X-Git-Url: http://git.droids-corp.org/?a=blobdiff_plain;f=drivers%2Fnet%2Fenic%2Fenic_main.c;h=cbd28394621ff8c404f6224c71377c974a9a8bd0;hb=0c9da7555da8c8373dfd69f798f832723ae6de71;hp=235ef59405c8d506918e83738d6e27f0f8074b1f;hpb=8d782f3f89e1dcd0c8af1c3c93501d7a06159d66;p=dpdk.git diff --git a/drivers/net/enic/enic_main.c b/drivers/net/enic/enic_main.c index 235ef59405..cbd2839462 100644 --- a/drivers/net/enic/enic_main.c +++ b/drivers/net/enic/enic_main.c @@ -8,7 +8,6 @@ #include #include #include -#include #include #include @@ -69,12 +68,12 @@ enic_rxmbuf_queue_release(__rte_unused struct enic *enic, struct vnic_rq *rq) } } -static void enic_free_wq_buf(struct vnic_wq_buf *buf) +static void enic_free_wq_buf(struct rte_mbuf **buf) { - struct rte_mbuf *mbuf = (struct rte_mbuf *)buf->mb; + struct rte_mbuf *mbuf = *buf; rte_pktmbuf_free_seg(mbuf); - buf->mb = NULL; + *buf = NULL; } static void enic_log_q_error(struct enic *enic) @@ -162,13 +161,12 @@ int enic_dev_stats_get(struct enic *enic, struct rte_eth_stats *r_stats) return 0; } -void enic_del_mac_address(struct enic *enic, int mac_index) +int enic_del_mac_address(struct enic *enic, int mac_index) { struct rte_eth_dev *eth_dev = enic->rte_dev; uint8_t *mac_addr = eth_dev->data->mac_addrs[mac_index].addr_bytes; - if (vnic_dev_del_addr(enic->vdev, mac_addr)) - dev_err(enic, "del mac addr failed\n"); + return vnic_dev_del_addr(enic->vdev, mac_addr); } int enic_set_mac_address(struct enic *enic, uint8_t *mac_addr) @@ -200,10 +198,15 @@ void enic_init_vnic_resources(struct enic *enic) { unsigned int error_interrupt_enable = 1; unsigned int error_interrupt_offset = 0; + unsigned int rxq_interrupt_enable = 0; + unsigned int rxq_interrupt_offset = ENICPMD_RXQ_INTR_OFFSET; unsigned int index = 0; unsigned int cq_idx; struct vnic_rq *data_rq; + if (enic->rte_dev->data->dev_conf.intr_conf.rxq) + rxq_interrupt_enable = 1; + for (index = 0; index < enic->rq_count; index++) { cq_idx = enic_cq_rq(enic, enic_rte_rq_idx_to_sop_idx(index)); @@ -225,11 +228,13 @@ void enic_init_vnic_resources(struct enic *enic) 0 /* cq_head */, 0 /* cq_tail */, 1 /* cq_tail_color */, - 0 /* interrupt_enable */, + rxq_interrupt_enable, 1 /* cq_entry_enable */, 0 /* cq_message_enable */, - 0 /* interrupt offset */, + rxq_interrupt_offset, 0 /* cq_message_addr */); + if (rxq_interrupt_enable) + rxq_interrupt_offset++; } for (index = 0; index < enic->wq_count; index++) { @@ -237,6 +242,9 @@ void enic_init_vnic_resources(struct enic *enic) enic_cq_wq(enic, index), error_interrupt_enable, error_interrupt_offset); + /* Compute unsupported ol flags for enic_prep_pkts() */ + enic->wq[index].tx_offload_notsup_mask = + PKT_TX_OFFLOAD_MASK ^ enic->tx_offload_mask; cq_idx = enic_cq_wq(enic, index); vnic_cq_init(&enic->cq[cq_idx], @@ -252,10 +260,12 @@ void enic_init_vnic_resources(struct enic *enic) (u64)enic->wq[index].cqmsg_rz->iova); } - vnic_intr_init(&enic->intr, - enic->config.intr_timer_usec, - enic->config.intr_timer_type, - /*mask_on_assertion*/1); + for (index = 0; index < enic->intr_count; index++) { + vnic_intr_init(&enic->intr[index], + enic->config.intr_timer_usec, + enic->config.intr_timer_type, + /*mask_on_assertion*/1); + } } @@ -276,7 +286,7 @@ enic_alloc_rx_queue_mbufs(struct enic *enic, struct vnic_rq *rq) rq->ring.desc_count); /* - * If *not* using scatter and the mbuf size is smaller than the + * If *not* using scatter and the mbuf size is greater than the * requested max packet size (max_rx_pkt_len), then