/* SPDX-License-Identifier: BSD-3-Clause * Copyright(c) 2018-2025 Beijing WangXun Technology Co., Ltd. * Copyright(c) 2010-2017 Intel Corporation */ #include #include #include #include #include #include #include "ngbe_logs.h" #include "base/ngbe.h" #include "ngbe_ethdev.h" #include "ngbe_rxtx.h" #include "ngbe_regs_group.h" static const struct reg_info ngbevf_regs_general[] = { {NGBE_VFRST, 1, 1, "NGBE_VFRST"}, {NGBE_VFSTATUS, 1, 1, "NGBE_VFSTATUS"}, {NGBE_VFMBCTL, 1, 1, "NGBE_VFMAILBOX"}, {NGBE_VFMBX, 16, 4, "NGBE_VFMBX"}, {NGBE_VFPBWRAP, 1, 1, "NGBE_VFPBWRAP"}, {0, 0, 0, ""} }; static const struct reg_info ngbevf_regs_interrupt[] = { {0, 0, 0, ""} }; static const struct reg_info ngbevf_regs_rxdma[] = { {0, 0, 0, ""} }; static const struct reg_info ngbevf_regs_tx[] = { {0, 0, 0, ""} }; /* VF registers */ static const struct reg_info *ngbevf_regs[] = { ngbevf_regs_general, ngbevf_regs_interrupt, ngbevf_regs_rxdma, ngbevf_regs_tx, NULL}; #define NGBEVF_PMD_NAME "rte_ngbevf_pmd" /* PMD name */ static int ngbevf_dev_close(struct rte_eth_dev *dev); static int ngbevf_dev_link_update(struct rte_eth_dev *dev, int wait_to_complete); static void ngbevf_intr_disable(struct rte_eth_dev *dev); static void ngbevf_intr_enable(struct rte_eth_dev *dev); static int ngbevf_dev_stats_reset(struct rte_eth_dev *dev); static int ngbevf_vlan_offload_config(struct rte_eth_dev *dev, int mask); static void ngbevf_set_vfta_all(struct rte_eth_dev *dev, bool on); static void ngbevf_configure_msix(struct rte_eth_dev *dev); static int ngbevf_dev_promiscuous_enable(struct rte_eth_dev *dev); static int ngbevf_dev_promiscuous_disable(struct rte_eth_dev *dev); static void ngbevf_remove_mac_addr(struct rte_eth_dev *dev, uint32_t index); static void ngbevf_dev_interrupt_handler(void *param); /* * The set of PCI devices this driver supports (for VF) */ static const struct rte_pci_id pci_id_ngbevf_map[] = { { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860AL_W_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860A2_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860A2S_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860A4_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860A4S_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860AL2_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860AL2S_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860AL4_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860AL4S_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860NCSI_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860A1_VF) }, { RTE_PCI_DEVICE(PCI_VENDOR_ID_WANGXUN, NGBE_DEV_ID_EM_WX1860A1L_VF) }, { .vendor_id = 0, /* sentinel */ }, }; static const struct rte_eth_desc_lim rx_desc_lim = { .nb_max = NGBE_RING_DESC_MAX, .nb_min = NGBE_RING_DESC_MIN, .nb_align = NGBE_RXD_ALIGN, }; static const struct rte_eth_desc_lim tx_desc_lim = { .nb_max = NGBE_RING_DESC_MAX, .nb_min = NGBE_RING_DESC_MIN, .nb_align = NGBE_TXD_ALIGN, .nb_seg_max = NGBE_TX_MAX_SEG, .nb_mtu_seg_max = NGBE_TX_MAX_SEG, }; static const struct eth_dev_ops ngbevf_eth_dev_ops; static const struct rte_ngbe_xstats_name_off rte_ngbevf_stats_strings[] = { {"rx_multicast_packets", offsetof(struct ngbevf_hw_stats, vfmprc)}, }; #define NGBEVF_NB_XSTATS (sizeof(rte_ngbevf_stats_strings) / \ sizeof(rte_ngbevf_stats_strings[0])) /* * Negotiate mailbox API version with the PF. * After reset API version is always set to the basic one (ngbe_mbox_api_10). * Then we try to negotiate starting with the most recent one. * If all negotiation attempts fail, then we will proceed with * the default one (ngbe_mbox_api_10). */ static void ngbevf_negotiate_api(struct ngbe_hw *hw) { int32_t i; /* start with highest supported, proceed down */ static const int sup_ver[] = { ngbe_mbox_api_13, ngbe_mbox_api_12, ngbe_mbox_api_11, ngbe_mbox_api_10, }; for (i = 0; i < ARRAY_SIZE(sup_ver); i++) { if (ngbevf_negotiate_api_version(hw, sup_ver[i]) == 0) break; } } static void generate_random_mac_addr(struct rte_ether_addr *mac_addr) { uint64_t random; /* Set Organizationally Unique Identifier (OUI) prefix. */ mac_addr->addr_bytes[0] = 