cq_db; + uint16_t rq_ci; + uint16_t cq_ci; + volatile struct mlx5_wqe_data_seg(*wqes)[]; + volatile struct mlx5_cqe(*cqes)[]; + struct rxq_zip zip; /* Compressed context. */ + struct rte_mbuf *(*elts)[]; + struct rte_mempool *mp; + struct mlx5_rxq_stats stats; +} __rte_cache_aligned; + +/* RX queue control descriptor. */ +struct rxq_ctrl { struct priv *priv; /* Back pointer to private data. */ - struct rte_mempool *mp; /* Memory Pool for allocations. */ - struct ibv_mr *mr; /* Memory Region (for mp). */ struct ibv_cq *cq; /* Completion Queue. */ - struct ibv_qp *qp; /* Queue Pair. */ - struct ibv_exp_qp_burst_family *if_qp; /* QP burst interface. */ - struct ibv_exp_cq_family *if_cq; /* CQ interface. */ - /* MAC flow steering rules. */ - struct ibv_flow *mac_flow[MLX5_MAX_MAC_ADDRESSES]; - unsigned int port_id; /* Port ID for incoming packets. */ - unsigned int elts_n; /* (*elts)[] length. */ - unsigned int elts_head; /* Current index in (*elts)[]. */ - union { - struct rxq_elt_sp (*sp)[]; /* Scattered RX elements. */ - struct rxq_elt (*no_sp)[]; /* RX elements. */ - } elts; - unsigned int sp:1; /* Use scattered RX elements. */ - uint32_t mb_len; /* Length of a mp-issued mbuf. */ - struct mlx5_rxq_stats stats; /* RX queue counters. */ - unsigned int socket; /* CPU socket ID for allocations. */ + struct ibv_exp_wq *wq; /* Work Queue. */ struct ibv_exp_res_domain *rd; /* Resource Domain. */ + struct fdir_queue *fdir_queue; /* Flow director queue. */ + struct ibv_mr *mr; /* Memory Region (for mp). */ + struct ibv_exp_wq_family *if_wq; /* WQ burst interface. */ + struct ibv_exp_cq_family_v1 *if_cq; /* CQ interface. */ + unsigned int socket; /* CPU socket ID for allocations. */ + struct rxq rxq; /* Data path structure. */ +}; + +/* Hash RX queue types. */ +enum hash_rxq_type { + HASH_RXQ_TCPV4, + HASH_RXQ_UDPV4, + HASH_RXQ_IPV4, + HASH_RXQ_TCPV6, + HASH_RXQ_UDPV6, + HASH_RXQ_IPV6, + HASH_RXQ_ETH, }; -/* TX element. */ -struct txq_elt { - struct rte_mbuf *buf; +/* Flow structure with Ethernet specification. It is packed to prevent padding + * between attr and spec as this layout is expected by libibverbs. */ +struct flow_attr_spec_eth { + struct ibv_exp_flow_attr attr; + struct ibv_exp_flow_spec_eth spec; +} __attribute__((packed)); + +/* Define a struct flow_attr_spec_eth object as an array of at least + * "size" bytes. Room after the first index is normally used to store + * extra flow specifications. */ +#define FLOW_ATTR_SPEC_ETH(name, size) \ + struct flow_attr_spec_eth name \ + [((size) / sizeof(struct flow_attr_spec_eth)) + \ + !!((size) % sizeof(struct flow_attr_spec_eth))] + +/* Initialization data for hash RX queue. */ +struct hash_rxq_init { + uint64_t hash_fields; /* Fields that participate in the hash. */ + uint64_t dpdk_rss_hf; /* Matching DPDK RSS hash fields. */ + unsigned int flow_priority; /* Flow priority to use. */ + union { + struct { + enum ibv_exp_flow_spec_type type; + uint16_t size; + } hdr; + struct ibv_exp_flow_spec_tcp_udp tcp_udp; + struct ibv_exp_flow_spec_ipv4 ipv4; + struct ibv_exp_flow_spec_ipv6 ipv6; + struct ibv_exp_flow_spec_eth eth; + } flow_spec; /* Flow specification template. */ + const struct hash_rxq_init *underlayer; /* Pointer to underlayer. */ }; -/* Linear buffer type. It is used when transmitting buffers with too many - * segments that do not fit the hardware queue (see max_send_sge). - * Extra segments are copied (linearized) in such buffers, replacing the - * last SGE during TX. - * The size is arbitrary but large enough to hold a jumbo frame with - * 8 segments considering mbuf.buf_len is about 2048 bytes. */ -typedef uint8_t linear_t[16384]; +/* Initialization data for indirection table. */ +struct ind_table_init { + unsigned int max_size; /* Maximum number of WQs. */ + /* Hash RX queues using this table. */ + unsigned int hash_types; + unsigned int hash_types_n; +}; + +/* Initialization data for special flows. */ +struct special_flow_init { + uint8_t dst_mac_val[6]; + uint8_t dst_mac_mask[6]; + unsigned int hash_types; + unsigned int per_vlan:1; +}; + +enum hash_rxq_flow_type { + HASH_RXQ_FLOW_TYPE_PROMISC, + HASH_RXQ_FLOW_TYPE_ALLMULTI, + HASH_RXQ_FLOW_TYPE_BROADCAST, + HASH_RXQ_FLOW_TYPE_IPV6MULTI, + HASH_RXQ_FLOW_TYPE_MAC, +}; + +#ifndef NDEBUG +static inline const char * +hash_rxq_flow_type_str(enum hash_rxq_flow_type flow_type) +{ + switch (flow_type) { + case HASH_RXQ_FLOW_TYPE_PROMISC: + return "promiscuous"; + case HASH_RXQ_FLOW_TYPE_ALLMULTI: + return "allmulticast"; + case HASH_RXQ_FLOW_TYPE_BROADCAST: + return "broadcast"; + case HASH_RXQ_FLOW_TYPE_IPV6MULTI: + return "IPv6 multicast"; + case HASH_RXQ_FLOW_TYPE_MAC: + return "MAC"; + } + return NULL; +} +#endif /* NDEBUG */ + +struct hash_rxq { + struct priv *priv; /* Back pointer to private data. */ + struct ibv_qp *qp; /* Hash RX QP. */ + enum hash_rxq_type type; /* Hash RX queue type. */ + /* MAC flow steering rules, one per VLAN ID. */ + struct ibv_exp_flow *mac_flow + [MLX5_MAX_MAC_ADDRESSES][MLX5_MAX_VLAN_IDS]; + struct ibv_exp_flow *special_flow + [MLX5_MAX_SPECIAL_FLOWS][MLX5_MAX_VLAN_IDS]; +}; /* TX queue descriptor. */ +RTE_STD_C11 struct txq { - struct priv *priv; /* Back pointer to private data. */ + uint16_t elts_head; /* Current index in (*elts)[]. */ + uint16_t elts_tail; /* First element awaiting completion. */ + uint16_t elts_comp; /* Counter since last completion request. */ + uint16_t cq_ci; /* Consumer index for completion queue. */ + uint16_t wqe_ci; /* Consumer index for work queue. */ + uint16_t elts_n:4; /* (*elts)[] length (in log2). */ + uint16_t cqe_n:4; /* Number of CQ elements (in log2). */ + uint16_t wqe_n:4; /* Number of of WQ elements (in log2). */ + uint16_t bf_buf_size:4; /* Log2 Blueflame size. */ + uint16_t bf_offset; /* Blueflame offset. */ + uint16_t max_inline; /* Multiple of RTE_CACHE_LINE_SIZE to inline. */ + uint32_t qp_num_8s; /* QP number shifted by 8. */ + volatile struct mlx5_cqe (*cqes)[]; /* Completion queue. */ + volatile struct mlx5_wqe64 (*wqes)[]; /* Work queue. */ + volatile uint32_t *qp_db; /* Work queue doorbell. */ + volatile uint32_t *cq_db; /* Completion queue doorbell. */ + volatile void *bf_reg; /* Blueflame register. */ struct { - struct rte_mempool *mp; /* Cached Memory Pool. */ + const struct rte_mempool *mp; /* Cached Memory Pool. */ struct ibv_mr *mr; /* Memory Region (for mp). */ - uint32_t lkey; /* mr->lkey */ + uint32_t lkey; /* htonl(mr->lkey) */ } mp2mr[MLX5_PMD_TX_MP_CACHE]; /* MP to MR translation table. */ + struct rte_mbuf *(*elts)[]; /* TX elements. */ + struct mlx5_txq_stats stats; /* TX queue counters. */ +} __rte_cache_aligned; + +/* TX queue control descriptor. */ +struct txq_ctrl { + struct priv *priv; /* Back pointer to private data. */ struct ibv_cq *cq; /* Completion Queue. */ struct ibv_qp *qp; /* Queue Pair. */ struct