reduce the * posted buffer size to max_rx_pkt_len. HW still receives packets * larger than max_rx_pkt_len, but they will be truncated, which we @@ -304,6 +314,26 @@ enic_alloc_rx_queue_mbufs(struct enic *enic, struct vnic_rq *rq) rq_buf_len); rq->mbuf_ring[i] = mb; } + /* + * Do not post the buffers to the NIC until we enable the RQ via + * enic_start_rq(). + */ + rq->need_initial_post = true; + /* Initialize fetch index while RQ is disabled */ + iowrite32(0, &rq->ctrl->fetch_index); + return 0; +} + +/* + * Post the Rx buffers for the first time. enic_alloc_rx_queue_mbufs() has + * allocated the buffers and filled the RQ descriptor ring. Just need to push + * the post index to the NIC. + */ +static void +enic_initial_post_rx(struct enic *enic, struct vnic_rq *rq) +{ + if (!rq->in_use || !rq->need_initial_post) + return; /* make sure all prior writes are complete before doing the PIO write */ rte_rmb(); @@ -316,11 +346,8 @@ enic_alloc_rx_queue_mbufs(struct enic *enic, struct vnic_rq *rq) dev_debug(enic, "port=%u, qidx=%u, Write %u posted idx, %u sw held\n", enic->port_id, rq->index, rq->posted_index, rq->rx_nb_hold); iowrite32(rq->posted_index, &rq->ctrl->posted_index); - iowrite32(0, &rq->ctrl->fetch_index); rte_rmb(); - - return 0; - + rq->need_initial_post = false; } static void * @@ -333,8 +360,8 @@ enic_alloc_consistent(void *priv, size_t size, struct enic *enic = (struct enic *)priv; struct enic_memzone_entry *mze; - rz = rte_memzone_reserve_aligned((const char *)name, - size, SOCKET_ID_ANY, 0, ENIC_ALIGN); + rz = rte_memzone_reserve_aligned((const char *)name, size, + SOCKET_ID_ANY, RTE_MEMZONE_IOVA_CONTIG, ENIC_ALIGN); if (!rz) { pr_err("%s : Failed to allocate memory requested for %s\n", __func__, name); @@ -393,16 +420,14 @@ enic_free_consistent(void *priv, int enic_link_update(struct enic *enic) { struct rte_eth_dev *eth_dev = enic->rte_dev; - int ret; - int link_status = 0; + struct rte_eth_link link; - link_status = enic_get_link_status(enic); - ret = (link_status == enic->link_status); - enic->link_status = link_status; - eth_dev->data->dev_link.link_status = link_status; - eth_dev->data->dev_link.link_duplex = ETH_LINK_FULL_DUPLEX; - eth_dev->data->dev_link.link_speed = vnic_dev_port_speed(enic->vdev); - return ret; + memset(&link, 0, sizeof(link)); + link.link_status = enic_get_link_status(enic); + link.link_duplex = ETH_LINK_FULL_DUPLEX; + link.link_speed = vnic_dev_port_speed(enic->vdev); + + return rte_eth_linkstatus_set(eth_dev, &link); } static void @@ -411,18 +436,139 @@ enic_intr_handler(void *arg) struct rte_eth_dev *dev = (struct rte_eth_dev *)arg; struct enic *enic = pmd_priv(dev); - vnic_intr_return_all_credits(&enic->intr); + vnic_intr_return_all_credits(&enic->intr[ENICPMD_LSC_INTR_OFFSET]); enic_link_update(enic); _rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_LSC, NULL); enic_log_q_error(enic); } +static int enic_rxq_intr_init(struct enic *enic) +{ + struct rte_intr_handle *intr_handle; + uint32_t rxq_intr_count, i; + int err; + + intr_handle = enic->rte_dev->intr_handle; + if (!enic->rte_dev->data->dev_conf.intr_conf.rxq) + return 0; + /* + * Rx queue interrupts only work when we have MSI-X interrupts, + * one per queue. Sharing one interrupt is technically + * possible with VIC, but it is not worth the complications it brings. + */ + if (!rte_intr_cap_multiple(intr_handle)) { + dev_err(enic, "Rx queue interrupts require MSI-X interrupts" + " (vfio-pci driver)\n"); + return -ENOTSUP; + } + rxq_intr_count = enic->intr_count - ENICPMD_RXQ_INTR_OFFSET; + err = rte_intr_efd_enable(intr_handle, rxq_intr_count); + if (err) { + dev_err(enic, "Failed to enable event fds for Rx queue" + " interrupts\n"); + return err; + } + intr_handle->intr_vec = rte_zmalloc("enic_intr_vec", + rxq_intr_count * sizeof(int), 0); + if (intr_handle->intr_vec == NULL) { + dev_err(enic, "Failed to allocate intr_vec\n"); + return -ENOMEM; + } + for (i = 0; i < rxq_intr_count; i++) + intr_handle->intr_vec[i] = i + ENICPMD_RXQ_INTR_OFFSET; + return 0; +} + +static void enic_rxq_intr_deinit(struct enic *enic) +{ + struct rte_intr_handle *intr_handle; + + intr_handle = enic->rte_dev->intr_handle; + rte_intr_efd_disable(intr_handle); + if (intr_handle->intr_vec != NULL) { + rte_free(intr_handle->intr_vec); + intr_handle->intr_vec = NULL; + } +} + +static void enic_prep_wq_for_simple_tx(struct enic *enic, uint16_t queue_idx) +{ + struct wq_enet_desc *desc; + struct vnic_wq *wq; + unsigned int i; + + /* + * Fill WQ descriptor fields that never change. Every descriptor is + * one packet, so set EOP. Also set CQ_ENTRY every ENIC_WQ_CQ_THRESH + * descriptors (i.e. request one completion update every 32 packets). + */ + wq = &enic->wq[queue_idx]; + desc = (struct wq_enet_desc *)wq->ring.descs; + for (i = 0; i < wq->ring.desc_count; i++, desc++) { + desc->header_length_flags = 1 << WQ_ENET_FLAGS_EOP_SHIFT; + if (i % ENIC_WQ_CQ_THRESH == ENIC_WQ_CQ_THRESH - 1) + desc->header_length_flags |= + (1 << WQ_ENET_FLAGS_CQ_ENTRY_SHIFT); + } +} + +/* + * The 'strong' version is in enic_rxtx_vec_avx2.c. This weak version is used + * used when that file is not compiled. + */ +__rte_weak bool +enic_use_vector_rx_handler(__rte_unused struct enic *enic) +{ + return false; +} + +static void pick_rx_handler(struct enic *enic) +{ + struct rte_eth_dev *eth_dev; + + /* + * Preference order: + * 1. The vectorized handler if possible and requested. + * 2. The non-scatter, simplified handler if scatter Rx is not used. + * 3. The default handler as a fallback. + */ + eth_dev = enic->rte_dev; + if (enic_use_vector_rx_handler(enic)) + return; + if (enic->rq_count > 0 && enic->rq[0].data_queue_enable == 0) { + PMD_INIT_LOG(DEBUG, " use the non-scatter Rx handler"); + eth_dev->rx_pkt_burst = &enic_noscatter_recv_pkts; + } else { + PMD_INIT_LOG(DEBUG, " use the normal Rx handler"); + eth_dev->rx_pkt_burst = &enic_recv_pkts; + } +} + int enic_enable(struct enic *enic) { unsigned int index; int err; struct rte_eth_dev *eth_dev = enic->rte_dev; + uint64_t simple_tx_offloads; + uintptr_t p; + + if (enic->enable_avx2_rx) { + struct rte_mbuf mb_def = { .buf_addr = 0 }; + + /* + * mbuf_initializer contains const-after-init fields of + * receive mbufs (i.e. 64 bits of fields from rearm_data). + * It is currently used by the vectorized handler. + */ + mb_def.nb_segs = 1; + mb_def.data_off = RTE_PKTMBUF_HEADROOM; + mb_def.port = enic->port_id; + rte_mbuf_refcnt_set(&mb_def, 1); + rte_compiler_barrier(); + p = (uintptr_t)&mb_def.rearm_data; + enic->mbuf_initializer = *(uint64_t *)p; + } eth_dev->data->dev_link.link_speed = vnic_dev_port_speed(enic->vdev); eth_dev->data->dev_link.link_duplex = ETH_LINK_FULL_DUPLEX; @@ -434,6 +580,9 @@ int enic_enable(struct enic *enic) if (eth_dev->data->dev_conf.intr_conf.lsc) vnic_dev_notify_set(enic->vdev, 0); + err = enic_rxq_intr_init(enic); + if (err) + return err; if (enic_clsf_init(enic)) dev_warning(enic, "Init of hash table for clsf failed."