X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fmlx5%2Fmlx5_rxq.c;h=889428f48ab95897a0f472e68e4e6affd2c44d93;hb=80f872ee0222e3936856e43a693168425b1c78ae;hp=dacffc92511767615c2901004bffdb5ceb2eb169;hpb=2fec07edd4bca7ead134a5a1b4672a682d6c6ad4;p=dpdk.git diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c index dacffc9251..889428f48a 100644 --- a/drivers/net/mlx5/mlx5_rxq.c +++ b/drivers/net/mlx5/mlx5_rxq.c @@ -21,13 +21,16 @@ #include #include +#include +#include #include "mlx5_defs.h" #include "mlx5.h" -#include "mlx5_tx.h" #include "mlx5_rx.h" #include "mlx5_utils.h" #include "mlx5_autoconf.h" +#include "mlx5_devx.h" +#include "rte_pmd_mlx5.h" /* Default RSS hash key also used for ConnectX-3. */ @@ -65,7 +68,7 @@ mlx5_rxq_cqe_num(struct mlx5_rxq_data *rxq_data) unsigned int wqe_n = 1 << rxq_data->elts_n; if (mlx5_rxq_mprq_enabled(rxq_data)) - cqe_n = wqe_n * (1 << rxq_data->strd_num_n) - 1; + cqe_n = wqe_n * RTE_BIT32(rxq_data->log_strd_num) - 1; else cqe_n = wqe_n - 1; return cqe_n; @@ -128,15 +131,17 @@ error: * Pointer to RX queue structure. * * @return - * 0 on success, errno value on failure. + * 0 on success, negative errno value on failure. */ static int rxq_alloc_elts_sprq(struct mlx5_rxq_ctrl *rxq_ctrl) { const unsigned int sges_n = 1 << rxq_ctrl->rxq.sges_n; unsigned int elts_n = mlx5_rxq_mprq_enabled(&rxq_ctrl->rxq) ? - (1 << rxq_ctrl->rxq.elts_n) * (1 << rxq_ctrl->rxq.strd_num_n) : - (1 << rxq_ctrl->rxq.elts_n); + RTE_BIT32(rxq_ctrl->rxq.elts_n) * + RTE_BIT32(rxq_ctrl->rxq.log_strd_num) : + RTE_BIT32(rxq_ctrl->rxq.elts_n); + bool has_vec_support = mlx5_rxq_check_vec_support(&rxq_ctrl->rxq) > 0; unsigned int i; int err; @@ -147,13 +152,20 @@ rxq_alloc_elts_sprq(struct mlx5_rxq_ctrl *rxq_ctrl) buf = rte_pktmbuf_alloc(seg->mp); if (buf == NULL) { - DRV_LOG(ERR, "port %u empty mbuf pool", - PORT_ID(rxq_ctrl->priv)); + if (rxq_ctrl->share_group == 0) + DRV_LOG(ERR, "port %u queue %u empty mbuf pool", + RXQ_PORT_ID(rxq_ctrl), + rxq_ctrl->rxq.idx); + else + DRV_LOG(ERR, "share group %u queue %u empty mbuf pool", + rxq_ctrl->share_group, + rxq_ctrl->share_qid); rte_errno = ENOMEM; goto error; } - /* Headroom is reserved by rte_pktmbuf_alloc(). */ - MLX5_ASSERT(DATA_OFF(buf) == RTE_PKTMBUF_HEADROOM); + /* Only vectored Rx routines rely on headroom size. */ + MLX5_ASSERT(!has_vec_support || + DATA_OFF(buf) >= RTE_PKTMBUF_HEADROOM); /* Buffer is supposed to be empty. */ MLX5_ASSERT(rte_pktmbuf_data_len(buf) == 0); MLX5_ASSERT(rte_pktmbuf_pkt_len(buf) == 0); @@ -166,7 +178,7 @@ rxq_alloc_elts_sprq(struct mlx5_rxq_ctrl *rxq_ctrl) (*rxq_ctrl->rxq.elts)[i] = buf; } /* If Rx vector is activated. */ - if (mlx5_rxq_check_vec_support(&rxq_ctrl->rxq) > 0) { + if (has_vec_support) { struct mlx5_rxq_data *rxq = &rxq_ctrl->rxq; struct rte_mbuf *mbuf_init = &rxq->fake_mbuf; struct rte_pktmbuf_pool_private *priv = @@ -178,9 +190,10 @@ rxq_alloc_elts_sprq(struct mlx5_rxq_ctrl *rxq_ctrl) mbuf_init->data_off = RTE_PKTMBUF_HEADROOM; rte_mbuf_refcnt_set(mbuf_init, 1); mbuf_init->nb_segs = 1; - mbuf_init->port = rxq->port_id; + /* For shared queues port is provided in CQE */ + mbuf_init->port = rxq->shared ? 0 : rxq->port_id; if (priv->flags & RTE_PKTMBUF_POOL_F_PINNED_EXT_BUF) - mbuf_init->ol_flags = EXT_ATTACHED_MBUF; + mbuf_init->ol_flags = RTE_MBUF_F_EXTERNAL; /* * prevent compiler reordering: * rearm_data covers previous fields. @@ -192,11 +205,16 @@ rxq_alloc_elts_sprq(struct mlx5_rxq_ctrl *rxq_ctrl) for (j = 0; j < MLX5_VPMD_DESCS_PER_LOOP; ++j) (*rxq->elts)[elts_n + j] = &rxq->fake_mbuf; } - DRV_LOG(DEBUG, - "port %u SPRQ queue %u allocated and configured %u segments" - " (max %u packets)", - PORT_ID(rxq_ctrl->priv), rxq_ctrl->rxq.idx, elts_n, - elts_n / (1 << rxq_ctrl->rxq.sges_n)); + if (rxq_ctrl->share_group == 0) + DRV_LOG(DEBUG, + "port %u SPRQ queue %u allocated and configured %u segments (max %u packets)", + RXQ_PORT_ID(rxq_ctrl), rxq_ctrl->rxq.idx, elts_n, + elts_n / (1 << rxq_ctrl->rxq.sges_n)); + else + DRV_LOG(DEBUG, + "share group %u SPRQ queue %u allocated and configured %u segments (max %u packets)", + rxq_ctrl->share_group, rxq_ctrl->share_qid, elts_n, + elts_n / (1 << rxq_ctrl->rxq.sges_n)); return 0; error: err = rte_errno; /* Save rte_errno before cleanup. */ @@ -206,8 +224,12 @@ error: rte_pktmbuf_free_seg((*rxq_ctrl->rxq.elts)[i]); (*rxq_ctrl->rxq.elts)[i] = NULL; } - DRV_LOG(DEBUG, "port %u SPRQ queue %u failed, freed everything", - PORT_ID(rxq_ctrl->priv), rxq_ctrl->rxq.idx); + if (rxq_ctrl->share_group == 0) + DRV_LOG(DEBUG, "port %u SPRQ queue %u failed, freed everything", + RXQ_PORT_ID(rxq_ctrl), rxq_ctrl->rxq.idx); + else + DRV_LOG(DEBUG, "share group %u SPRQ queue %u failed, freed everything", + rxq_ctrl->share_group, rxq_ctrl->share_qid); rte_errno = err; /* Restore rte_errno. */ return -rte_errno; } @@ -219,7 +241,7 @@ error: * Pointer to RX queue structure. * * @return - * 0 on success, errno value on failure. + * 0 on success, negative errno value on failure. */ int rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl) @@ -232,7 +254,9 @@ rxq_alloc_elts(struct mlx5_rxq_ctrl *rxq_ctrl) */ if (mlx5_rxq_mprq_enabled(&rxq_ctrl->rxq)) ret = rxq_alloc_elts_mprq(rxq_ctrl); - return (ret || rxq_alloc_elts_sprq(rxq_ctrl)); + if (ret == 0) + ret = rxq_alloc_elts_sprq(rxq_ctrl); + return ret; } /** @@ -273,16 +297,20 @@ rxq_free_elts_sprq(struct mlx5_rxq_ctrl *rxq_ctrl) { struct mlx5_rxq_data *rxq = &rxq_ctrl->rxq; const uint16_t q_n = mlx5_rxq_mprq_enabled(&rxq_ctrl->rxq) ? - (1 << rxq->elts_n) * (1 << rxq->strd_num_n) : - (1 << rxq->elts_n); + RTE_BIT32(rxq->elts_n) * RTE_BIT32(rxq->log_strd_num) : + RTE_BIT32(rxq->elts_n); const uint16_t q_mask = q_n - 1; uint16_t elts_ci = mlx5_rxq_mprq_enabled(&rxq_ctrl->rxq) ? rxq->elts_ci : rxq->rq_ci; uint16_t used = q_n - (elts_ci - rxq->rq_pi); uint16_t i; - DRV_LOG(DEBUG, "port %u Rx queue %u freeing %d WRs", - PORT_ID(rxq_ctrl->priv), rxq->idx, q_n); + if (rxq_ctrl->share_group == 0) + DRV_LOG(DEBUG, "port %u Rx queue %u freeing %d WRs", + RXQ_PORT_ID(rxq_ctrl), rxq->idx, q_n); + else + DRV_LOG(DEBUG, "share group %u Rx queue %u freeing %d WRs", + rxq_ctrl->share_group, rxq_ctrl->share_qid, q_n); if (rxq->elts == NULL) return; /** @@ -332,24 +360,22 @@ uint64_t mlx5_get_rx_queue_offloads(struct rte_eth_dev *dev) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_dev_config *config = &priv->config; - uint64_t offloads = (DEV_RX_OFFLOAD_SCATTER | - DEV_RX_OFFLOAD_TIMESTAMP | - DEV_RX_OFFLOAD_JUMBO_FRAME | - DEV_RX_OFFLOAD_RSS_HASH); + uint64_t offloads = (RTE_ETH_RX_OFFLOAD_SCATTER | + RTE_ETH_RX_OFFLOAD_TIMESTAMP | + RTE_ETH_RX_OFFLOAD_RSS_HASH); - if (!config->mprq.enabled) + if (!priv->config.mprq.enabled) offloads |= RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT; - if (config->hw_fcs_strip) - offloads |= DEV_RX_OFFLOAD_KEEP_CRC; - if (config->hw_csum) - offloads |= (DEV_RX_OFFLOAD_IPV4_CKSUM | - DEV_RX_OFFLOAD_UDP_CKSUM | - DEV_RX_OFFLOAD_TCP_CKSUM); - if (config->hw_vlan_strip) - offloads |= DEV_RX_OFFLOAD_VLAN_STRIP; - if (MLX5_LRO_SUPPORTED(dev)) - offloads |= DEV_RX_OFFLOAD_TCP_LRO; + if (priv->sh->config.hw_fcs_strip) + offloads |= RTE_ETH_RX_OFFLOAD_KEEP_CRC; + if (priv->sh->dev_cap.hw_csum) + offloads |= (RTE_ETH_RX_OFFLOAD_IPV4_CKSUM | + RTE_ETH_RX_OFFLOAD_UDP_CKSUM | + RTE_ETH_RX_OFFLOAD_TCP_CKSUM); + if (priv->sh->dev_cap.hw_vlan_strip) + offloads |= RTE_ETH_RX_OFFLOAD_VLAN_STRIP; + if (priv->sh->dev_cap.lro_supported) + offloads |= RTE_ETH_RX_OFFLOAD_TCP_LRO; return offloads; } @@ -363,7 +389,7 @@ mlx5_get_rx_queue_offloads(struct rte_eth_dev *dev) uint64_t mlx5_get_rx_port_offloads(void) { - uint64_t offloads = DEV_RX_OFFLOAD_VLAN_FILTER; + uint64_t offloads = RTE_ETH_RX_OFFLOAD_VLAN_FILTER; return offloads; } @@ -384,15 +410,13 @@ mlx5_get_rx_port_offloads(void) static int mlx5_rxq_releasable(struct rte_eth_dev *dev, uint16_t idx) { - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_ctrl *rxq_ctrl; + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); - if (!