+
+ struct ibv_device *ibv_match[ret + 1];
+
+ while (ret-- > 0) {
+ struct rte_pci_addr pci_addr;
+
+ DRV_LOG(DEBUG, "checking device \"%s\"", ibv_list[ret]->name);
+ if (mlx5_ibv_device_to_pci_addr(ibv_list[ret], &pci_addr))
+ continue;
+ if (pci_dev->addr.domain != pci_addr.domain ||
+ pci_dev->addr.bus != pci_addr.bus ||
+ pci_dev->addr.devid != pci_addr.devid ||
+ pci_dev->addr.function != pci_addr.function)
+ continue;
+ DRV_LOG(INFO, "PCI information matches for device \"%s\"",
+ ibv_list[ret]->name);
+ ibv_match[n++] = ibv_list[ret];
+ }
+ ibv_match[n] = NULL;
+
+ struct mlx5_dev_spawn_data list[n];
+ int nl_route = n ? mlx5_nl_init(NETLINK_ROUTE) : -1;
+ int nl_rdma = n ? mlx5_nl_init(NETLINK_RDMA) : -1;
+ unsigned int i;
+ unsigned int u;
+
+ /*
+ * The existence of several matching entries (n > 1) means port
+ * representors have been instantiated. No existing Verbs call nor
+ * /sys entries can tell them apart, this can only be done through
+ * Netlink calls assuming kernel drivers are recent enough to
+ * support them.
+ *
+ * In the event of identification failure through Netlink, try again
+ * through sysfs, then either:
+ *
+ * 1. No device matches (n == 0), complain and bail out.
+ * 2. A single IB device matches (n == 1) and is not a representor,
+ * assume no switch support.
+ * 3. Otherwise no safe assumptions can be made; complain louder and
+ * bail out.
+ */
+ for (i = 0; i != n; ++i) {
+ list[i].ibv_dev = ibv_match[i];
+ list[i].eth_dev = NULL;
+ if (nl_rdma < 0)
+ list[i].ifindex = 0;
+ else
+ list[i].ifindex = mlx5_nl_ifindex
+ (nl_rdma, list[i].ibv_dev->name);
+ if (nl_route < 0 ||
+ !list[i].ifindex ||
+ mlx5_nl_switch_info(nl_route, list[i].ifindex,
+ &list[i].info) ||
+ ((!list[i].info.representor && !list[i].info.master) &&
+ mlx5_sysfs_switch_info(list[i].ifindex, &list[i].info))) {
+ list[i].ifindex = 0;
+ memset(&list[i].info, 0, sizeof(list[i].info));
+ continue;
+ }
+ }
+ if (nl_rdma >= 0)
+ close(nl_rdma);
+ if (nl_route >= 0)
+ close(nl_route);
+ /* Count unidentified devices. */
+ for (u = 0, i = 0; i != n; ++i)
+ if (!list[i].info.master && !list[i].info.representor)
+ ++u;
+ if (u) {
+ if (n == 1 && u == 1) {
+ /* Case #2. */
+ DRV_LOG(INFO, "no switch support detected");
+ } else {
+ /* Case #3. */
+ DRV_LOG(ERR,
+ "unable to tell which of the matching devices"
+ " is the master (lack of kernel support?)");
+ n = 0;
+ }
+ }
+ /*
+ * Sort list to probe devices in natural order for users convenience
+ * (i.e. master first, then representors from lowest to highest ID).