0x00; mac_addr->addr_bytes[1] = 0x09; mac_addr->addr_bytes[2] = 0xC0; /* Force indication of locally assigned MAC address. */ mac_addr->addr_bytes[0] |= RTE_ETHER_LOCAL_ADMIN_ADDR; /* Generate the last 3 bytes of the MAC address with a random number. */ random = rte_rand(); memcpy(&mac_addr->addr_bytes[3], &random, 3); } /* * Virtual Function device init */ static int eth_ngbevf_dev_init(struct rte_eth_dev *eth_dev) { int err; uint32_t tc, tcs; struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(eth_dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; struct ngbe_hw *hw = ngbe_dev_hw(eth_dev); struct ngbe_vfta *shadow_vfta = NGBE_DEV_VFTA(eth_dev); struct ngbe_hwstrip *hwstrip = NGBE_DEV_HWSTRIP(eth_dev); struct rte_ether_addr *perm_addr = (struct rte_ether_addr *)hw->mac.perm_addr; PMD_INIT_FUNC_TRACE(); eth_dev->dev_ops = &ngbevf_eth_dev_ops; eth_dev->rx_descriptor_status = ngbe_dev_rx_descriptor_status; eth_dev->tx_descriptor_status = ngbe_dev_tx_descriptor_status; eth_dev->rx_pkt_burst = &ngbe_recv_pkts; eth_dev->tx_pkt_burst = &ngbe_xmit_pkts; /* for secondary processes, we don't initialise any further as primary * has already done this work. Only check we don't need a different * RX function */ if (rte_eal_process_type() != RTE_PROC_PRIMARY) { struct ngbe_tx_queue *txq; uint16_t nb_tx_queues = eth_dev->data->nb_tx_queues; /* TX queue function in primary, set by last queue initialized * Tx queue may not initialized by primary process */ if (eth_dev->data->tx_queues) { txq = eth_dev->data->tx_queues[nb_tx_queues - 1]; ngbe_set_tx_function(eth_dev, txq); } else { /* Use default TX function if we get here */ PMD_INIT_LOG(NOTICE, "No TX queues configured yet. Using default TX function."); } ngbe_set_rx_function(eth_dev); return 0; } rte_eth_copy_pci_info(eth_dev, pci_dev); hw->device_id = pci_dev->id.device_id; hw->vendor_id = pci_dev->id.vendor_id; hw->sub_system_id = pci_dev->id.subsystem_device_id; ngbe_map_device_id(hw); hw->hw_addr = (void *)pci_dev->mem_resource[0].addr; /* initialize the vfta */ memset(shadow_vfta, 0, sizeof(*shadow_vfta)); /* initialize the hw strip bitmap*/ memset(hwstrip, 0, sizeof(*hwstrip)); /* Initialize the shared code (base driver) */ err = ngbe_init_shared_code(hw); if (err != 0) { PMD_INIT_LOG(ERR, "Shared code init failed for ngbevf: %d", err); return -EIO; } /* init_mailbox_params */ hw->mbx.init_params(hw); /* Reset the hw statistics */ ngbevf_dev_stats_reset(eth_dev); /* Disable the interrupts for VF */ ngbevf_intr_disable(eth_dev); hw->mac.num_rar_entries = 32; /* The MAX of the underlying PF */ err = hw->mac.reset_hw(hw); /* * The VF reset operation returns the NGBE_ERR_INVALID_MAC_ADDR when * the underlying PF driver has not assigned a MAC address to the VF. * In this case, assign a random MAC address. */ if (err != 0 && err != NGBE_ERR_INVALID_MAC_ADDR) { PMD_INIT_LOG(ERR, "VF Initialization Failure: %d", err); /* * This error code will be propagated to the app by * rte_eth_dev_reset, so use a public error code rather than * the internal-only NGBE_ERR_RESET_FAILED */ return -EAGAIN; } /* negotiate mailbox API version to use with the PF. */ ngbevf_negotiate_api(hw); /* Get Rx/Tx queue count via mailbox, which is ready after reset_hw */ ngbevf_get_queues(hw, &tcs, &tc); /* Allocate memory for storing MAC addresses */ eth_dev->data->mac_addrs = rte_zmalloc("ngbevf", RTE_ETHER_ADDR_LEN * hw->mac.num_rar_entries, 0); if (eth_dev->data->mac_addrs == NULL) { PMD_INIT_LOG(ERR, "Failed to allocate %u bytes needed to store MAC addresses", RTE_ETHER_ADDR_LEN * hw->mac.num_rar_entries); return -ENOMEM; } /* Generate a random MAC address, if none was assigned by PF. */ if (rte_is_zero_ether_addr(perm_addr)) { generate_random_mac_addr(perm_addr); err = ngbe_set_rar_vf(hw, 1, perm_addr->addr_bytes, 0, 1); if (err) { rte_free(eth_dev->data->mac_addrs); eth_dev->data->mac_addrs = NULL; return err; } PMD_INIT_LOG(INFO, "\tVF MAC address not assigned by Host PF"); PMD_INIT_LOG(INFO, "\tAssign randomly generated MAC address " "%02x:%02x:%02x:%02x:%02x:%02x", perm_addr->addr_bytes[0], perm_addr->addr_bytes[1], perm_addr->addr_bytes[2], perm_addr->addr_bytes[3], perm_addr->addr_bytes[4], perm_addr->addr_bytes[5]); } /* Copy the permanent MAC address */ rte_ether_addr_copy(perm_addr, ð_dev->data->mac_addrs[0]); /* reset the hardware with the new settings */ err = hw->mac.start_hw(hw); if (err) { PMD_INIT_LOG(ERR, "VF Initialization Failure: %d", err); return -EIO; } /* enter promiscuous mode */ ngbevf_dev_promiscuous_enable(eth_dev); rte_intr_callback_register(intr_handle, ngbevf_dev_interrupt_handler, eth_dev); rte_intr_enable(intr_handle); ngbevf_intr_enable(eth_dev); PMD_INIT_LOG(DEBUG, "port %d vendorID=0x%x deviceID=0x%x mac.type=%s", eth_dev->data->port_id, pci_dev->id.vendor_id, pci_dev->id.device_id, "ngbe_mac_sp_vf"); return 0; } /* Virtual Function device uninit */ static int eth_ngbevf_dev_uninit(struct rte_eth_dev *eth_dev) { PMD_INIT_FUNC_TRACE(); if (rte_eal_process_type() != RTE_PROC_PRIMARY) return 0; ngbevf_dev_close(eth_dev); return 0; } static int eth_ngbevf_pci_probe(struct rte_pci_driver *pci_drv __rte_unused, struct rte_pci_device *pci_dev) { return rte_eth_dev_pci_generic_probe(pci_dev, sizeof(struct ngbe_adapter), eth_ngbevf_dev_init); } static int eth_ngbevf_pci_remove(struct rte_pci_device *pci_dev) { return rte_eth_dev_pci_generic_remove(pci_dev, eth_ngbevf_dev_uninit); } /* * virtual function driver struct */ static struct rte_pci_driver rte_ngbevf_pmd = { .id_table = pci_id_ngbevf_map, .drv_flags = RTE_PCI_DRV_NEED_MAPPING, .probe = eth_ngbevf_pci_probe, .remove = eth_ngbevf_pci_remove, }; static int ngbevf_dev_xstats_get_names(__rte_unused struct rte_eth_dev *dev, struct rte_eth_xstat_name *xstats_names, unsigned int limit) { unsigned int i; if (limit < NGBEVF_NB_XSTATS && xstats_names != NULL) return -ENOMEM; if (xstats_names != NULL) for (i = 0; i < NGBEVF_NB_XSTATS; i++) snprintf(xstats_names[i].name, sizeof(xstats_names[i].name), "%s", rte_ngbevf_stats_strings[i].name); return NGBEVF_NB_XSTATS; } static void ngbevf_update_stats(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); struct ngbevf_hw_stats *hw_stats = (struct ngbevf_hw_stats *) NGBE_DEV_STATS(dev); /* Good Rx packet, include VF loopback */ NGBE_UPDCNT32(NGBE_QPRXPKT(0), hw_stats->last_vfgprc, hw_stats->vfgprc); /* Good Rx octets, include VF loopback */ NGBE_UPDCNT36(NGBE_QPRXOCTL(0), hw_stats->last_vfgorc, hw_stats->vfgorc); /* Rx Multicst Packet */ NGBE_UPDCNT32(NGBE_QPRXMPKT(0), hw_stats->last_vfmprc, hw_stats->vfmprc); /* Rx Broadcast Packet */ NGBE_UPDCNT32(NGBE_QPRXBPKT(0), hw_stats->last_vfbprc, hw_stats->vfbprc); hw->rx_loaded = 0; /* Good Tx packet, include VF loopback */ NGBE_UPDCNT32(NGBE_QPTXPKT(0), hw_stats->last_vfgptc, hw_stats->vfgptc); /* Good Tx octets, include VF loopback */ NGBE_UPDCNT36(NGBE_QPTXOCTL(0), hw_stats->last_vfgotc, hw_stats->vfgotc); /* Tx Multicst Packet */ NGBE_UPDCNT32(NGBE_QPTXMPKT(0), hw_stats->last_vfmprc, hw_stats->vfmprc); /* Tx Broadcast Packet */ NGBE_UPDCNT32(NGBE_QPTXBPKT(0), hw_stats->last_vfbptc, hw_stats->vfbptc); hw->offset_loaded = 0; } static int ngbevf_dev_xstats_get(struct rte_eth_dev *dev, struct rte_eth_xstat *xstats, unsigned int n) { struct ngbevf_hw_stats *hw_stats = (struct ngbevf_hw_stats *) NGBE_DEV_STATS(dev); unsigned int i; if (n < NGBEVF_NB_XSTATS) return NGBEVF_NB_XSTATS; ngbevf_update_stats(dev); if (!xstats) return 0; /* Extended stats */ for (i = 0; i < NGBEVF_NB_XSTATS; i++) { xstats[i].id = i; xstats[i].value = *(uint64_t *)(((char *)hw_stats) + rte_ngbevf_stats_strings[i].offset); } return NGBEVF_NB_XSTATS; } static int ngbevf_dev_stats_get(struct rte_eth_dev *dev, struct rte_eth_stats *stats) { struct ngbevf_hw_stats *hw_stats = (struct ngbevf_hw_stats *) NGBE_DEV_STATS(dev); ngbevf_update_stats(dev); if (stats == NULL) return -EINVAL; stats->ipackets = hw_stats->vfgprc; stats->ibytes = hw_stats->vfgorc; stats->opackets = hw_stats->vfgptc; stats->obytes = hw_stats->vfgotc; return 0; } static int ngbevf_dev_stats_reset(struct rte_eth_dev *dev) { struct ngbevf_hw_stats *hw_stats = (struct ngbevf_hw_stats *) NGBE_DEV_STATS(dev); /* Sync HW register to the last stats */ ngbevf_dev_stats_get(dev, NULL); /* reset