(*priv->rxqs)[idx]) { + if (rxq == NULL) { rte_errno = EINVAL; return -rte_errno; } - rxq_ctrl = container_of((*priv->rxqs)[idx], struct mlx5_rxq_ctrl, rxq); - return (__atomic_load_n(&rxq_ctrl->refcnt, __ATOMIC_RELAXED) == 1); + return (__atomic_load_n(&rxq->refcnt, __ATOMIC_RELAXED) == 1); } /* Fetches and drops all SW-owned and error CQEs to synchronize CQ. */ @@ -452,13 +476,13 @@ int mlx5_rx_queue_stop_primary(struct rte_eth_dev *dev, uint16_t idx) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_data *rxq = (*priv->rxqs)[idx]; - struct mlx5_rxq_ctrl *rxq_ctrl = - container_of(rxq, struct mlx5_rxq_ctrl, rxq); + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); + struct mlx5_rxq_ctrl *rxq_ctrl = rxq->ctrl; int ret; + MLX5_ASSERT(rxq != NULL && rxq_ctrl != NULL); MLX5_ASSERT(rte_eal_process_type() == RTE_PROC_PRIMARY); - ret = priv->obj_ops.rxq_obj_modify(rxq_ctrl->obj, MLX5_RXQ_MOD_RDY2RST); + ret = priv->obj_ops.rxq_obj_modify(rxq, MLX5_RXQ_MOD_RDY2RST); if (ret) { DRV_LOG(ERR, "Cannot change Rx WQ state to RESET: %s", strerror(errno)); @@ -466,7 +490,7 @@ mlx5_rx_queue_stop_primary(struct rte_eth_dev *dev, uint16_t idx) return ret; } /* Remove all processes CQEs. */ - rxq_sync_cq(rxq); + rxq_sync_cq(&rxq_ctrl->rxq); /* Free all involved mbufs. */ rxq_free_elts(rxq_ctrl); /* Set the actual queue state. */ @@ -538,26 +562,26 @@ int mlx5_rx_queue_start_primary(struct rte_eth_dev *dev, uint16_t idx) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_data *rxq = (*priv->rxqs)[idx]; - struct mlx5_rxq_ctrl *rxq_ctrl = - container_of(rxq, struct mlx5_rxq_ctrl, rxq); + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); + struct mlx5_rxq_data *rxq_data = &rxq->ctrl->rxq; int ret; - MLX5_ASSERT(rte_eal_process_type() == RTE_PROC_PRIMARY); + MLX5_ASSERT(rxq != NULL && rxq->ctrl != NULL); + MLX5_ASSERT(rte_eal_process_type() == RTE_PROC_PRIMARY); /* Allocate needed buffers. */ - ret = rxq_alloc_elts(rxq_ctrl); + ret = rxq_alloc_elts(rxq->ctrl); if (ret) { DRV_LOG(ERR, "Cannot reallocate buffers for Rx WQ"); rte_errno = errno; return ret; } rte_io_wmb(); - *rxq->cq_db = rte_cpu_to_be_32(rxq->cq_ci); + *rxq_data->cq_db = rte_cpu_to_be_32(rxq_data->cq_ci); rte_io_wmb(); /* Reset RQ consumer before moving queue to READY state. */ - *rxq->rq_db = rte_cpu_to_be_32(0); + *rxq_data->rq_db = rte_cpu_to_be_32(0); rte_io_wmb(); - ret = priv->obj_ops.rxq_obj_modify(rxq_ctrl->obj, MLX5_RXQ_MOD_RST2RDY); + ret = priv->obj_ops.rxq_obj_modify(rxq, MLX5_RXQ_MOD_RST2RDY); if (ret) { DRV_LOG(ERR, "Cannot change Rx WQ state to READY: %s", strerror(errno)); @@ -565,8 +589,8 @@ mlx5_rx_queue_start_primary(struct rte_eth_dev *dev, uint16_t idx) return ret; } /* Reinitialize RQ - set WQEs. */ - mlx5_rxq_initialize(rxq); - rxq->err_state = MLX5_RXQ_ERR_STATE_NO_ERROR; + mlx5_rxq_initialize(rxq_data); + rxq_data->err_state = MLX5_RXQ_ERR_STATE_NO_ERROR; /* Set actual queue state. */ dev->data->rx_queue_state[idx] = RTE_ETH_QUEUE_STATE_STARTED; return 0; @@ -614,14 +638,19 @@ mlx5_rx_queue_start(struct rte_eth_dev *dev, uint16_t idx) * RX queue index. * @param desc * Number of descriptors to configure in queue. + * @param[out] rxq_ctrl + * Address of pointer to shared Rx queue control. * * @return * 0 on success, a negative errno value otherwise and rte_errno is set. */ static int -mlx5_rx_queue_pre_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t *desc) +mlx5_rx_queue_pre_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t *desc, + struct mlx5_rxq_ctrl **rxq_ctrl) { struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_rxq_priv *rxq; + bool empty; if (!rte_is_power_of_2(*desc)) { *desc = 1 << log2above(*desc); @@ -638,16 +667,146 @@ mlx5_rx_queue_pre_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t *desc) rte_errno = EOVERFLOW; return -rte_errno; } - if (!mlx5_rxq_releasable(dev, idx)) { - DRV_LOG(ERR, "port %u unable to release queue index %u", - dev->data->port_id, idx); - rte_errno = EBUSY; - return -rte_errno; + if (rxq_ctrl == NULL || *rxq_ctrl == NULL) + return 0; + if (!(*rxq_ctrl)->rxq.shared) { + if (!mlx5_rxq_releasable(dev, idx)) { + DRV_LOG(ERR, "port %u unable to release queue index %u", + dev->data->port_id, idx); + rte_errno = EBUSY; + return -rte_errno; + } + mlx5_rxq_release(dev, idx); + } else { + if ((*rxq_ctrl)->obj != NULL) + /* Some port using shared Rx queue has been started. */ + return 0; + /* Release all owner RxQ to reconfigure Shared RxQ. */ + do { + rxq = LIST_FIRST(&(*rxq_ctrl)->owners); + LIST_REMOVE(rxq, owner_entry); + empty = LIST_EMPTY(&(*rxq_ctrl)->owners); + mlx5_rxq_release(ETH_DEV(rxq->priv), rxq->idx); + } while (!empty); + *rxq_ctrl = NULL; } - mlx5_rxq_release(dev, idx); return 0; } +/** + * Get the shared Rx queue object that matches group and queue index. + * + * @param dev + * Pointer to Ethernet device structure. + * @param group + * Shared RXQ group. + * @param share_qid + * Shared RX queue index. + * + * @return + * Shared RXQ object that matching, or NULL if not found. + */ +static struct mlx5_rxq_ctrl * +mlx5_shared_rxq_get(struct rte_eth_dev *dev, uint32_t group, uint16_t share_qid) +{ + struct mlx5_rxq_ctrl *rxq_ctrl; + struct mlx5_priv *priv = dev->data->dev_private; + + LIST_FOREACH(rxq_ctrl, &priv->sh->shared_rxqs, share_entry) { + if (rxq_ctrl->share_group == group && + rxq_ctrl->share_qid == share_qid) + return rxq_ctrl; + } + return NULL; +} + +/** + * Check whether requested Rx queue configuration matches shared RXQ. + * + * @param rxq_ctrl + * Pointer to shared RXQ. + * @param dev + * Pointer to Ethernet device structure. + * @param idx + * Queue index. + * @param desc + * Number of descriptors to configure in queue. + * @param socket + * NUMA socket on which memory must be allocated. + * @param[in] conf + * Thresholds parameters. + * @param mp + * Memory pool for buffer allocations. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +static bool +mlx5_shared_rxq_match(struct mlx5_rxq_ctrl *rxq_ctrl, struct rte_eth_dev *dev, + uint16_t idx, uint16_t desc, unsigned int socket, + const struct rte_eth_rxconf *conf, + struct rte_mempool *mp) +{ + struct mlx5_priv *spriv = LIST_FIRST(&rxq_ctrl->owners)->priv; + struct mlx5_priv *priv = dev->data->dev_private; + unsigned int i; + + RTE_SET_USED(conf); + if (rxq_ctrl->socket != socket) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: socket mismatch", + dev->data->port_id, idx); + return false; + } + if (rxq_ctrl->rxq.elts_n != log2above(desc)) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: descriptor number mismatch", + dev->data->port_id, idx); + return false; + } + if (priv->mtu != spriv->mtu) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: mtu mismatch", + dev->data->port_id, idx); + return false; + } + if (priv->dev_data->dev_conf.intr_conf.rxq != + spriv->dev_data->dev_conf.intr_conf.rxq) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: interrupt mismatch", + dev->data->port_id, idx); + return false; + } + if (mp != NULL && rxq_ctrl->rxq.mp != mp) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: mempool mismatch", + dev->data->port_id, idx); + return false; + } else if (mp == NULL) { + if (conf->rx_nseg != rxq_ctrl->rxseg_n) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: segment number mismatch", + dev->data->port_id, idx); + return false; + } + for (i = 0; i < conf->rx_nseg; i++) { + if (memcmp(&conf->rx_seg[i].split, &rxq_ctrl->rxseg[i], + sizeof(struct rte_eth_rxseg_split))) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: segment %u configuration mismatch", + dev->data->port_id, idx, i); + return false; + } + } + } + if (priv->config.hw_padding != spriv->config.hw_padding) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: padding mismatch", + dev->data->port_id, idx); + return false; + } + if (priv->config.cqe_comp != spriv->config.cqe_comp || + (priv->config.cqe_comp && + priv->config.cqe_comp_fmt != spriv->config.cqe_comp_fmt)) { + DRV_LOG(ERR, "port %u queue index %u failed to join shared group: CQE compression mismatch", + dev->data->port_id, idx); + return false; + } + return true; +} + /** * * @param dev @@ -672,14 +831,15 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, struct rte_mempool *mp) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_data *rxq = (*priv->rxqs)[idx]; - struct mlx5_rxq_ctrl *rxq_ctrl = - container_of(rxq, struct mlx5_rxq_ctrl, rxq); + struct mlx5_rxq_priv *rxq; + struct mlx5_rxq_ctrl *rxq_ctrl = NULL; struct rte_eth_rxseg_split *rx_seg = (struct rte_eth_rxseg_split *)conf->rx_seg; struct rte_eth_rxseg_split rx_single = {.mp = mp}; uint16_t n_seg = conf->rx_nseg; int res; + uint64_t offloads = conf->offloads | + dev->data->dev_conf.rxmode.offloads; if (mp) { /* @@ -691,11 +851,8 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, n_seg = 1; } if (n_seg > 1) { - uint64_t offloads = conf->offloads | - dev->data->dev_conf.rxmode.offloads; - /* The offloads should be checked on rte_eth_dev layer. */ - MLX5_ASSERT(offloads & DEV_RX_OFFLOAD_SCATTER); + MLX5_ASSERT(offloads & RTE_ETH_RX_OFFLOAD_SCATTER); if (!(offloads & RTE_ETH_RX_OFFLOAD_BUFFER_SPLIT)) { DRV_LOG(ERR, "port %u queue index %u split " "offload not configured", @@ -705,19 +862,75 @@ mlx5_rx_queue_setup(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, } MLX5_ASSERT(n_seg < MLX5_MAX_RXQ_NSEG); } - res = mlx5_rx_queue_pre_setup(dev, idx, &desc); + if (conf->share_group > 0) { + if (!priv->sh->cdev->config.hca_attr.mem_rq_rmp) { + DRV_LOG(ERR, "port %u queue index %u shared Rx queue not supported by fw", + dev->data->port_id, idx); + rte_errno = EINVAL; + return -rte_errno; + } + if (priv->obj_ops.rxq_obj_new != devx_obj_ops.rxq_obj_new) { + DRV_LOG(ERR, "port %u queue index %u shared Rx queue needs DevX api", + dev->data->port_id, idx); + rte_errno = EINVAL; + return -rte_errno; + } + if (conf->share_qid >= priv->rxqs_n) { + DRV_LOG(ERR, "port %u shared Rx queue index %u > number of Rx queues %u", + dev->data->port_id, conf->share_qid, + priv->rxqs_n); + rte_errno = EINVAL; + return -rte_errno; + } + if (priv->config.mprq.enabled) { + DRV_LOG(ERR, "port %u shared Rx queue index %u: not supported when MPRQ enabled", + dev->data->port_id, conf->share_qid); + rte_errno = EINVAL; + return -rte_errno; + } + /* Try to reuse shared RXQ. */ + rxq_ctrl = mlx5_shared_rxq_get(dev, conf->share_group, + conf->share_qid); + if (rxq_ctrl != NULL && + !mlx5_shared_rxq_match(rxq_ctrl, dev, idx, desc, socket, + conf, mp)) { + rte_errno = EINVAL; + return -rte_errno; + } + } + res = mlx5_rx_queue_pre_setup(dev, idx, &desc, &rxq_ctrl); if (res) return res; - rxq_ctrl = mlx5_rxq_new(dev, idx, desc, socket, conf, rx_seg, n_seg); - if (!rxq_ctrl) { - DRV_LOG(ERR, "port %u unable to allocate queue index %u", + /* Allocate RXQ. */ + rxq = mlx5_malloc(MLX5_MEM_RTE | MLX5_MEM_ZERO, sizeof(*rxq), 0, + SOCKET_ID_ANY); + if (!rxq) { + DRV_LOG(ERR, "port %u unable to allocate rx queue index %u private data", dev->data->port_id, idx); rte_errno = ENOMEM; return -rte_errno; } + if (rxq_ctrl == NULL) { + rxq_ctrl = mlx5_rxq_new(dev, idx, desc, socket, conf, rx_seg, + n_seg); + if (rxq_ctrl == NULL) { + DRV_LOG(ERR, "port %u unable to allocate rx queue index %u", + dev->data->port_id, idx); + mlx5_free(rxq); + rte_errno = ENOMEM; + return -rte_errno; + } + } + rxq->priv = priv; + rxq->idx = idx; + (*priv->rxq_privs)[idx] = rxq; + /* Join owner list. */ + LIST_INSERT_HEAD(&rxq_ctrl->owners, rxq, owner_entry); + rxq->ctrl = rxq_ctrl; + mlx5_rxq_ref(dev, idx); DRV_LOG(DEBUG, "port %u adding Rx queue %u to list", dev->data->port_id, idx); - (*priv->rxqs)[idx] = &rxq_ctrl->rxq; + dev->data->rx_queues[idx] = &rxq_ctrl->rxq; return 0; } @@ -741,12 +954,11 @@ mlx5_rx_hairpin_queue_setup(struct rte_eth_dev *dev, uint16_t idx, const struct rte_eth_hairpin_conf *hairpin_conf) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_data *rxq = (*priv->rxqs)[idx]; - struct mlx5_rxq_ctrl *rxq_ctrl = - container_of(rxq, struct mlx5_rxq_ctrl, rxq); + struct mlx5_rxq_priv *rxq; + struct mlx5_rxq_ctrl *rxq_ctrl; int res; - res = mlx5_rx_queue_pre_setup(dev, idx, &desc); + res = mlx5_rx_queue_pre_setup(dev, idx, &desc, NULL); if (res) return res; if (hairpin_conf->peer_count != 1) { @@ -778,41 +990,51 @@ mlx5_rx_hairpin_queue_setup(struct rte_eth_dev *dev, uint16_t idx, return -rte_errno; } } - rxq_ctrl = mlx5_rxq_hairpin_new(dev, idx, desc, hairpin_conf); + rxq = mlx5_malloc(MLX5_MEM_RTE | MLX5_MEM_ZERO, sizeof(*rxq), 0, + SOCKET_ID_ANY); + if (!rxq) { + DRV_LOG(ERR, "port %u unable to allocate hairpin rx queue index %u private data", + dev->data->port_id, idx); + rte_errno = ENOMEM; + return -rte_errno; + } + rxq->priv = priv; + rxq->idx = idx; + (*priv->rxq_privs)[idx] = rxq; + rxq_ctrl = mlx5_rxq_hairpin_new(dev, rxq, desc, hairpin_conf); if (!rxq_ctrl) { - DRV_LOG(ERR, "port %u unable to allocate queue index %u", + DRV_LOG(ERR, "port %u unable to allocate hairpin queue index %u", dev->data->port_id, idx); + mlx5_free(rxq); + (*priv->rxq_privs)[idx] = NULL; rte_errno = ENOMEM; return -rte_errno; } - DRV_LOG(DEBUG, "port %u adding Rx queue %u to list", + DRV_LOG(DEBUG, "port %u adding hairpin Rx queue %u to list", dev->data->port_id, idx); - (*priv->rxqs)[idx] = &rxq_ctrl->rxq; + dev->data->rx_queues[idx] = &rxq_ctrl->rxq; return 0; } /** * DPDK callback to release a RX queue. * - * @param dpdk_rxq - * Generic RX queue pointer. + * @param dev + * Pointer to Ethernet device structure. + * @param qid + * Receive queue index. */ void -mlx5_rx_queue_release(void *dpdk_rxq) +mlx5_rx_queue_release(struct rte_eth_dev *dev, uint16_t qid) { - struct mlx5_rxq_data *rxq = (struct mlx5_rxq_data *)dpdk_rxq; - struct mlx5_rxq_ctrl *rxq_ctrl; - struct mlx5_priv *priv; + struct mlx5_rxq_data *rxq = dev->data->rx_queues[qid]; if (rxq == NULL) return; - rxq_ctrl = container_of(rxq, struct mlx5_rxq_ctrl, rxq); - priv = rxq_ctrl->priv; - if (!mlx5_rxq_releasable(ETH_DEV(priv), rxq_ctrl->rxq.idx)) + if (!mlx5_rxq_releasable(dev, qid)) rte_panic("port %u Rx queue %u is still used by a flow and" - " cannot be removed\n", - PORT_ID(priv), rxq->idx); - mlx5_rxq_release(ETH_DEV(priv), rxq_ctrl->rxq.idx); + " cannot be removed\n", dev->data->port_id, qid); + mlx5_rxq_release(dev, qid); } /** @@ -834,16 +1056,10 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev) unsigned int count = 0; struct rte_intr_handle *intr_handle = dev->intr_handle; - /* Representor shares dev->intr_handle with PF. */ - if (priv->representor) - return 0; if (!dev->data->dev_conf.intr_conf.rxq) return 0; mlx5_rx_intr_vec_disable(dev); - intr_handle->intr_vec = mlx5_malloc(0, - n * sizeof(intr_handle->intr_vec[0]), - 0, SOCKET_ID_ANY); - if (intr_handle->intr_vec == NULL) { + if (rte_intr_vec_list_alloc(intr_handle, NULL, n)) { DRV_LOG(ERR, "port %u failed to allocate memory for interrupt" " vector, Rx interrupts will not be supported", @@ -851,25 +1067,26 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev) rte_errno = ENOMEM; return -rte_errno; } - intr_handle->type = RTE_INTR_HANDLE_EXT; + + if (rte_intr_type_set(intr_handle, RTE_INTR_HANDLE_EXT)) + return -rte_errno; + for (i = 0; i != n; ++i) { /* This rxq obj must not be released in this function. */ - struct mlx5_rxq_ctrl *rxq_ctrl = mlx5_rxq_get(dev, i); - struct mlx5_rxq_obj *rxq_obj = rxq_ctrl ? rxq_ctrl->obj : NULL; + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, i); + struct mlx5_rxq_obj *rxq_obj = rxq ? rxq->ctrl->obj : NULL; int rc; /* Skip queues that cannot request interrupts. */ if (!rxq_obj || (!rxq_obj->ibv_channel && !rxq_obj->devx_channel)) { /* Use invalid intr_vec[] index to disable entry. */ - intr_handle->intr_vec[i] = - RTE_INTR_VEC_RXTX_OFFSET + - RTE_MAX_RXTX_INTR_VEC_ID; - /* Decrease the rxq_ctrl's refcnt */ - if (rxq_ctrl) - mlx5_rxq_release(dev, i); + if (rte_intr_vec_list_index_set(intr_handle, i, + RTE_INTR_VEC_RXTX_OFFSET + RTE_MAX_RXTX_INTR_VEC_ID)) + return -rte_errno; continue; } + mlx5_rxq_ref(dev, i); if (count >= RTE_MAX_RXTX_INTR_VEC_ID) { DRV_LOG(ERR, "port %u too many Rx queues for interrupt" @@ -891,14 +1108,19 @@ mlx5_rx_intr_vec_enable(struct rte_eth_dev *dev) mlx5_rx_intr_vec_disable(dev); return -rte_errno; } - intr_handle->intr_vec[i] = RTE_INTR_VEC_RXTX_OFFSET + count; - intr_handle->efds[count] = rxq_obj->fd; + + if (rte_intr_vec_list_index_set(intr_handle, i, + RTE_INTR_VEC_RXTX_OFFSET + count)) + return -rte_errno; + if (rte_intr_efds_index_set(intr_handle, count, + rxq_obj->fd)) + return -rte_errno; count++; } if (!count) mlx5_rx_intr_vec_disable(dev); - else - intr_handle->nb_efd = count; + else if (rte_intr_nb_efd_set(intr_handle, count)) + return -rte_errno; return 0; } @@ -917,29 +1139,26 @@ mlx5_rx_intr_vec_disable(struct rte_eth_dev *dev) unsigned int rxqs_n = priv->rxqs_n; unsigned int n = RTE_MIN(rxqs_n, (uint32_t)RTE_MAX_RXTX_INTR_VEC_ID); - /* Representor shares dev->intr_handle with PF. */ - if (priv->representor) - return; if (!dev->data->dev_conf.intr_conf.rxq) return; - if (!intr_handle->intr_vec) + if (rte_intr_vec_list_index_get(intr_handle, 0) < 0) goto free; for (i = 0; i != n; ++i) { - if (intr_handle->intr_vec[i] == RTE_INTR_VEC_RXTX_OFFSET + - RTE_MAX_RXTX_INTR_VEC_ID) + if (rte_intr_vec_list_index_get(intr_handle, i) == + RTE_INTR_VEC_RXTX_OFFSET + RTE_MAX_RXTX_INTR_VEC_ID) continue; /** * Need to access directly the queue to release the reference * kept in mlx5_rx_intr_vec_enable(). */ - mlx5_rxq_release(dev, i); + mlx5_rxq_deref(dev, i); } free: rte_intr_free_epoll_fd(intr_handle); - if (intr_handle->intr_vec) - mlx5_free(intr_handle->intr_vec); - intr_handle->nb_efd = 0; - intr_handle->intr_vec = NULL; + + rte_intr_vec_list_free(intr_handle); + + rte_intr_nb_efd_set(intr_handle, 0); } /** @@ -956,15 +1175,13 @@ mlx5_arm_cq(struct mlx5_rxq_data *rxq, int sq_n_rxq) int sq_n = 0; uint32_t doorbell_hi; uint64_t doorbell; - void *cq_db_reg = (char *)rxq->cq_uar + MLX5_CQ_DOORBELL; sq_n = sq_n_rxq & MLX5_CQ_SQN_MASK; doorbell_hi = sq_n << MLX5_CQ_SQN_OFFSET | (rxq->cq_ci & MLX5_CI_MASK); doorbell = (uint64_t)doorbell_hi << 32; doorbell |= rxq->cqn; - rxq->cq_db[MLX5_CQ_ARM_DB] = rte_cpu_to_be_32(doorbell_hi); - mlx5_uar_write64(rte_cpu_to_be_64(doorbell), - cq_db_reg, rxq->uar_lock_cq); + mlx5_doorbell_ring(&rxq->uar_data, rte_cpu_to_be_64(doorbell), + doorbell_hi, &rxq->cq_db[MLX5_CQ_ARM_DB], 0); } /** @@ -981,19 +1198,14 @@ mlx5_arm_cq(struct mlx5_rxq_data *rxq, int sq_n_rxq) int mlx5_rx_intr_enable(struct rte_eth_dev *dev, uint16_t rx_queue_id) { - struct mlx5_rxq_ctrl *rxq_ctrl; - - rxq_ctrl = mlx5_rxq_get(dev, rx_queue_id); - if (!rxq_ctrl) + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, rx_queue_id); + if (!rxq) goto error; - if (rxq_ctrl->irq) { - if (!rxq_ctrl->obj) { - mlx5_rxq_release(dev, rx_queue_id); + if (rxq->ctrl->irq) { + if (!rxq->ctrl->obj) goto error; - } - mlx5_arm_cq(&rxq_ctrl->rxq, rxq_ctrl->rxq.cq_arm_sn); + mlx5_arm_cq(&rxq->ctrl->rxq, rxq->ctrl->rxq.cq_arm_sn); } - mlx5_rxq_release(dev, rx_queue_id); return 0; error: rte_errno = EINVAL; @@ -1015,23 +1227,21 @@ int mlx5_rx_intr_disable(struct rte_eth_dev *dev, uint16_t rx_queue_id) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_ctrl *rxq_ctrl; + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, rx_queue_id); int ret = 0; - rxq_ctrl = mlx5_rxq_get(dev, rx_queue_id); - if (!rxq_ctrl) { + if (!rxq) { rte_errno = EINVAL; return -rte_errno; } - if (!rxq_ctrl->obj) + if (!rxq->ctrl->obj) goto error; - if (rxq_ctrl->irq) { - ret = priv->obj_ops.rxq_event_get(rxq_ctrl->obj); + if (rxq->ctrl->irq) { + ret = priv->obj_ops.rxq_event_get(rxq->ctrl->obj); if (ret < 0) goto error; - rxq_ctrl->rxq.cq_arm_sn++; + rxq->ctrl->rxq.cq_arm_sn++; } - mlx5_rxq_release(dev, rx_queue_id); return 0; error: /** @@ -1042,12 +1252,9 @@ error: rte_errno = errno; else rte_errno = EINVAL; - ret = rte_errno; /* Save rte_errno before cleanup. */ - mlx5_rxq_release(dev, rx_queue_id); - if (ret != EAGAIN) + if (rte_errno != EAGAIN) DRV_LOG(WARNING, "port %u unable to disable interrupt on Rx queue %d", dev->data->port_id, rx_queue_id); - rte_errno = ret; /* Restore rte_errno. */ return -rte_errno; } @@ -1068,6 +1275,11 @@ mlx5_rxq_obj_verify(struct rte_eth_dev *dev) struct mlx5_rxq_obj *rxq_obj; LIST_FOREACH(rxq_obj, &priv->rxqsobj, next) { + if (rxq_obj->rxq_ctrl == NULL) + continue; + if (rxq_obj->rxq_ctrl->rxq.shared && + !LIST_EMPTY(&rxq_obj->rxq_ctrl->owners)) + continue; DRV_LOG(DEBUG, "port %u Rx queue %u still referenced", dev->data->port_id, rxq_obj->rxq_ctrl->rxq.idx); ++ret; @@ -1136,7 +1348,7 @@ mlx5_mprq_free_mp(struct rte_eth_dev *dev) rte_mempool_free(mp); /* Unset mempool for each Rx queue. */ for (i = 0; i != priv->rxqs_n; ++i) { - struct mlx5_rxq_data *rxq = (*priv->rxqs)[i]; + struct mlx5_rxq_data *rxq = mlx5_rxq_data_get(dev, i); if (rxq == NULL) continue; @@ -1167,34 +1379,37 @@ mlx5_mprq_alloc_mp(struct rte_eth_dev *dev) unsigned int buf_len; unsigned int obj_num; unsigned int obj_size; - unsigned int strd_num_n = 0; - unsigned int strd_sz_n = 0; + unsigned int log_strd_num = 0; + unsigned int log_strd_sz = 0; unsigned int i; unsigned int n_ibv = 0; + int ret; if (!mlx5_mprq_enabled(dev)) return 0; /* Count the total number of descriptors configured. */ for (i = 0; i != priv->rxqs_n; ++i) { - struct mlx5_rxq_data *rxq = (*priv->rxqs)[i]; - struct mlx5_rxq_ctrl *rxq_ctrl = container_of - (rxq, struct mlx5_rxq_ctrl, rxq); + struct mlx5_rxq_ctrl *rxq_ctrl = mlx5_rxq_ctrl_get(dev, i); + struct mlx5_rxq_data *rxq; - if (rxq == NULL || rxq_ctrl->type != MLX5_RXQ_TYPE_STANDARD) + if (rxq_ctrl == NULL || rxq_ctrl->is_hairpin) continue; + rxq = &rxq_ctrl->rxq; n_ibv++; desc += 1 << rxq->elts_n; /* Get the max number of strides. */ - if (strd_num_n < rxq->strd_num_n) - strd_num_n = rxq->strd_num_n; + if (log_strd_num < rxq->log_strd_num) + log_strd_num = rxq->log_strd_num; /* Get the max size of a stride. */ - if (strd_sz_n < rxq->strd_sz_n) - strd_sz_n = rxq->strd_sz_n; - } - MLX5_ASSERT(strd_num_n && strd_sz_n); - buf_len = (1 << strd_num_n) * (1 << strd_sz_n); - obj_size = sizeof(struct mlx5_mprq_buf) + buf_len + (1 << strd_num_n) * - sizeof(struct rte_mbuf_ext_shared_info) + RTE_PKTMBUF_HEADROOM; + if (log_strd_sz < rxq->log_strd_sz) + log_strd_sz = rxq->log_strd_sz; + } + MLX5_ASSERT(log_strd_num && log_strd_sz); + buf_len = RTE_BIT32(log_strd_num) * RTE_BIT32(log_strd_sz); + obj_size = sizeof(struct mlx5_mprq_buf) + buf_len + + RTE_BIT32(log_strd_num) * + sizeof(struct rte_mbuf_ext_shared_info) + + RTE_PKTMBUF_HEADROOM; /* * Received packets can be either memcpy'd or externally referenced. In * case that the packet is attached to an mbuf as an external buffer, as @@ -1240,7 +1455,7 @@ mlx5_mprq_alloc_mp(struct rte_eth_dev *dev) snprintf(name, sizeof(name), "port-%u-mprq", dev->data->port_id); mp = rte_mempool_create(name, obj_num, obj_size, MLX5_MPRQ_MP_CACHE_SZ, 0, NULL, NULL, mlx5_mprq_buf_init, - (void *)((uintptr_t)1 << strd_num_n), + (void *)((uintptr_t)1 << log_strd_num), dev->device->numa_node, 0); if (mp == NULL) { DRV_LOG(ERR, @@ -1250,17 +1465,24 @@ mlx5_mprq_alloc_mp(struct rte_eth_dev *dev) rte_errno = ENOMEM; return -rte_errno; } + ret = mlx5_mr_mempool_register(priv->sh->cdev, mp, false); + if (ret < 0 && rte_errno != EEXIST) { + ret = rte_errno; + DRV_LOG(ERR, "port %u failed to register a mempool for Multi-Packet RQ", + dev->data->port_id); + rte_mempool_free(mp); + rte_errno = ret; + return -rte_errno; + } priv->mprq_mp = mp; exit: /* Set mempool for each Rx queue. */ for (i = 0; i != priv->rxqs_n; ++i) { - struct mlx5_rxq_data *rxq = (*priv->rxqs)[i]; - struct mlx5_rxq_ctrl *rxq_ctrl = container_of - (rxq, struct mlx5_rxq_ctrl, rxq); + struct mlx5_rxq_ctrl *rxq_ctrl = mlx5_rxq_ctrl_get(dev, i); - if (rxq == NULL || rxq_ctrl->type != MLX5_RXQ_TYPE_STANDARD) + if (rxq_ctrl == NULL || rxq_ctrl->is_hairpin) continue; - rxq->mprq_mp = mp; + rxq_ctrl->rxq.mprq_mp = mp; } DRV_LOG(INFO, "port %u Multi-Packet RQ is configured", dev->data->port_id); @@ -1291,7 +1513,7 @@ mlx5_max_lro_msg_size_adjust(struct rte_eth_dev *dev, uint16_t idx, { struct mlx5_priv *priv = dev->data->dev_private; - if (priv->config.hca_attr.lro_max_msg_sz_mode == + if (priv->sh->cdev->config.hca_attr.lro_max_msg_sz_mode == MLX5_LRO_MAX_MSG_SIZE_START_FROM_L4 && max_lro_size > MLX5_MAX_TCP_HDR_OFFSET) max_lro_size -= MLX5_MAX_TCP_HDR_OFFSET; @@ -1309,6 +1531,128 @@ mlx5_max_lro_msg_size_adjust(struct rte_eth_dev *dev, uint16_t idx, priv->max_lro_msg_size * MLX5_LRO_SEG_CHUNK_SIZE); } +/** + * Prepare both size and number of stride for Multi-Packet RQ. + * + * @param dev + * Pointer to Ethernet device. + * @param idx + * RX queue index. + * @param desc + * Number of descriptors to configure in queue. + * @param rx_seg_en + * Indicator if Rx segment enables, if so Multi-Packet RQ doesn't enable. + * @param min_mbuf_size + * Non scatter min mbuf size, max_rx_pktlen plus overhead. + * @param actual_log_stride_num + * Log number of strides to configure for this queue. + * @param actual_log_stride_size + * Log stride size to configure for this queue. + * + * @return + * 0 if Multi-Packet RQ is supported, otherwise -1. + */ +static int +mlx5_mprq_prepare(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, + bool rx_seg_en, uint32_t min_mbuf_size, + uint32_t *actual_log_stride_num, + uint32_t *actual_log_stride_size) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_port_config *config = &priv->config; + struct mlx5_dev_cap *dev_cap = &priv->sh->dev_cap; + uint32_t log_min_stride_num = dev_cap->mprq.log_min_stride_num; + uint32_t log_max_stride_num = dev_cap->mprq.log_max_stride_num; + uint32_t log_def_stride_num = + RTE_MIN(RTE_MAX(MLX5_MPRQ_DEFAULT_LOG_STRIDE_NUM, + log_min_stride_num), + log_max_stride_num); + uint32_t log_min_stride_size = dev_cap->mprq.log_min_stride_size; + uint32_t log_max_stride_size = dev_cap->mprq.log_max_stride_size; + uint32_t log_def_stride_size = + RTE_MIN(RTE_MAX(MLX5_MPRQ_DEFAULT_LOG_STRIDE_SIZE, + log_min_stride_size), + log_max_stride_size); + uint32_t log_stride_wqe_size; + + if (mlx5_check_mprq_support(dev) != 1 || rx_seg_en) + goto unsupport; + /* Checks if chosen number of strides is in supported range. */ + if (config->mprq.log_stride_num > log_max_stride_num || + config->mprq.log_stride_num < log_min_stride_num) { + *actual_log_stride_num = log_def_stride_num; + DRV_LOG(WARNING, + "Port %u Rx queue %u number of strides for Multi-Packet RQ is out of range, setting default value (%u)", + dev->data->port_id, idx, RTE_BIT32(log_def_stride_num)); + } else { + *actual_log_stride_num = config->mprq.log_stride_num; + } + if (config->mprq.log_stride_size) { + /* Checks if chosen size of stride is in supported range. */ + if (config->mprq.log_stride_size > log_max_stride_size || + config->mprq.log_stride_size < log_min_stride_size) { + *actual_log_stride_size = log_def_stride_size; + DRV_LOG(WARNING, + "Port %u Rx queue %u size of a stride for Multi-Packet RQ is out of range, setting default value (%u)", + dev->data->port_id, idx, + RTE_BIT32(log_def_stride_size)); + } else { + *actual_log_stride_size = config->mprq.log_stride_size; + } + } else { + if (min_mbuf_size <= RTE_BIT32(log_max_stride_size)) + *actual_log_stride_size = log2above(min_mbuf_size); + else + goto unsupport; + } + log_stride_wqe_size = *actual_log_stride_num + *actual_log_stride_size; + /* Check if WQE buffer size is supported by hardware. */ + if (log_stride_wqe_size < dev_cap->mprq.log_min_stride_wqe_size) { + *actual_log_stride_num = log_def_stride_num; + *actual_log_stride_size = log_def_stride_size; + DRV_LOG(WARNING, + "Port %u Rx queue %u size of WQE buffer for Multi-Packet RQ is too small, setting default values (stride_num_n=%u, stride_size_n=%u)", + dev->data->port_id, idx, RTE_BIT32(log_def_stride_num), + RTE_BIT32(log_def_stride_size)); + log_stride_wqe_size = log_def_stride_num + log_def_stride_size; + } + MLX5_ASSERT(log_stride_wqe_size >= + dev_cap->mprq.log_min_stride_wqe_size); + if (desc <= RTE_BIT32(*actual_log_stride_num)) + goto unsupport; + if (min_mbuf_size > RTE_BIT32(log_stride_wqe_size)) { + DRV_LOG(WARNING, "Port %u Rx queue %u " + "Multi-Packet RQ is unsupported, WQE buffer size (%u) " + "is smaller than min mbuf size (%u)", + dev->data->port_id, idx, RTE_BIT32(log_stride_wqe_size), + min_mbuf_size); + goto unsupport; + } + DRV_LOG(DEBUG, "Port %u Rx queue %u " + "Multi-Packet RQ is enabled strd_num_n = %u, strd_sz_n = %u", + dev->data->port_id, idx, RTE_BIT32(*actual_log_stride_num), + RTE_BIT32(*actual_log_stride_size)); + return 0; +unsupport: + if (config->mprq.enabled) + DRV_LOG(WARNING, + "Port %u MPRQ is requested but cannot be enabled\n" + " (requested: pkt_sz = %u, desc_num = %u," + " rxq_num = %u, stride_sz = %u, stride_num = %u\n" + " supported: min_rxqs_num = %u, min_buf_wqe_sz = %u" + " min_stride_sz = %u, max_stride_sz = %u).