\ "Flow director feature will not work\n"); @@ -457,6 +606,29 @@ int enic_enable(struct enic *enic) } } + /* + * Use the simple TX handler if possible. Only checksum offloads + * and vlan insertion are supported. + */ + simple_tx_offloads = enic->tx_offload_capa & + (DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM | + DEV_TX_OFFLOAD_VLAN_INSERT | + DEV_TX_OFFLOAD_IPV4_CKSUM | + DEV_TX_OFFLOAD_UDP_CKSUM | + DEV_TX_OFFLOAD_TCP_CKSUM); + if ((eth_dev->data->dev_conf.txmode.offloads & + ~simple_tx_offloads) == 0) { + PMD_INIT_LOG(DEBUG, " use the simple tx handler"); + eth_dev->tx_pkt_burst = &enic_simple_xmit_pkts; + for (index = 0; index < enic->wq_count; index++) + enic_prep_wq_for_simple_tx(enic, index); + } else { + PMD_INIT_LOG(DEBUG, " use the default tx handler"); + eth_dev->tx_pkt_burst = &enic_xmit_pkts; + } + + pick_rx_handler(enic); + for (index = 0; index < enic->wq_count; index++) enic_start_wq(enic, index); for (index = 0; index < enic->rq_count; index++) @@ -471,7 +643,8 @@ int enic_enable(struct enic *enic) enic_intr_handler, (void *)enic->rte_dev); rte_intr_enable(&(enic->pdev->intr_handle)); - vnic_intr_unmask(&enic->intr); + /* Unmask LSC interrupt */ + vnic_intr_unmask(&enic->intr[ENICPMD_LSC_INTR_OFFSET]); return 0; } @@ -479,17 +652,21 @@ int enic_enable(struct enic *enic) int enic_alloc_intr_resources(struct enic *enic) { int err; + unsigned int i; dev_info(enic, "vNIC resources used: "\ "wq %d rq %d cq %d intr %d\n", enic->wq_count, enic_vnic_rq_count(enic), enic->cq_count, enic->intr_count); - err = vnic_intr_alloc(enic->vdev, &enic->intr, 0); - if (err) - enic_free_vnic_resources(enic); - - return err; + for (i = 0; i < enic->intr_count; i++) { + err = vnic_intr_alloc(enic->vdev, &enic->intr[i], i); + if (err) { + enic_free_vnic_resources(enic); + return err; + } + } + return 0; } void enic_free_rq(void *rxq) @@ -504,6 +681,19 @@ void enic_free_rq(void *rxq) enic = vnic_dev_priv(rq_sop->vdev); rq_data = &enic->rq[rq_sop->data_queue_idx]; + if (rq_sop->free_mbufs) { + struct rte_mbuf **mb; + int i; + + mb = rq_sop->free_mbufs; + for (i = ENIC_RX_BURST_MAX - rq_sop->num_free_mbufs; + i < ENIC_RX_BURST_MAX; i++) + rte_pktmbuf_free(mb[i]); + rte_free(rq_sop->free_mbufs); + rq_sop->free_mbufs = NULL; + rq_sop->num_free_mbufs = 0; + } + enic_rxmbuf_queue_release(enic, rq_sop); if (rq_data->in_use) enic_rxmbuf_queue_release(enic, rq_data); @@ -553,10 +743,13 @@ void enic_start_rq(struct enic *enic, uint16_t queue_idx) rq_data = &enic->rq[rq_sop->data_queue_idx]; struct rte_eth_dev *eth_dev = enic->rte_dev; - if (rq_data->in_use) + if (rq_data->in_use) { vnic_rq_enable(rq_data); + enic_initial_post_rx(enic, rq_data); + } rte_mb(); vnic_rq_enable(rq_sop); + enic_initial_post_rx(enic, rq_sop); eth_dev->data->rx_queue_state[queue_idx] = RTE_ETH_QUEUE_STATE_STARTED; } @@ -645,7 +838,7 @@ int enic_alloc_rq(struct enic *enic, uint16_t queue_idx, * See enic_alloc_rx_queue_mbufs(). */ if (max_rx_pkt_len < - enic_mtu_to_max_rx_pktlen(enic->rte_dev->data->mtu)) { + enic_mtu_to_max_rx_pktlen(enic->max_mtu)) { dev_warning(enic, "rxmode.max_rx_pkt_len is ignored" " when scatter rx mode is in use.