HW current stats*/ hw_stats->vfgprc = 0; hw_stats->vfgorc = 0; hw_stats->vfgptc = 0; hw_stats->vfgotc = 0; return 0; } static int ngbevf_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info) { struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct ngbe_hw *hw = ngbe_dev_hw(dev); dev_info->max_rx_queues = (uint16_t)hw->mac.max_rx_queues; dev_info->max_tx_queues = (uint16_t)hw->mac.max_tx_queues; dev_info->min_rx_bufsize = 1024; dev_info->max_rx_pktlen = NGBE_FRAME_SIZE_MAX; dev_info->max_mac_addrs = hw->mac.num_rar_entries; dev_info->max_hash_mac_addrs = NGBE_VMDQ_NUM_UC_MAC; dev_info->max_vfs = pci_dev->max_vfs; dev_info->max_vmdq_pools = RTE_ETH_64_POOLS; dev_info->rx_queue_offload_capa = ngbe_get_rx_queue_offloads(dev); dev_info->rx_offload_capa = (ngbe_get_rx_port_offloads(dev) | dev_info->rx_queue_offload_capa); dev_info->tx_queue_offload_capa = 0; dev_info->tx_offload_capa = ngbe_get_tx_port_offloads(dev); dev_info->hash_key_size = NGBE_HKEY_MAX_INDEX * sizeof(uint32_t); dev_info->reta_size = RTE_ETH_RSS_RETA_SIZE_128; dev_info->default_rxconf = (struct rte_eth_rxconf) { .rx_thresh = { .pthresh = NGBE_DEFAULT_RX_PTHRESH, .hthresh = NGBE_DEFAULT_RX_HTHRESH, .wthresh = NGBE_DEFAULT_RX_WTHRESH, }, .rx_free_thresh = NGBE_DEFAULT_RX_FREE_THRESH, .rx_drop_en = 0, .offloads = 0, }; dev_info->default_txconf = (struct rte_eth_txconf) { .tx_thresh = { .pthresh = NGBE_DEFAULT_TX_PTHRESH, .hthresh = NGBE_DEFAULT_TX_HTHRESH, .wthresh = NGBE_DEFAULT_TX_WTHRESH, }, .tx_free_thresh = NGBE_DEFAULT_TX_FREE_THRESH, .offloads = 0, }; dev_info->rx_desc_lim = rx_desc_lim; dev_info->tx_desc_lim = tx_desc_lim; return 0; } static int ngbevf_dev_link_update(struct rte_eth_dev *dev, int wait_to_complete) { return ngbe_dev_link_update_share(dev, wait_to_complete); } static void ngbevf_intr_disable(struct rte_eth_dev *dev) { struct ngbe_interrupt *intr = ngbe_dev_intr(dev); struct ngbe_hw *hw = ngbe_dev_hw(dev); PMD_INIT_FUNC_TRACE(); /* Clear interrupt mask to stop from interrupts being generated */ wr32(hw, NGBE_VFIMS, NGBE_VFIMS_MASK); ngbe_flush(hw); /* Clear mask value. */ intr->mask_misc = NGBE_VFIMS_MASK; } static void ngbevf_intr_enable(struct rte_eth_dev *dev) { struct ngbe_interrupt *intr = ngbe_dev_intr(dev); struct ngbe_hw *hw = ngbe_dev_hw(dev); PMD_INIT_FUNC_TRACE(); /* VF enable interrupt autoclean */ wr32(hw, NGBE_VFIMC, NGBE_VFIMC_MASK); ngbe_flush(hw); intr->mask_misc = 0; } static int ngbevf_dev_configure(struct rte_eth_dev *dev) { struct rte_eth_conf *conf = &dev->data->dev_conf; struct ngbe_adapter *adapter = ngbe_dev_adapter(dev); PMD_INIT_LOG(DEBUG, "Configured Virtual Function port id: %d", dev->data->port_id); /* * VF has no ability to enable/disable HW CRC * Keep the persistent behavior the same as Host PF */ #ifndef RTE_LIBRTE_NGBE_PF_DISABLE_STRIP_CRC if (conf->rxmode.offloads & RTE_ETH_RX_OFFLOAD_KEEP_CRC) { PMD_INIT_LOG(NOTICE, "VF can't disable HW CRC Strip"); conf->rxmode.offloads &= ~RTE_ETH_RX_OFFLOAD_KEEP_CRC; } #else if (!(conf->rxmode.offloads & RTE_ETH_RX_OFFLOAD_KEEP_CRC)) { PMD_INIT_LOG(NOTICE, "VF can't enable HW CRC Strip"); conf->rxmode.offloads |= RTE_ETH_RX_OFFLOAD_KEEP_CRC; } #endif /* * Initialize to TRUE. If any of Rx queues doesn't meet the bulk * allocation or vector Rx preconditions we will reset it. */ adapter->rx_bulk_alloc_allowed = true; return 0; } static int ngbevf_dev_start(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); uint32_t intr_vector = 0; struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; int err, mask = 0; PMD_INIT_FUNC_TRACE(); err = hw->mac.reset_hw(hw); if (err) { PMD_INIT_LOG(ERR, "Unable to reset vf hardware (%d)", err); return err; } hw->mac.get_link_status = true; /* negotiate mailbox API version to use with the PF. */ ngbevf_negotiate_api(hw); ngbevf_dev_tx_init(dev); /* This can fail when allocating mbufs for descriptor rings */ err = ngbevf_dev_rx_init(dev); /** * In this case, reuses the MAC address assigned by VF * initialization. */ if (err != 0 && err != NGBE_ERR_INVALID_MAC_ADDR) { PMD_INIT_LOG(ERR, "Unable to initialize RX hardware (%d)", err); ngbe_dev_clear_queues(dev); return