\n" + "Rx segment is %senable.", + dev->data->port_id, min_mbuf_size, desc, priv->rxqs_n, + RTE_BIT32(config->mprq.log_stride_size), + RTE_BIT32(config->mprq.log_stride_num), + config->mprq.min_rxqs_num, + RTE_BIT32(dev_cap->mprq.log_min_stride_wqe_size), + RTE_BIT32(dev_cap->mprq.log_min_stride_size), + RTE_BIT32(dev_cap->mprq.log_max_stride_size), + rx_seg_en ? "" : "not "); + return -1; +} + /** * Create a DPDK Rx queue. * @@ -1332,53 +1676,60 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_rxq_ctrl *tmpl; unsigned int mb_len = rte_pktmbuf_data_room_size(rx_seg[0].mp); - struct mlx5_dev_config *config = &priv->config; + struct mlx5_port_config *config = &priv->config; uint64_t offloads = conf->offloads | dev->data->dev_conf.rxmode.offloads; - unsigned int lro_on_queue = !!(offloads & DEV_RX_OFFLOAD_TCP_LRO); - unsigned int max_rx_pkt_len = lro_on_queue ? + unsigned int lro_on_queue = !!(offloads & RTE_ETH_RX_OFFLOAD_TCP_LRO); + unsigned int max_rx_pktlen = lro_on_queue ? dev->data->dev_conf.rxmode.max_lro_pkt_size : - dev->data->dev_conf.rxmode.max_rx_pkt_len; - unsigned int non_scatter_min_mbuf_size = max_rx_pkt_len + + dev->data->mtu + (unsigned int)RTE_ETHER_HDR_LEN + + RTE_ETHER_CRC_LEN; + unsigned int non_scatter_min_mbuf_size = max_rx_pktlen + RTE_PKTMBUF_HEADROOM; unsigned int max_lro_size = 0; unsigned int first_mb_free_size = mb_len - RTE_PKTMBUF_HEADROOM; - const int mprq_en = mlx5_check_mprq_support(dev) > 0 && n_seg == 1 && - !rx_seg[0].offset && !rx_seg[0].length; - unsigned int mprq_stride_nums = config->mprq.stride_num_n ? - config->mprq.stride_num_n : MLX5_MPRQ_STRIDE_NUM_N; - unsigned int mprq_stride_size = non_scatter_min_mbuf_size <= - (1U << config->mprq.max_stride_size_n) ? - log2above(non_scatter_min_mbuf_size) : MLX5_MPRQ_STRIDE_SIZE_N; - unsigned int mprq_stride_cap = (config->mprq.stride_num_n ? - (1U << config->mprq.stride_num_n) : (1U << mprq_stride_nums)) * - (config->mprq.stride_size_n ? - (1U << config->mprq.stride_size_n) : (1U << mprq_stride_size)); + uint32_t mprq_log_actual_stride_num = 0; + uint32_t mprq_log_actual_stride_size = 0; + bool rx_seg_en = n_seg != 1 || rx_seg[0].offset || rx_seg[0].length; + const int mprq_en = !mlx5_mprq_prepare(dev, idx, desc, rx_seg_en, + non_scatter_min_mbuf_size, + &mprq_log_actual_stride_num, + &mprq_log_actual_stride_size); /* * Always allocate extra slots, even if eventually * the vector Rx will not be used. */ uint16_t desc_n = desc + config->rx_vec_en * MLX5_VPMD_DESCS_PER_LOOP; + size_t alloc_size = sizeof(*tmpl) + desc_n * sizeof(struct rte_mbuf *); const struct rte_eth_rxseg_split *qs_seg = rx_seg; unsigned int tail_len; - tmpl = mlx5_malloc(MLX5_MEM_RTE | MLX5_MEM_ZERO, - sizeof(*tmpl) + desc_n * sizeof(struct rte_mbuf *) + - (!!mprq_en) * - (desc >> mprq_stride_nums) * sizeof(struct mlx5_mprq_buf *), - 0, socket); + if (mprq_en) { + /* Trim the number of descs needed. */ + desc >>= mprq_log_actual_stride_num; + alloc_size += desc * sizeof(struct mlx5_mprq_buf *); + } + tmpl = mlx5_malloc(MLX5_MEM_RTE | MLX5_MEM_ZERO, alloc_size, 0, socket); if (!tmpl) { rte_errno = ENOMEM; return NULL; } + LIST_INIT(&tmpl->owners); MLX5_ASSERT(n_seg && n_seg <= MLX5_MAX_RXQ_NSEG); + /* + * Save the original segment configuration in the shared queue + * descriptor for the later check on the sibling queue creation. + */ + tmpl->rxseg_n = n_seg; + rte_memcpy(tmpl->rxseg, qs_seg, + sizeof(struct rte_eth_rxseg_split) * n_seg); /* * Build the array of actual buffer offsets and lengths. * Pad with the buffers from the last memory pool if * needed to handle max size packets, replace zero length * with the buffer length from the pool. */ - tail_len = max_rx_pkt_len; + tail_len = max_rx_pktlen; do { struct mlx5_eth_rxseg *hw_seg = &tmpl->rxq.rxseg[tmpl->rxq.rxseg_n]; @@ -1416,7 +1767,7 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, "port %u too many SGEs (%u) needed to handle" " requested maximum packet size %u, the maximum" " supported are %u", dev->data->port_id, - tmpl->rxq.rxseg_n, max_rx_pkt_len, + tmpl->rxq.rxseg_n, max_rx_pktlen, MLX5_MAX_RXQ_NSEG); rte_errno = ENOTSUP; goto error; @@ -1437,66 +1788,42 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, } while (tail_len || !rte_is_power_of_2(tmpl->rxq.rxseg_n)); MLX5_ASSERT(tmpl->rxq.rxseg_n && tmpl->rxq.rxseg_n <= MLX5_MAX_RXQ_NSEG); - if (tmpl->rxq.rxseg_n > 1 && !(offloads & DEV_RX_OFFLOAD_SCATTER)) { + if (tmpl->rxq.rxseg_n > 1 && !(offloads & RTE_ETH_RX_OFFLOAD_SCATTER)) { DRV_LOG(ERR, "port %u Rx queue %u: Scatter offload is not" " configured and no enough mbuf space(%u) to contain " "the maximum RX packet length(%u) with head-room(%u)", - dev->data->port_id, idx, mb_len, max_rx_pkt_len, + dev->data->port_id, idx, mb_len, max_rx_pktlen, RTE_PKTMBUF_HEADROOM); rte_errno = ENOSPC; goto error; } - tmpl->type = MLX5_RXQ_TYPE_STANDARD; - if (mlx5_mr_btree_init(&tmpl->rxq.mr_ctrl.cache_bh, - MLX5_MR_BTREE_CACHE_N, socket)) { + tmpl->is_hairpin = false; + if (mlx5_mr_ctrl_init(&tmpl->rxq.mr_ctrl, + &priv->sh->cdev->mr_scache.dev_gen, socket)) { /* rte_errno is already set. */ goto error; } tmpl->socket = socket; if (dev->data->dev_conf.intr_conf.rxq) tmpl->irq = 1; - /* - * This Rx queue can be configured as a Multi-Packet RQ if all of the - * following conditions are met: - * - MPRQ is enabled. - * - The number of descs is more than the number of strides. - * - max_rx_pkt_len plus overhead is less than the max size - * of a stride or mprq_stride_size is specified by a user. - * Need to make sure that there are enough strides to encap - * the maximum packet size in case mprq_stride_size is set. - * Otherwise, enable Rx scatter if necessary. - */ - if (mprq_en && desc > (1U << mprq_stride_nums) && - (non_scatter_min_mbuf_size <= - (1U << config->mprq.max_stride_size_n) || - (config->mprq.stride_size_n && - non_scatter_min_mbuf_size <= mprq_stride_cap))) { + if (mprq_en) { /* TODO: Rx scatter isn't supported yet. */ tmpl->rxq.sges_n = 0; - /* Trim the number of descs needed. */ - desc >>= mprq_stride_nums; - tmpl->rxq.strd_num_n = config->mprq.stride_num_n ? - config->mprq.stride_num_n : mprq_stride_nums; - tmpl->rxq.strd_sz_n = config->mprq.stride_size_n ? - config->mprq.stride_size_n : mprq_stride_size; + tmpl->rxq.log_strd_num = mprq_log_actual_stride_num; + tmpl->rxq.log_strd_sz = mprq_log_actual_stride_size; tmpl->rxq.strd_shift_en = MLX5_MPRQ_TWO_BYTE_SHIFT; tmpl->rxq.strd_scatter_en = - !!(offloads & DEV_RX_OFFLOAD_SCATTER); + !!(offloads & RTE_ETH_RX_OFFLOAD_SCATTER); tmpl->rxq.mprq_max_memcpy_len = RTE_MIN(first_mb_free_size, config->mprq.max_memcpy_len); - max_lro_size = RTE_MIN(max_rx_pkt_len, - (1u << tmpl->rxq.strd_num_n) * - (1u << tmpl->rxq.strd_sz_n)); - DRV_LOG(DEBUG, - "port %u Rx queue %u: Multi-Packet RQ is enabled" - " strd_num_n = %u, strd_sz_n = %u", - dev->data->port_id, idx, - tmpl->rxq.strd_num_n, tmpl->rxq.strd_sz_n); + max_lro_size = RTE_MIN(max_rx_pktlen, + RTE_BIT32(tmpl->rxq.log_strd_num) * + RTE_BIT32(tmpl->rxq.log_strd_sz)); } else if (tmpl->rxq.rxseg_n == 1) { - MLX5_ASSERT(max_rx_pkt_len <= first_mb_free_size); + MLX5_ASSERT(max_rx_pktlen <= first_mb_free_size); tmpl->rxq.sges_n = 0; - max_lro_size = max_rx_pkt_len; - } else if (offloads & DEV_RX_OFFLOAD_SCATTER) { + max_lro_size = max_rx_pktlen; + } else if (offloads & RTE_ETH_RX_OFFLOAD_SCATTER) { unsigned int sges_n; if (lro_on_queue && first_mb_free_size < @@ -1517,32 +1844,14 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, "port %u too many SGEs (%u) needed to handle" " requested maximum packet size %u, the maximum" " supported are %u", dev->data->port_id, - 1 << sges_n, max_rx_pkt_len, + 1 << sges_n, max_rx_pktlen, 1u << MLX5_MAX_LOG_RQ_SEGS); rte_errno = ENOTSUP; goto error; } tmpl->rxq.sges_n = sges_n; - max_lro_size = max_rx_pkt_len; + max_lro_size = max_rx_pktlen; } - if (config->mprq.enabled && !mlx5_rxq_mprq_enabled(&tmpl->rxq)) - DRV_LOG(WARNING, - "port %u MPRQ is requested but cannot be enabled\n" - " (requested: pkt_sz = %u, desc_num = %u," - " rxq_num = %u, stride_sz = %u, stride_num = %u\n" - " supported: min_rxqs_num = %u," - " min_stride_sz = %u, max_stride_sz = %u).", - dev->data->port_id, non_scatter_min_mbuf_size, - desc, priv->rxqs_n, - config->mprq.stride_size_n ? - (1U << config->mprq.stride_size_n) : - (1U << mprq_stride_size), - config->mprq.stride_num_n ? - (1U << config->mprq.stride_num_n) : - (1U << mprq_stride_nums), - config->mprq.min_rxqs_num, - (1U << config->mprq.min_stride_size_n), - (1U << config->mprq.max_stride_size_n)); DRV_LOG(DEBUG, "port %u maximum number of segments per packet: %u", dev->data->port_id, 1 << tmpl->rxq.sges_n); if (desc % (1 << tmpl->rxq.sges_n)) { @@ -1557,9 +1866,9 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, } mlx5_max_lro_msg_size_adjust(dev, idx, max_lro_size); /* Toggle RX checksum offload if hardware supports it. */ - tmpl->rxq.csum = !!(offloads & DEV_RX_OFFLOAD_CHECKSUM); + tmpl->rxq.csum = !!(offloads & RTE_ETH_RX_OFFLOAD_CHECKSUM); /* Configure Rx timestamp. */ - tmpl->rxq.hw_timestamp = !!(offloads & DEV_RX_OFFLOAD_TIMESTAMP); + tmpl->rxq.hw_timestamp = !!(offloads & RTE_ETH_RX_OFFLOAD_TIMESTAMP); tmpl->rxq.timestamp_rx_flag = 0; if (tmpl->rxq.hw_timestamp && rte_mbuf_dyn_rx_timestamp_register( &tmpl->rxq.timestamp_offset, @@ -1568,12 +1877,12 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, goto error; } /* Configure VLAN stripping. */ - tmpl->rxq.vlan_strip = !!(offloads & DEV_RX_OFFLOAD_VLAN_STRIP); + tmpl->rxq.vlan_strip = !!