\n"); } @@ -657,20 +850,20 @@ int enic_alloc_rq(struct enic *enic, uint16_t queue_idx, } /* number of descriptors have to be a multiple of 32 */ - nb_sop_desc = (nb_desc / mbufs_per_pkt) & ~0x1F; - nb_data_desc = (nb_desc - nb_sop_desc) & ~0x1F; + nb_sop_desc = (nb_desc / mbufs_per_pkt) & ENIC_ALIGN_DESCS_MASK; + nb_data_desc = (nb_desc - nb_sop_desc) & ENIC_ALIGN_DESCS_MASK; rq_sop->max_mbufs_per_pkt = mbufs_per_pkt; rq_data->max_mbufs_per_pkt = mbufs_per_pkt; if (mbufs_per_pkt > 1) { - min_sop = 64; + min_sop = ENIC_RX_BURST_MAX; max_sop = ((enic->config.rq_desc_count / - (mbufs_per_pkt - 1)) & ~0x1F); + (mbufs_per_pkt - 1)) & ENIC_ALIGN_DESCS_MASK); min_data = min_sop * (mbufs_per_pkt - 1); max_data = enic->config.rq_desc_count; } else { - min_sop = 64; + min_sop = ENIC_RX_BURST_MAX; max_sop = enic->config.rq_desc_count; min_data = 0; max_data = 0; @@ -741,10 +934,21 @@ int enic_alloc_rq(struct enic *enic, uint16_t queue_idx, goto err_free_sop_mbuf; } + rq_sop->free_mbufs = (struct rte_mbuf **) + rte_zmalloc_socket("rq->free_mbufs", + sizeof(struct rte_mbuf *) * + ENIC_RX_BURST_MAX, + RTE_CACHE_LINE_SIZE, rq_sop->socket_id); + if (rq_sop->free_mbufs == NULL) + goto err_free_data_mbuf; + rq_sop->num_free_mbufs = 0; + rq_sop->tot_nb_desc = nb_desc; /* squirl away for MTU update function */ return 0; +err_free_data_mbuf: + rte_free(rq_data->mbuf_ring); err_free_sop_mbuf: rte_free(rq_sop->mbuf_ring); err_free_cq: @@ -784,25 +988,15 @@ int enic_alloc_wq(struct enic *enic, uint16_t queue_idx, static int instance; wq->socket_id = socket_id; - if (nb_desc) { - if (nb_desc > enic->config.wq_desc_count) { - dev_warning(enic, - "WQ %d - number of tx desc in cmd line (%d)"\ - "is greater than that in the UCSM/CIMC adapter"\ - "policy. Applying the value in the adapter "\ - "policy (%d)\n", - queue_idx, nb_desc, enic->config.wq_desc_count); - } else if (nb_desc != enic->config.wq_desc_count) { - enic->config.wq_desc_count = nb_desc; - dev_info(enic, - "TX Queues - effective number of descs:%d\n", - nb_desc); - } - } + /* + * rte_eth_tx_queue_setup() checks min, max, and alignment. So just + * print an info message for diagnostics. + */ + dev_info(enic, "TX Queues - effective number of descs:%d\n", nb_desc); /* Allocate queue resources */ err = vnic_wq_alloc(enic->vdev, &enic->wq[queue_idx], queue_idx, - enic->config.wq_desc_count, + nb_desc, sizeof(struct wq_enet_desc)); if (err) { dev_err(enic, "error in allocation of wq\n"); @@ -810,7 +1004,7 @@ int enic_alloc_wq(struct enic *enic, uint16_t queue_idx, } err = vnic_cq_alloc(enic->vdev, &enic->cq[cq_index], cq_index, - socket_id, enic->config.wq_desc_count, + socket_id, nb_desc, sizeof(struct cq_enet_wq_desc)); if (err) { vnic_wq_free(wq); @@ -823,9 +1017,8 @@ int enic_alloc_wq(struct enic *enic, uint16_t queue_idx, instance++); wq->cqmsg_rz = rte_memzone_reserve_aligned((const char *)name, - sizeof(uint32_t), - SOCKET_ID_ANY, 0, - ENIC_ALIGN); + sizeof(uint32_t), SOCKET_ID_ANY, + RTE_MEMZONE_IOVA_CONTIG, ENIC_ALIGN); if (!wq->cqmsg_rz) return -ENOMEM; @@ -837,8 +1030,11 @@ int enic_disable(struct enic *enic) unsigned int i; int err; - vnic_intr_mask(&enic->intr); - (void)vnic_intr_masked(&enic->intr); /* flush write */ + for (i = 0; i < enic->intr_count; i++) { + vnic_intr_mask(&enic->intr[i]); + (void)vnic_intr_masked(&enic->intr[i]); /* flush write */ + } + enic_rxq_intr_deinit(enic); rte_intr_disable(&enic->pdev->intr_handle); rte_intr_callback_unregister(&enic->pdev->intr_handle, enic_intr_handler, @@ -881,7 +1077,8 @@ int enic_disable(struct enic *enic) vnic_rq_clean(&enic->rq[i], enic_free_rq_buf); for (i = 0; i < enic->cq_count; i++) vnic_cq_clean(&enic->cq[i]); - vnic_intr_clean(&enic->intr); + for (i = 0; i < enic->intr_count; i++) + vnic_intr_clean(&enic->intr[i]); return 0; } @@ -914,9 +1111,10 @@ static int enic_dev_wait(struct vnic_dev *vdev, static int enic_dev_open(struct enic *enic) { int err; + int flags = CMD_OPENF_IG_DESCCACHE; err = enic_dev_wait(enic->vdev, vnic_dev_open, - vnic_dev_open_done, 0); + vnic_dev_open_done, flags); if (err) dev_err(enic_get_dev(enic), "vNIC device open failed, err %d\n", err); @@ -1101,18 +1299,32 @@ int enic_set_rss_conf(struct enic *enic, struct rte_eth_rss_conf *rss_conf) (eth_dev->data->dev_conf.rxmode.mq_mode & ETH_MQ_RX_RSS_FLAG) && rss_hf != 0) { rss_enable = 1; - if (rss_hf & ETH_RSS_IPV4) + if (rss_hf & (ETH_RSS_IPV4 | ETH_RSS_FRAG_IPV4 | + ETH_RSS_NONFRAG_IPV4_OTHER)) rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_IPV4; if (rss_hf & ETH_RSS_NONFRAG_IPV4_TCP) rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV4; - if (rss_hf & ETH_RSS_IPV6) + if (rss_hf & ETH_RSS_NONFRAG_IPV4_UDP) { + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_UDP_IPV4; + if (enic->udp_rss_weak) { + /* + * 'TCP' is not a typo. The "weak" version of + * UDP RSS requires both the TCP and UDP bits + * be set. It does enable TCP RSS as well. + */ + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV4; + } + } + if (rss_hf & (ETH_RSS_IPV6 | ETH_RSS_IPV6_EX | + ETH_RSS_FRAG_IPV6 | ETH_RSS_NONFRAG_IPV6_OTHER)) rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_IPV6; - if (rss_hf & ETH_RSS_NONFRAG_IPV6_TCP) + if (rss_hf & (ETH_RSS_NONFRAG_IPV6_TCP | ETH_RSS_IPV6_TCP_EX)) rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV6; - if (rss_hf & ETH_RSS_IPV6_EX) - rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_IPV6_EX; - if (rss_hf & ETH_RSS_IPV6_TCP_EX) - rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV6_EX; + if (rss_hf & (ETH_RSS_NONFRAG_IPV6_UDP | ETH_RSS_IPV6_UDP_EX)) { + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_UDP_IPV6; + if (enic->udp_rss_weak) + rss_hash_type |= NIC_CFG_RSS_HASH_TYPE_TCP_IPV6; + } } else { rss_enable = 0; rss_hf = 0; @@ -1134,8 +1346,11 @@ int enic_set_rss_conf(struct enic *enic, struct rte_eth_rss_conf *rss_conf) enic->rss_hf = rss_hf; enic->rss_hash_type = rss_hash_type; enic->rss_enable = rss_enable; + } else { + dev_err(enic, "Failed to update RSS configurations." + " hash=0x%x\n", rss_hash_type); } - return 0; + return ret; } int enic_set_vlan_strip(struct enic *enic) @@ -1163,13 +1378,12 @@ int enic_get_link_status(struct enic *enic) static void enic_dev_deinit(struct enic *enic) { - struct rte_eth_dev *eth_dev = enic->rte_dev; - /* stop link status checking */ vnic_dev_notify_unset(enic->vdev); - rte_free(eth_dev->data->mac_addrs); + /* mac_addrs is freed by rte_eth_dev_release_port() */ rte_free(enic->cq); + rte_free(enic->intr); rte_free(enic->rq); rte_free(enic->wq); } @@ -1179,12 +1393,16 @@ int enic_set_vnic_res(struct enic *enic) { struct rte_eth_dev *eth_dev = enic->rte_dev; int rc = 0; - unsigned int required_rq, required_wq, required_cq; + unsigned int required_rq, required_wq, required_cq, required_intr; /* Always use two vNIC RQs per eth_dev RQ, regardless of Rx scatter. */ required_rq = eth_dev->data->nb_rx_queues * 2; required_wq = eth_dev->data->nb_tx_queues; required_cq = eth_dev->data->nb_rx_queues + eth_dev->data->nb_tx_queues; + required_intr = 1; /* 1 for LSC even if intr_conf.lsc is 0 */ + if (eth_dev->data->dev_conf.intr_conf.rxq) { + required_intr += eth_dev->data->nb_rx_queues; + } if (enic->conf_rq_count < required_rq) { dev_err(dev, "Not enough Receive queues. Requested:%u which uses %d RQs on VIC, Configured:%u\n", @@ -1203,11 +1421,18 @@ int enic_set_vnic_res(struct enic *enic) required_cq, enic->conf_cq_count); rc = -EINVAL; } + if (enic->conf_intr_count < required_intr) { + dev_err(dev, "Not enough Interrupts to support Rx queue" + " interrupts. Required:%u, Configured:%u\n", + required_intr, enic->conf_intr_count); + rc = -EINVAL; + } if (rc == 0) { enic->rq_count = eth_dev->data->nb_rx_queues; enic->wq_count = eth_dev->data->nb_tx_queues; enic->cq_count = enic->rq_count + enic->wq_count; + enic->intr_count = required_intr; } return rc; @@ -1303,20 +1528,26 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) "MTU (%u) is greater than value configured in NIC (%u)\n", new_mtu, config_mtu); - /* The easy case is when scatter is disabled. However if the MTU - * becomes greater than the mbuf data size, packet drops will ensue. + /* Update the MTU and maximum packet length */ + eth_dev->data->mtu = new_mtu; + eth_dev->data->dev_conf.rxmode.max_rx_pkt_len = + enic_mtu_to_max_rx_pktlen(new_mtu); + + /* + * If the device has not started (enic_enable), nothing to do. + * Later, enic_enable() will set up RQs reflecting the new maximum + * packet length. */ - if (!(enic->rte_dev->data->dev_conf.rxmode.offloads & - DEV_RX_OFFLOAD_SCATTER)) { - eth_dev->data->mtu = new_mtu; + if (!eth_dev->data->dev_started) goto set_mtu_done; - } - /* Rx scatter is enabled so reconfigure RQ's on the fly. The point is to - * change Rx scatter mode if necessary for better performance. I.e. if - * MTU was greater than the mbuf size and now it's less, scatter Rx - * doesn't have to be used and vice versa. - */ + /* + * The device has started, re-do RQs on the fly. In the process, we + * pick up the new maximum packet length. + * + * Some applications rely on the ability to change MTU without stopping + * the device. So keep this behavior for now. + */ rte_spinlock_lock(&enic->mtu_lock); /* Stop traffic on all RQs */ @@ -1341,12 +1572,12 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) /* now it is safe to reconfigure the RQs */ - /* update the mtu */ - eth_dev->data->mtu = new_mtu; /* free and reallocate RQs with the new MTU */ for (rq_idx = 0; rq_idx < enic->rq_count; rq_idx++) { rq = &enic->rq[enic_rte_rq_idx_to_sop_idx(rq_idx)]; + if (!rq->in_use) + continue; enic_free_rq(rq); rc = enic_alloc_rq(enic, rq_idx, rq->socket_id, rq->mp, @@ -1367,7 +1598,7 @@ int enic_set_mtu(struct enic *enic, uint16_t new_mtu) /* put back the real receive function */ rte_mb(); - eth_dev->rx_pkt_burst = enic_recv_pkts; + pick_rx_handler(enic); rte_mb(); /* restart Rx traffic */ @@ -1409,6 +1640,8 @@ static int enic_dev_init(struct enic *enic) /* Queue counts may be zeros. rte_zmalloc returns NULL in that case. */ enic->cq = rte_zmalloc("enic_vnic_cq", sizeof(struct vnic_cq) * enic->conf_cq_count, 8); + enic->intr = rte_zmalloc("enic_vnic_intr", sizeof(struct vnic_intr) * + enic->conf_intr_count, 8); enic->rq = rte_zmalloc("enic_vnic_rq", sizeof(struct vnic_rq) * enic->conf_rq_count, 8); enic->wq = rte_zmalloc("enic_vnic_wq", sizeof(struct vnic_wq) * @@ -1417,6 +1650,10 @@ static int enic_dev_init(struct enic *enic) dev_err(enic, "failed to allocate vnic_cq, aborting.\n"); return -1; } + if (enic->conf_intr_count > 0 && enic->intr == NULL) { + dev_err(enic, "failed to allocate vnic_intr, aborting.\n"); + return -1; + } if (enic->conf_rq_count > 0 && enic->rq == NULL) { dev_err(enic, "failed to allocate vnic_rq, aborting.