err; } /* Set vfta */ ngbevf_set_vfta_all(dev, 1); /* Set HW strip */ mask = RTE_ETH_VLAN_STRIP_MASK | RTE_ETH_VLAN_FILTER_MASK | RTE_ETH_VLAN_EXTEND_MASK; err = ngbevf_vlan_offload_config(dev, mask); if (err) { PMD_INIT_LOG(ERR, "Unable to set VLAN offload (%d)", err); ngbe_dev_clear_queues(dev); return err; } ngbevf_dev_rxtx_start(dev); /* check and configure queue intr-vector mapping */ if (rte_intr_cap_multiple(intr_handle) && dev->data->dev_conf.intr_conf.rxq) { /* According to datasheet, only vector 0/1/2 can be used, * now only one vector is used for Rx queue */ intr_vector = 1; if (rte_intr_efd_enable(intr_handle, intr_vector)) return -1; } if (rte_intr_dp_is_en(intr_handle)) { if (rte_intr_vec_list_alloc(intr_handle, "intr_vec", dev->data->nb_rx_queues)) { PMD_INIT_LOG(ERR, "Failed to allocate %d rx_queues intr_vec", dev->data->nb_rx_queues); return -ENOMEM; } } ngbevf_configure_msix(dev); /* When a VF port is bound to VFIO-PCI, only miscellaneous interrupt * is mapped to VFIO vector 0 in eth_ngbevf_dev_init( ). * If previous VFIO interrupt mapping setting in eth_ngbevf_dev_init( ) * is not cleared, it will fail when following rte_intr_enable( ) tries * to map Rx queue interrupt to other VFIO vectors. * So clear uio/vfio intr/evevnfd first to avoid failure. */ rte_intr_disable(intr_handle); rte_intr_enable(intr_handle); /* Re-enable interrupt for VF */ ngbevf_intr_enable(dev); /* * Update link status right before return, because it may * start link configuration process in a separate thread. */ ngbevf_dev_link_update(dev, 0); hw->adapter_stopped = false; return 0; } static int ngbevf_dev_stop(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); struct ngbe_adapter *adapter = ngbe_dev_adapter(dev); struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; if (hw->adapter_stopped) return 0; PMD_INIT_FUNC_TRACE(); ngbevf_intr_disable(dev); hw->adapter_stopped = 1; hw->mac.stop_hw(hw); /* * Clear what we set, but we still keep shadow_vfta to * restore after device starts */ ngbevf_set_vfta_all(dev, 0); /* Clear stored conf */ dev->data->scattered_rx = 0; ngbe_dev_clear_queues(dev); /* Clean datapath event and queue/vec mapping */ rte_intr_efd_disable(intr_handle); rte_intr_vec_list_free(intr_handle); adapter->rss_reta_updated = 0; return 0; } static int ngbevf_dev_close(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; int ret; PMD_INIT_FUNC_TRACE(); if (rte_eal_process_type() != RTE_PROC_PRIMARY) return 0; hw->mac.reset_hw(hw); ret = ngbevf_dev_stop(dev); ngbe_dev_free_queues(dev); /** * Remove the VF MAC address ro ensure * that the VF traffic goes to the PF * after stop, close and detach of the VF **/ ngbevf_remove_mac_addr(dev, 0); dev->rx_pkt_burst = NULL; dev->tx_pkt_burst = NULL; /* Disable the interrupts for VF */ ngbevf_intr_disable(dev); rte_free(dev->data->mac_addrs); dev->data->mac_addrs = NULL; rte_intr_disable(intr_handle); rte_intr_callback_unregister(intr_handle, ngbevf_dev_interrupt_handler, dev); return ret; } /* * Reset VF device */ static int ngbevf_dev_reset(struct rte_eth_dev *dev) { int ret; ret = eth_ngbevf_dev_uninit(dev); if (ret) return ret; ret = eth_ngbevf_dev_init(dev); return ret; } static void ngbevf_set_vfta_all(struct rte_eth_dev *dev, bool on) { struct ngbe_hw *hw = ngbe_dev_hw(dev); struct ngbe_vfta *shadow_vfta = NGBE_DEV_VFTA(dev); int i = 0, j = 0, vfta = 0, mask = 1; for (i = 0; i < NGBE_VFTA_SIZE; i++) { vfta = shadow_vfta->vfta[i]; if (vfta) { mask = 1; for (j = 0; j < 32; j++) { if (vfta & mask) hw->mac.set_vfta(hw, (i << 5) + j, 0, on, false); mask <<= 1; } } } } static int ngbevf_vlan_filter_set(struct rte_eth_dev *dev, uint16_t vlan_id, int on) { struct ngbe_hw *hw = ngbe_dev_hw(dev); struct ngbe_vfta *shadow_vfta = NGBE_DEV_VFTA(dev); uint32_t vid_idx = 0; uint32_t vid_bit = 0; int ret = 0; PMD_INIT_FUNC_TRACE(); /* vind is not used in VF driver, set to 0, check ngbe_set_vfta_vf */ ret = hw->mac.set_vfta(hw, vlan_id, 0, !!on, false); if (ret) { PMD_INIT_LOG(ERR, "Unable to set VF vlan"); return ret; } vid_idx = (uint32_t)((vlan_id >> 5) & 0x7F); vid_bit = (uint32_t)(1 << (vlan_id & 0x1F)); /* Save what we set and restore it after device reset */ if (on) shadow_vfta->vfta[vid_idx] |= vid_bit; else shadow_vfta->vfta[vid_idx] &= ~vid_bit; return 0; } static void ngbevf_vlan_strip_queue_set(struct rte_eth_dev *dev, uint16_t queue, int on) { struct ngbe_hw *hw = ngbe_dev_hw(dev); uint32_t ctrl; PMD_INIT_FUNC_TRACE(); if (queue >= hw->mac.max_rx_queues) return; ctrl = rd32(hw, NGBE_RXCFG(queue)); if (on) ctrl |= NGBE_RXCFG_VLAN; else ctrl &= ~NGBE_RXCFG_VLAN; wr32(hw, NGBE_RXCFG(queue), ctrl); ngbe_vlan_hw_strip_bitmap_set(dev, queue, on); } static int ngbevf_vlan_offload_config(struct rte_eth_dev *dev, int mask) { struct ngbe_rx_queue *rxq; uint16_t i; int on = 0; /* VF function only support hw strip feature, others are not support */ if (mask & RTE_ETH_VLAN_STRIP_MASK) { for (i = 0; i < dev->data->nb_rx_queues; i++) { rxq = dev->data->rx_queues[i]; on = !!(rxq->offloads & RTE_ETH_RX_OFFLOAD_VLAN_STRIP); ngbevf_vlan_strip_queue_set(dev, i, on); } } return 0; } static int ngbevf_vlan_offload_set(struct rte_eth_dev *dev, int mask) { ngbe_config_vlan_strip_on_all_queues(dev, mask); ngbevf_vlan_offload_config(dev, mask); return 0; } static int ngbevf_dev_rx_queue_intr_enable(struct rte_eth_dev *dev, uint16_t queue_id) { struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; struct ngbe_interrupt *intr = ngbe_dev_intr(dev); struct ngbe_hw *hw = ngbe_dev_hw(dev); uint32_t vec = NGBE_MISC_VEC_ID; if (rte_intr_allow_others(intr_handle)) vec = NGBE_RX_VEC_START; intr->mask_misc &= ~(1 << vec); RTE_SET_USED(queue_id); wr32(hw, NGBE_VFIMC, ~intr->mask_misc); rte_intr_enable(intr_handle); return 0; } static int ngbevf_dev_rx_queue_intr_disable(struct rte_eth_dev *dev, uint16_t queue_id) { struct ngbe_interrupt *intr = ngbe_dev_intr(dev); struct ngbe_hw *hw = ngbe_dev_hw(dev); struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; uint32_t vec = NGBE_MISC_VEC_ID; if (rte_intr_allow_others(intr_handle)) vec = NGBE_RX_VEC_START; intr->mask_misc |= (1 << vec); RTE_SET_USED(queue_id); wr32(hw, NGBE_VFIMS, intr->mask_misc); return 0; } static void ngbevf_set_ivar_map(struct ngbe_hw *hw, int8_t direction, uint8_t queue, uint8_t msix_vector) { uint32_t tmp, idx; if (direction == -1) { /* other causes */ msix_vector |= NGBE_VFIVAR_VLD; tmp = rd32(hw, NGBE_VFIVARMISC); tmp &= ~0xFF; tmp |= msix_vector; wr32(hw, NGBE_VFIVARMISC, tmp); } else { /* rx or tx cause */ /* Workaround for ICR lost */ idx = ((16 * (queue & 1)) + (8 * direction)); tmp = rd32(hw, NGBE_VFIVAR(queue >> 1)); tmp &= ~(0xFF << idx); tmp |= (msix_vector << idx); wr32(hw, NGBE_VFIVAR(queue >> 1), tmp); } } static void ngbevf_configure_msix(struct rte_eth_dev *dev) { struct rte_pci_device *pci_dev = RTE_ETH_DEV_TO_PCI(dev); struct rte_intr_handle *intr_handle = pci_dev->intr_handle; struct ngbe_hw *hw = ngbe_dev_hw(dev); uint32_t q_idx; uint32_t vector_idx = NGBE_MISC_VEC_ID; uint32_t base = NGBE_MISC_VEC_ID; /* Configure VF other cause ivar */ ngbevf_set_ivar_map(hw, -1, 1, vector_idx); /* won't configure msix register if no mapping is done * between intr vector and event fd. */ if (!rte_intr_dp_is_en(intr_handle)) return; if (rte_intr_allow_others(intr_handle)) { base = NGBE_RX_VEC_START; vector_idx = NGBE_RX_VEC_START; } /* Configure all RX queues of VF */ for (q_idx = 0; q_idx < dev->data->nb_rx_queues; q_idx++) { /* Force all queue use vector 0, * as NGBE_VF_MAXMSIVECOTR = 1 */ ngbevf_set_ivar_map(hw, 0, q_idx, vector_idx); rte_intr_vec_list_index_set(intr_handle, q_idx, vector_idx); if (vector_idx < base + rte_intr_nb_efd_get(intr_handle) - 1) vector_idx++; } /* As RX queue setting above show, all queues use the vector 0. * Set only the ITR value of NGBE_MISC_VEC_ID. */ wr32(hw, NGBE_ITR(NGBE_MISC_VEC_ID), NGBE_ITR_IVAL(NGBE_QUEUE_ITR_INTERVAL_DEFAULT) | NGBE_ITR_WRDSA); } static int ngbevf_add_mac_addr(struct rte_eth_dev *dev, struct rte_ether_addr *mac_addr, __rte_unused uint32_t index, __rte_unused uint32_t pool) { struct ngbe_hw *hw = ngbe_dev_hw(dev); int err; /* * On a VF, adding again the same MAC addr is not an idempotent * operation. Trap this case to avoid exhausting