(offloads & RTE_ETH_RX_OFFLOAD_VLAN_STRIP); /* By default, FCS (CRC) is stripped by hardware. */ tmpl->rxq.crc_present = 0; tmpl->rxq.lro = lro_on_queue; - if (offloads & DEV_RX_OFFLOAD_KEEP_CRC) { - if (config->hw_fcs_strip) { + if (offloads & RTE_ETH_RX_OFFLOAD_KEEP_CRC) { + if (priv->sh->config.hw_fcs_strip) { /* * RQs used for LRO-enabled TIRs should not be * configured to scatter the FCS. @@ -1600,24 +1909,24 @@ mlx5_rxq_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, dev->data->port_id, tmpl->rxq.crc_present ? "disabled" : "enabled", tmpl->rxq.crc_present << 2); - /* Save port ID. */ tmpl->rxq.rss_hash = !!priv->rss_conf.rss_hf && - (!!(dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS)); + (!!(dev->data->dev_conf.rxmode.mq_mode & RTE_ETH_MQ_RX_RSS)); + /* Save port ID. */ tmpl->rxq.port_id = dev->data->port_id; - tmpl->priv = priv; + tmpl->sh = priv->sh; tmpl->rxq.mp = rx_seg[0].mp; tmpl->rxq.elts_n = log2above(desc); - tmpl->rxq.rq_repl_thresh = - MLX5_VPMD_RXQ_RPLNSH_THRESH(desc_n); - tmpl->rxq.elts = - (struct rte_mbuf *(*)[desc_n])(tmpl + 1); + tmpl->rxq.rq_repl_thresh = MLX5_VPMD_RXQ_RPLNSH_THRESH(desc_n); + tmpl->rxq.elts = (struct rte_mbuf *(*)[desc_n])(tmpl + 1); tmpl->rxq.mprq_bufs = (struct mlx5_mprq_buf *(*)[desc])(*tmpl->rxq.elts + desc_n); -#ifndef RTE_ARCH_64 - tmpl->rxq.uar_lock_cq = &priv->sh->uar_lock_cq; -#endif tmpl->rxq.idx = idx; - __atomic_fetch_add(&tmpl->refcnt, 1, __ATOMIC_RELAXED); + if (conf->share_group > 0) { + tmpl->rxq.shared = 1; + tmpl->share_group = conf->share_group; + tmpl->share_qid = conf->share_qid; + LIST_INSERT_HEAD(&priv->sh->shared_rxqs, tmpl, share_entry); + } LIST_INSERT_HEAD(&priv->rxqsctrl, tmpl, next); return tmpl; error: @@ -1631,8 +1940,8 @@ error: * * @param dev * Pointer to Ethernet device. - * @param idx - * RX queue index. + * @param rxq + * RX queue. * @param desc * Number of descriptors to configure in queue. * @param hairpin_conf @@ -1642,9 +1951,11 @@ error: * A DPDK queue object on success, NULL otherwise and rte_errno is set. */ struct mlx5_rxq_ctrl * -mlx5_rxq_hairpin_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, +mlx5_rxq_hairpin_new(struct rte_eth_dev *dev, struct mlx5_rxq_priv *rxq, + uint16_t desc, const struct rte_eth_hairpin_conf *hairpin_conf) { + uint16_t idx = rxq->idx; struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_rxq_ctrl *tmpl; @@ -1654,22 +1965,67 @@ mlx5_rxq_hairpin_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, rte_errno = ENOMEM; return NULL; } - tmpl->type = MLX5_RXQ_TYPE_HAIRPIN; + LIST_INIT(&tmpl->owners); + rxq->ctrl = tmpl; + LIST_INSERT_HEAD(&tmpl->owners, rxq, owner_entry); + tmpl->is_hairpin = true; tmpl->socket = SOCKET_ID_ANY; tmpl->rxq.rss_hash = 0; tmpl->rxq.port_id = dev->data->port_id; - tmpl->priv = priv; + tmpl->sh = priv->sh; tmpl->rxq.mp = NULL; tmpl->rxq.elts_n = log2above(desc); tmpl->rxq.elts = NULL; tmpl->rxq.mr_ctrl.cache_bh = (struct mlx5_mr_btree) { 0 }; - tmpl->hairpin_conf = *hairpin_conf; tmpl->rxq.idx = idx; - __atomic_fetch_add(&tmpl->refcnt, 1, __ATOMIC_RELAXED); + rxq->hairpin_conf = *hairpin_conf; + mlx5_rxq_ref(dev, idx); LIST_INSERT_HEAD(&priv->rxqsctrl, tmpl, next); return tmpl; } +/** + * Increase Rx queue reference count. + * + * @param dev + * Pointer to Ethernet device. + * @param idx + * RX queue index. + * + * @return + * A pointer to the queue if it exists, NULL otherwise. + */ +struct mlx5_rxq_priv * +mlx5_rxq_ref(struct rte_eth_dev *dev, uint16_t idx) +{ + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); + + if (rxq != NULL) + __atomic_fetch_add(&rxq->refcnt, 1, __ATOMIC_RELAXED); + return rxq; +} + +/** + * Dereference a Rx queue. + * + * @param dev + * Pointer to Ethernet device. + * @param idx + * RX queue index. + * + * @return + * Updated reference count. + */ +uint32_t +mlx5_rxq_deref(struct rte_eth_dev *dev, uint16_t idx) +{ + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); + + if (rxq == NULL) + return 0; + return __atomic_sub_fetch(&rxq->refcnt, 1, __ATOMIC_RELAXED); +} + /** * Get a Rx queue. * @@ -1681,18 +2037,51 @@ mlx5_rxq_hairpin_new(struct rte_eth_dev *dev, uint16_t idx, uint16_t desc, * @return * A pointer to the queue if it exists, NULL otherwise. */ -struct mlx5_rxq_ctrl * +struct mlx5_rxq_priv * mlx5_rxq_get(struct rte_eth_dev *dev, uint16_t idx) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_data *rxq_data = (*priv->rxqs)[idx]; - struct mlx5_rxq_ctrl *rxq_ctrl = NULL; - if (rxq_data) { - rxq_ctrl = container_of(rxq_data, struct mlx5_rxq_ctrl, rxq); - __atomic_fetch_add(&rxq_ctrl->refcnt, 1, __ATOMIC_RELAXED); - } - return rxq_ctrl; + MLX5_ASSERT(priv->rxq_privs != NULL); + return (*priv->rxq_privs)[idx]; +} + +/** + * Get Rx queue shareable control. + * + * @param dev + * Pointer to Ethernet device. + * @param idx + * RX queue index. + * + * @return + * A pointer to the queue control if it exists, NULL otherwise. + */ +struct mlx5_rxq_ctrl * +mlx5_rxq_ctrl_get(struct rte_eth_dev *dev, uint16_t idx) +{ + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); + + return rxq == NULL ? NULL : rxq->ctrl; +} + +/** + * Get Rx queue shareable data. + * + * @param dev + * Pointer to Ethernet device. + * @param idx + * RX queue index. + * + * @return + * A pointer to the queue data if it exists, NULL otherwise. + */ +struct mlx5_rxq_data * +mlx5_rxq_data_get(struct rte_eth_dev *dev, uint16_t idx) +{ + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); + + return rxq == NULL ? NULL : &rxq->ctrl->rxq; } /** @@ -1710,29 +2099,46 @@ int mlx5_rxq_release(struct rte_eth_dev *dev, uint16_t idx) { struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_rxq_priv *rxq; struct mlx5_rxq_ctrl *rxq_ctrl; + uint32_t refcnt; - if (!(*priv->rxqs)[idx]) + if (priv->rxq_privs == NULL) return 0; - rxq_ctrl = container_of((*priv->rxqs)[idx], struct mlx5_rxq_ctrl, rxq); - if (__atomic_sub_fetch(&rxq_ctrl->refcnt, 1, __ATOMIC_RELAXED) > 1) + rxq = mlx5_rxq_get(dev, idx); + if (rxq == NULL || rxq->refcnt == 0) + return 0; + rxq_ctrl = rxq->ctrl; + refcnt = mlx5_rxq_deref(dev, idx); + if (refcnt > 1) { return 1; - if (rxq_ctrl->obj) { - priv->obj_ops.rxq_obj_release(rxq_ctrl->obj); - LIST_REMOVE(rxq_ctrl->obj, next); - mlx5_free(rxq_ctrl->obj); - rxq_ctrl->obj = NULL; - } - if (rxq_ctrl->type == MLX5_RXQ_TYPE_STANDARD) { - rxq_free_elts(rxq_ctrl); - dev->data->rx_queue_state[idx] = RTE_ETH_QUEUE_STATE_STOPPED; - } - if (!__atomic_load_n(&rxq_ctrl->refcnt, __ATOMIC_RELAXED)) { - if (rxq_ctrl->type == MLX5_RXQ_TYPE_STANDARD) - mlx5_mr_btree_free(&rxq_ctrl->rxq.mr_ctrl.cache_bh); - LIST_REMOVE(rxq_ctrl, next); - mlx5_free(rxq_ctrl); - (*priv->rxqs)[idx] = NULL; + } else if (refcnt == 1) { /* RxQ stopped. */ + priv->obj_ops.rxq_obj_release(rxq); + if (!rxq_ctrl->started && rxq_ctrl->obj != NULL) { + LIST_REMOVE(rxq_ctrl->obj, next); + mlx5_free(rxq_ctrl->obj); + rxq_ctrl->obj = NULL; + } + if (!rxq_ctrl->is_hairpin) { + if (!rxq_ctrl->started) + rxq_free_elts(rxq_ctrl); + dev->data->rx_queue_state[idx] = + RTE_ETH_QUEUE_STATE_STOPPED; + } + } else { /* Refcnt zero, closing device. */ + LIST_REMOVE(rxq, owner_entry); + if (LIST_EMPTY(&rxq_ctrl->owners)) { + if (!rxq_ctrl->is_hairpin) + mlx5_mr_btree_free + (&rxq_ctrl->rxq.mr_ctrl.cache_bh); + if (rxq_ctrl->rxq.shared) + LIST_REMOVE(rxq_ctrl, share_entry); + LIST_REMOVE(rxq_ctrl, next); + mlx5_free(rxq_ctrl); + } + dev->data->rx_queues[idx] = NULL; + mlx5_free(rxq); + (*priv->rxq_privs)[idx] = NULL; } return 0; } @@ -1762,7 +2168,7 @@ mlx5_rxq_verify(struct rte_eth_dev *dev) } /** - * Get a Rx queue type. + * Check whether RxQ type is Hairpin. * * @param dev * Pointer to Ethernet device. @@ -1770,21 +2176,15 @@ mlx5_rxq_verify(struct rte_eth_dev *dev) * Rx queue index. * * @return - * The Rx queue type. + * True if Rx queue type is Hairpin, otherwise False. */ -enum mlx5_rxq_type -mlx5_rxq_get_type(struct rte_eth_dev *dev, uint16_t idx) +bool +mlx5_rxq_is_hairpin(struct rte_eth_dev *dev, uint16_t idx) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_ctrl *rxq_ctrl = NULL; + struct mlx5_rxq_ctrl *rxq_ctrl = mlx5_rxq_ctrl_get(dev, idx); - if (idx < priv->rxqs_n && (*priv->rxqs)[idx]) { - rxq_ctrl = container_of((*priv->rxqs)[idx], - struct mlx5_rxq_ctrl, - rxq); - return rxq_ctrl->type; - } - return MLX5_RXQ_TYPE_UNDEFINED; + return (idx < priv->rxqs_n && rxq_ctrl != NULL && rxq_ctrl->is_hairpin); } /* @@ -1801,17 +2201,9 @@ mlx5_rxq_get_type(struct rte_eth_dev *dev, uint16_t idx) const struct rte_eth_hairpin_conf * mlx5_rxq_get_hairpin_conf(struct rte_eth_dev *dev, uint16_t idx) { - struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_rxq_ctrl *rxq_ctrl = NULL; + struct mlx5_rxq_priv *rxq = mlx5_rxq_get(dev, idx); - if (idx < priv->rxqs_n && (*priv->rxqs)[idx]) { - rxq_ctrl = container_of((*priv->rxqs)[idx], - struct mlx5_rxq_ctrl, - rxq); - if (rxq_ctrl->type == MLX5_RXQ_TYPE_HAIRPIN) - return &rxq_ctrl->hairpin_conf; - } - return NULL; + return mlx5_rxq_is_hairpin(dev, idx) ? &rxq->hairpin_conf : NULL; } /** @@ -1826,7 +2218,7 @@ mlx5_rxq_get_hairpin_conf(struct rte_eth_dev *dev, uint16_t idx) * Number of queues in the array. * * @return - * 1 if all queues in indirection table match 0 othrwise. + * 1 if all queues in indirection table match 0 otherwise. */ static int mlx5_ind_table_obj_match_queues(const struct mlx5_ind_table_obj *ind_tbl, @@ -1879,8 +2271,9 @@ mlx5_ind_table_obj_get(struct rte_eth_dev *dev, const uint16_t *queues, * Pointer to Ethernet device. * @param