// SPDX-License-Identifier: GPL-2.0-or-later /* * Userspace interface * Linux ethernet bridge * * Authors: * Lennert Buytenhek */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "br_private.h" /* * Determine initial path cost based on speed. * using recommendations from 802.1d standard * * Since driver might sleep need to not be holding any locks. */ static int port_cost(struct net_device *dev) { struct ethtool_link_ksettings ecmd; if (!__ethtool_get_link_ksettings(dev, &ecmd)) { switch (ecmd.base.speed) { case SPEED_10000: return 2; case SPEED_1000: return 4; case SPEED_100: return 19; case SPEED_10: return 100; } } /* Old silly heuristics based on name */ if (!strncmp(dev->name, "lec", 3)) return 7; if (!strncmp(dev->name, "plip", 4)) return 2500; return 100; /* assume old 10Mbps */ } /* Check for port carrier transitions. */ void br_port_carrier_check(struct net_bridge_port *p, bool *notified) { struct net_device *dev = p->dev; struct net_bridge *br = p->br; if (!(p->flags & BR_ADMIN_COST) && netif_running(dev) && netif_oper_up(dev)) p->path_cost = port_cost(dev); *notified = false; if (!netif_running(br->dev)) return; spin_lock_bh(&br->lock); if (netif_running(dev) && netif_oper_up(dev)) { if (p->state == BR_STATE_DISABLED) { br_stp_enable_port(p); *notified = true; } } else { if (p->state != BR_STATE_DISABLED) { br_stp_disable_port(p); *notified = true; } } spin_unlock_bh(&br->lock); } static void br_port_set_promisc(struct net_bridge_port *p) { int err = 0; if (br_promisc_port(p)) return; err = dev_set_promiscuity(p->dev, 1); if (err) return; br_fdb_unsync_static(p->br, p); p->flags |= BR_PROMISC; } static void br_port_clear_promisc(struct net_bridge_port *p) { int err; /* Check if the port is already non-promisc or if it doesn't * support UNICAST filtering. Without unicast filtering support * we'll end up re-enabling promisc mode anyway, so just check for * it here. */ if (!br_promisc_port(p) || !(p->dev->priv_flags & IFF_UNICAST_FLT)) return; /* Since we'll be clearing the promisc mode, program the port * first so that we don't have interruption in traffic. */ err = br_fdb_sync_static(p->br, p); if (err) return; dev_set_promiscuity(p->dev, -1); p->flags &= ~BR_PROMISC; } /* When a port is added or removed or when certain port flags * change, this function is called to automatically manage * promiscuity setting of all the bridge ports. We are always called * under RTNL so can skip using rcu primitives. */ void br_manage_promisc(struct net_bridge *br) { struct net_bridge_port *p; bool set_all = false; /* If vlan filtering is disabled or bridge interface is placed * into promiscuous mode, place all ports in promiscuous mode. */ if ((br->dev->flags & IFF_PROMISC) || !br_vlan_enabled(br->dev)) set_all = true; list_for_each_entry(p, &br->port_list, list) { if (set_all) { br_port_set_promisc(p); } else { /* If the number of auto-ports is <= 1, then all other * ports will have their output configuration * statically specified through fdbs. Since ingress * on the auto-port becomes forwarding/egress to other * ports and egress configuration is statically known, * we can say that ingress configuration of the * auto-port is also statically known. * This lets us disable promiscuous mode and write * this config to hw. */ if (br->auto_cnt == 0 || (br->auto_cnt == 1 && br_auto_port(p))) br_port_clear_promisc(p); else br_port_set_promisc(p); } } } int nbp_backup_change(struct net_bridge_port *p, struct net_device *backup_dev) { struct net_bridge_port *old_backup = rtnl_dereference(p->backup_port); struct net_bridge_port *backup_p = NULL; ASSERT_RTNL(); if (backup_dev) { if (!netif_is_bridge_port(backup_dev)) return -ENOENT; backup_p = br_port_get_rtnl(backup_dev); if (backup_p->br != p->br) return -EINVAL; } if (p == backup_p) return -EINVAL; if (old_backup == backup_p) return 0; /* if the backup link is already set, clear it */ if (old_backup) old_backup->backup_redirected_cnt--; if (backup_p) backup_p->backup_redirected_cnt++; rcu_assign_pointer(p->backup_port, backup_p); return 0; } static void nbp_backup_clear(struct net_bridge_port *p) { nbp_backup_change(p, NULL); if (p->backup_redirected_cnt) { struct