the [very limited] * set of PF resources used to store VF MAC addresses. */ if (memcmp(hw->mac.perm_addr, mac_addr, sizeof(struct rte_ether_addr)) == 0) return -1; err = ngbevf_set_uc_addr_vf(hw, 2, mac_addr->addr_bytes); if (err != 0) PMD_DRV_LOG(ERR, "Unable to add MAC address " "%02x:%02x:%02x:%02x:%02x:%02x - err=%d", mac_addr->addr_bytes[0], mac_addr->addr_bytes[1], mac_addr->addr_bytes[2], mac_addr->addr_bytes[3], mac_addr->addr_bytes[4], mac_addr->addr_bytes[5], err); return err; } static void ngbevf_remove_mac_addr(struct rte_eth_dev *dev, uint32_t index) { struct ngbe_hw *hw = ngbe_dev_hw(dev); struct rte_ether_addr *perm_addr = (struct rte_ether_addr *)hw->mac.perm_addr; struct rte_ether_addr *mac_addr; uint32_t i; int err; /* * The NGBE_VF_SET_MACVLAN command of the ngbe-pf driver does * not support the deletion of a given MAC address. * Instead, it imposes to delete all MAC addresses, then to add again * all MAC addresses with the exception of the one to be deleted. */ (void)ngbevf_set_uc_addr_vf(hw, 0, NULL); /* * Add again all MAC addresses, with the exception of the deleted one * and of the permanent MAC address. */ for (i = 0, mac_addr = dev->data->mac_addrs; i < hw->mac.num_rar_entries; i++, mac_addr++) { /* Skip the deleted MAC address */ if (i == index) continue; /* Skip NULL MAC addresses */ if (rte_is_zero_ether_addr(mac_addr)) continue; /* Skip the permanent MAC address */ if (memcmp(perm_addr, mac_addr, sizeof(struct rte_ether_addr)) == 0) continue; err = ngbevf_set_uc_addr_vf(hw, 2, mac_addr->addr_bytes); if (err != 0) PMD_DRV_LOG(ERR, "Adding again MAC address " "%02x:%02x:%02x:%02x:%02x:%02x failed " "err=%d", mac_addr->addr_bytes[0], mac_addr->addr_bytes[1], mac_addr->addr_bytes[2], mac_addr->addr_bytes[3], mac_addr->addr_bytes[4], mac_addr->addr_bytes[5], err); } } static int ngbevf_set_default_mac_addr(struct rte_eth_dev *dev, struct rte_ether_addr *addr) { struct ngbe_hw *hw = ngbe_dev_hw(dev); hw->mac.set_rar(hw, 0, (void *)addr, 0, 0); return 0; } static int ngbevf_dev_set_mtu(struct rte_eth_dev *dev, uint16_t mtu) { struct ngbe_hw *hw; uint32_t max_frame = mtu + RTE_ETHER_HDR_LEN + RTE_ETHER_CRC_LEN; struct rte_eth_dev_data *dev_data = dev->data; hw = ngbe_dev_hw(dev); if (mtu < RTE_ETHER_MIN_MTU || max_frame > RTE_ETHER_MAX_JUMBO_FRAME_LEN) return -EINVAL; /* If device is started, refuse mtu that requires the support of * scattered packets when this feature has not been enabled before. */ if (dev_data->dev_started && !dev_data->scattered_rx && (max_frame + 2 * RTE_VLAN_HLEN > dev->data->min_rx_buf_size - RTE_PKTMBUF_HEADROOM)) { PMD_INIT_LOG(ERR, "Stop port first."); return -EINVAL; } /* * When supported by the underlying PF driver, use the NGBE_VF_SET_MTU * request of the version 2.0 of the mailbox API. * For now, use the NGBE_VF_SET_LPE request of the version 1.0 * of the mailbox API. */ if (ngbevf_rlpml_set_vf(hw, max_frame)) return -EINVAL; return 0; } static int ngbevf_get_reg_length(struct rte_eth_dev *dev __rte_unused) { int count = 0; int g_ind = 0; const struct reg_info *reg_group; while ((reg_group = ngbevf_regs[g_ind++])) count += ngbe_regs_group_count(reg_group); return count; } static int ngbevf_get_regs(struct rte_eth_dev *dev, struct rte_dev_reg_info *regs) { struct ngbe_hw *hw = ngbe_dev_hw(dev); uint32_t *data = regs->data; int g_ind = 0; int count = 0; const struct reg_info *reg_group; if (data == NULL) { regs->length = ngbevf_get_reg_length(dev); regs->width = sizeof(uint32_t); return 0; } /* Support only full register dump */ if (regs->length == 0 || regs->length == (uint32_t)ngbevf_get_reg_length(dev)) { regs->version = hw->mac.type << 24 | hw->revision_id << 16 | hw->device_id; while ((reg_group = ngbevf_regs[g_ind++])) count += ngbe_read_regs_group(dev, &data[count], reg_group); return 0; } return -ENOTSUP; } static int ngbevf_dev_promiscuous_enable(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); int ret; switch (hw->mac.update_xcast_mode(hw, NGBEVF_XCAST_MODE_PROMISC)) { case 0: ret = 0; break; case NGBE_ERR_FEATURE_NOT_SUPPORTED: ret = -ENOTSUP; break; default: ret = -EAGAIN; break; } return