+ */
+ if (n)
+ qsort(list, n, sizeof(*list), mlx5_dev_spawn_data_cmp);
+ /* Default configuration. */
+ dev_config = (struct mlx5_dev_config){
+ .mps = MLX5_ARG_UNSET,
+ .tx_vec_en = 1,
+ .rx_vec_en = 1,
+ .txq_inline = MLX5_ARG_UNSET,
+ .txqs_inline = MLX5_ARG_UNSET,
+ .txqs_vec = MLX5_ARG_UNSET,
+ .inline_max_packet_sz = MLX5_ARG_UNSET,
+ .vf_nl_en = 1,
+ .mprq = {
+ .enabled = 0, /* Disabled by default. */
+ .stride_num_n = MLX5_MPRQ_STRIDE_NUM_N,
+ .max_memcpy_len = MLX5_MPRQ_MEMCPY_DEFAULT_LEN,
+ .min_rxqs_num = MLX5_MPRQ_MIN_RXQS,
+ },
+ };
+ /* Device speicific configuration. */
+ switch (pci_dev->id.device_id) {
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5BF:
+ dev_config.txqs_vec = MLX5_VPMD_MAX_TXQS_BLUEFIELD;
+ break;
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX4VF:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX4LXVF:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5VF:
+ case PCI_DEVICE_ID_MELLANOX_CONNECTX5EXVF:
+ dev_config.vf = 1;
+ break;
+ default:
+ break;
+ }
+ /* Set architecture-dependent default value if unset. */
+ if (dev_config.txqs_vec == MLX5_ARG_UNSET)
+ dev_config.txqs_vec = MLX5_VPMD_MAX_TXQS;
+ for (i = 0; i != n; ++i) {
+ uint32_t restore;
+
+ list[i].eth_dev = mlx5_dev_spawn(&pci_dev->device,
+ list[i].ibv_dev, dev_config,
+ &list[i].info);
+ if (!list[i].eth_dev) {
+ if (rte_errno != EBUSY && rte_errno != EEXIST)
+ break;
+ /* Device is disabled or already spawned. Ignore it. */
+ continue;
+ }
+ restore = list[i].eth_dev->data->dev_flags;
+ rte_eth_copy_pci_info(list[i].eth_dev, pci_dev);
+ /* Restore non-PCI flags cleared by the above call. */
+ list[i].eth_dev->data->dev_flags |= restore;
+ rte_eth_dev_probing_finish(list[i].eth_dev);
+ }
+ mlx5_glue->free_device_list(ibv_list);
+ if (!n) {
+ DRV_LOG(WARNING,
+ "no Verbs device matches PCI device " PCI_PRI_FMT ","
+ " are kernel drivers loaded?",
+ pci_dev->addr.domain, pci_dev->addr.bus,
+ pci_dev->addr.devid, pci_dev->addr.function);
+ rte_errno = ENOENT;
+ ret = -rte_errno;
+ } else if (i != n) {
+ DRV_LOG(ERR,
+ "probe of PCI device " PCI_PRI_FMT " aborted after"
+ " encountering an error: %s",
+ pci_dev->addr.domain, pci_dev->addr.bus,
+ pci_dev->addr.devid, pci_dev->addr.function,
+ strerror(rte_errno));
+ ret = -rte_errno;
+ /* Roll back. */
+ while (i--) {
+ if (!list[i].eth_dev)
+ continue;
+ mlx5_dev_close(list[i].eth_dev);
+ /* mac_addrs must not be freed because in dev_private */
+ list[i].eth_dev->data->mac_addrs = NULL;
+ claim_zero(rte_eth_dev_release_port(list[i].eth_dev));
+ }
+ /* Restore original error. */
+ rte_errno = -ret;
+ } else {
+ ret = 0;
+ }
+ return ret;
+}
+
+/**
+ * DPDK callback to remove a PCI device.
+ *
+ * This function removes all Ethernet devices belong to a given PCI device.
+ *
+ * @param[in] pci_dev
+ * Pointer to the PCI device.
+ *
+ * @return
+ * 0 on success, the function cannot fail.
+ */
+static int
+mlx5_pci_remove(struct rte_pci_device *pci_dev)
+{
+ uint16_t port_id;
+ struct rte_eth_dev *port;
+
+ for (port_id = 0; port_id < RTE_MAX_ETHPORTS; port_id++) {
+ port = &rte_eth_devices[port_id];
+ if (port->state != RTE_ETH_DEV_UNUSED &&
+ port->device == &pci_dev->device)
+ rte_eth_dev_close(port_id);
+ }