ibv_exp_qp_burst_family *if_qp; /* QP burst interface. */ struct ibv_exp_cq_family *if_cq; /* CQ interface. */ -#if MLX5_PMD_MAX_INLINE > 0 - uint32_t max_inline; /* Max inline send size <= MLX5_PMD_MAX_INLINE. */ -#endif - unsigned int elts_n; /* (*elts)[] length. */ - struct txq_elt (*elts)[]; /* TX elements. */ - unsigned int elts_head; /* Current index in (*elts)[]. */ - unsigned int elts_tail; /* First element awaiting completion. */ - unsigned int elts_comp; /* Number of completion requests. */ - unsigned int elts_comp_cd; /* Countdown for next completion request. */ - unsigned int elts_comp_cd_init; /* Initial value for countdown. */ - struct mlx5_txq_stats stats; /* TX queue counters. */ - linear_t (*elts_linear)[]; /* Linearized buffers. */ - struct ibv_mr *mr_linear; /* Memory Region for linearized buffers. */ - unsigned int socket; /* CPU socket ID for allocations. */ struct ibv_exp_res_domain *rd; /* Resource Domain. */ + unsigned int socket; /* CPU socket ID for allocations. */ + struct txq txq; /* Data path structure. */ }; /* mlx5_rxq.c */ -void rxq_cleanup(struct rxq *); -int rxq_rehash(struct rte_eth_dev *, struct rxq *); -int rxq_setup(struct rte_eth_dev *, struct rxq *, uint16_t, unsigned int, - const struct rte_eth_rxconf *, struct rte_mempool *); +extern const struct hash_rxq_init hash_rxq_init[]; +extern const unsigned int hash_rxq_init_n; + +extern uint8_t rss_hash_default_key[]; +extern const size_t rss_hash_default_key_len; + +size_t priv_flow_attr(struct priv *, struct ibv_exp_flow_attr *, + size_t, enum hash_rxq_type); +int priv_create_hash_rxqs(struct priv *); +void priv_destroy_hash_rxqs(struct priv *); +int priv_allow_flow_type(struct priv *, enum hash_rxq_flow_type); +int priv_rehash_flows(struct priv *); +void rxq_cleanup(struct rxq_ctrl *); +int rxq_rehash(struct rte_eth_dev *, struct rxq_ctrl *); +int rxq_ctrl_setup(struct rte_eth_dev *, struct rxq_ctrl *, uint16_t, + unsigned int, const struct rte_eth_rxconf *, + struct rte_mempool *); int mlx5_rx_queue_setup(struct rte_eth_dev *, uint16_t, uint16_t, unsigned int, const struct rte_eth_rxconf *, struct rte_mempool *); void mlx5_rx_queue_release(void *); +uint16_t mlx5_rx_burst_secondary_setup(void *, struct rte_mbuf **, uint16_t); /* mlx5_txq.c */ -void txq_cleanup(struct txq *); +void txq_cleanup(struct txq_ctrl *); +int txq_ctrl_setup(struct rte_eth_dev *, struct txq_ctrl *, uint16_t, + unsigned int, const struct rte_eth_txconf *); int mlx5_tx_queue_setup(struct rte_eth_dev *, uint16_t, uint16_t, unsigned int, const struct rte_eth_txconf *); void mlx5_tx_queue_release(void *); +uint16_t mlx5_tx_burst_secondary_setup(void *, struct rte_mbuf **, uint16_t); /* mlx5_rxtx.c */ uint16_t mlx5_tx_burst(void *, struct rte_mbuf **, uint16_t); -uint16_t mlx5_rx_burst_sp(void *, struct rte_mbuf **, uint16_t); +uint16_t mlx5_tx_burst_mpw(void *, struct rte_mbuf **, uint16_t); +uint16_t mlx5_tx_burst_mpw_inline(void *, struct rte_mbuf **, uint16_t); uint16_t mlx5_rx_burst(void *, struct rte_mbuf **, uint16_t); uint16_t removed_tx_burst(void *, struct rte_mbuf **, uint16_t); uint16_t removed_rx_burst(void *, struct rte_mbuf **, uint16_t); +/* mlx5_mr.c */ + +struct ibv_mr *mlx5_mp2mr(struct ibv_pd *, struct rte_mempool *); +void txq_mp2mr_iter(struct rte_mempool *, void *); +uint32_t txq_mp2mr_reg(struct txq *, struct rte_mempool *, unsigned int); + #endif /* RTE_PMD_MLX5_RXTX_H_ */