ret; } static int ngbevf_dev_promiscuous_disable(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); int ret; switch (hw->mac.update_xcast_mode(hw, NGBEVF_XCAST_MODE_NONE)) { case 0: ret = 0; break; case NGBE_ERR_FEATURE_NOT_SUPPORTED: ret = -ENOTSUP; break; default: ret = -EAGAIN; break; } return ret; } static int ngbevf_dev_allmulticast_enable(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); int ret; if (dev->data->promiscuous == 1) return 0; switch (hw->mac.update_xcast_mode(hw, NGBEVF_XCAST_MODE_ALLMULTI)) { case 0: ret = 0; break; case NGBE_ERR_FEATURE_NOT_SUPPORTED: ret = -ENOTSUP; break; default: ret = -EAGAIN; break; } return ret; } static int ngbevf_dev_allmulticast_disable(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); int ret; switch (hw->mac.update_xcast_mode(hw, NGBEVF_XCAST_MODE_MULTI)) { case 0: ret = 0; break; case NGBE_ERR_FEATURE_NOT_SUPPORTED: ret = -ENOTSUP; break; default: ret = -EAGAIN; break; } return ret; } static void ngbevf_mbx_process(struct rte_eth_dev *dev) { struct ngbe_hw *hw = ngbe_dev_hw(dev); u32 in_msg = 0; /* peek the message first */ in_msg = rd32(hw, NGBE_VFMBX); /* PF reset VF event */ if (in_msg == NGBE_PF_CONTROL_MSG) { /* dummy mbx read to ack pf */ if (ngbe_read_mbx(hw, &in_msg, 1, 0)) return; rte_eth_dev_callback_process(dev, RTE_ETH_EVENT_INTR_RESET, NULL); } } static int ngbevf_dev_interrupt_get_status(struct rte_eth_dev *dev) { uint32_t eicr; struct ngbe_hw *hw = ngbe_dev_hw(dev); struct ngbe_interrupt *intr = ngbe_dev_intr(dev); ngbevf_intr_disable(dev); /* read-on-clear nic registers here */ eicr = rd32(hw, NGBE_VFICR); intr->flags = 0; /* only one misc vector supported - mailbox */ eicr &= NGBE_VFICR_MASK; /* Workaround for ICR lost */ intr->flags |= NGBE_FLAG_MAILBOX; /* To avoid compiler warnings set eicr to used. */ RTE_SET_USED(eicr); return 0; } static int ngbevf_dev_interrupt_action(struct rte_eth_dev *dev) { struct ngbe_interrupt *intr = ngbe_dev_intr(dev); if (intr->flags & NGBE_FLAG_MAILBOX) { ngbevf_mbx_process(dev); intr->flags &= ~NGBE_FLAG_MAILBOX; } ngbevf_intr_enable(dev); return 0; } static void ngbevf_dev_interrupt_handler(void *param) { struct rte_eth_dev *dev = (struct rte_eth_dev *)param; ngbevf_dev_interrupt_get_status(dev); ngbevf_dev_interrupt_action(dev); } /* * dev_ops for virtual function, bare necessities for basic vf * operation have been implemented */ static const struct eth_dev_ops ngbevf_eth_dev_ops = { .dev_configure = ngbevf_dev_configure, .dev_start = ngbevf_dev_start, .dev_stop = ngbevf_dev_stop, .link_update = ngbevf_dev_link_update, .stats_get = ngbevf_dev_stats_get, .xstats_get = ngbevf_dev_xstats_get, .stats_reset = ngbevf_dev_stats_reset, .xstats_reset = ngbevf_dev_stats_reset, .xstats_get_names = ngbevf_dev_xstats_get_names, .dev_close = ngbevf_dev_close, .dev_reset = ngbevf_dev_reset, .promiscuous_enable = ngbevf_dev_promiscuous_enable, .promiscuous_disable = ngbevf_dev_promiscuous_disable, .allmulticast_enable = ngbevf_dev_allmulticast_enable, .allmulticast_disable = ngbevf_dev_allmulticast_disable, .dev_infos_get = ngbevf_dev_info_get, .dev_supported_ptypes_get = ngbe_dev_supported_ptypes_get, .mtu_set = ngbevf_dev_set_mtu, .vlan_filter_set = ngbevf_vlan_filter_set, .vlan_strip_queue_set = ngbevf_vlan_strip_queue_set, .vlan_offload_set = ngbevf_vlan_offload_set, .rx_queue_setup = ngbe_dev_rx_queue_setup, .rx_queue_release = ngbe_dev_rx_queue_release, .tx_queue_setup = ngbe_dev_tx_queue_setup, .tx_queue_release = ngbe_dev_tx_queue_release, .rx_queue_intr_enable = ngbevf_dev_rx_queue_intr_enable, .rx_queue_intr_disable = ngbevf_dev_rx_queue_intr_disable, .mac_addr_add = ngbevf_add_mac_addr, .mac_addr_remove = ngbevf_remove_mac_addr, .set_mc_addr_list = ngbe_dev_set_mc_addr_list, .rxq_info_get = ngbe_rxq_info_get, .txq_info_get = ngbe_txq_info_get, .mac_addr_set = ngbevf_set_default_mac_addr, .get_reg = ngbevf_get_regs, .tx_done_cleanup = ngbe_dev_tx_done_cleanup, }; RTE_PMD_REGISTER_PCI(net_ngbe_vf, rte_ngbevf_pmd); RTE_PMD_REGISTER_PCI_TABLE(net_ngbe_vf, pci_id_ngbevf_map); RTE_PMD_REGISTER_KMOD_DEP(net_ngbe_vf, "* igb_uio | vfio-pci");