\n"); return -1; @@ -1429,23 +1666,75 @@ static int enic_dev_init(struct enic *enic) /* Get the supported filters */ enic_fdir_info(enic); - eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr", ETH_ALEN - * ENIC_MAX_MAC_ADDR, 0); + eth_dev->data->mac_addrs = rte_zmalloc("enic_mac_addr", + sizeof(struct rte_ether_addr) * + ENIC_UNICAST_PERFECT_FILTERS, 0); if (!eth_dev->data->mac_addrs) { dev_err(enic, "mac addr storage alloc failed, aborting.\n"); return -1; } - ether_addr_copy((struct ether_addr *) enic->mac_addr, + rte_ether_addr_copy((struct rte_ether_addr *)enic->mac_addr, eth_dev->data->mac_addrs); vnic_dev_set_reset_flag(enic->vdev, 0); LIST_INIT(&enic->flows); rte_spinlock_init(&enic->flows_lock); + enic->max_flow_counter = -1; /* set up link status checking */ vnic_dev_notify_set(enic->vdev, -1); /* No Intr for notify */ + enic->overlay_offload = false; + if (enic->disable_overlay && enic->vxlan) { + /* + * Explicitly disable overlay offload as the setting is + * sticky, and resetting vNIC does not disable it. + */ + if (vnic_dev_overlay_offload_ctrl(enic->vdev, + OVERLAY_FEATURE_VXLAN, + OVERLAY_OFFLOAD_DISABLE)) { + dev_err(enic, "failed to disable overlay offload\n"); + } else { + dev_info(enic, "Overlay offload is disabled\n"); + } + } + if (!enic->disable_overlay && enic->vxlan && + /* 'VXLAN feature' enables VXLAN, NVGRE, and GENEVE. */ + vnic_dev_overlay_offload_ctrl(enic->vdev, + OVERLAY_FEATURE_VXLAN, + OVERLAY_OFFLOAD_ENABLE) == 0) { + enic->tx_offload_capa |= + DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM | + DEV_TX_OFFLOAD_GENEVE_TNL_TSO | + DEV_TX_OFFLOAD_VXLAN_TNL_TSO; + enic->tx_offload_mask |= + PKT_TX_OUTER_IPV6 | + PKT_TX_OUTER_IPV4 | + PKT_TX_OUTER_IP_CKSUM | + PKT_TX_TUNNEL_MASK; + enic->overlay_offload = true; + dev_info(enic, "Overlay offload is enabled\n"); + } + /* + * Reset the vxlan port if HW vxlan parsing is available. It + * is always enabled regardless of overlay offload + * enable/disable. + */ + if (enic->vxlan) { + enic->vxlan_port = ENIC_DEFAULT_VXLAN_PORT; + /* + * Reset the vxlan port to the default, as the NIC firmware + * does not reset it automatically and keeps the old setting. + */ + if (vnic_dev_overlay_offload_cfg(enic->vdev, + OVERLAY_CFG_VXLAN_PORT_UPDATE, + ENIC_DEFAULT_VXLAN_PORT)) { + dev_err(enic, "failed to update vxlan port\n"); + return -EINVAL; + } + } + return 0; } @@ -1479,14 +1768,20 @@ int enic_probe(struct enic *enic) enic_free_consistent); /* - * Allocate the consistent memory for stats upfront so both primary and - * secondary processes can dump stats. + * Allocate the consistent memory for stats and counters upfront so + * both primary and secondary processes can access them. */ err = vnic_dev_alloc_stats_mem(enic->vdev); if (err) { dev_err(enic, "Failed to allocate cmd memory, aborting\n"); goto err_out_unregister; } + err = vnic_dev_alloc_counter_mem(enic->vdev); + if (err) { + dev_err(enic, "Failed to allocate counter memory, aborting\n"); + goto err_out_unregister; + } + /* Issue device open to get device in known state */ err = enic_dev_open(enic); if (err) { @@ -1495,8 +1790,10 @@ int enic_probe(struct enic *enic) } /* Set ingress vlan rewrite mode before vnic initialization */ + dev_debug(enic, "Set ig_vlan_rewrite_mode=%u\n", + enic->ig_vlan_rewrite_mode); err = vnic_dev_set_ig_vlan_rewrite_mode(enic->vdev, - IG_VLAN_REWRITE_MODE_PASS_THRU); + enic->ig_vlan_rewrite_mode); if (err) { dev_err(enic, "Failed to set ingress vlan rewrite mode, aborting.\n");