net_bridge_port *cur_p; list_for_each_entry(cur_p, &p->br->port_list, list) { struct net_bridge_port *backup_p; backup_p = rtnl_dereference(cur_p->backup_port); if (backup_p == p) nbp_backup_change(cur_p, NULL); } } WARN_ON(rcu_access_pointer(p->backup_port) || p->backup_redirected_cnt); } static void nbp_update_port_count(struct net_bridge *br) { struct net_bridge_port *p; u32 cnt = 0; list_for_each_entry(p, &br->port_list, list) { if (br_auto_port(p)) cnt++; } if (br->auto_cnt != cnt) { br->auto_cnt = cnt; br_manage_promisc(br); } } static void nbp_delete_promisc(struct net_bridge_port *p) { /* If port is currently promiscuous, unset promiscuity. * Otherwise, it is a static port so remove all addresses * from it. */ dev_set_allmulti(p->dev, -1); if (br_promisc_port(p)) dev_set_promiscuity(p->dev, -1); else br_fdb_unsync_static(p->br, p); } static void release_nbp(struct kobject *kobj) { struct net_bridge_port *p = container_of(kobj, struct net_bridge_port, kobj); kfree(p); } static void brport_get_ownership(struct kobject *kobj, kuid_t *uid, kgid_t *gid) { struct net_bridge_port *p = kobj_to_brport(kobj); net_ns_get_ownership(dev_net(p->dev), uid, gid); } static struct kobj_type brport_ktype = { #ifdef CONFIG_SYSFS .sysfs_ops = &brport_sysfs_ops, #endif .release = release_nbp, .get_ownership = brport_get_ownership, }; static void destroy_nbp(struct net_bridge_port *p) { struct net_device *dev = p->dev; p->br = NULL; p->dev = NULL; dev_put(dev); kobject_put(&p->kobj); } static void destroy_nbp_rcu(struct rcu_head *head) { struct net_bridge_port *p = container_of(head, struct net_bridge_port, rcu); destroy_nbp(p); } static unsigned get_max_headroom(struct net_bridge *br) { unsigned max_headroom = 0; struct net_bridge_port *p; list_for_each_entry(p, &br->port_list, list) { unsigned dev_headroom = netdev_get_fwd_headroom(p->dev); if (dev_headroom > max_headroom) max_headroom = dev_headroom; } return max_headroom; } static void update_headroom(struct net_bridge *br, int new_hr) { struct net_bridge_port *p; list_for_each_entry(p, &br->port_list, list) netdev_set_rx_headroom(p->dev, new_hr); br->dev->needed_headroom = new_hr; } /* Delete port(interface) from bridge is done in two steps. * via RCU. First step, marks device as down. That deletes * all the timers and stops new packets from flowing through. * * Final cleanup doesn't occur until after all CPU's finished * processing packets. * * Protected from multiple admin operations by RTNL mutex */ static void del_nbp(struct net_bridge_port *p) { struct net_bridge *br = p->br; struct net_device *dev = p->dev; sysfs_remove_link(br->ifobj, p->dev->name); nbp_delete_promisc(p); spin_lock_bh(&br->lock); br_stp_disable_port(p); spin_unlock_bh(&br->lock); br_ifinfo_notify(RTM_DELLINK, NULL, p); list_del_rcu(&p->list); if (netdev_get_fwd_headroom(dev) == br->dev->needed_headroom) update_headroom(br, get_max_headroom(br)); netdev_reset_rx_headroom(dev); nbp_vlan_flush(p); br_fdb_delete_by_port(br, p, 0, 1); switchdev_deferred_process(); nbp_backup_clear(p); nbp_update_port_count(br); netdev_upper_dev_unlink(dev, br->dev); dev->priv_flags &= ~IFF_BRIDGE_PORT; netdev_rx_handler_unregister(dev); br_multicast_del_port(p); kobject_uevent(&p->kobj, KOBJ_REMOVE); kobject_del(&p->kobj); br_netpoll_disable(p); call_rcu(&p->rcu, destroy_nbp_rcu); } /* Delete bridge device */ void br_dev_delete(struct net_device *dev, struct list_head *head) { struct net_bridge *br = netdev_priv(dev); struct net_bridge_port *p, *n; list_for_each_entry_safe(p, n, &br->port_list, list) { del_nbp(p); } br_recalculate_neigh_suppress_enabled(br); br_fdb_delete_by_port(br, NULL, 0, 1); cancel_delayed_work_sync(&br->gc_work); br_sysfs_delbr(br->dev); unregister_netdevice_queue(br->dev, head); } /* find an available port number */ static int find_portno(struct net_bridge *br) { int index; struct net_bridge_port *p; unsigned long *inuse; inuse = bitmap_zalloc(BR_MAX_PORTS, GFP_KERNEL); if (!inuse) return -ENOMEM; set_bit(0, inuse); /* zero is reserved */ list_for_each_entry(p, &br->port_list, list) { set_bit(p->port_no, inuse); } index = find_first_zero_bit(inuse, BR_MAX_PORTS); bitmap_free(inuse); retu