ind_table * Indirection table to release. - * @param standalone - * Indirection table for Standalone queue. + * @param deref_rxqs + * If true, then dereference RX queues related to indirection table. + * Otherwise, no additional action will be taken. * * @return * 1 while a reference on it exists, 0 when freed. @@ -1888,21 +2281,23 @@ mlx5_ind_table_obj_get(struct rte_eth_dev *dev, const uint16_t *queues, int mlx5_ind_table_obj_release(struct rte_eth_dev *dev, struct mlx5_ind_table_obj *ind_tbl, - bool standalone) + bool deref_rxqs) { struct mlx5_priv *priv = dev->data->dev_private; unsigned int i, ret; rte_rwlock_write_lock(&priv->ind_tbls_lock); ret = __atomic_sub_fetch(&ind_tbl->refcnt, 1, __ATOMIC_RELAXED); - if (!ret && !standalone) + if (!ret) LIST_REMOVE(ind_tbl, next); rte_rwlock_write_unlock(&priv->ind_tbls_lock); if (ret) return 1; priv->obj_ops.ind_table_destroy(ind_tbl); - for (i = 0; i != ind_tbl->queues_n; ++i) - claim_nonzero(mlx5_rxq_release(dev, ind_tbl->queues[i])); + if (deref_rxqs) { + for (i = 0; i != ind_tbl->queues_n; ++i) + claim_nonzero(mlx5_rxq_deref(dev, ind_tbl->queues[i])); + } mlx5_free(ind_tbl); return 0; } @@ -1941,39 +2336,45 @@ mlx5_ind_table_obj_verify(struct rte_eth_dev *dev) * Pointer to Ethernet device. * @param ind_table * Indirection table to modify. + * @param ref_qs + * Whether to increment RxQ reference counters. * * @return * 0 on success, a negative errno value otherwise and rte_errno is set. */ int mlx5_ind_table_obj_setup(struct rte_eth_dev *dev, - struct mlx5_ind_table_obj *ind_tbl) + struct mlx5_ind_table_obj *ind_tbl, + bool ref_qs) { struct mlx5_priv *priv = dev->data->dev_private; uint32_t queues_n = ind_tbl->queues_n; uint16_t *queues = ind_tbl->queues; - unsigned int i, j; + unsigned int i = 0, j; int ret = 0, err; const unsigned int n = rte_is_power_of_2(queues_n) ? log2above(queues_n) : - log2above(priv->config.ind_table_max_size); - - for (i = 0; i != queues_n; ++i) { - if (!mlx5_rxq_get(dev, queues[i])) { - ret = -rte_errno; - goto error; + log2above(priv->sh->dev_cap.ind_table_max_size); + + if (ref_qs) + for (i = 0; i != queues_n; ++i) { + if (mlx5_rxq_ref(dev, queues[i]) == NULL) { + ret = -rte_errno; + goto error; + } } - } ret = priv->obj_ops.ind_table_new(dev, n, ind_tbl); if (ret) goto error; __atomic_fetch_add(&ind_tbl->refcnt, 1, __ATOMIC_RELAXED); return 0; error: - err = rte_errno; - for (j = 0; j < i; j++) - mlx5_rxq_release(dev, ind_tbl->queues[j]); - rte_errno = err; + if (ref_qs) { + err = rte_errno; + for (j = 0; j < i; j++) + mlx5_rxq_deref(dev, queues[j]); + rte_errno = err; + } DRV_LOG(DEBUG, "Port %u cannot setup indirection table.", dev->data->port_id); return ret; @@ -1990,20 +2391,27 @@ error: * Number of queues in the array. * @param standalone * Indirection table for Standalone queue. + * @param ref_qs + * Whether to increment RxQ reference counters. * * @return * The Verbs/DevX object initialized, NULL otherwise and rte_errno is set. */ -static struct mlx5_ind_table_obj * +struct mlx5_ind_table_obj * mlx5_ind_table_obj_new(struct rte_eth_dev *dev, const uint16_t *queues, - uint32_t queues_n, bool standalone) + uint32_t queues_n, bool standalone, bool ref_qs) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_ind_table_obj *ind_tbl; int ret; + /* + * Allocate maximum queues for shared action as queue number + * maybe modified later. + */ ind_tbl = mlx5_malloc(MLX5_MEM_ZERO, sizeof(*ind_tbl) + - queues_n * sizeof(uint16_t), 0, SOCKET_ID_ANY); + (standalone ? priv->rxqs_n : queues_n) * + sizeof(uint16_t), 0, SOCKET_ID_ANY); if (!ind_tbl) { rte_errno = ENOMEM; return NULL; @@ -2011,19 +2419,41 @@ mlx5_ind_table_obj_new(struct rte_eth_dev *dev, const uint16_t *queues, ind_tbl->queues_n = queues_n; ind_tbl->queues = (uint16_t *)(ind_tbl + 1); memcpy(ind_tbl->queues, queues, queues_n * sizeof(*queues)); - ret = mlx5_ind_table_obj_setup(dev, ind_tbl); + ret = mlx5_ind_table_obj_setup(dev, ind_tbl, ref_qs); if (ret < 0) { mlx5_free(ind_tbl); return NULL; } - if (!standalone) { - rte_rwlock_write_lock(&priv->ind_tbls_lock); + rte_rwlock_write_lock(&priv->ind_tbls_lock); + if (!standalone) LIST_INSERT_HEAD(&priv->ind_tbls, ind_tbl, next); - rte_rwlock_write_unlock(&priv->ind_tbls_lock); - } + else + LIST_INSERT_HEAD(&priv->standalone_ind_tbls, ind_tbl, next); + rte_rwlock_write_unlock(&priv->ind_tbls_lock); + return ind_tbl; } +static int +mlx5_ind_table_obj_check_standalone(struct rte_eth_dev *dev __rte_unused, + struct mlx5_ind_table_obj *ind_tbl) +{ + uint32_t refcnt; + + refcnt = __atomic_load_n(&ind_tbl->refcnt, __ATOMIC_RELAXED); + if (refcnt <= 1) + return 0; + /* + * Modification of indirection tables having more than 1 + * reference is unsupported. + */ + DRV_LOG(DEBUG, + "Port %u cannot modify indirection table %p (refcnt %u > 1).", + dev->data->port_id, (void *)ind_tbl, refcnt); + rte_errno = EINVAL; + return -rte_errno; +} + /** * Modify an indirection table. * @@ -2037,6 +2467,10 @@ mlx5_ind_table_obj_new(struct rte_eth_dev *dev, const uint16_t *queues, * Number of queues in the array. * @param standalone * Indirection table for Standalone queue. + * @param ref_new_qs + * Whether to increment new RxQ set reference counters. + * @param deref_old_qs + * Whether to decrement old RxQ set reference counters. * * @return * 0 on success, a negative errno value otherwise and rte_errno is set. @@ -2045,54 +2479,113 @@ int mlx5_ind_table_obj_modify(struct rte_eth_dev *dev, struct mlx5_ind_table_obj *ind_tbl, uint16_t *queues, const uint32_t queues_n, - bool standalone) + bool standalone, bool ref_new_qs, bool deref_old_qs) { struct mlx5_priv *priv = dev->data->dev_private; - unsigned int i, j; + unsigned int i = 0, j; int ret = 0, err; const unsigned int n = rte_is_power_of_2(queues_n) ? log2above(queues_n) : - log2above(priv->config.ind_table_max_size); + log2above(priv->sh->dev_cap.ind_table_max_size); MLX5_ASSERT(standalone); RTE_SET_USED(standalone); - if (__atomic_load_n(&ind_tbl->refcnt, __ATOMIC_RELAXED) > 1) { - /* - * Modification of indirection ntables having more than 1 - * reference unsupported. Intended for standalone indirection - * tables only. - */ - DRV_LOG(DEBUG, - "Port %u cannot modify indirection table (refcnt> 1).", - dev->data->port_id); - rte_errno = EINVAL; + if (mlx5_ind_table_obj_check_standalone(dev, ind_tbl) < 0) return -rte_errno; - } - for (i = 0; i != queues_n; ++i) { - if (!mlx5_rxq_get(dev, queues[i])) { - ret = -rte_errno; - goto error; + if (ref_new_qs) + for (i = 0; i != queues_n; ++i) { + if (!mlx5_rxq_ref(dev, queues[i])) { + ret = -rte_errno; + goto error; + } } - } MLX5_ASSERT(priv->obj_ops.ind_table_modify); ret = priv->obj_ops.ind_table_modify(dev, n, queues, queues_n, ind_tbl); if (ret) goto error; - for (j = 0; j < ind_tbl->queues_n; j++) - mlx5_rxq_release(dev, ind_tbl->queues[j]); + if (deref_old_qs) + for (i = 0; i < ind_tbl->queues_n; i++) + claim_nonzero(mlx5_rxq_deref(dev, ind_tbl->queues[i])); ind_tbl->queues_n = queues_n; ind_tbl->queues = queues; return 0; error: - err = rte_errno; - for (j = 0; j < i; j++) - mlx5_rxq_release(dev, ind_tbl->queues[j]); - rte_errno = err; + if (ref_new_qs) { + err = rte_errno; + for (j = 0; j < i; j++) + mlx5_rxq_deref(dev, queues[j]); + rte_errno = err; + } DRV_LOG(DEBUG, "Port %u cannot setup indirection table.", dev->data->port_id); return ret; } +/** + * Attach an indirection table to its queues. + * + * @param dev + * Pointer to Ethernet device. + * @param ind_table + * Indirection table to attach. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +int +mlx5_ind_table_obj_attach(struct rte_eth_dev *dev, + struct mlx5_ind_table_obj *ind_tbl) +{ + int ret; + + ret = mlx5_ind_table_obj_modify(dev, ind_tbl, ind_tbl->queues, + ind_tbl->queues_n, + true /* standalone */, + true /* ref_new_qs */, + false /* deref_old_qs */); + if (ret != 0) + DRV_LOG(ERR, "Port %u could not modify indirect table obj %p", + dev->data->port_id, (void *)ind_tbl); + return ret; +} + +/** + * Detach an indirection table from its queues. + * + * @param dev + * Pointer to Ethernet device. + * @param ind_table + * Indirection table to detach. + * + * @return + * 0 on success, a negative errno value otherwise and rte_errno is set. + */ +int +mlx5_ind_table_obj_detach(struct rte_eth_dev *dev, + struct mlx5_ind_table_obj *ind_tbl) +{ + struct mlx5_priv *priv = dev->data->dev_private; + const unsigned int n = rte_is_power_of_2(ind_tbl->queues_n) ? + log2above(ind_tbl->queues_n) : + log2above(priv->sh->dev_cap.ind_table_max_size); + unsigned int i; + int ret; + + ret = mlx5_ind_table_obj_check_standalone(dev, ind_tbl); + if (ret != 0) + return ret; + MLX5_ASSERT(priv->obj_ops.ind_table_modify); + ret = priv->obj_ops.ind_table_modify(dev, n, NULL, 0, ind_tbl); + if (ret != 0) { + DRV_LOG(ERR, "Port %u could not modify indirect table obj %p", + dev->data->port_id, (void *)ind_tbl); + return ret; + } + for (i = 0; i < ind_tbl->queues_n; i++) + mlx5_rxq_release(dev, ind_tbl->queues[i]); + return ret; +} + int mlx5_hrxq_match_cb(void *tool_ctx __rte_unused, struct mlx5_list_entry *entry, void *cb_ctx) @@ -2103,6 +2596,7 @@ mlx5_hrxq_match_cb(void *tool_ctx __rte_unused, struct mlx5_list_entry *entry, return (hrxq->rss_key_len != rss_desc->key_len || memcmp(hrxq->rss_key, rss_desc->key, rss_desc->key_len) || + hrxq->hws_flags != rss_desc->hws_flags || hrxq->hash_fields != rss_desc->hash_fields || hrxq->ind_table->queues_n != rss_desc->queue_num || memcmp(hrxq->ind_table->queues, rss_desc->queue, @@ -2142,6 +2636,7 @@ mlx5_hrxq_modify(struct rte_eth_dev *dev, uint32_t hrxq_idx, struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_hrxq *hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], hrxq_idx); + bool dev_started = !!dev->data->dev_started; int ret; if (!hrxq) { @@ -2162,7 +2657,7 @@ mlx5_hrxq_modify(struct rte_eth_dev *dev, uint32_t hrxq_idx, if (hrxq->standalone) { /* * Replacement of indirection table unsupported for - * stanalone hrxq objects (used by shared RSS). + * standalone hrxq objects (used by shared RSS). */ rte_errno = ENOTSUP; return -rte_errno; @@ -2170,7 +2665,8 @@ mlx5_hrxq_modify(struct rte_eth_dev *dev, uint32_t hrxq_idx, ind_tbl = mlx5_ind_table_obj_get(dev, queues, queues_n); if (!ind_tbl) ind_tbl = mlx5_ind_table_obj_new(dev, queues, queues_n, - hrxq->standalone); + hrxq->standalone, + dev_started); } if (!ind_tbl) { rte_errno = ENOMEM; @@ -2185,8 +2681,7 @@ mlx5_hrxq_modify(struct rte_eth_dev *dev, uint32_t hrxq_idx, } if (ind_tbl != hrxq->ind_table) { MLX5_ASSERT(!hrxq->standalone); - mlx5_ind_table_obj_release(dev, hrxq->ind_table, - hrxq->standalone); + mlx5_ind_table_obj_release(dev, hrxq->ind_table, true); hrxq->ind_table = ind_tbl; } hrxq->hash_fields = hash_fields; @@ -2196,7 +2691,7 @@ error: err = rte_errno; if (ind_tbl != hrxq->ind_table) { MLX5_ASSERT(!hrxq->standalone); - mlx5_ind_table_obj_release(dev, ind_tbl, hrxq->standalone); + mlx5_ind_table_obj_release(dev, ind_tbl, true); } rte_errno = err; return -rte_errno; @@ -2208,12 +2703,16 @@ __mlx5_hrxq_remove(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq) struct mlx5_priv *priv = dev->data->dev_private; #ifdef HAVE_IBV_FLOW_DV_SUPPORT - mlx5_glue->destroy_flow_action(hrxq->action); + if (hrxq->hws_flags) + mlx5dr_action_destroy(hrxq->action); + else + mlx5_glue->destroy_flow_action(hrxq->action); #endif priv->obj_ops.hrxq_destroy(hrxq); if (!hrxq->standalone) { mlx5_ind_table_obj_release(dev, hrxq->ind_table, - hrxq->standalone); + hrxq->hws_flags ? + (!!dev->data->dev_started) : true); } mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_HRXQ], hrxq->idx); } @@ -2257,11 +2756,13 @@ __mlx5_hrxq_create(struct rte_eth_dev *dev, int ret; queues_n = rss_desc->hash_fields ? queues_n : 1; - if (!ind_tbl) + if (!ind_tbl && !rss_desc->hws_flags) ind_tbl = mlx5_ind_table_obj_get(dev, queues, queues_n); if (!ind_tbl) ind_tbl = mlx5_ind_table_obj_new(dev, queues, queues_n, - standalone); + standalone || + rss_desc->hws_flags, + !!dev->data->dev_started); if (!ind_tbl) return NULL; hrxq = mlx5_ipool_zmalloc(priv->sh->ipool[MLX5_IPOOL_HRXQ], &hrxq_idx); @@ -2272,6 +2773,7 @@ __mlx5_hrxq_create(struct rte_eth_dev *dev, hrxq->ind_table = ind_tbl; hrxq->rss_key_len = rss_key_len; hrxq->hash_fields = rss_desc->hash_fields; + hrxq->hws_flags = rss_desc->hws_flags; memcpy(hrxq->rss_key, rss_key, rss_key_len); ret = priv->obj_ops.hrxq_new(dev, hrxq, rss_desc->tunnel); if (ret < 0) @@ -2279,7 +2781,7 @@ __mlx5_hrxq_create(struct rte_eth_dev *dev, return hrxq; error: if (!rss_desc->ind_tbl) - mlx5_ind_table_obj_release(dev, ind_tbl, standalone); + mlx5_ind_table_obj_release(dev, ind_tbl, true); if (hrxq) mlx5_ipool_free(priv->sh->ipool[MLX5_IPOOL_HRXQ], hrxq_idx); return NULL; @@ -2333,13 +2835,13 @@ mlx5_hrxq_clone_free_cb(void *tool_ctx, struct mlx5_list_entry *entry) * RSS configuration for the Rx hash queue. * * @return - * An hash Rx queue index on success. + * An hash Rx queue on success. */ -uint32_t mlx5_hrxq_get(struct rte_eth_dev *dev, +struct mlx5_hrxq *mlx5_hrxq_get(struct rte_eth_dev *dev, struct mlx5_flow_rss_desc *rss_desc) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_hrxq *hrxq; + struct mlx5_hrxq *hrxq = NULL; struct mlx5_list_entry *entry; struct mlx5_flow_cb_ctx ctx = { .data = rss_desc, @@ -2350,12 +2852,10 @@ uint32_t mlx5_hrxq_get(struct rte_eth_dev *dev, } else { entry = mlx5_list_register(priv->hrxqs, &ctx); if (!entry) - return 0; + return NULL; hrxq = container_of(entry, typeof(*hrxq), entry); } - if (hrxq) - return hrxq->idx; - return 0; + return hrxq; } /** @@ -2364,17 +2864,15 @@ uint32_t mlx5_hrxq_get(struct rte_eth_dev *dev, * @param dev * Pointer to Ethernet device. * @param hrxq_idx - * Index to Hash Rx queue to release. + * Hash Rx queue to release. * * @return * 1 while a reference on it exists, 0 when freed. */ -int mlx5_hrxq_release(struct rte_eth_dev *dev, uint32_t hrxq_idx) +int mlx5_hrxq_obj_release(struct rte_eth_dev *dev, struct mlx5_hrxq *hrxq) { struct mlx5_priv *priv = dev->data->dev_private; - struct mlx5_hrxq *hrxq; - hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], hrxq_idx); if (!hrxq) return 0; if (!hrxq->standalone) @@ -2383,6 +2881,26 @@ int mlx5_hrxq_release(struct rte_eth_dev *dev, uint32_t hrxq_idx) return 0; } +/** + * Release the hash Rx queue with index. + * + * @param dev + * Pointer to Ethernet device. + * @param hrxq_idx + * Index to Hash Rx queue to release. + * + * @return + * 1 while a reference on it exists, 0 when freed. + */ +int mlx5_hrxq_release(struct rte_eth_dev *dev, uint32_t hrxq_idx) +{ + struct mlx5_priv *priv = dev->data->dev_private; + struct mlx5_hrxq *hrxq; + + hrxq = mlx5_ipool_get(priv->sh->ipool[MLX5_IPOOL_HRXQ], hrxq_idx); + return mlx5_hrxq_obj_release(dev, hrxq); +} + /** * Create a drop Rx Hash queue. * @@ -2480,14 +2998,130 @@ mlx5_rxq_timestamp_set(struct rte_eth_dev *dev) { struct mlx5_priv *priv = dev->data->dev_private; struct mlx5_dev_ctx_shared *sh = priv->sh; - struct mlx5_rxq_data *data; unsigned int i; for (i = 0; i != priv->rxqs_n; ++i) { - if (!(*priv->rxqs)[i]) + struct mlx5_rxq_data *data = mlx5_rxq_data_get(dev, i); + + if (data == NULL) continue; - data = (*priv->rxqs)[i]; data->sh = sh; - data->rt_timestamp = priv->config.rt_timestamp; + data->rt_timestamp = sh->dev_cap.rt_timestamp; + } +} + +/** + * Validate given external RxQ rte_plow index, and get pointer to concurrent + * external RxQ object to map/unmap. + * + * @param[in] port_id + * The port identifier of the Ethernet device. + * @param[in] dpdk_idx + * Queue index in rte_flow. + * + * @return + * Pointer to concurrent external RxQ on success, + * NULL otherwise and rte_errno is set. + */ +static struct mlx5_external_rxq * +mlx5_external_rx_queue_get_validate(uint16_t port_id, uint16_t dpdk_idx) +{ + struct rte_eth_dev *dev; + struct mlx5_priv *priv; + + if (dpdk_idx < MLX5_EXTERNAL_RX_QUEUE_ID_MIN) { + DRV_LOG(ERR, "Queue index %u should be in range: [%u, %u].", + dpdk_idx, MLX5_EXTERNAL_RX_QUEUE_ID_MIN, UINT16_MAX); + rte_errno = EINVAL; + return NULL; + } + if (rte_eth_dev_is_valid_port(port_id) < 0) { + DRV_LOG(ERR, "There is no Ethernet device for port %u.", + port_id); + rte_errno = ENODEV; + return NULL; + } + dev = &rte_eth_devices[port_id]; + priv = dev->data->dev_private; + if (!mlx5_imported_pd_and_ctx(priv->sh->cdev)) { + DRV_LOG(ERR, "Port %u " + "external RxQ isn't supported on local PD and CTX.", + port_id); + rte_errno = ENOTSUP; + return NULL; + } + if (!mlx5_devx_obj_ops_en(priv->sh)) { + DRV_LOG(ERR, + "Port %u external RxQ isn't supported by Verbs API.", + port_id); + rte_errno = ENOTSUP; + return NULL; + } + /* + * When user configures remote PD and CTX and device creates RxQ by + * DevX, external RxQs array is allocated. + */ + MLX5_ASSERT(priv->ext_rxqs != NULL); + return &priv->ext_rxqs[dpdk_idx - MLX5_EXTERNAL_RX_QUEUE_ID_MIN]; +} + +int +rte_pmd_mlx5_external_rx_queue_id_map(uint16_t port_id, uint16_t dpdk_idx, + uint32_t hw_idx) +{ + struct mlx5_external_rxq *ext_rxq; + uint32_t unmapped = 0; + + ext_rxq = mlx5_external_rx_queue_get_validate(port_id, dpdk_idx); + if (ext_rxq == NULL) + return -rte_errno; + if (!__atomic_compare_exchange_n(&ext_rxq->refcnt, &unmapped, 1, false, + __ATOMIC_RELAXED, __ATOMIC_RELAXED)) { + if (ext_rxq->hw_id != hw_idx) { + DRV_LOG(ERR, "Port %u external RxQ index %u " + "is already mapped to HW index (requesting is " + "%u, existing is %u).", + port_id, dpdk_idx, hw_idx, ext_rxq->hw_id); + rte_errno = EEXIST; + return -rte_errno; + } + DRV_LOG(WARNING, "Port %u external RxQ index %u " + "is already mapped to the requested HW index (%u)", + port_id, dpdk_idx, hw_idx); + + } else { + ext_rxq->hw_id = hw_idx; + DRV_LOG(DEBUG, "Port %u external RxQ index %u " + "is successfully mapped to the requested HW index (%u)", + port_id, dpdk_idx, hw_idx); + } + return 0; +} + +int +rte_pmd_mlx5_external_rx_queue_id_unmap(uint16_t port_id, uint16_t dpdk_idx) +{ + struct mlx5_external_rxq *ext_rxq; + uint32_t mapped = 1; + + ext_rxq = mlx5_external_rx_queue_get_validate(port_id, dpdk_idx); + if (ext_rxq == NULL) + return -rte_errno; + if (ext_rxq->refcnt > 1) { + DRV_LOG(ERR, "Port %u external RxQ index %u still referenced.", + port_id, dpdk_idx); + rte_errno = EINVAL; + return -rte_errno; + } + if (!__atomic_compare_exchange_n(&ext_rxq->refcnt, &mapped, 0, false, + __ATOMIC_RELAXED, __ATOMIC_RELAXED)) { + DRV_LOG(ERR, "Port %u external RxQ index %u doesn't exist.", + port_id, dpdk_idx); + rte_errno = EINVAL; + return -rte_errno; } + DRV_LOG(DEBUG, + "Port %u external RxQ index %u is successfully unmapped.", + port_id, dpdk_idx); + return 0; }